Machsupport Forum
Third party software and hardware support forums. => CS-Lab => Topic started by: Hood on April 03, 2013, 06:17:58 PM
-
This is my first impressions of the analogue version of the CSMIO/IP. I have previously used the Step/Dir version and loved it but for me the analogue version is even nicer. I am fitting it to a Chiron FZ12S and although I have not fully finished the retrofit I have seen enough to know it is working extremely well and beyond what I had hoped for.
In addition to all the things I like about the Step/Dir version the analogue version has the Index homing feature and it is very nice and works exactly as it should. The Step/dir version can also do it but it needs more setup and possibly additional hardware (depends on servo drives being used) The analogue CSMIO doesn’t require anything extra and all you have to do is enter the Home switch input and encoder count for that axis. When you home it moves to the switch at the speed you have set in Homing and Limits, it then backs off the switch at a much lower rate and then once the switch is again closed it seeks the index pulse of the encoder and sets machine cords.
Another great thing about the analogue version is the encoders are obviously constantly monitored (to close the loop) but they are also used to update the DROs in Mach, so once homed at start up Mach always knows exactly where the axis is, even after an E-Stop or if you disable the drives and move the axis manually.
The auto tuning also seems to work very well for me, so good in fact that I have not even bothered trying to tweak with manual tuning. Initially I was just using the default encoder simulation from my servo drives (4096ppr) but now I have increased that in the drives to 15,000ppr and things are even nicer.
The reason I increased the encoder resolution was I was wondering if it would help give me smoother MPG motion. I had noticed that in later versions of the plugin the motion has not been quite so smooth on the lathe that I have the CSMIO/IP-S on, it used to be silky smooth but now it is slightly rough. It is still way ahead of multistep motion of the parallel port or smoothstepper but was just annoying me that it had been as smooth as silk previously.
On the Chiron however is was much rougher and was actually getting to the point where it was too rough to be useable. The reason it was worse on the Chiron, I think, is the combination of fast rapids\ fast acceleration\heavy machine. I contacted CS-Lab regarding later plugins having rougher MPG motion and it seems some users were having issues with the axes running on slightly after stopping turning the MPG so they altered things.. This happens in Mach also if you have slow acceleration so I suspect the people with that issue had slow accelerations.
Anyway increasing the encoder counts seems to have helped as things are almost as smooth as the previous versions were on the lathe and now I feel the MPG is very useable. The reason I think it has made a difference is the CSMIO is able to respond quicker due to the higher encoder count, but that’s just my thinking, it may be some other reason.
So all in all I am extremely happy with the analogue version and I can see me updating my other machines in the future with the CSMIO/IP-A rather than the CSMIO/IP-S.
Hood
-
Hood
We are still in the wiring stage of hook up. Should I have a constant 24v out put on pin 1 of my digital out put(24v supply for outputs 0-3)? And thus does my relay coil connect across pin 1 and pin 2(output 0)?
Also do I need to download additional drivers? My computer see's the controller but will not communicate?
Do you know if there is a schematic available for the this unit(CSMIO/IP-A)?
Thanks
Apollo ll
-
If you have the IP-A controller then you connect a 24v Supply to Pin 1, 4, 7, 10 and pin 14, 17, 20, 23 (positive to 1 etc, 0v to 14 etc) you then connect your load between the output pin and 0v
Same goes for the other outputs, see if pic below helps.
I you then call an output to become active current will flow from that pin, for example in the drawing below if you told Mach to set the IP-A's OutPut 0 active then the current would flow from Pin 2 through coil to 0v of the supply.
The only schematics available at this time are the pinouts of the connections, I think a full manual similar to the IP-S is being done at the moment but basically you connect things in a similar fashion to the IP-S just paying attention to the different pin layout in the IP-A.
Hood
-
Hood,
That sure does help.
Thank you
Apollo ll
-
I'm planning on ordering the CSMIO/IP-A this week, so it's good to read positive information about it. It seems like there really isn't much info out there yet.
I've got a 1991 Shinx router with Sanyo servos and drives, but the original controller just gave up the ghost. It was run by an Epson 286 that apparently is completely proprietary, and stopped reading the boot up disk (I've tried replacing drives, different disks, etc.). I look forward to reading more and perhaps contributing once I have one in my own machine.
