Messages

5661

Mach4 General Discussion / Re: Limit switches

« on: December 16, 2017, 06:32:27 PM »

Hi,

My home would be the corner of my material in that instance.  Does that sound right?

No. The corner of your material is 0,0,0 in WORK CO-ORDINATES. Homing is about defining a location 0,0,0 in MACHINE CO-ORDINATES.

Note that the 0,0,0 home location doesn't have to be in one corner or at the end of travel. It can in fact be anywhere within the machine boundaries.
Its normal to arrange your machine home to be at one corner of the machine limits because it causes less confusion for human operators.
What's important is that every time you home your machine it goes to exactly the same spot. With my roller plunger microswitches I can get to
within 0.02mm. If I had indexing on my steppers I could use both the home switches and index homing to get within 2um!

In the first instance FORGET limit switches, work on the best, most accurate, and most reliable home switches you can manage.

Craig

5662

Mach4 General Discussion / Re: Limit switches

« on: December 16, 2017, 06:21:03 PM »

Hi,
homing establishes the machine co-ordinate 0,0,0. What your proposing for the Z axis will result in uncertainty in the location of that position.

Lets say that you home your z axis by touching off and it happens to be 50mm lower than the extreme uppermost of travel. If your softlimt is -50mm
then all is well. If a program/MDI or jog attempts to drive the z axis higher than the -50mm machine co-ord limit it will stop and prevent a crash.

Now imagine the next session of Mach and you have a different tool fitted. Now your z axis machine zero may be 45mm below the extreme uppermost of
travel. If the softlimits are still -50mm then an upward MDI of 49mm would crash your axis.

The point is that for your softlimits to provide any protection the zero location MUST be repeatably defined. Quite frankly if it comes to a choice between
limit switches and home switches I choose home switches. It took nearly a year before I got around to fitting home switches and had suffered a number of
crashes. Since I fitted them I've had one only when I forgot to reference the machine....dummy! Fitting good home switches has proven to be the single best
thing I've done to improve the robustness of my machine operations and even better can come back to the machine a couple of days later and carry on the
same part from where I left off. With good home switches you can get away without limit switches. Without good home switches you'll need good limits
because you're going to crash into them a lot.

Craig

5663

General Mach Discussion / Re: Not connection (Breakout board te421+TB6600+NEMA23) with my computer

« on: December 16, 2017, 05:30:31 PM »

Hi,
run DriverTest.exe and post a screen shot of the result page.

Craig

5664

Mach4 General Discussion / Re: Limit switches

« on: December 16, 2017, 03:30:46 PM »

Hi,
you need all three switches. There are means to cause one switch to do two jobs and many do to economise on switches. In the early days of Mach
when everyone had a parallel port with only 5 inputs it was necessary to be a bit crafty to combine switches together. You have a PoKeys board with
lots of inputs so you don't have to do that nor would I recommend it. Combining switches can be confusing and if for instance you require a switch to behave
as a home switch then Mach will ignore it as a limit switch while homing is in progress. So just when you are likely to need it you suddenly have no limit switch,
not good.

The limit switches are at each end of the axes and really are safety devices. Provided they work their quality is not critical. Further if you have properly homed
(referenced) your machine and have correctly configured soft limits the limit switches will never be operated.

The home switch is a different matter. You will reference your machine on every start up and sometimes several times throughout a session. To have a repeatable
home position requires a top quality switch. Note that while its common to put the home switch at or near the end of an axis you don't have to, in fact the roller
plunger switches I'd recommend are best placed within the axis extents, where is a matter of mounting convenience.

I have three roller plunger switches as home switches, one for each axis and each on a dedicated input. The limit switches, all 6 of them are all hooked in series
and are monitored by one input on the BoB. If an over limit excursion occurs Mach will know that its happened and shut down but it won't know which axis
or whether its a positive or negative over limit. Given the number of inputs that I have to spare that's a bit silly....still there is 'nothing so permanent as a temporary
solution'.

