Machsupport Forum
Mach Discussion => Mach4 General Discussion => Topic started by: joeaverage on January 17, 2018, 01:28:52 AM
-
Hi All,
I've been using Mach4Hobby for about a year and I luv it. I especially like the flexibility that comes with the modular structure and how Lua has been integrated.
I would admit that at the beginning I had my doubts but the more I learn the better I like it.
I have noticed however that many who try Mach4 are less that happy...comments like "I'll wait for NFS to make Mach4 like Mach3 used to be...' are common.
It strikes me that if half (this is my wild guess) of people who try Mach4 who don't carry on with it then is Mach4 really a hobby product?
I am happy to make the effort to program Mach4 to my liking but maybe you are not. I'm hoping that this post might generate a few comments and maybe a poll
is in order to establish what people like and dislike about Mach4.
Craig
-
It strikes me that if half (this is my wild guess) of people who try Mach4 who don't carry on with it then is Mach4 really a hobby product?
There have been quite a few long established Mach3 users (and notable forum contributors) who have tried Mach4, not carried on with it, and then moved to using an entirely different software.
Tweakie.
-
I am happy to make the effort to program Mach4 to my liking but maybe you are not.
Craig
IMHO, this should be an option (very nice, BTW), NOT a necessity !
It should work as intended ... in its most basic form.
Russ
-
For us, we have a lot of special equipment. When we used/use Mach3, we wrote a lot of macros to carry out programs, with that, we had several very custom screen sets. Mach4 took some time to learn (and still learning), but it is a far better product and it has been well worth the time we've taken to learn it. Mach4 would work in its most basic form for us in some applications, but when you can customize something to do pretty much anything you want, it's hard not to make modifications to suit your needs.
I Can see why others, especially hobbyist, don't want to use Mach4 if they don't intend to take a little time to learn the program, but IMHO if someone were to take just a little bit of time to figure out the software and LUA they would be very happy they did. I do think there are some things that NFS could make standard to make it easier to set up but I also understand that it is currently working and they have better things to adjust and improve before making it more plug and play.
Mach3 isn't even an option to us anymore, Mach4 is just so much better.
-
It should work as intended ... in its most basic form.
And building on what Russ said,
And that must also include the external motion controller.
One needs a horse and a wagon to join the wagon train to venture out West.
RICH
-
Please don't hesitate to reach out to our support team using the 'Support' tab at machsupport.com with any specific issues or concerns. Mach4 (unlike Mach3) is still being improved and issues are being addressed as they arise. I am working on improving and creating new training material for users who are brand new to CNC or just new to Mach4. I will be the first to admit that it is different than Mach3 and there is a learning curve, but my team is available to help bridge that gap.
Thank you,
Bryanna
-
The switch from Mach3 lathe to mach4 lathe has been painful. Learning another programming language, errors in the screen sets, lack of documentation as to Mach4's model (lua running screen, lua running core, lua running scripts all in separate 'threads' that don't talk to each other nicely), lack of documentation on the detail of the wizards...
Having said that I wouldn't go back to Mach3 lathe. Standardizing on Fanuc style Gcode and macros means there is a lot of information available. The screen editor, once you learn where to find properties and actions and events allows you to customize to your hearts content. Hardly anyone uses Mach4 lathe yet so it is much more difficult to get forum support but thanks to guys like Craig who are willing to go above and beyond their own interests to help teach and explain has allowed me to have made great progress in understanding the new paradigm.
It is unfortunate that we have lost some of the knowledgeable users like Franco who has moved on to Centroid. He was a valuable contributor to Mach3 lathe users but when I looked at Centroid I found that you could not customize their screens in any fashion, they didn't have an API available either so that was not an option for me. It is heartening to see Rich post here with his vast knowledge on CNC lathes. I understand that when you have invested hundreds of hours making things work the way you want the thought of moving to something new means not only redoing everything but learning how to do those same things in a new way. Hope he sticks around.
If, as a hobbyist, all you want to do is run Gcode files then Mach4 does that. The gripe on hardware integration is valid but getting better, but, unless your willing to buy ALL of the hardware, including the machine, as what happens when you be from an OEM, you will always have to configure according to what you have put together. Of course you could always switch over to the Linux based CNC, but then you would have to learn another language, operating system, screen paradigm, hardware setup and configuration and rely solely on user support ;^)
MTCW
RT
-
We are using Mach4 lathe in a production shop. It is working far better than Mach3. We are about finished with another Mach4 lathe retrofit. One thing that is great about Mach4 is the Modbus, a year ago I knew nothing about modbus, it was relatively easy to learn and now we have 3 going on 4 machines using a PLC to handle most of our I/O communicating with Mach4.
To be honest, I did look into Centroid, but the fact that we can customize screens and scripts is a huge advantage to us. We have retrofitted OD Grinders that have completely customized screens, we are using Macro Programming to make our grinding more efficient with In Process Gauging. We couldn't do that with Mach3 and we wouldn't have the flexibility with other software like Centroid to accomplish our end goal, which is basically a conversational grinding operation.
-
Bryanna,
Out of curiosity, what's the preferred method of getting help, here or email support? If it's email support, no one learns anything other than the emailer. If it's here, there wouldn't be ANY zero reply posts or threads that just die with no resolution. Where did these people go who got no reply to what seems a legit post? Did they email support and get an answer that no one else knows about, or did they walk away?
Tom
-
"Out of curiosity, what's the preferred method of getting help, here or email support?"
That would depend on who you want to get an answer from. There are plenty of users on this forum who are knowledgeable and willing to help and the information shared becomes public knowledge. I think this forum is a wonderful way for users to collaborate and share ideas and I don't mean to take anything away from it by mentioning the support option available at machsupport.com.
However, many issues require a decent exchange of information (profile, g code files, macros, screenshots, etc) in order to solve and this isn't always the best platform for that kind of conversation. The support team responds to every single ticket that we receive, but we cannot monitor every thread on the forum and respond to each post. I'd like to be able to, but it just isn't feasible.
My general recommendation would be for users who are not getting the answers they need here to reach out to our team. Some users do bring their answers back to the forum so they can be shared and passed along, but that's up to them.
-
IMHO, I would and do use this forum as a first option. Growing this forum grows everyone's knowledge who utilizes it. I have reached out to Mach Support for issues that pertained to glitches in the screen editor, and have referenced that talk in the post on this forum to keep the thread up to date.
-
Hi Russ,
I am happy to make the effort to program Mach4 to my liking but maybe you are not.
Craig
IMHO, this should be an option (very nice, BTW), NOT a necessity !
It should work as intended ... in its most basic form.
This is the problem, that for most users Mach4 does not work 'out of the box' and modest adjustments have to be made
to get it to work. Your post regarding the GoToWorkZero function is an example. Not in itself a huge obstacle but
does put new users off.
There is also a perception that if, say Mach3, works a certain way then Mach4 should work identically. This is not the case
in practice. There are a few reasons. One among them is that Mach3 used shortcuts and workarounds to do certain things
but they run contrary to Fanuc21 interpretation. An example is repeated G82 lines in drill files, Mach3 will accomodate
such code whereas Mach4 which adheres to the Fanuc standard does not.
One forum contributor was perplexed, if not outright annoyed, that a signal provided in Mach4 (ISIG_SPINDLE_AT_SPEED)
but didn't actually do anything. His reasoning goes that Mach should behave in a certain way on receipt of such a signal
when in fact Mach does absolutely nothing UNLESS the user programs the desired behavior. This is an example of what I have
found to be an often repeated complaint...'Mach does not behave the way I expect....I want...demand that NFS fix it!'
I have chosen my words carefully, for if you have detected that there is a level of frustration expressed in that comment
that is correct. While some , including the OP in this case overcame that frustration and programmed the behaviour he desired,
many have not, they throw their hands up in despair and drop Mach4 altogether.
We all have ideas about how a CNC program ought to behave and more often than not contrary to someone elses interpretation.
Human nature!
Craig
-
Losing your customer base while improving your product may result in a great product but is a really bad business decision.
It is amazing how not getting a user manual in the hands of customers even while everything is changing causes everyone heartache. You waste your time answering the same questions over and over, the customer gets frustrated not finding an answer on his time schedule not yours, so walks away from your product.
I am glad to hear those that have persisted are getting good results with Mach 4. The strength of a Windows based CNC besides a nice looking interface is the fantastic productivity you can get by running a Cam program alongside. Yes I know that isn’t recommended, and I don’t switch to the Cam program while machining. There is no denying how effective this is.
I personally think the future is more like the past. A dedicated non-PC running the machine completely while a program like Mach supplies only the user interface. Computer hardware is getting more powerful and cheaper at the same time.
-
Hello Craig. I understand what you are saying ... and I agree with you.
I need to emphasize "..... in its most basic form."
Like, the code for a mill (2 nice options btw) under a button on a lathe screen.
And a work coord DRO set as un-editable.
Not aware of any others at this point, but must suspect that there are more.
Customization is great, but these are basics that could be expected to respond as expected, that's all.
With many folks, "A first impression ... is a lasting impression" be it good, or bad.
These teeny "most basic" issues contribute to that impression.
Regards,
Russ
-
I am happy to make the effort to program Mach4 to my liking but maybe you are not.
There's more to it then that.
Mach4 is not popular for a lot of reasons.
1. It's too difficult to set up. There's a lot of competition out there now, and you can argue that all of Mach4's competitors are easier to setup and use, and don't require any programming for the vat majority of users. It's possible to make a powerful program that's not overly complicated.
2. Cost. Mach4 is more expensive than it's competitors. But there is a legitimate reason for this. The real issue here is their business model.
Most of the competition offers a package of both hardware and software. Because they are making money on their hardware, they sell their software for a lot less money, or in some cases give it away for free.
Artsoft has to charge more for their software, because they don't sell software. Which leads to another issue.
3. 3rd party hardware. Mach4 decided to stick with the same flawed system they had with Mach3, where the user has to rely on a 3rd party to stay up to date with their plugins, or to even implement features that the user may need. How many motion controllers can be considered fully functional at this time, supporting all of Mach4's features? Remember, Mach4 was released over 3 years ago.
With the competiton's products, the hardware and software development is always in sync. New features work immediately, and you don't need to figure out who's at fault when something doesn't work.
I'm one of the people that has moved on to other software, although my old machine still runs Mach3.
I was an early purchaser of Mach4, over 3 years ago. I watched the (slow) development for over a year, before deciding it was time to move on.
In the same way that different people prefer different CAD and CAM programs, they also prefer different control software. Because Mach4 is so very different from Mach3, many of the tens of thousands of Mach3 users find that it's just too different from what they are used to.
-
This is just my 2 cents.
When setting up a new machine there are so many different things to set up. Having to learn LUA on top of all this just really sucks. I can understand more custom stuff but i think set up for more basic stuff should be a lot easier.
I have stuck with Mach 4 and i like it now but i think having to learn LUA is just way too much.
Not putting down the Mach 4 guys because i realise how much work is involved in creating brand new software so i thank the team for this.