-
Well here is a post detailing the first parts made on the Chiron, couldnt be happier with the CSMIO, best controller I have used :)
http://www.machsupport.com/forum/index.php/topic,22495.new.html#new
Hood
-
Very Nice Hood!, Nice all around! Makes me jealous! ;D
Brett
-
Hood, Got another question regarding the wiring with my CSMIO/IP-A. My axis motion, jog and tool clamp inputs all have three wires. How do each of the hook up? Do they hook to the digital input or the analog? Thanks a lot for your help.
Apollo ll
-
What kind of inputs are they? ie the clamp, what is it?, a solenoid or a limit switch to say it has clamped ? or?
Hood
-
Hello Hood,
The axis motion is the jog hand wheel. The jog button is what changes the amount the machine moves with every click of the hand wheel (.001, .0001 and so on).
Im not certain about the tool clamped function, maybe some sort of safety?
I will try to send a picture of the diagram.
Thanks for all your help!
Apollo ll
-
Can you take a closer up pic of each part as afraid I can not really make out what it is in that pic.
Hood
-
Hope these give you a better idea.
Thanks,
Apollo ll
-
Ok the wiring for jog and motion still dont really help much but the clamp looks like it may be a prox switch. You would really need to find out the part number of that prox to see what voltages it can handle.
The pic of the other two controls however help a bit. The Jog +- looks like it is a potentiometer and the Axis motion looks like it is a MPG. The pot should be easy enough as you would connect that to an analogue input, ie one side to 10v, other to 0v (Gnd) and the middle to an analogue input.
The MPG, you would really need to see if you could find a part number on it to make sure it is indeed a MPG that is putting out ttl signals. If it is then really you need the MPG module for the CSMIO, the other opyion would be to convert the 5v signals to 24v and put into the CSMIO itself but I am not sure how well that would work, so try and find out what all components are first by looking at their part numbers or any other info they have on them directly.
Hood
-
Hello Hood,
As far as the axis motion, and the jog switch Im going to order the MPG Module and pendant from CS Labs.
You were right it is a proximity switch. Part number EGT12x02AP024-SEM4. It looks like it will handle 30volts, and 250ma. Im not sure how to hook it into the CSMIO though?
Any ideas?
Thanks for all your help!
Apollo ll
-
Do you have a link to the data sheet for the prox?
Hood
-
The web site I found it on is, www.datasheets.globalspec.com
Part # EGT 12x02AP024-SEM4
Apollo ll
-
Can you attach the data sheet as I am not really finding much with that exact part number.
Hood
-
Hello Hood,
Here is a link for the part.
http://datasheets.globalspec.com/ds/52/EuchnerUSA/95C10632-D76F-44BC-98F7-5BD973F69129
-
Afraid I would have to sign up to download that data sheet. It took me a long time to get rid of the constant bombardment of emails from Globalspec before, so not going to do it again.
If you have downloaded, please attach here using the additional options button.
Hood
-
here is the pdf
-
Hey Hood, Sorry to bombard you with all these questions but I sure do appreciate the help. It will be great to see this in action once it is done. Along with the Proximity switch I am trying to tie in the last of the spindle control. The spindle speed control signal says it takes a 0 to +/- 12v analog but the CSMIO only has +/- 10v analog. Does that hook to the analog +/- 10v or the analog 0-10v output. And will it affect the actual speed since it is 10v instead of the 12v that it was originally.
Apollo ll
-
Ok looked at the pdf, thanks Ya-Nvr-No. On page 3 it looks like you would want diagram 1 for your prox, ie NO Positive.
So
br1 to 24v supply
bk4 to input on CSMIO
bl3 to 0v
Regarding the spindle drive, if it is looking for a +/- analogue signal then you would use the Spindle Axis options found under Special Functions tab, see below for screenshot of my setup.
Also make sure you do not have Spindle Speed Averaging set on Ports and Pins, Spindle Setup, if you do the RPM will not read correctly in Mach.
Regarding the +/- 12 for your drive, it is likely you can scale the input voltage in the drives parameters so that 10v would be full rpm.
Hood
-
Hood,
Can you tell me what the wires leading from the CMOSIO CAN bus on your Chiron are for (in the photos they are leading off the encoder module CAN)? I thought that they only connected to more modules, so I was surprised to see a cable coming off it, but no obvious destination. I've got the IP-A on it's way, along with four I/O modules, so I'm starting to map things out and figure out what the possibilities are with these controllers (perhaps beyond what my machine is already set up for). If that CAN connection can be useful for something other than modules, I'd be curious to know.