If it comes to it don't worry about limit switches for the Z axis, if you just can't fit them in you can't fit them in, but you absolutely need a decent home switch.
If you can't reference the Z axis soft limits are meaningless and you WILL crash your Z axis.

Craig

5665

General Mach Discussion / Re: Delta servo and uc 300 5 lpt motion controller.

« on: December 16, 2017, 02:53:05 PM »

Hi,
the BoB you've already got will probably work. The documentation is very light on detail. In particular it does not specify the characteristics of the outputs 2-9.
To hook direct to your servo you want open collector outputs capable of sinking current and can withstand 24V. I suspect that they can't withstand that voltage
and will have to be buffered before it can be used. I think also that you'll probably need two of them.

Each one has only 5 inputs, you'll need two for the fault lines which leaves only three for all the limits, homes etc.

My I suggest that you try the one you have on the bench to see if will work OK and then decide if you want or need another one. The only way to
know if it can handle the 24V opencircuit voltage of the servo drive input is to hook it up and try it. It is quite possible that it will damage the BoB.
The safe way would be to hook up a transistor/mosfet or buffer IC. How are your electronic skills?

There are BoBs on the market that will handle 24V, they would have to be chosen with care but they are out there.

Craig

5666

General Mach Discussion / Re: Delta servo and uc 300 5 lpt motion controller.

« on: December 16, 2017, 12:28:26 AM »

Hi,
I think that if your servo has a 17 bit encoder then it will be absolute and if memory serves it has a resolution of 131,072 counts per rev.

Lets work backwards and say that for simplicities sake you wish to keep the pulse rate to 100kHz. I imagine you'll still wish your servos to achieve max speed,
ie 50 rev/sec. Therefore the resolution is limited to:
100,000 / 50 =2000 counts per rev.

The electronic gear ratio is 131072/2000= 65.536   If you read the manual the gear ratio can be selected with two numbers, a numerator and a denominator.
Both numerator and denominator have to be integer and less than 65536. The exact ratio could be achieved by selecting numerator as 65536 and denominator
as 1000.

Just as a matter of interest with a 2:1 reduction and your servos having 2000 count per rev resolution your linear resolution would be:
5 (mm/rev) / (2 (gear reduction) * 2000 (servo count per rev)) = 1.25um, pretty damn good. Your Steps Per, that you would put in the motor tuning
would be 800 step/mm. If the servos reach their max speed the axis velocity in the motor tuning page would be 7500 mm/min, there again that's great
speed for a hobbyist machine.

I imagine you'll have a breakout board between the UC300 and the servo drive. Unless you bought an absolute dog any breakout board should do 100kHz
in a canter. ger21's recommendation, a UB1 sounds good.
There should be no extra circuitry required between the BoB and the drive. Probably a good idea to use a shielded cable and don't make the cable
any longer than it need be. Microphone cable would be a good idea, its beautifully flexible and shielded to preserve low level signals, two wires plus the shield.
That would cover your Step and Direction signals.

As I posted earlier you will need some extra wires, they will not be high speed signalling wires. One will be a Mach output from the UC300/BoB to the drive
to enable it. There will be one wire from the drive to a digital input on your BoB/UC300 to Mach to flag a fault. The last wire will be an output from Mach
to rest the fault condition. Note this last wire could probably go to the fault reset input of both servo drives.

You'll need to program your drive. I see in the manual that Delta produce software to allow you to easily program you drive. Do you have it? My Allen Bradley
servo can only be programmed this way, and its very very good. It has a database of all the servo motor models with all the limiting values, physical, thermal
and electrical characteristics and makes choosing and defining parameters for your application easy.  Delta drives can also be programmed with the digital
keypad, tedious but doable.

The essential parameters to set are that you wish to use step/direction position control. You need to set the electronic gearing numbers. There will also be the
limiting values, things like max speed, max current, overload duration and others. Hopefully Delta's setup software will provide all of those.
You will need to assign the digital inputs, your enable and fault reset signals and one digital output, drive fault signal.