-
Remember, the vast majority (95%+?) of users are not programmers, and have no desire to become programmers. They just want to run their machines.
For every Mach4 users that likes it, there are probably 25 people that aren't going to bother learning Lua, which they shouldn't have to do.
They basically eliminated 90% of their potential user base.
-
Hi Gerry,
good straight talking stuff...I like it.
You are correct very few CNCers are programmers or want to be. As it stands, effectively, you cannot use Mach4
without at least a little programming. As you also point out given the third party hardware model, set up is far from
straight forward counting against widespread adoption by hobbyists.
That I stuck with it until I became familiar enough to start programming, which I might add is VERY satisfying, is more
about my own interest and skillset than being representative. With that said I now have a very flexible CNC solution
but recognise that many may never arrive at that situation.
That brings me back to the question I posed that started this thread 'Is Mach4 really a hobby grade material?'.
If it is accepted that hobbyists are not by and large programmers then the answer is no.
If you are prepared to learn Lua/Mach/wxWidgets then the answer is yes.
Craig
-
Hmm....this may turn out to be a very interesting thread.
I am thinking that I'll just waite for Russ to figure Mach 4 all out and then can just call him to bring me up to speed in short order. >:D
RICH
-
Graig,
Can you provide a wild ask guess on long it may take a current Mach 3 user to:
- Transition from Mach3 mill and lathe to Mach 4 mill and lathe and be capable with both?
- Learn to create say a rather complex screen page?
- Be able to do the Lua coding / programing as compared to what was used for Mach 3?
Just some wild ask time frames,
RICH
-
Hi Rich,
I can't answer the lathe question, I don't use either Mach3 or Mach4.
The initial set-up problems, minor programming tweaks like Russ is dealing with at the moment, I will guess take
20 odd hours, in fits and starts over a couple of weeks, as least that's what I did.
I would estimate another 100 hours or so studying before Lua starts to make sense and probably another 100 hours
or so to get comfortable with the modular structure of Mach. How soon that can be achieved depends on how dedicated
you are. It took me about three months to assimilate all that info. While its still only scratching the surface I found by
that time I was deriving immense satisfaction from the pursuit. While its not the same as making parts, if you consider
the value of a hobby to be what you learn in its pursuit the 'learning' Mach has been good for me.
200-300 hours solid learning is not an insurmountable investment in time but neither is it insubstantial.
Craig
-
Hi Rich,
by the way with the screen editor built in making your own screens is pretty straight forward and easy.
Craig
-
I think the first thing to do is define Hobby. I will define them as
Group 1: Enjoys building the machines.
Group 2: Only interest in building a machine is that their other hobby (or business) requires a CNC and they assume that building a machine is cheaper than buying one.
Group 1 will choose Mach4 over Mach3 hands down I think. They will research, read manuals, etc.
Group 2 will be upset if they have to take the time to learn anything. After all, this is taking time form their true hobby.
Folks are passionate about their hobbies. Its only the stuff they consider work they want to skip through........ so they can get back to their hobby. How many times have you seen folks ask questions that are covered in the manuals, posted in the forum, covered in videos, etc.? I'm not talking about different questions either. I'm talking about the same question answered in all of those places and then some. If group 2 were considered hobbyist then HotRod, Fast Ford, Golf Digest and a plethora of other magazines would have gone out of business years ago. After all they earn a living selling publications related to hobbies. Mach provides some pretty decent manuals for free but many can't be bothered reading a manual, watching a video or reading a post pertaining to their hobby. Nope, it just doesn't add up. On what planet is it ok to ask others to put more effort into their hobby than than they themselves will.
I'm firmly in group 1. I love making machines run good, faster, more efficient, more intuitive, nicer interfaces, etc. That's what drives me. I have zero interest in running them after that though. And I remember very well what brought me to Mach. I was doing CNC maintenance for years. I asked reps. of various controller manufacturers every so often how much the controls would cost to build or retrofit a machine for myself. Back then the answer was a pretty consistent $10,000 an axis. Well, that was well beyond my reach. Then a friend told me about Mach3 and I have been hooked since. It didn't cost me $30,000 but I have been learning for almost 15 years. And if I used any thing other than what I know already, that would be another learning curve. And Mach3 to Mach4 is the same. Its different than 3 but leaps and bounds better IMO.
If your in group 2 I will say that you can Do It Yourself and potentially save a lot of money. This is exactly what Hobby is for. But it will cost in time (and that is true for Mach3, Mach4, or any other). You must realize that Mach is software. It is not smart enough to change the way it works to suite you. You will have to be smart enough to change the way you work to suit it (again, true for any controller until we have AI). However, if you find yourself in a position that you just cant change the way your doing something more often than not Mach can be taught (through customization) to do it your way. Mach4s capacity to meet these situations are vastly improved over Mach3s. But customization requires more effort, time and/or money depending on how you go about it. You have to weigh it all out. I encourage people to build their own system myself but it isn't for everyone.
The best I can tell, most newcomers like Mach4 much better than Mach3.
-
And yes, I agree with Russ. The most basic stuff should (and to my knowledge will be) shored up. But I will also say.......
expected to respond as expected
is relative. Depends a lot on where the expectations come from.
-
Hi Chaoticone,
I'm group 1 and proud of it...all group 1's unite...all go insane together and beat the rush! LOL
As you say genuine CNC enthusiasts will eventually come to Mach4 being so much more capable eventually.
The group 2 types are very likely to hate Mach4 as Mach4 definitely requires some study and application.
For these individuals having all the most basic stuff work seamlessly is important. In this regard I believe that NFS
are guilty of not making sure that their experience is as trouble free as possible. Various forum contributors
have made available solutions to many of the simplest Mach4 shortcomings and yet NFS have not incorperated those
solutions into later builds. Likewise having the documentation fall behind the actual state of Mach4 is causing some
to lose interest.
I understand that NFS need to concentrate on those matters which result in OEM and Industrial sales but to leave
a chunk of Hobby users adrift invites less uptake which may have long term consequences. NFS is a private business
and it is certainly not my place to cast aspersions on the business plans of others. Please accept my apologies if it
appears I trying to do just that. My needs are best met by NFS prospering and continuing to develop market
leading products, widespread uptake of Mach4 by hobby users seems to me to be a worthwhile business goal.
Craig
-
Lol, the crazy train left the station a long time ago........
I agree it needs to be as smooth as possible. But again, that is very dependent on several factors. NFS can't carry the weight of all of it. There is a very big difference in something not working, and not working as expected too. A lot of times i see issues with the expectation. Not all the time but lots of times it doesn't work as expected but works exactly as it was meant to and is verified by the docs.
Docs will always lag behind. no way around it but needs to be minimized as much as possible.
Various forum contributors
have made available solutions to many of the simplest Mach4 shortcomings
Where at? If someone wants to make a list and get it to me, I'll have a look. Can't make any promises but will look and go from there.
-
Hi,
your right there is a big difference between not working and not working as expected. I have seen others fall into that trap and have done so myself.
Would you consider small additions to the API notes. Often the notes are so short and terse that they are not as meaningful as they could be.
I had occasion to raise a support ticket because I couldn't get an API to run (mc.SignalWait) to find that there is a restriction on the signal ID. My query was answered
but no addition was made to the notes on that particular API and I find that several months later I'm answering the same question on the forum from some other
frustrated user. Had a simple addition been made to the notes at the time the problem would not have recurred. I think RT has a similar example, a one keystroke edit
would save future users some frustration.
We are lead to believe that Mach4 is evolving yet very simple additions and tweaks are not finding their way into later builds.
Craig
-
Hi,
I didn't run m4 for about a year. Before I loved to work on my own screen, learned lua and created extensions ...
Now I installed the most recent hobby-version and m4 comes up, but as soon as I change to edit mode, I see a crash.
As my message stayed without feedback, I thought: may be it crashes only with my screen.
... and yes: with the provided screens I can enter edit mode, but with my selfbuilt screen it crashes.
So I looked for hints about changes, I have to care about, but there's no history file, no readme and no caution.txt ...
Seems as if I have to throw away my work and start over again. That's not really attractive.
From my point of view, I expect professional software to offer an upgrade-path like the app is able to read an older screen format and writes the current version.
I don't see any way to find out, what tears m4 down - even if I compare both xml-files with a diff-viewer.
So for me it looks like m4 is developed for industrial customers and hobby-users have to live with the giveaways :(
Reinhard
-
Hi Reinhard,
as far as I can tell you can't read a screen .set file, its compiled and so there is no way to compare the files side by side to see which one differs.
I to am using the latest build of Mach but the screenset is a personalised copy of wx4.set from a build about 14-15 months ago and it works with the current
build. I would guess therefore that your screenset is corrupt rather than incompatible.
I am intrigued... what is it about this situation that causes you to blame the difference between Hobby and Industrial? If a file is corrupt surely it will be corrupt
and unusable by both?
Craig
-
Hi Craig,
as far as I can tell you can't read a screen .set file
the screen-file is just a zip archive with the screen.xml at root level and a directory containing the images. It can be read by all tools, that can handle zip archives and there's no problem, change anything inside the screen file. On linux systems, the file will be recognized as zip-archive without the need of renaming it.
I would guess therefore that your screenset is corrupt rather than incompatible.
No, I'm really sure, that the screen file is not corrupt!
So there has to be some incompatibility. But I don't have the tools to figure it out.
what is it about this situation that causes you to blame the difference between Hobby and Industrial?
I reported several bugs here in forum, as well as at support-mail - the time I dedicated at m4 none of my issues has been adressed by an update and I never got feedback from support-mail. I wanted to check the most recent version to see, whether any issue has been resolved, but if my whole work is rubbish now, I'm not very attracted to test any more with m4.
Additionally I translated the german files and sent them at support mail - but those translations did not find the way back in hobby-installation.
XP is said to be dead and after awaiting a whole year, there's stil no linux-version arising ...
So I'm a bit disappointed.
Reinhard
-
Hi Reinhard,
you know more about the screen file than I do. I can tell you that I'm using the latest build of Mach4 3481 but using a copy of wx4 from build 3233....I think, certainly
the build current as of October 2016. The evidence I have suggests that its not an incompatibility issue.
I have considerable sympathy where you have provided information to NFS and that has not been reflected in the later builds. You may have noted that I have commented
about that in this thread. I would have to agree that NFS have not made the most of the solutions that have come from hobby users. It may be that because of finacial
imperatives that NFS is obligated to ignore the hobby market I believe however the long term effect of that policy will be that hobby users will desert Mach.
Can you find an earlier build of Mach4? I seem to recall that PMDX maintained an archive of early releases.
Craig
-
When Mach4 version 2 was introduced it made almost everything users developed under the original version useless. Is this what is being described here ?
Tweakie.
-
Hi Craig,
Sorry, I should have mentioned the build numbers :(
I worked on my screenset with 3390 and the most recent build, I have is 3633.
Do you have lua extensions in your screen and did you use functions from a selfwritten module file?
What do you mean with NFS? For me NFS stands for network file system and I don't know, how that could be related to m4
Can you find an earlier build of Mach4?
