Messages

61

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 15, 2018, 05:27:55 PM »

Mick

I think you are doing the best thing by checking in the Denford forum.

The CSMIO was updated by CS Labs a couple of years ago, and the new version was externally labelled FP4. If you purchased it recently you will have the FP4 version and it will be so marked on the front. Either way you are presumeably using plugin version 2.910, which is the latest that supports Mach3.

None of this is material to your issue, I was just interested. I have the FP4 but use the v3 plugin for Mach4.

Allan

62

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 15, 2018, 04:37:23 PM »

I also had a rather frustrating afternoon. Went to use the bandsaw for a bit of domestic renovation and to my surprise and dismay  it wouldn't go. Opening it up I found a mouse nest in the base and the little so and so's had completely chewed through one of the internal cables: sheath, insulation and copper, the lot. It will be quite an awkward job to fix. Oh well, the joys of living in the English countryside.

Craig

Thanks for stepping in. I suspected the encoders would be 10000ppr. I have no experience of these dc servos and drives at all, so could do with your knowledge of this stuff.

Mick

The CSMIO cannot produce more than +/-10V however it is tuned, and that should cause your motor to reach maximum speed. So either the servo's gain is lower than expected, or the motor is current limiting for some reason. Does the servo drive have any diagnostics or  error indications? I guess I've been spoilt by modern ac servos and drives.

You need to be confident the drive is working and tuned correctly before you can sensibly tune the CSMIO. Your current settings sound very strange to me, I would expect the proporthional gain setting to be a lot (10x ?) higher than the integral gain.  But I would leave the CSMIO until you have resolved why the drive cannot produce the motor speed you expect.

Do you have the original or the FP4 version of the CSMIO ?

Allan

63

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 15, 2018, 09:38:12 AM »

Mick

Glad you are moving in a forward direction.

As you have movement you might like to do a quick check that you are moving the correct distance. For example, enter asimple move of 50mm using G1 G91 X50 F100 or something similar into Mach3's MDI line and check that movement is the correct distance. I suggest this just to check that your encoders deliver 2500 pulses per rotation as opposed to 2500 lines, as the latter would give you 10000 pulser per rev. Otherwise your arithmetic is fine. Be sure to use the pulses, rather than lines per rev in the CMIO plugin as well.

Allan

64

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 15, 2018, 06:24:22 AM »

Mick

I've just looked back and see that your system is totally different to what I have: You use a lathe, mine is a mill. You have dc servos, and mine are the more modern ac type. So I'm still hoping someone with a similar setup may jump in here and help.

One thing I notice is that you make no mention of your basic motor settings. Be sure before attempting any tuning that you have correctly entered your encoder counts per revolution in the CSMIO plugin, and have the counts per unit set in Mach4's motor settings.

Allan

65

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 15, 2018, 06:06:29 AM »

Hi Mick

OK, it sounds as if your drives are elderly and don't feature auto-tune. Also, as they were working properly before, you are quite right to assert that they should already be optimally tuned, so we will assume that this is so.

Yes, the CSMIO will need tuning to match your system: there is no default here as all systems differ. As I suggested above, you should begin by increasing the proportional gain. This serves 3 purposes: It demonstrates that the system is functioning at a basic level, it enables you to ascertain that the encoder feedback is of the correct polarity, and it provides a basis from which you can proceed to auto-tune or manually tune the controller.

Allan

66

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 14, 2018, 07:34:29 PM »

OK. I had a quick peep at the manual and see that the ports and pins you mention are ignored by the controller.

It could be tuning, as it sounds as if the drives are powered and enabled. You should tune the servo drives first. This can normally be done using an auto-tune utility.

Once you have the servo drives tuned, begin with the IP-A. Progressively increase the proportional gain. For now the other coefficients can be set to 0. The motors should start to feel stiffer as the servo takes control. If the motor starts and continues to the end of travel, you probably have the encoder feedback set incorrectly and need to reverse its direction in the plugin. Once you get to the point that the motors feel reasonably stiff yet remain stable, try using the IP-A's auto-tune feature, which I have found works quite well. You can always tweak things manually later if need be.