The tool changer on that Chiron is about the fastest I've seen, particularly because of how fast you have the rapids set up on that thing! It certainly does not look it's age, so great job! I used to have a machine that had two opposing heads, which also had a very fast tool change. In that case, it put the tool in one head, which was upside down and away from the work, while the other head was busy cutting. As soon as the "in use" head finished, the heads flipped over and the new tool was ready to go while the other head got the next tool pre-loaded. It was a bit of a pain to program, since you had to make a tool call for the next tool you will need, not the one you currently need since whatever is in the upside down head is what will cut next. Basically, you had to call for the first two tools before you could even start cutting by doing two consecutive tool changes (T1 and T2 for example). That would get you started with T1 while T2 would be waiting in the other head. When you finally need T2, you had to make a tool change to T3 in order to bring T2 down to work while T1 is replaced with T3 (arbitrary tool numbers of course).
-Mike
-
Mike you can only use the CanBUS for connecting to modules. I have a MPG module and an extra I/O module up in the operator panel, see pic below of partially done panel.
Sounds like a very interesting machine you had, what make/model was it?
Hood
-
Thanks Hood! It had occurred to me after the last post that there may have been a remote module somewhere requiring a longer cable that I didn't see in the other pictures. Thanks for the info.
The machine was a late 80's SCM Routomat (router, I deal mostly in wood and plastic fabrication). They came with a lot of different head configurations and I think that I would have preferred a standard 10 tool carousel/single head config in retrospect. Mine had a 6 tool linear rack positioned above the upside-down head. I tried many different software packages within my budget ($5000 or less) and none of them could properly post-process code for the changer, so the carousel would have made life a little easier. I think your changer is faster than mine was, mostly because the Z rapid on the original NUM 750 controller left something to be desired. ;) If the Z could have rapid traversed up to 0 at 400ipm, it would be pretty similar. As it was, 200ipm was about the limit with that controller and servo combination. It's really amazing to see how fast your machine is moving the Z axis given the amount of weight I suspect it has to position (I assume there is some amount of pneumatics helping do the actual lifting?). That tool changer has got to weight a lot!
-Mike
-
No there is only a servo motor on the Z, no air assist or counterweight.
The motor is geared 2:1 to the screw and the screw is 10mm pitch so 4000rpm motor and 20,000mm/min rapids as per the original setup.
Toolchanges can actually be quicker as the last owner screwed the dampers on the main up/down cylinders up a bit and I have just left it at that. Also I could programme each Z move for each particular tool so that I do not have to move so far to change but as I am not doing production work the extra second or two per toolchange doesnt matter to me.
Hood
-
I think most people would be thrilled to have a tool change that fast as it is. I once had a job where I cut out about 1000 oval cribbage boards, some made out of walnut and some out of man-made for paint (MDF). It required 5 tools to run and there were 128 holes (including both two tracks and some peg holders at each end). There was also a storage pocket with thumb wells to pick the lid off the pocket, which had a second pocket creating a ledge for the lid. There was also a detail around the top outer rim of the oval. The fastest I could produce them ended up being approx. 5 minutes each, and with a 1300x2600 table I could fit quite a few on there at a time given they were about 8 inches at the widest part of the ellipse. The tool change made all the difference in that case since saving 4 or 5 seconds per tool adds up to almost 7 hours of production time. I imagine a tool changer like yours would have saved another 3 hours on top of that (and I really thought mine was quite fast). My main obstacle ended up being the code, since the NUM 750 only had about 80kb of memory available and no drip feed option. The NUM is a conversational controller though, and I was able to program a loop back through some incremental code (the cribbage board portion), and then switch to absolute to reposition to the next blank.
My current machine is set up with 1000RPM Sanyo San-Driver servos. Both my last machine and this one have had the servos maxed out at far less than the full rated RPMs. This current machine runs in rapid at a max of 750rpms in X and Y, with only 650rpms in the Z. I've always wondered if the manufacturers are just playing it exceedingly safe, or are the servos more likely to fail if run at full rated speed? I also figured that there may be other reasons for the lower RPMs, such as the encoders used combined with the age of the controllers. Do you think that it would be safe to run the servos at full rated RPMs with the CMSIO? I'm sure that from a computing power standpoint, it should be light-years faster than what it's replacing, and I would think that if a company such as Sanyo rates the motor at 1000RPMs, then it probably already has some built in headroom over what they could actually withstand reliably. I would love to be able to run the machine at 20,000mm/min (X,Y) and 10,000mm/min (Z) as that would be a bit more than an incremental improvement over the current speeds with no additional cost.