I have several versions, but not a complete archive. Why do you ask for earlier versions?
... effect of that policy will be that hobby users will desert Mach.
Yes, that's very poor. I know about several people, that don't want to try m4 - some of them stay with m3, but lot of them are looking for other solutions. Lot of potential hobby users don't understand and don't estimate the fanuc compatibility. And from what I can see - there's no affordable motion controller for power-users (like pci/pcie-mesa cards for linuxcnc) that works with m4.
Reinhard
-
NFS = Newfangled Solutions = Artsoft
-
When Mach4 version 2 was introduced it made almost everything users developed under the original version useless. Is this what is being described here ?
Wow, good thing I didn't know that, and I didn't develop anything under Version 1, never used version 1 other than to look at it. Now that I'm working on stuff in Version 2, when Version 3 comes out will "almost everything" I've done be rendered useless or have things stabilized to where we'll be good moving forward? If a new version is likely to break everything I've done, I'm probably out of here right now!!
-
when Version 3 comes out will "almost everything" I've done be rendered useless
I don't think so but if I knew the answer to that I would be buying a lottery ticket. :D
Tweakie
-
Hello, django013, please send in in a support ticket with your screen set and we will help figure out what is preventing you from being able to load run or edit it in later versions of Mach4. Please include a copy of your profile that you can get by clicking in the help then support tabs and select package current profile. Best regards Jim.
http://support.machsupport.com/
-
A Lot has been said since yesterday.
I think Craig and Chaoticone have struck a huge chord; users who EXPECT Mach4 to behave a certain way will get frustrated and leave. I personally think certain aspects of Mach4 should behave in a certain way, but I understand they don't and I find a work around. Some hobbyist just won't put the time in to fix a button or two.
The first machine we used Mach4 on was a Profile Knife Grinder, it took a long time to figure out LUA and Mach4 to get it working how we wanted it to, and to be honest it was a lot of bells and whistles over just getting it to work. However, the next machine I used Mach4 on a retrofit it took 1/8 of that time. I made a folder with all the custom code and button scripts to aid future retrofits. For us this is beneficial, but how many hobbyists will have 6 CNC Machines (and counting) to set up.
Chaoticone,
I find myself in Group 1 and Group 2.
We are not hobbyists when it comes to Mach4, we are a machine shop that make wood cutting tools, so we dig in and make Mach4 listen to us and has proved to be very worth while.
We just bought a Haas VF4 to up our capacity and some of the features on the Haas are incredible. What I love about Mach4 is the fact that I can take the features from the Haas control (which is awesome IMHO) and integrate them into future retrofits.
But getting back on topic. Is Mach4 really hobby material?
I agree that NFS could update the API more frequently to aid other users, I think certain configuration settings should be improved to be more plug and play (external buttons, scripts, etc), and the folder that has example macros and scripts should be updated more often.
The example folder can be filled with examples from this forum. If I had such a comprehensive source of information (a more extensive example folder) Most of my questions starting out would have been answered. I think Mach4 has great hobby potential, I just needs to be a little more streamlined.
-
I'm not aware of any changes between version 1 and 2 that would break others screens or profiles. The default screens and profiles were drastically improved though.
I think the problem was when some modified the default screens or profiles for their own use. Why they would do that, I have no idea. Anyway, when they updated the defaults were updated. It wasn't that the update broke anything of theirs it just updated the defaults. If the defaults were what they chose to use for their own then yeah, they would be overwritten. The fix is to use loader to copy a default profile and create a custom profile and configure that custom profile for your use. But that is covered in the manual so I'm sure no one reading this did not already know that.
When customizing, always start by copying and saving the closest matching default as something custom. This applies to everything you customize in Mach. This is also more often than not a good computing practice. If this bit you I'm sorry to hear that. But remember, it is a lesson you learned that is advancing you in your hobby. Also be careful who you blame for what you didn't know.
Another good practice when customizing is to use Mach4s constants (listed in the API doc). These are there to facilitate updates to the core and not break custom scripts (provided they are written correctly of course). But if folks doing customizing don't use them, updates to the core may break their scripts. But the tools are there to prevent this. I have been bit by this more than once myself (old habits die hard). I just had to ask myself...... "do I get mad at myself for not using the tool or do I get mad at the guys fixing bugs, adding features, adding tools etc.?".
There is yet another step you can take to protect your customization even more if you want to. Put your custom scripts in a custom lua module. The tool is there.
Now the question is, do these things (and lots of things not listed here) make Mach4 less hobby friendly or more? I say more but some also make it more different (if taken advantage of). But, the ones that make it more different (better IMO) don't have to be used. The ones that don't make it better are pretty much the exact same. Pretty much everything I compare from Mach3 to Mach4 ends with this same conclusion.
-
Hi Reinhard,
all my customisations are Lua extensions, I haven't written any modules.
From your post it sounds like you have written modules and other than Lua?
Craig
-
Also be careful who you blame for what you didn't know.
Not sure if this was aimed at me, but I didn't blame anybody. I'd never heard this thing about Mach4 version 2 breaking stuff, it was new to me. I never used v1 I guess.
That said, and not that it matters now, but what build, or version number, did Mach4 v2 start with and Mach4 v1 end with, and when? Just curious, not blaming anyone for not knowing:)
Tom
-
Having Mach4 so you can run any machine type all in one, would there be conflicts with the G and M codes macros and scripts if you were saying useing it to run a lathe would the built in G and M codes macros whatever for a Mill or a plasma cutter interfere with it's operation of the lathe.
-
Graig,
Thanks for the time estimates since they give a "rough idea" on what a convert to Mach 4
may need to invest. My only comment, and only a guess, would be that a "newbie" may add another 100 hours.
RICH
-
Not sure if this was aimed at me, but I didn't blame anybody
Nope, not aimed at you or anyone else in particular. The same thing I have told my sons when they do something and in hind sight realize they could have done things differently and saved some time, money, effort, etc. They are usually mad at their self. But no reason to be mad at their self or anyone else. They didn't know what they didn't know and no one who knew it had any idea they didn't know or needed to know. Bottom line is everyone is responsible for knowing what it is they need to know to do what they want. If they want it bad enough, they will learn everything they have to to get to where they want to be. It's just life.
-
Is Mach4 really Hobby Material?
The question is rather thought provoking. A hobby is something one does for personal pleasure for whatever reasons and the reason need not make any sense to anybody except the one enjoying the hobby. The level pursued, time spent enjoying it, choice of activities associated with the hobby provides an arena to satsify the individual. So the answer is yes..... Mach 4 is hobby material and is just one part of CNC.
Personally can't get emotional about Mach 4 because of how I view it.
RICH
-
The question is rather thought provoking. A hobby is something one does for personal pleasure for whatever reasons and the reason need not make any sense to anybody except the one enjoying the hobby. The level pursued, time spent enjoying it, choice of activities associated with the hobby provides an arena to satsify the individual. So the answer is yes..... Mach 4 is hobby material and is just one part of CNC.
Exactly Rich. That's it in a nut shell. I would like to emphasise ONE PART.
-
I just read this thread and I must say I'm certainly confused in a lot of ways. During the days of Mach 3, we compiled a list of things that people were asking for that Mach 3 simply could never do. That list became Mach 4. So Mach 4 was the direct result of us listening to what our users wanted! But now they don't want it? Thus my confused state of mind.
Ii used to be, or I used to think, that hobbyist had simple machines to do simple things. 3 axis mill without a tool changer, for example. But no. Just take a look at the feature request thread!!!! 90% of that thread is people wanting Mach 4 to do something special. But also, 90% of that can be accomplished with using the tools that are already provided with Mach 4!
So there is a strange dichotomy going on here. People want shrink wrapped software simplicity but also want it to do something very special that THEY want it to do. So I don't buy the "Mach 4 won't work without a lot of programming" complaint. Mach 4 will run my Matsuura MC500 "out of the box" with absolutely no LUA programming! Just by mapping signals to I/O and setting up the motors. All of this is done in the Mach configuration dialog. The only thing that requires LUA is the tool changer. It required VB in Mach 3. And since no two tool changers are the same (assuming some are built, some are converted, etc..), this becomes one of the "special things". But I could run that machine with manual tool changes with Mach 4 "out of the box" all day long. In fact, I don't use the tool changer all that often with the things I do.
"But I have brand 'X' joystick that I want to use..." Guess what? You are wanting something special.
"But I want to touch off a tool at a certain location on my custom built machine, set the height, and restart with one button. And I might have to learn LUA?" The answer is yes!!! The good news is that you can make that work. The bad news is that we will never be able to provide that "out of the box".
I could go on and on, but I digress.
As to the hobbyist definition... If you end up buying a Centroid controller, are you still a hobbyist? A point to debate. I'm a ham radio guy. That is another one of my "hobbies". But you know what? Now I find myself collecting expensive test gear like signal generators and spectrum analyzers. I like fixing radios. I no longer consider it a hobby and I now consider it as a semi-profession venture. I make money fixing radios. It is no longer something I do just for myself. The tools I use are something far greater than what a normal ham radio hobbyist uses. BTW, I wish ANY ONE PIECE of that test gear was as affordable as Mach 4. So ask yourself the question "Am I truly still a hobbyist?", and be honest.
Also, I have a 3 axis X2 machine running Mach 4 and ESS. It runs Mach 4 out of the box without any LUA code at all. Now THAT is THE quintessential hobby machine. So yeah, I think Mach 4 is hobby material. But isn't it nice that Mach 4 will also run my Matsuura production type machine? Make no mistake, I have no illusions about the differences between those two machines. One is a hobby machine and the other could be put to work in any professional machine shop. One didn't required a bit of LUA and the other did.
Steve
-
Dude1
This thread was probably not the best one to post your question ;D
If I understand you correctly you're talking about using Mach4 to run different operations (mill, turn) on one machine. You can using different profiles.
Mach4 has different subdirectories for each 'Profile' you create and each profile has a 'macros' subdirectory that only its profile accesses. Your m4Lathe profile will only use the macros within the m4Lathe\macros directory, etc.
If your trying to do combine mill and lathe operations using the same profile then there can only be on m2000 macro.
HTH
RT
-
rhtuttle
More than likely It's just a question I have always wondered if there is a potential problem.
To add Mach4 to a basic 3 axis mill is easy if you have good clear instructions that come with the external devices.
-
Hi All,
the purpose of this thread was to 'take the pulse' as it were of where Mach4 is at.
I think Chatoicone has summed it up extremely well, there are largely two groups of Mach4 users, the first
group being CNC enthusiasts whom find Mach4 a good fit. The other group are those who consider CNC as a means
to an ends, be it business or another hobby...motorcycles or models or....For this group of users they wish Mach4
to work their machine with the least hassles.
Given the wide variety of hardware out there and the apparent wish for users to have individualised tweaks then
Mach4 can certainly do it but perhaps rather unfairly cops a bit of flack.
New users should really consider what sort of software solution they want...if they want least hassles then Mach can do
the business but as soon as you want to customise your machine the effort required goes up dramatically.