Most servo drives, I haven't read the manual closely enough yet, allow you to hook limit switches direct to the drive. As you know it is common in Mach to hook
limit switches to your BoB/UC300 to signal Mach and Mach shuts down as a result of a over limit excursion. With a servo drive you can, within limits, program
what you want to happen. The limit switches would be hooked direct to the drive and would stop the servo from exceeding the limit  and would prevent any jogging
which would take the axis further out of bounds. It will if you wish signal Mach on an output of your designation of the condition. It could trigger some automatic
response like rehoming as well. I've not heard any argument that suggests that hooking limits to the drive offers any compelling advantage over hooking the limits
to Mach. As I'm using my servo as a spindle I've not had to worry about it. My suggestion would be to hook your limits to Mach in the normal way to start with.
If a reason comes up to change that strategy do it then. Keeping it simple to start with is probably the way to go.

You may have noticed in the manual that you can directly jog a drive by using two inputs per your designation and those inputs could be hooked to physical buttons
via your UC300/BoB. Likewise you can program a number of positions that can be selected by reading a number of digital inputs of your designation. In short the
servo drive has many more options than is required for your lathe but could be a great deal of use in huge printing press for instance which could have hundreds
of servos for any number of tasks. That those options exist does not mean you have to use them but could be useful to you at a future time.

Craig

5667

General Mach Discussion / Re: Z axis loss of cutting depth after multiple Z movements

« on: December 15, 2017, 05:41:49 PM »

Hi,
to be honest I think your G540 and Leadshine drivers are perfectly adequate.

The use of a good external controller which will get you away from the XP/parallel port bottleneck is the best use of your resources.

I use an Ethernet Smooth Stepper and two Homan Designs MB 026V boards. I have a very tight budget also and what I cant afford is to buy something
that doesn't work. The PMDX-424 is a highly practical board, it has its BoB interface builtin which includes it own power supply and has two port worth
of IO, so despite being $50 odd more than the ESS it probably works out cheaper in that it doesn't require the extras.

Craig

5668

General Mach Discussion / Re: Delta servo and uc 300 5 lpt motion controller.

« on: December 15, 2017, 02:54:20 PM »

Hi ger21,
whoops, I try to stay comfortably within the specs for this reason.

Do you have a scope? Would be nice to see the waveform at max speed, could it be input capacitance of the drive dragging the output of the BoB down
or is the BoB output just not plain fast enough?

Craig

5669

General Mach Discussion / Re: Delta servo and uc 300 5 lpt motion controller.

« on: December 15, 2017, 02:34:51 PM »

Hi ger21,
I have an ESS which is supposedly good to 4MHz! I suspect that this really only of significance as an input, who tries to signal to a servo drive
at 4Mhz?. I found it entirely enough of a battle to differentially signal my Allen Bradley servo at 480kHz let alone MHz. Even then I realised that I was
trying to achieve best possible resolution which is far far better than I could ever use and so made a rational decision about resolution and got back to
open collector signalling rates.

Natural inclination is 'I'm going to build this sucker to go at max speed with max resolution at max power' and then realise its just not possible or practical.
Rational decisions can reduce those hurdles to manageable.

Craig

5670

General Mach Discussion / Re: Spindle index

« on: December 15, 2017, 02:16:10 PM »

Hi George,
lathe threading is largely enacted by the motion controller, not Mach.

I have been using Mach4 and an Ethernet Smooth Stepper for a year and its still not quite capable of lathe threading. The issue is not Mach4 but the
Smooth Stepper, the plugin is not capable of it yet. Large chunks of the code to get it to thread are written and in place but there is a bug. The expectation
is that the bug will be fixed over the next month or so.

To my knowledge only Vital Systems Hicon board is currently capable of lathe threading. I may be proven wrong and other manufacturers have got it working.
Still, the upshot is that the plugins required for Mach4 operation are still developing and lathe threading is at the leading edge of that development.
Mach4 will not advance your ability to thread with a lathe....yet.

Craig