-
I would imagine if a manufacturer rates a speed as continuous then that is exactly what it is and really shouldnt shorten the life of the motor.
What may have been the problem is the control had a max frequency it could read, ie if your encoder input was too high due to the RPMs then the control may not be able to handle that.
The CSMIO has 3MHz capability for the encoder inputs, on the Chiron I am using only 1MHz (15,000 counts per rev and 4000rpm) and on the spindle I am using just over 2MHz
Originally I had the drives at the default 4096 lines per rev encoder output and although it worked well things were definitely smoother and tighter when I increased the resolution as the control could correct much more quickly.
Hood
-
I won't be making the CSMIO even breath hard then. My encoders are Sumtak Optcoders with only 500p/r, though I'm not sure if a 4x multiplier would apply to them as I'm having trouble finding any documentation on them. Even then, I'd only be producing 333khz at 1000rpms. I think it would be surprising if the original controller was able to keep up with half of that, which probably does explain the feedrates.
Mike
-
Yes I would think it will be 2000 counts per rev.
Hood
-
I'll probably end up going under that assumption. ;)
My math was off earlier, it would only be 33.3 khz. The IP-A may fall asleep waiting for information from the encoders.
-Mike
-
LOL.
Do the encoders feed back to the Amps?
If they do you can maybe multiply the output from the amp to the CSMIO, all depends on the age of the Amps I suppose, I have some Osai ones here that are just fixed output, ie whatever encoder count is in it is the same out but more modern ones you can multiply or divide the encoder outputs.
Hood
-
The encoder feeds back only to the controller with A, A-, B, B-, Z, and Z- leads. The machine and servo system is circa 1991, so not ancient by analog standards I would suppose, but does not have that functionality.
-Mike
-
There must be a tach on the motors then I would think that feeds back to the amps for speed reference?
Hood
-
There are 2 tachometer wires running to/from the motors/amps. Does it use those as part of the feedback loop? There do appear to be some other features that I'm trying to decipher from the manual.
It looks like my IP-A and I/O modules arrive on Thursday or Friday depending on customs. Hopefully I'll see some basic motion in the next two weeks. :D
-Mike
-
The amp will be using the tach for Velocity control and the CSMIO will use the encoder feedback to close the position loop.
Modern drives tend to use encoders for both.
Hood
-
Since the encoder only feeds back to the controller, would it make sense to update the encoder to something with higher resolution? If so, does the encoder have to be the same type of encoder that was there originally? For example, could an encoder like that in the PDF below be used which has only A, B, and X (index), even though the original had a A, B, Z, A-, B-, and Z-. They both operate on the same 5v power supply. Would the newer encoder be necessary to home on index or would the old encoder provide an index signal the controller would recognize? The CUI encoder looks to be capable of over 8000 ppr after quadrature encoding, and given the slow speed of the motors and the CSMIO's capacity for reading the signal, perhaps it would be worth the upgrade? I would assume that it would help generate smoother motion, even if it doesn't really produce an practical mechanical resolution increase.
http://www.cui.com/product/resource/amt10x.pdf
-
You want differential signalling, whether that is direct from the encoder or a line driver close to the encoder so if you go with the AMT you will have to either build a line driver circuit or buy one.
The old encoder has index signals, that is the Z and Z- and you will be able to do index homing with it, assuming of course it is ttl output, it likely will be.
Higher counts are definitely an advantage, the encoders I have on the motors are sine/cosine encoders, what they call smart encoders, and have a standard count of 1024 lines (4096 quad), as they are sine waves they can be accurately interpolated and my drives can provide that to 256x.
I think personally I would set up with the original encoders first and see how things are, you will likely find they are at least as good if not better than now.
Hood
-
Howdy Hood, I have several issues that I could use your advise on.
1) On a CS-MIO--IP-A to an Apollo II hook up what is the wire to pin configuration when connecting the analog X, Y and Z (+/-10v) to the Axis Drive Boards? I am not getting ten volts from the outputs to control the Axis Drive Boards. Am I supposed to supply voltage some how or is the CS-MIO supposed to supply it?
2) What do I hook up the spindle break to?