Craig
-
I think you are about as plain and accurate as possible Craig.
I would also like to point out that your statement..... "but as soon as you want to customise your machine the effort required goes up dramatically" is true for almost any I know of if customizing is even an option.
-
Just take a look at the feature request thread!!!! 90% of that thread is people wanting Mach 4 to do something special. But also, 90% of that can be accomplished with using the tools that are already provided with Mach 4!
Steve
Hello Steve.
I agree with your post. I'd say its fair. It addresses that 90% you mentioned.
I'd be in that other 10% ... maybe the lower 50% of it. (you can probably tell ::) ) ;D
Picture the guy with a homemade lathe made from surplus pallets, plywood, allthread and one E-Stop button.
I wouldn't expect him to have to do anything internally in order to have the program run as described.
Might sound a bit picky, but these little thing can really discourage and intimidate the lesser educated of us tinkerers. And they appear to be adding up.
I'm ok with it, I can muddle through with the help of the forum and you guys. I'll buy it.
I've learned to expect the unexpected. Just sayin'.
If there was a button in my new self-driving car that said "Take me to Church", I wouldn't expect it to deliver me to the whorehouse. :)
Russ
-
Hi Russ,
I wasn't aware that the destinations differed....LOL
Craig
-
This discussion reminds me of a conversation I had with a customer who bought one of the robotic greenhouse watering/spraying machines that I used to build. He calls up all pissed off because the machine crashed into the end wall, got the boom all twisted up in the motorized vent louvers, and continued spraying water damaging a lot of plants. I asked him what the machine said, because our machines had voice synthesizers and would tell you exactly what the problem was. I asked did it say "Collision, Boom twisted' or anything like that? He said no and I thought about it for a moment. Then I said "YOU were driving, you had it in manual and weren't watching it. Why didn't you run it in auto?" He "Oh I haven't had time to learn how to do that!" I hung up on him.
Two months later he comes into my booth at a trade show and says "Well, that was embarassing!" I said "You mean the part where all you had to do was flip the switch to auto and press start and the machine tells you how to enter a speed so it knows how much water to apply then does what you ask with no crashes?" He says "Yeah that part too." So this machine was capable of doing useful work right out of the box despite the fact it was extremely programmable for those who were interested.
-
If there was a button in my new self-driving car that said "Take me to Church", I wouldn't expect it to deliver me to the whorehouse. Smiley
I wouldn't expect it to but if you ever get one that does just that let me know. I want it. It is obviously capable of reading minds. ;D
-
Steve,
-. -.-- ...-- .--. , yep, I certainly know about other hobby costs. That's one of the things that all hobbies have, namely, cost.
Cost is just a matter of degree one wants to expend on his hobby and everybody is different. What is expensive to one person is peanuts to another. The farmers barn is sided and roofed with Al but the house siding is in shambles. I know folks who spend more on an accurate gun / sniper rifle than some pay for a car. Heck, some youngsters spend more on one mag wheel for their junk car than the cost to convert to Mach4. The neighbor bought a $3k planer for his wood working hobby and the bird houses he makes are not exactly flying off the shelves in his shop. All depends on what one wants and once you decide you have to pay for it.
COST- if your going to convert to Mach4 one will need a license, most likely an external motion device and maybe a
breakout board .....so $500 to play in the new arena.
If one does not want to spend $500 then don't upgrade to Mach 4. Period!
Only the individual can rationalize it out.
I like to think of Mach4 similar to what an undertaker once told me ............... there is time and we are patient, but, sooner or later you all come to us, maybe not me, but the services will be needed! ::)
So much for COST.
RICH
-
Hi Russ,
I wasn't aware that the destinations differed....LOL
Craig
Well, there ya go. Just like Mach4, that would probably be perfect ...... for some users. :)
I stand corrected.
Nevermind ;D
-
If there was a button in my new self-driving car that said "Take me to Church", I wouldn't expect it to deliver me to the whorehouse. Smiley
I wouldn't expect it to but if you ever get one that does just that let me know. I want it. It is obviously capable of reading minds. ;D
You're probably right Brett. I had the preacher with me. ;D
-
Rich,
-. -.-- ...-- .--. , .-- - ----- -.- I'm sure you do know about hobby costs! $500 to play in the ham world is chump. And I'm also sure you know that we are super nerds, right? :)
I use my CNC machine to help build solid state amplifiers. Check out my QRZ page to see it. The copper heat spreader has the MOSFET holes machined into it. And of course, the case was CNC'd as well. CNC and ham radio go well together. I use my little desk CNC machine to make PCBs on occasion.
People that are just in the CNC hobby are lucky and don't have to take tests and get licensed.
Steve
-
You're probably right Brett. I had the preacher with me. ;D
I got to go on that one. ;D
-
Hi,
So Mach 4 was the direct result of us listening to what our users wanted! But now they don't want it? Thus my confused state of mind.
Ii used to be, or I used to think, that hobbyist had simple machines to do simple things. 3 axis mill without a tool changer, for example. But no. Just take a look at the feature request thread!!!! 90% of that thread is people wanting Mach 4 to do something special. But also, 90% of that can be accomplished with using the tools that are already provided with Mach 4!
I completely agree with Steve: mach4 is feature-rich and powerful enuf to drive any hobby machine.
... and I agree on him too, that the one who makes money with his cnc is no more a hobby user.
For me, m4 is well suited to drive any special machine beyond hobby interests too.
... and I don't have any problem in having to learn m4 programming and a new programming language to solve special needs.
Even more - I don't care, if api or fileformat changes - as long as there exist an upgrade path, with reasonable costs and instructions on how to reach current api level.
What I miss most is communication from artsoft to hobby-users. Let me tell, what I thought during last year:
This time last year I was convinced, that m4 will be my way. But then I became unsure, as I didn't see any reaction on bug reports. I followed the updates, but could not see any issue solved.
Then from august until january this year there was no update - and I thought: well, they might be working on other products or for industrial customers ...
When the last update came out, it was clear: there has been folks working on m4. Nearly 200 commits is a bunch of changes.
But where can I read what has been done?
The installation does not contain any information about that.
So from my point of view, best way to bring hobby-users back to artsoft and document the commitment of artsoft to hobby-users would be in establishing a ticket system public-readable like this forum. This way, hobbyists can enter wishes and bug-report and developers can answer by setting issue-priority (like "we're working on it" or "closed. Not our business").
Hobbists that can search the tickets and see, what happened to reported bugs (resolved with build ####, need more information).
I think, that will create transparency and feedback with little cost to developers and it could be a way, to receive user-submissions like translated files or the like.
The ticket system will allow to create some statistics and history without the need to have a developer write a novel.
I think, a ticket system will create communication from artsoft to hobby-users at little to no extra costs for the developers.
Reinhard
-
Rienhold, we have a change log in the Mach4 download tab on our website that covers changes in the latest releases. http://www.machsupport.com/wp-content/uploads/2014/04/Change%20log%203481.pdf
We created the ticketing system and have added staff here at Artsoft specifically to work with hobby users. Our tech support team responds to and works with hobby users every day. Tuesday morning alone of this week I assisted 3 users in getting their machines up and running with Mach4 and a compatible motion controller right out of the box. No custom scripting required. In fact we have profiles for the common 3020 6040 table top CNC machines that we share with users and literally they are up and running within 20-30 minutes depending on their ability and knowledge.
We created the help desk to better track and ensure that tickets are responded to and users are helped. Some folks do not want to post publicly what they have for questions, and to be honest I don't think that you would want to read 10 posts a day about how to license the software or what OS the software can run on. ;D Some users when they get answers post them on the forums some do not. We have kept the forums open for those that want to share publicly what they are doing or working on and that is the individuals choice to make and we respect that choice. We are providing support options for users in addition to forums rather then instead of the forums and in the end its the comfort level and choice of the end user. We will continue to add how to videos to our youtube channel so if there are specific things that you would like to see please let us know! Best regards, Jim.
-
Hi Jim,
we have a change log in the Mach4 download tab
Yeah - pretty outdated and it looks like a developer has to sit down and write a novel :(
Such things could be automated from ticket system so its a one-click operation at build time.
... and it would be nice, if you package that file into the installer, so any user can read it after installation.
Tuesday morning alone of this week I assisted 3 users in getting their machines up and running with Mach4 and a compatible motion controller right out of the box. No custom scripting required.
Sure you do!
But who knows about it?
In germany we have a saying: do good things and talk about it
I don't doubt, that your doing good - but I miss your talking about it ;)
A whole year without any change (yes, I realized the blog entry about lua) on your website is a very long time.
We created the help desk to better track and ensure that tickets are responded to and users are helped. Some folks do not want to post publicly what they have for questions, and to be honest I don't think that you would want to read 10 posts a day about how to license the software or what OS the software can run on.
You're right.
But beside that - look at the forum thread like "mach 4 feature request" and "mach 4 bug reports" ...
They are dead, unsearchable and nobody cares about it =:O
I believe, that a ticket system as used by open source hosters will reduce your work (respect to feedback) and rise your reputation for hobby-users.
Well my 2 cent ;)
Reinhard
-
Thank you for the feed back and you are correct that we need to get the word out that we are here to help!
So getting to your screen set, thank you for sending that in. I have loaded it here in 3 different Mach4 builds and it all opens and I can edit it as you can see in the response and screenshots that I sent you on your ticket. Run dependency walker on your XP machine and see if there is a file that the later build needs I believe that is your issue as I am able to edit it with out crashing or having any error messages here in Win10.
Best regards, Jim
-
Hi Jim,
thank you for your support and the hint to dependency walker.
There are 2 dlls missing: ieshims.dll and wer.dll - wich looks like a well known problem (https://social.technet.microsoft.com/Forums/windows/de-DE/8a751f65-ade9-4b8b-a3d3-c720ccbd3d2c/ieshimsdll-missing?forum=w7itproinstall) without solution.
I'll try to find that dlls somewhere, but if I don't find them, I'm done with m4 :(
Reinhard
-
Mach4 uses c++ and that XP os that you are running is not even supported by Microsoft any longer. You can find the C++ redistributable from Microsoft and that will most likely address your issue. Trying run the latest software versions on a 15 years old un-supported OS is going to come at a price I am afraid. Best regards Jim.
https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads
-
Thanks again, Jim!
the dll's from technet discussion did not work and the redistributable C++-runtime did not solve it either.
So I think, that's (finally?) the end of XP - and I'm very sad to have to say good bye to m4.
So whenever you decide to start the linux trail - please drop me a mail and I would be happy in helping.
best regards
Reinhard
-
Reinhard, there is a development version on the FTP that may work better for you if your interested in trying. Version 4.2.0.3643. If you do, let me know how it goes. The only way i will ever know how something works in XP again is to hear about it.