Allan

67

Mach4 General Discussion / Re: Problem running XY Offset Calibration Probe

« on: July 14, 2018, 03:39:14 PM »

The default is just an arbitrary number. In my system it would have been 1mm! Seriously, the diameter can be any size you choose, and the time it takes doesn't detract from correct operation.

Allan

68

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 14, 2018, 03:33:24 PM »

Sorry, I have no knowledge of Mach3 so maybe someone else can help here.

You sound to have the basic connections correct, and the drives must be enabled as you say the battery causes movement. But why are you using step and direction Pins? The CSMIO/IP-A is an analogue controller, so don't you need to set up the Ch0 analogue output within the CSMIO plugin to get it working?  Similarly, what is the port 10 you mention for the encoder? Again the encoder outputs from the drive connect directly to the IP-A and  settings are handled within the plugin, not Mach3.

I may be wrong, but the pins you mention seem appropreate to the CSMIO/IP-S rather than the IP-A. Presumably you are using the correct plugin and it gives you the required settings for the IP-A?

Allan

69

Mach4 General Discussion / Re: Problem running XY Offset Calibration Probe

« on: July 14, 2018, 02:43:57 PM »

Did you look in the History window to see if an error such as "Probe obstructed" or "No contact with probe" immediately preceded "Can't transition from PROBING to OPERATION"? The latter may well have occurred simply because Mach4 had by then been disabled.

The controller is bound fairly closely with probing, and I don't use the ESS, but I have found that with certain probes the technique used in the mcProbing.lua module to determine the success of a probe strike does not work. This might explain why approach speed etc. seems to be a factor for some users. The correct way to test for a strike is to compare the commanded destination with the reported strike point. If a strike occurs within the commanded distance, these will differ. If no strike occurs, they will be (approximately) equal. I say approximately because I have found that they can be subject to small rounding errors such that testing for equality will invariably fail.

If you would like to try this approach, here is the code I personally have found reliable:

Code: [Select]

function Probing.CheckProbeStrike(ok, ProbeTo, ProbeReg, Hit)

-- AW Test for strike using position data as Probe signal can give the wrong indication
-- AW ProbeTo is commanded probe position
-- AW ProbeReg is the register containing the recorded strike point.
-- AW Units of ProbeTo and @ProbeReg must match i.e. user or machine units
-- AW Hit = non-zero indicates that a strike is expected,Hit = 0 that probe should remain clear

if not ok then; do return false; end; end
local inst = mc.mcGetInstance()
local EndPoint = mc.mcCntlGetPoundVar(inst, ProbeReg)
--local probeHit, rc = mc.mcCntlProbeGetStrikeStatus(inst) -- This doesn't work
--if (EndPoint == ProbeTo) then -- This doesn't work
if (math.abs(EndPoint - ProbeTo) < 0.0001) then
-- Probe failed to strike
if (Hit ~= 0) then
mc.mcCntlSetLastError(inst, "ProbeTo: " .. tostring(ProbeTo) .. "  EndPoint: " .. tostring(EndPoint))
mc.mcCntlSetLastError(inst, "Probe: No contact with probe")
ok = false
end
elseif (Hit == 0) then
-- Probe struck
mc.mcCntlSetLastError(inst, "ProbeTo: " .. tostring(ProbeTo) .. "  EndPoint: " .. tostring(EndPoint))
mc.mcCntlSetLastError(inst, "Probe: Probe was obstructed")
ok = false
end
return ok
end

My probing module is highly customised, so you would need to adapt this and the existing CheckProbe function calls but it should suffice to give you the general idea of what is needed. Take note that this must only be used where a G31 has been issued: at other times it is  appropriate to simply test the probe signal as is done in the standard module. You could initially try this approach for the calibration functions and expand it to others if you have any success with it.

Allan

70

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: July 12, 2018, 04:55:53 PM »

No, you don't need to supply 10V. The 10V line you mention is an output that you can optionally use to power potentiometers or the likes.

The CSMIO servo outputs are the 6 +/-10V outputs and you should connect one of these and its associated 0V to the drive you are using.

You will also need to wire the drive's encoder outputs to the IP-A's encoder inputs for the same channel.

Allan