3)I have my eject tool function hooked to output 1. How do I tell the Mach III program what that port is so it can control it, and the same with all the outputs?
Thanks a ton.
Apollo ll
-
What is ApolloII?
No you dont have to supply 10v, that will be supplied by the CSMIO. You will only get 10v when the CSMIO thinks it needs to send 10v to get the axis to where it is meant, normally you will see a much lower voltage. If you attach your xml I can check that you have Mach and the Plugin set correctly.
Normally a servo drive will have a brake connection on it so the drive can be set up to control the brake. If it doesnt then probably an enable in Mach or maybe even the Enables in the CSMIO would be used to control a relay that will power/unpower your brake. How exactly you do wil depend on the brake setup, normally they are spring loaded so that with no power they are applied, is yours that way?
I think the outputs on the main CSMIO are port 10 but would have to check up on that as its been a while since I set it up.
Hood
-
Hello Hood,
I can not send the xml file. It says I already submitted that post. Any ideas how to send it?
Thanks,
Apollo ll
-
I will try once more.
Apollo ll
-
That (going by size etc) would seem to be the Mach3.exe, it is the xml I need. If the profile you are using is called Mach3Mill then look for a file called that which has a .xml file extension rather than a .exe extension.
Also can you clarify what the ApolloII is please.
Hood
-
Hello Hood,
Trying again.
Thanks
-
Hood,
The machine I have is a Bridgeport Apollo ll. It is a 3 axis cnc knee mill. 1993 vintage.
Thanks,
Apollo ll
-
Ok, that xml attached has the printer port as the controller, if it is definitely the one you are using then you need to go to Function Configs menu then choose Reset Device Select. You then restart Mach and you will be given the option to choose the CSMIO as your controller. If you dont then you do not have the plugin for the CSMIO correctly installed.
Hood
-
Hood,
Iam sorry but I think I sent the wrong one.
See this one looks better?
Thanks,
Apollo ll
-
Ok you dont need to set port and pin numbers in Machs Motor Setup page, all that is needed is you enable the axis. It should not do any harm having numbers there however.
What you have to do is go to Config menu then Config Plugins then Click on the Config at the side of CSMIO/IP-A plugin.
You can then enable the axes and tell it what outputs you are using and also set up your encoder ins etc.
Hood
-
Mike you can only use the CanBUS for connecting to modules. I have a MPG module and an extra I/O module up in the operator panel, see pic below of partially done panel.
Sounds like a very interesting machine you had, what make/model was it?
Hood
Some very interesting reading here, I'm like a sponge learning things today.
I like the keybaord Hood, do you remember where you purchasaed it from?
-
I got that one for free, have purchased the others, I use on Machines, from eBay. There is also a place that sells cheap (relatively speaking ) stainless ones, I was about to buy one but remembered I still had one sitting in the cupboard :)
http://www.kestronics.com/catalog/keyboards-c-4.html?osCsid=0mgcaj8jfqauoogdll86e92gn2
Hood
-
Thanks hood, more toys for my Christmas list
-
Hood, what computer board are you in use of? It seems very compact when it comes to size.
-
Its a micro atx mobo, based on the Intel Atom. The one I have is now discontinued
http://www.aria.co.uk/VendorStores/Asus+AT5NM10-I+Intel+Atom+D510+Micro-ATX+Motherboard+?productId=40767
but there are plenty of similar mobos around.
I have used Via Pico mobos in another couple of machines, they are really tiny (100mm x 72mm, see pic)
Hood
-
What are your impressions from the Intel Atom? As you know, CSMIO products take out the pulse generation load from the PC and hence we only need a general PC with Ethernet capability to run Mach.
I am now offered a second-hand Dell Optiplex 160 for as low as 65 Euro.
The specs are:
CPU: Intel Atom 1.60GHz
Memory: 2GB DDR2
Hard Drive: 60GB SATA
According to the Dell website:
Dimensions:
Weight: 4.6lbs/2.09kg
Height: 9.02"/22.9cm
Width: 1.85"/4.7cm
Depth: 8.86"22.5cm
Maybe it is a good machine for start. I couldn't save the space for a PC inside the electrical panel so I will be mounting a second one with a PC and protection circuit for it inside. But it's much better to be something compact and the best is something like your model but they are no big availabilities locally and I don't have much time left to wait for a PC lol :)
-
Should work fine, the via pico I use on two machines is only 1Gig CPU and the Atom I use is 1.6Gig.
Hood