-
there is a development version on the FTP that may work better for you if your interested in trying. Version 4.2.0.3643
Thank you so much! You're my hero :)
Looks like 3643 works again (as 3485 did before). So only 3633 fails on XP
Well, don't take me wrong - I only used m4 with the screen editor as I don't have a working cnc yet. So I can't say if it will fail on real work. But at the moment I'm happy, that I can continue working on my screens :)
Reinhard
-
there is a development version on the FTP that may work better for you if your interested in trying. Version 4.2.0.3643
Thank you so much! You're my hero :)
Looks like 3643 works again (as 3485 did before). So only 3633 fails on XP
Well, don't take me wrong - I only used m4 with the screen editor as I don't have a working cnc yet. So I can't say if it will fail on real work. But at the moment I'm happy, that I can continue working on my screens :)
Reinhard
:)
-
Hi,
I see a number of CNC users that do not think Mach4 is for them.
Yes there are still some outstanding issues like A axis not being able to maintain surface speed as the diameter controlled by the Z axis changes like Mach3 does.
Also Mach4 still cannot correctly display Toolpaths for 4/5 Axis simultaneous G-code as well as simple cylinders correctly like Mach3.
But both of these issues can be taken care of by the Mach4 programmers.
Apart from these issues Mach4 is an improvement over Mach3 in many ways especially machining very fine detail with small cutters down to a .05mm Rad tapered without any cutter damage using Hard brass.
Mach4 V3481 and HiCON are rock solid and can machine all 3 Axis G-codes perfectly, I have not tried 4 Axis as we need the above 2 issues included into Mach4.
We use Rhino/madCAM and Aspire to generate the G-code and both have no issues with Mach4.
Yes you may have to do a little Lua code to make it suit your choice of Controller Card and MPG, but that is no different to what you had to do with Mach3.
Both HiCON and ESS cards are in many cases a lot easier to set up in Mach4 than Mach3.
I cannot speak about the other controller cards as I only use these two.
Mach4 V3481 works fine with ESS and XP with 2GB Ram and with Windows 10 using HiCON.
Regards,
Mauri.
-
Hi Guys,
Just to add fuel to the fire, back in the Mach3 world you would have had to either use VB script (remember all the problems???) or Brains (Ever try to edit a long Brain???) to accomplish anything that was not included with Mach 3.
In Mach 4 you use Lua and the PMC. I don't hear a lot about people using the PMC but I just was able to connect 6 physical buttons on my lathe retrofit without know how to use LUA at all. LOVE IT!!!!!!!!!!!!!!!!!
It is a ladder logic program that compiles LUA script in the background.
My opinion is that Mach4 for a hobbyist is fine especially when you want to run a 3 axis machine or a 2 axis lathe. The hardest part for me was going from Mach3 to Mach4 I had all kinds of preconceived ideas about how it "should" work as opposed as to how it does work. Yes it has some idiosyncrasies but please don't tell me that Mach3 and LinuxCNC don't.
Mach3 has made a lot of parts for me over the years and I learned a lot making it do the things I want. The two major problems with Mach3 is that VB gets out of sync once in a while and I have broken my probe a couple of time. Problem two is that I get .XML file corruption after a few months of hard use and have to restore Mach3 from a backup. Once the space bar stops working as feedhold it is time to restore it. What else has stopped working properly that I don't know about in the mean time??????????
I have used LinuxCNC in the past and I am actually putting it on a machine as we speak because I don't have the electronics to do what I want in Mach4 and I currently do have all the electronics to do it in LinuxCNC.
I however don't like the look and "feel" of it but sometimes you gotta do what you gotta do.
I did however leave my self the option to upgrade to Mach4 in the future since I want to standardize on one control for all my machines.
I consider myself a high end hobbyist in that I have some large pieces of iron (quite a few of them actually) which are beyond what most would consider hobby size machines.
I am converting all my machines to Mach4 from Mach3 just because it works better for me.
Enough rambling! Time for more coffee and getting to work.
-
In Mach 4 you use Lua and the PMC. I don't hear a lot about people using the PMC but I just was able to connect 6 physical buttons on my lathe retrofit without know how to use LUA at all. LOVE IT!!!!!!!!!!!!!!!!!
Good on you Mike, that's awesome. Also a very good point I did not mention either. But you also either already knew or had to learn ladder logic. The question for the purpose of this topic is......... (and this is for everyone reading that wants to play along, not necessarily you Mike).
I know better but for arguments sake lets say you did not already know or understand ladder logic and had to take the time to learn it.
Would it be your fault for not already knowing ladder logic?
Would it be no ones fault and just part of the learning curve and even enjoyable as part of your choosen hobby?
Would it be Mach4s developers fault for providing the tool that gives you that option in Hobby?
If the answer you come to is option 3, let me know what options and/or features you think we should disable in Hobby. Or maybe we leave them in and it just helps us determine what questions to ask in a litmus test to determine if you really are a Hobby (realize doing your customization may take considerable amounts of your time and effort) or Industrial (wanting customization but don't have the time to do it myself but also willing to pay for it) user.
If Mach4 Hobby isn't considered Hobby material I'm sure we can make adjustments to fix that.
-
Hi,
Yes there are still some outstanding issues like A axis not being able to maintain surface speed as the diameter controlled by the Z axis changes like Mach3 does.
Hm, is this an mach4 issue? Shouldn't that be calculated by the cam process?
Mach4 V3481 and HiCON are rock solid and can machine all 3 Axis G-codes perfectly, I have not tried 4 Axis as we need the above 2 issues included into Mach4.
Hm, that's the point, where I became unsure.
I bought an ESS, but when I started configuration, I was astonished by the update-frequency of 40Hz, which already is 4 times faster than with mach3.
But ...
My cnc will be a 4-axis gantry, where I start with CL-steppers. The stepper-drivers handles the closed loop, but I'm interested in moving CL up to mach4. But I think, 40Hz is way to slow to start with CL.
Next doubt: I want feed-override in hardware - no matter whether to use potentionmeter or mpg - and I'm quite sure, that 40Hz is to slow for it too.
Therefore I wrote, that I miss decent motion controller, that ist affordable (for me less than 500 bucks).
But that's not an issue for artsoft - it's just coupled to the decision mach4 or not mach4, as any software requires its own motion controller and each is incompatible to the rest.
Reinhard
-
django013,
I have tested many CAM programs in the past.
I do not know of any CAM program that can output surface speed and the A Axis using the Z Axis as the radius to control the speed.
If you want a controller that can in all conditions you need to speed the $$, HiCON cards can do it all.
Regards,
Mauri.
-
Hi Mauri,
I do not know of any CAM program that can output surface speed and the A Axis using the Z Axis as the radius to control the speed.
I think, the radius in the picture is just a helper information for human beings.
CAM, as well as motion controller has a path with starting- and endpoint. Cutting edge of the drill bit is the Z-Axis, so CAM and/or motion controller need to calculate a path with possibly 4-axis moving. The surface speed is the speed working out that path. The path may be radius based, but need not be so.
If you want a controller that can in all conditions you need to speed the $$, HiCON cards can do it all.
Hm, if I get the picture right, Hicon does not raise closed loop to mach4. Closed loop works at motion controller level.
I guess, same would be possible with ESS too. Not sure, if that makes any difference to closed loop at stepper controller level.
Closed loop at application level may have an impact at path optimizer.
... and I did not see any special input for feed override.
So I'm not convinced, that Hicon offers better solution than ESS does. Hicon is just way more expensive.
Reinhard
-
Have a look at Pokeys
-
Hi Reinhard,
So I'm not convinced, that Hicon offers better solution than ESS does. Hicon is just way more expensive.
Reinhard
The Hiconn is not way more expensive.
With the ESS you need to add breakout boards. The Hiconn has them built in.
The ESS and most other motion controllers are 5v I/O. The hiconn is 24V.
The Hiconn now comes in two flavors, the 6 axis high speed version and the 4 axis 125 KHz hobby version for significantly less. Both of the Hiconn's feedhold almost instantly.
If you are building a machine that would have run fine of the PPort then the hobby version should work great for you. Mach3 peaked at 100Khz and most computers wouldn't go over 50 KHz.
The 6 axis version does 4 MHz IIRC. Great for digital servos.
I just picked up the Hiconn hobby version for my lathe and I am looking forward to single point threading as soon as I get my encoder hooked up. Not sure how many other boards can thread yet.
I also picked up a Pokeys 57CNC board and I am very impressed. It will be used to run a small machine that will have a laser on it at some point. The initial price was very fair but I will keep an eye on what breakout boards and other items needed to be added for full functionality.
-
Fwiw, you can add an MB2 breakout board to an ESS for $150 and have 24V I/O, for a little over 1/2 the price of the Hicon.
And the Hicon can get really expensive when you start adding on the options.
I've never used either, but from what I've read over the years, I think the Hicon is better supported, and a more "industrial" solution.
-
Already posted about cost of upgrading to Mach 4 and thought it appropriate to post about TIME.
The first thing that comes to my mind is when and if one should upgrade at all.
Only the individual can determine when, recall my undertaker comment, one will upgrade in
the future.
- If currently Mach 3 does everything you want it to do then no hurry
- If Mach 4 offers you something you need and desire that is lacking in Mach3 then maybe start the transition.
- If you are a newbie to CNC, then chances are you don't know what you really need, the learning time may be about the same, you
won't be biased by something you don't know about in the first place, Mach 4 may be the better choice.
Matters not if you are a hobbiest or commercial from a practical point of usage.
Some idea of time to convert was already posted and it is subjective. A few things come to mind as someone who only fooled with Mach 4 some three years ago. I'll keep track of of time when I start looking into Mach4 just out of curiuosity. But being a little wiser these days I will assume some things and accept some things as follows:
- Mach 4 hobby works for every command that is in the current manuals and accept the fact that there are differences in the
commands and they agree with books by Smid.
- The motion controller can provide for the instructions given it by Mach4 for every command. Understand that I just don't have any
tolerance to waite for "second party" to make their controller functional, said differentlt and candidly, have no tolerance for
waiting years for some command to be implemented!
- Documentation must be available and complete for the above two comments relative to a basic mill or lathe. Currently there are 10
Mach 4 documents available for Mach of which one probably needs to "study" 5 of them. Two are basicaly gcode programming
and the time to transition should be relatively short for me. The gcode manuals are not meant to teach you how to write / program gcode and if one deisres understanding surely you will find it in the Smid series of books. Smids CNC Handbook is 500 pages, so figure around 30-40 hours to read it, then another 120 hours to study it, or take a two courses at a tech school for say 20 weeks! The TIME required to learn gcode is just a matter of degree of the level required to the what you need to do. And yes, if you don't use it you will loose it, but the refresh time certainly is shorter.
- I would say allow an hour for each for the Mach 4 manuals to get a "flavor" and can't really say anything about an external controller manual. If the status of completness is not there then it could be a show stopper for me personaly at this time. Hmm ....the manual for Mach 3 mill took 6 months to do, and yes have done a few so I understand the investment of time for a manual.
BUT
After all these years Mach 4 should have good manuals and there is NO reason not to!
- If you are more advanced and will desire screen customization additional time frames come into play. Any one who has done a
custom screen, even a specific custom page using VB in Mach 3, and knew how to use a screen designer, knows the effort required.
I can only "assume" that the current screen designer is still rather intuitive and the transition won't take to much time. It will be just
like learning a new CAM program since you need to use the new tools that are provided.
SO
No use bitching about Lua ( I think it's still something they do in Hawaii ). :D
That's the tool necessary, and as Graig pointed out, one will need to understand Mach 4's structure. Thus the user will need to invest
time into knowing what they want to do, learn a new screen designer, study / learn the new structure, and lastly be able and willing
to transition to a new programming language. This is the one that reflect's back onto "Mach 3 doing what I want it to do". I
personally will need to invest a fair amount of time just getting Mach 4 Hobby to do replace what I can do in Mach 3.
I must say the following and mean no offense to anyone. :P
If one does not want to invest the TIME, good grief READ or STUDY, then please pursue a different hobby, since a few parts of CNC are NOT plug and play!
Just some thoughts about Mach 4, my other hobby is calling, :) ;)
RICH
-
G-code provides for sophisticated CAM programs to directly control feed rate for
non-Cartesian motion (rotary axes) by specifying the Time to complete the move
instead of the Feed rate. The code for this mode is G93 and the Rate is specified
as a number equal to 1 divided by the desired time to complete in minutes. This is
called "Inverse Time Mode".
While in G93 mode, every block that contains feed motion must have an F word,
I.E. feed rate is not modal while in G93 mode. For a simple rotary axis this places the
responsibility for compensating for Z radius effects on tool tip travel as the A axis
turns on the CAM program generating the code. It also allows for much more
complicated situations than just a simple rotary axis. The Mach3 software did
have a mode that would work fairly well for a simple rotary axis, but it was not
industry standard and is not implemented in Mach4.
Steve Stallings
PMDX
django013,
I have tested many CAM programs in the past.
I do not know of any CAM program that can output surface speed and the A Axis using the Z Axis as the radius to control the speed.
If you want a controller that can in all conditions you need to speed the $$, HiCON cards can do it all.
Regards,
Mauri.
-
Newbie question save me some time.
Is Mach4 lathe threading currently 100% functional as described in the manual for the hobby version?
and
What external controllers currently provide 100% functionality for lathe threading at this time?
Thanks,
RICH
-
Hi,
Mauri and I have had a conversation about how this might work.
My contention is that once a coordinated move is dispatched to the controller it is no longer possible to change the A axis rotation rate, it would break the move.
My suggestion was that the A axis be temporarily assigned as an out of band axis, Mach4 Hobby allows only one out of band axis and it is always the spindle.
Because the A axis is ansynchronous the speed of rotation can now be changed during a coordinated move of the other axes. In particular if the Z axis is equivelent
to radius then the speed of the A axis can be calculated and repeatedly updated by a short PLC script.
During this time however the main spindle is uncontrolled, it is a shortcoming of Mach4Hobby. Mach4Industrial allows up to six out of band axes.
Craig
-
Hi,
I have found a possible solution to the A axis rotary speed issue, but have not been able to make it work yet on my files, keep on getting an error.
"Implicit move command given before listing an actual move command (G01/2/3). (Line 4)"
Line 4 = N13G01 Z50.000 F300
The site below has a Jar program that can convert a G94 G-Code file into a G93 G-Code file.
https://www.ganotechnologies.com/cnc/rapidrotary/
If anyone can get it working let us know and provide a G-Code example.
Regards,
Mauri.
-
Hi,
The Hiconn is not way more expensive.
With the ESS you need to add breakout boards. The Hiconn has them built in.
The ESS and most other motion controllers are 5v I/O. The hiconn is 24V.
When I look at the closed loop option with more than 3 axis - then the Hiconn solution charges about 2k bucks, which is ten times an ESS.
With little knowledge of electronics a bob is not hard to build.
My contention is that once a coordinated move is dispatched to the controller it is no longer possible to change the A axis rotation rate, it would break the move.
Well, that probabely depends on what communication is possible betwenn mach4 and the motion controller - and of cause it depends on the communication refresh rate.
Because the A axis is ansynchronous the speed of rotation can now be changed during a coordinated move of the other axes. In particular if the Z axis is equivelent
to radius then the speed of the A axis can be calculated and repeatedly updated by a short PLC script.
Well, that requires a fast PLC thread, as well as fast communication between motion controller and mach4.
Don't know the interface speed of other motion controllers. ESS has 40Hz - which is quite slow for such operations.
Reinhard
-
Mach4 threading support is complete as described in the manual, BUT it is dependent
on the motion controller to implement threading features.
As of 22 January 2018, and as best I know, threading is fully implemented by
Hycon, PMDX SmartBOBs, and PoKeys.
Warp9 (Ethernet Smooth Stepper) is currently working on threading support
but it is not available for test yet.
CS Labs web site reports that they have a Beta release that supports threading.
An add-on option module is required for encoder interface. I am not aware of
any mention on the forum from anyone who has used it.
I thought CNC Drive had released a Mach4 plug-in that was supposed to support
threading, but I can find no mention of threading on their web site nor have
there been any forum mentions from users.
Newbie question save me some time.
Is Mach4 lathe threading currently 100% functional as described in the manual for the hobby version?
and
What external controllers currently provide 100% functionality for lathe threading at this time?
Thanks,
RICH
-
Newbie question save me some time.
Is Mach4 lathe threading currently 100% functional as described in the manual for the hobby version?
and
What external controllers currently provide 100% functionality for lathe threading at this time?
Thanks,
RICH
Warp9 is expecting Mach4 threading with the ESS to be available soon.
Tweakie.
-
I thought CNC Drive had released a Mach4 plug-in that was supposed to support
threading, but I can find no mention of threading on their web site nor have
there been any forum mentions from users.
The last I heard, the UC100 plugin did not support threading, but that was a few months ago. And I don't know if the plugins for the various CNC drive devices all have the same functionality. I'd send them an email to find out their current status.
-
I thought CNC Drive had released a Mach4 plug-in that was supposed to support
threading, but I can find no mention of threading on their web site nor have
there been any forum mentions from users.
The last I heard, the UC100 plugin did not support threading, but that was a few months ago. And I don't know if the plugins for the various CNC drive devices all have the same functionality. I'd send them an email to find out their current status.
I mailed in an inquiry several days ago.
No reply yet.
Russ
-
Hi Reinhard,
Quote
My contention is that once a coordinated move is dispatched to the controller it is no longer possible to change the A axis rotation rate, it would break the move.
Well, that probabely depends on what communication is possible betwenn mach4 and the motion controller - and of cause it depends on the communication refresh rate.
I don't think that is the case. The trajectory calculated by Mach and the position/velocity/time data once its dispatched to
the controller cannot be altered. The controller may do so, THC is an obvious example where the controller adds or subtracts pulses
from the Z axis pulse stream. In general though it can't be done...if you attempt to change the pulse rate of one axis say, then
the eventual endpoint of the move will be different than Machs assumed endpoint. Unless the controller can report back to
Mach and Mach can accommodate moves that don't go where they are supposed to then Mach has no chance of controlling
the machine. That requires an out of band axis which can be altered WITHOUT affecting a coordinated move of the remaining
axes.
Craig
-
Hi,
Mach4 threading support is complete as described in the manual, BUT it is dependent
on the motion controller to implement threading features.
As of 22 January 2018, and as best I know, threading is fully implemented by
Hycon, PMDX SmartBOBs, and PoKeys.
That information is very valueable! Thank you!
When you read propaganda from hardware builder, than it looks like everyone has best product on market. So its very hard for those that don't know that hardware, find out differences. And if you ask at forum, you have to separate answers from fanboys and wise guys.
Would be great, if artsoft would publish a table, that compares hardware support for special mach4 themes like threading. Support for closed loop and pid parameters could be another question. Or hardware support for feed override during motion ...
I guess, such comparision could speed up competition of hardware builders ;)
Unless the controller can report back to Mach and Mach can accommodate moves that don't go where they are supposed to then Mach has no chance of controlling the machine. That requires an out of band axis which can be altered WITHOUT affecting a coordinated move of the remaining axes.
Sure, out of band axis is one requirement. But if mach4 should recalculate speed of turning axis, it need to process motion feedback of an axis (during motion). At least, if speed depends on position.
So the position has to be sampled (if using an encoder) and transmitted quite fast.
If you look at the 40Hz of an ESS: an axis moved at 12m/min moves 5mm between to communication triggers, so no chance to calculate any finegrained speed changes.
Reinhard
-
Hi Reinhard,
But if mach4 should recalculate speed of turning axis, it need to process motion feedback of an axis (during motion). At least, if speed depends on position.
I think the situation is even more complex than that. If you wish to alter one axis, in this case we are talking about the rotational
speed of the A axis, but any axis that is one of the coordinated axes, it would be necessary for Mach to abort the remainder of the
move, hoping against hope that the sudden deceleration doesn't screw up your homing, read the actual location of the controlled
point and then recalculate the move for all axes including the variation of the one axis that you wished to alter.
I suppose its not impossible to imagine a control system that would do so but that is far away from where Mach4 is. Once the
position/velocity/time data is dispatched to the controller it can't be changed by Mach. The controller can but as you know the
circumstances where the controller effectively 'overwrites' the PVT table are to support specific functions, eg THC, homing etc.
No controller that I am aware of can alter a coordinated axis velocity during a coordinated move. Maybe PoKeys or Galil with
their internal programming could be coaxed into doing so but not otherwise.
Craig
-
Sorry guy's,
that i have to disturb.
you are talking here about 200-300-400$ Software, not about "real" (sorry to use this word),
CNC controllers (10000-20000$ aeria).
Tom
PS that is all "nice to have" option
-
Tom,
Not disturbing at all and only wish more of the 25000+ members would chime in......
The real ones are expensive, but, frankly no one from a hobby point of view is going to shell out that kind of money. SO will just echo the following:
Mach 4 hobby works for every command that is in the current manuals and part two is the motion controller
can provide for the instructions given it by Mach4 for every command. One without the other is useless.
Maybe a better spec would be specifically what a motion controller does not currently do relative to Mach 4 software.
Steve
Thanks for the info about which external motion controllers provide for threading.
If a external controller does not work "yet" then it is of no use use. Promises don't count any more!
RICH
-
If a external controller does not work "yet" then it is of no use use. Promises don't count any more!
RICH
Hi Rich,
Well said. By this time if the controller can't do what Mach4 can do then Houston, we have a problem!
-
Hi,
all of the readily available controllers do all the basic stuff.
There are a number of functions which require specific support/behaviour from the controller to be enacted.
The simplest is homing, but literally if a controller can't support homing I can't see it being on the market either.
The features which determine whether a controller is mature or still under development are probing, THC, backlash
compensation and lathe threading.
All Mach4 ready boards that I'm aware of support probing, so much so that it should be considered entry level.
The rest are all over the place. Vital Sytems Hicon board has always enjoyed a reputation for leading the development
curve. Given the cost of a Hicon board perhaps it should come as no surprise, Vital Systems probably enjoy a development
budget greater than all the others.
Amongst the 'value' manufacturers PMDX and PoKeys seem to be offering a greater range of features than the ESS say,
and the CNCDrive products (UC100, UC300, UC400) are only so recently released with a Mach4 plugin it comes as no
surprise that the features offered lag behind so of the more experienced Mach4 manufacturers.
When external motion controllers were first contemplated and then manufactured to suit Mach3 there was a wide spread
in what was offed as extra features and quite some years passed before all manufacturers implemented all of them.
That situation is repeating itself with Mach4. One comment in this thread was that NFS should maintain a list of compatible
boards and features offered. It would be nice but NFS is absolutely not the right organisation to do it...they're too close
to the manufactures concerned and any suggestion of bias or tardiness of behalf of NFS to update it list could result in a nasty
stouch behind the scenes.
A suitably respected and/or qualified individual could maintain such a list but it would rely on the support of the manufacturers
to provide the information. I could well imagine that some or even all manufacturers might be very reluctant to pass information
to another person whom would have a significant role in wether the market took up their products or not.
In some respects I think the manufacturers should be responsible. For instance Steve has chimed in that his PMDX smartBoBs
support threading and encoder spindle control. Why the hell should I be unsure, I mean I've been to the PMDX website on
many occasions yet had not gotten that message. SHOUT OUT THE GOOD STUFF Steve!
A comment earlier in this thread suggested that the third party hardware supplier model has made uptake of Mach 3 or 4
somewhat more difficult than it might have been. That overlooks the advantages to customers re competition for those
same products. Notwithstanding the confusion and varied approach to extras that comes with multiple manufacturers I
feel overall that I, as a customer, am well supported.
Craig
-
No controller that I am aware of can alter a coordinated axis velocity during a coordinated move.
Really?
For me its such a basic requirement, that I can't think how to live without it.
I worked on industrieal cncs and there feed-override is vital. Every cnc-job will be tested by using feed-override.
Well, I build a hobby cnc, but I'm used to professional work and I'd like to do so with my hobby cnc too :O
you are talking here about 200-300-400$ Software, not about "real" (sorry to use this word),
CNC controllers (10000-20000$ aeria).
That's right, but I think, mach4 is not the problem :)
The features which determine whether a controller is mature or still under development are probing, THC, backlash
compensation and lathe threading.
Well, probing is so essential, that I think, a controller that can't do probing or which causes problems at probing should be retired from market and blamed any time you write a new message.
In the feature-list I miss the feed override by hardware during movement!
I heard about hobby controllers, that support it, but it was not a controller for mach4 :(
THC and lathe threading might be the same software problem - but as already said, I'm convinced, that its not a problem of mach4 but of the controllers.
backlash is separate goal.
From my point of view, there's a path from entry-level-cncs (driven by simple steppers) until professional cncs driven by what ever motors controlled by closed loop with absolut positioning system (which for me is an industrial issue).
For me, the path looks like
- simple steppers
- steppers with closed loop drivers
- encoders moved from motor to ballscrew - closed loop by application (mach4)
- absolute encoders and closed loop by application
I start with level 2 (closed loop stepper drivers) on my cnc and I'm interested in level 3 - which (for me) is the answer to backlash and the like.
But to enable these functions, you need a different interface between mach4 and motion controller. For me, the refresh rate between mach4 and motion controller should be at least double the supported steprate. As most hobby-level stepper drivers support a max. steprate of 200kHz - so 400kHz sounds like a reasonable limit for hobbylevel. Industrial solutions might require refreshrates of about 1MHz or more, but that could be a feature of industrial mach4.
I don't know the amount of data, that has to be exchanged between mach4 and a motion controller, when mach4 controlls the closed loop, but I think, 100MBit Ethernet might not be fast enuf. So may be, those controllers need to be attached to workstations internal bus (like mesa 6i24 anything I/O)...
Reinhard
[/list]
-
I'm interested in level 3 - which (for me) is the answer to backlash and the like.
I can't imagine any controller at any level being the answer to backlash. Backlash is simply a mechanical problem. No controller can truly fix anything it doesn't have control over (which is the definition of backlash).
Just my 2 cents.
-
No controller can truly fix anything it doesn't have control over (which is the definition of backlash).
Agree. So looks like I didn't understand the meaning of backlash :(
Reinhard
-
Hi Reinhard,
if you attempt to change the pulse rate of one axis which is part of a coordinated move the eventual endpoint will no longer be the point which the trajectory planner
calculated.
All the other axes will reach their desired end points but the axis you altered mid flight will not. How does your software now plan the next move?
Feedrate on the other hand can be changed because ALL axes involved in the move vary and by the same proportion, so the move may be slower but at any
given point along the trajectory all axes are coordinated.
Craig
-
Hi Craig,
if you attempt to change the pulse rate of one axis which is part of a coordinated move the eventual endpoint will no longer be the point which the trajectory planner calculated.
Hm, guess u got me wrong.
Feed override does not change a single axis, but all coordinated axis at the same time.
The industrial controllers I worked with had a dial for G0 moves and a separate dial for all working moves. With that controllers it is possible to start a move with velocity of 0 and increase velocity at will (and so change velocity during move).
For that to happen with mach4 the controller needs to sample an analog value (or an encoder) and pass that value to mach4 fast enuf that the motion can be worked out based on that value.
The only case, where I can imagine to change velocity of a single axis is on lathe when you plane the front surface. In that case the axis don't has an endpoint and the speed of the turning axis depends on the x-position - thus mach4 has to know x-position during move.
Threading (for me) is just a matter of synchronization and for so very close to the closed loop goal ;)
Reinhard
-
Graig,
How about letting the manufacture's of the external motion devices do it. Create a new topic under
the Mach 4 board ( External Motion Controllers for Mach 4) where they can clearly list what is appropriate relative to Mach 4 and users can caomment. That surely would provide folks an intial consideration for selection. If they don't participate then they don't have anything to offer relative to Mach 4 and are not interested in the buisness!
Mach 4 is a two part deal ( Mach 4 and external motion controller) in consideration of whether one should convert or choose to start with Mach 4. That's why I deliberately listed them separately back in one of my posts.
Neither one should be falted for the others problem. ;)
I agree with Brett about backlash.... BUT.... If BC is not provided for there will be some percentage of hobbiests that should not use the new software.
RICH
-
My understanding from back in the initial talk of Mach4, was that Artsoft would maintain a list of certified hardware that worked with Mach4, and what functions they supported.
This should have been available from day one, and the fact that it didn't happen is a big negative.
All of Mach4's competition makes their own hardware, and don't have any software-hardware compatibility issues.
-
This should have been available from day one, and the fact that it didn't happen is a big negative.
YES Sir! That's the point!
-
Hi Reinhard,
The only case, where I can imagine to change velocity of a single axis is on lathe when you plane the front surface. In that case the axis don't has an endpoint and the speed of the turning axis depends on the x-position - thus mach4 has to know x-position during move.
This question about changing the velocity of an axis within a coordinated move came up because a member wished to alter
the A axis rotation speed in proportion to radius. A very similar situation to that you proposed.
Mach4 can change the feedrate by changing all axes within the corordinated move. It can update the feedrate 40 times per second.
Is that update rate not adequate?
I understand the PLC runs faster than that, about 12.5ms. You can program it to run faster although I believe there are practical
limits which suggest about 5ms. The PLC could read an analogue vale at around 200Hz, is that fast enough?
Craig
-
Mach4 can change the feedrate by changing all axes within the corordinated move.
I know that :)
... but add to that the capability to handle an external control
It can update the feedrate 40 times per second.
I blamed ESS for this frequency and I very hope, that its not mach4 that forces that slow communication.
Is that update rate not adequate?
Ask yourself! For me 40Hz is extremly slow.
So just take the time (1s / 40) and calculate how far your cnc might move at i.e. 12m/min (result: 5mm)
For me, that stepsize is too big.
I understand the PLC runs faster than that, about 12.5ms. You can program it to run faster although I believe there are practical
limits which suggest about 5ms.
I don't know, where you take such asumptions from. 5ms is damn slow.
Lets calculate the requirements for hobbylevel close loop with common steppers:
A 2-phase stepper takes 200 Steps for one turn and has an encoder with 1000 lines. The encoder has 2 channels and is decoded by graycode. Hobbylevel stepper drivers support a max. steprate of 200kHz
Lets calculate the encoder: 200 steps means 1000 encoderlines or 4000 graycode pulses, so 200kHz steprate means 4MHz signals coming from encoder.
AFAIK ESS is capable to sample encoders with 4MHz.
4MHz means 125 nanoseconds between two pulses (5ms is 40,000 times bigger than 125 nanoseconds).
Your PC runs at some thousand GHz/s and a current SATA interface is rated at 6GBit/s - so with a motion controller, that is attached to internal PC-bus, that timing is no big thing - and I'm quite sure, that it isn't for mach4 neither.
Reinhard
-
Lets calculate the requirements for hobbylevel close loop with common steppers:
Mach4 is not a closed loop control, so this is irrelevant.
The only closed loop hobby level control available is LinuxCNC.
-
Mach4 is not a closed loop control, so this is irrelevant.
That's a clear statement.
Thank you
-
Hi,
Mach4 is not a closed loop control, so this is irrelevant.
Hm, for me, lathe threading is closed loop based on external encoder signals.
What I don't understand: if the functionality is already in mach4, why use it for little use case only and don't use it for common use case?
Reinhard
-
Reinhard,
You can have closed loop with HiCON card system and Mach4.
http://www.machsupport.com/forum/index.php/topic,35162.0.html
Regards,
Mauri.
-
Thank you Mauri,
I already read about that possibility. The point is, that that solution will consume 2thousand bucks, so I look around if I find alternatives ...
Reinhard
-
Hi,
Mach4 is not a closed loop control, so this is irrelevant.
Hm, for me, lathe threading is closed loop based on external encoder signals.
What I don't understand: if the functionality is already in mach4, why use it for little use case only and don't use it for common use case?
Reinhard
That seems moer like just "encoder following", not at all like a closed loop system.
-
Hi,
well, to follow an encoder, you have to process encoder signals. That's the hard part of the closed loop.
A pid controller in software is nothing complicated, so if you already have the data ... ?!?
... just my 2 cent
Reinhard
-
Lathe threading and regular closed loop motion are not that similar.
As mentioned lathe threading is actually encoder following. On top of that the
following is implemented by the motion controller and its plug-in, not directly
by the Mach4 core. The technique used by Mach4 to control lathe threading
cannot be applied to more than one axis at a time.
There are motion controllers, like the Hicon, that can manage servos with
encoders and provide a locally closed loop between the servo and the
internal system of the motion controller. They can even cause Mach4 to
display actual position instead of commanded location, but the decision
of what to do about following error and fault detection is still up to the
motion controller. This results in a system that is only marginally better
that using servo drivers that accept step and direction commands and
internally flag a fault if the following error exceeds a preset threshold.
-
Windows is not real time and thus any application running on it will never close the control loop. However, Mach can be part of a closed loop system. Some motion controllers operate in a closed loop manner with Mach. Galil, Vital System DSPMC, etc... For a servo system, the loop can be closed in three places, depending on the setup.
1. On the control itself (LunuxCNC).
2. On the motion controller.
3. On the servo drive.
It you want instantaneous feed rate override, it has to be done at the point in the system where the loop is closed. That leaves points 1 and 2, as I have never seen it implemented on point 3 (the servo drive).
Hardware based feed rate override is possible on the Galil with a pot connected to an analog input and a bit of Galil DMC programming. You change the sample period time base (TM) based on the analog value. However, if you are running servos, this also changes the PID loop sample period so you had better change the PID values to match the new time base as well (just a mathematical calculation). If you are running steppers with the Galil instead of servos, dealing with the PID values become irrelevant.
All that being said, I would not consider the Galil a hobby controller. It is more of an industrial motion controller due to its' complexity and price point. But the option is there. The higher priced controllers can usually accommodate these types of requirements. But we are talking about Hobby level machines and controllers, right? Professional features come at a price.
Threading... it needs to be done in the real time component. No encoder need be involved though. The motion profile can be, and is, calculated. The start and end point of the thread never changes. The challenge is the spindle speed. If the spindle speed could remain constant, a perfect thread could be produced with a predetermined motion profile. However, that is never the case in the real world. So the real-time component of the motion controller must monitor the speed of the spindle and change the time base in which the thread profile is executed to match. So how does a motion controller monitor the spindle speed? It can use an encoder. But a single pulse per rev is usually quite adequate. It turns out that producing a class A thread doesn't require THAT much monitoring of the spindle speed. If you have an encoder on the spindle, then use it. But it is definitely not required.
Steve
-
Steve,
To add to your point on threading Encoder vs single pulse. We have lathe that is using an encoder with 300 PPR and it threads just fine we also have 2 lathes in that use a single pule per rev and they thread just as well.
I Also agree with your point that certain features are industrial and not hobby, therefore you'll have to shell over some money for them.
-
Hi Steve,
thank you for pointing out the higher level controllers. Didn't heard about galil before - so have to read a bit to get familiar with them. Pricing is high, but looks interesting.
About the question linuxCNC versus mach4 - I'm quite sure, that mach4 could be much better than linuxCNC - you only have to try :D
... linuxcnc does not work on linux standard distributions - it needs a special (realtime-)kernel.
May be, there's something similar possible on windows systems too?
Threading ... No encoder need be involved though.
Some pages before I was told, that you can not guarantee, what you don't have under control.
For me, position of Z-Axis is related to position of the turning axis. If you know the position, velocity is irrelevant - so I don't know, how you can guarantee precision without measuring?
AFAIK threading is worked out by multiple passes usually. How can you match the same starting point without measuring?
Anyway - your decision.
Reinhard
@Steve
I skimmed galil pages for the instantaneous feed rate override. Did not found any.
Do you have a link, where I can read more about it?
-
Oups, I'm confused. Prices I saw at first sight where used prices at ebay :(
-
Some motion controllers operate in a closed loop manner with Mach. Galil, Vital System DSPMC, etc...
Vital System DSPMC looks what I was searching for. Great!
So I might have to eat dog food for a while and beg for bigger budget ;)
Reinhard
-
Reinhard,
On the Galil, you would use the TM command. TM is set to 1000 (approx 1ms) by default. That is the time base that everything uses. Increasing or decreasing the value of TM would be tantamount to feed rate override. The challenge then becomes to tie the value of TM to an analog input. This can be done in Galil DMC programming where a thread is dedicated to watching the analog and adjusting the values of TM (and possibly the PID values to if using servos. As I said, the Galil is a complex controller. They have good documentation, but it still takes years to master every facet. Meaning it is up to you to read the documentation and implement what you want to do. You will NEVER see a reference to feed rate override in their manuals. They give you the tools, you just have to figure out how to use them. However, if you buy a Galil, they do offer great support. They have good application engineers to help you get what you want out of the controller.
In order to run a Galil with Mach, you have to know how to run a Galil by itself first. I run a Galil on one of my machines. I consider it a viable hobby level controller IF you are willing to learn about it and IF you can/want to afford it. Budgets being what they are, everybody has their own levels to which they draw a line. Buying a Galil on eBay is not for the crowd that knows nothing about them. There are so many different models and a LOT of them on eBay were custom designs made for a particular purpose. And Galil offers no support for used controllers. So you better know exactly what you want to pull the trigger on an eBay Galil.
Multiple pass threading starts each thread pass on the index pulse on the encoder or the pulse per rev input. The infeed amount is already there. All that remains is synchronizing the Z feed rate with the actual spindle speed. Z always moves from the start point to the end point and never needs to be re-planned. It is NEVER as complicated as people make it. It is actually quite simple to produce a high quality thread with just a spindle pulse and a calculated trajectory. But it does take a real-time environment to implement the control loop.
There are real-time extensions for Windows. The problem is that all of them cost big bucks. It would more than double the price of Mach. There are two types for Windows, mainly; Hypervisor and HAL. A Hypervisor is running a RTOS on the hardware that then partitions a CPU core (or more) to runs Windows in a VM. The HAL method uses a custom Windows HAL. The HAL type turns Windows into a true RTOS and can be implemented with a single core processor. Of the two, I prefer the HAL type. Interval Zero https://www.intervalzero.com/ is one that I like. In fact, they have a product called KingStar that implements a software EtherCAT controller. They are developing a plugin which is nearing completion. They demonstrated it running a machine at our shop a few weeks back. But this stuff, while cool, is way out of the realm of Hobby land.
Steve
-
Thank you Steve, for your extended reply!
And Galil offers no support for used controllers.
good to know! Thanks. I appreciate support.
As I already wrote: DSPMC seems to fulfill all my requirements. It looks easier to use and is smarter to my budget ;)
So most probably it will be my upgrade path. I already have an ESS, but I consider feed override so important, that I'm not sure, whether to start with ESS or sell it and go for DSPMC from start on.
IF you are willing to learn about it and IF you can/want to afford it.
The first is no question. Learning is my main hobby :)
... but the second is.
Multiple pass threading starts each thread pass on the index pulse on the encoder or the pulse per rev input.
Ok, that sounds to me like there is an external signal sampled for threading synchronization.
I agree, that a threading pass is atomic and does not need to be recalculated.
The point is, I don't see a way to synchronize after return between two passes.
But hey - I you made it, it is good to know :)
There are real-time extensions for Windows. The problem is that all of them cost big bucks.
one more reason for mach4linux to arise :)
Reinhard
-
This dialog has gotten focused on some fairly technical capabilities. Going back to the original question, I think its a great program that is hampered by a lack of good documentation and access to a good set of examples. I have a basic DIY 3 axis machine. I had no trouble getting my system running with the ESS controller. Things only got complicated when I wanted to customize things for my unique system. I am willing to learn some LUA to add some features. When I try to get into it though I quickly get lost in the layers of things I need to know that are independant of each other. Some of the layers I see are: Mach4 partitioning of functionality, Mach4 APIs, LUA, wx.widgets and 3rd party addons. Each has some level of documentation, but nothing tying them together. The Mach 4 layer especially would really benefit from more info given its complex API structure. When taken as a whole, even a pro would probably struggle for awhile in this environment. I have persisted and have modified my screen, and added some macros, but I still don't understand a lot and in some cases don't know where to find the info I need.
Docs are hard work and become a hugh anchor as you change things, but are needed if your target market includes hobbiests. There are some interesting examples on ways to do this out there. Arduino is one. When the docs get confusing, I can look at a multitude of examples to understand usage. Arduino has a larger user base, but NFS can help by making it easier to find code withouut browsing the forum and finding something useful on page 11 of a thread. Fusion 360 has weak documentation, but lots of tutorials that band aid this deficiency.
-
Hi,
to understand mach4 programming it is helpful, if you already know multithread programming.
If you never had to deal with multiple threads, there's a nightmare of pitfalls you can't be aware off.
But that's not an issue for artsoft!
After reading the provided pdfs, I took the core-api-docs and startet to read it as a book. That helped a lot. Even if I can't remember any prototype, I have an idea, what's there and so I find the function very quick.
Then to understand the big picture, it helped me, read the screen.xml - extracted from the screen set. At first look most of it looks very strange. But if you try to get into it, you get a feeling for the big picture. You realize the power of mach4, its advanced archtecture and much more.
More documentation would be fine - but we don't live in a perfect world.
... and if you ever get stuck and trapped - that's the point to come here and ask for help.
That way worked for me and it will be fine for others too :)
Cheers Reinhard
-
Hi,
better documentation would be great. I have some sympathy for NFS it that there is a huge investment in time and resources to generate good documenation.
Simply put they don't have the resources to do it, they need to concentrate on those development issues that result in sales, particularly OEM sales.
Having said that, NFS has also been guilty of not even annotating the documents that they do have with issues that arise and are bought to thier attention.
I have one example where in the end I opened a support ticket to resolve a question, NFS duly got back to me and answered the question and could have annotated
the API.chm very simply but did not do so. Sure enough a few months later I was answering the same question for someone else on the forum.
I think Reinhard has a similar example where he put in an effort but no-one from NFS bothered to incorperate that contribution into the 'database'.
Eventually NFS are going to have to document their program....taking advantage of those tidbits of analysis/insight/experience developed by users seems useful.
Craig
-
Hi Reinhard,
I already have an ESS, but I consider feed override so important, that I'm not sure, whether to start with ESS or sell it and go for DSPMC from start on.
Mach4 with the ESS controller has feedrate override and as far as I can tell by standing at the machine operating it seems to be instantaneous.
If there is anything that the ESS lacks is analogue inputs. PoKeys do a good job and PMDX also have a couple of analogue input channels. Of course Galil
do them as well but at a cost as do Vital Systems.
Vital Systems boards offer closed loop operation, if you pay the extra, but if I understand it reads an encoder and produces a high resolution analogue voltage
to drive a DC amplifier which in turn drives a DC servo. DC servos are perfectly capable but pretty old school by now....additionally have you seen how expensive
they've become? Bang for your buck go for AC servos, the loop is closed by the drive, and usually by the same manufacturer with very VERY good results, better
than I could program a PID loop anyway. If you accept as I do that the drive/AC servo loop do as good or better job than any controller then you have paid
a premium for closed loop feedback which is not really required.
A worthwhile feature offered by Galil, Vital Systems and Pokeys is the ability to micro program the controller. You pay extra to have Vital Systems boards do it.
The advantage is that you can use that fine grained programming ability to close control loops OTHER than motor loops. Galil have lead the field in this regard
for many years.
Warp9 have a plan to release an API for the ESS. To be honest I can't see that it will happen quickly but it would add some of the ability of the above. It doesn't seem
likely that it will ever match Galil for this style of programming.
If you already have an ESS I would carry on with it. The ESS plugin lacks at this time THC and lathe threading but are being worked on.
Those extra features including closed loop servo control and microprogramming on a Vital Systems board are going to cost $1500 or more.
The first is no question. Learning is my main hobby Smiley
... but the second is.
The ESS will provide plenty of learning opportunities!
Craig