Position loss with UC-300ETH

[Derek]

Hi Derek,

Yes, exactly.
That firmware acts as a counter and it does nothing else just counts the step dirs.
So, if you have a UC100 you can use it as a counter device and so it can count the step dirs and so it can tell you if the other motion controller, another UC100 or UC300 or UC300ETH etc. sent the correct amount of step dir signals or not.

Do you think it would be possible to feed the step dirs to the UC-100 and the drive at the same time having the UC-100 counting while the system is moving?

[cncdrive]

Hi Derek,

I did not think about this yet, but yes, it should work.
An idea I just had is that you could make a crimped cable with one IDC connector for a UC300ETH-5LPT or DSUB25 male for a UC100 and this would be the controller used to generate the motion,
and you could crimp a second DSUB25 male connector on the same ribbon cable to connect a UC100 which measures the step/dirs and you could crimp a 3rd connector to connect to your BOB.

The UC100 inputs have pretty low impedance, so one UC300 or UC100 etc. can easily drive out many of them and it can also easily drive an optocoupler plus an input of another UC100.
So, I see no problem doing it.

There is one possible problem I see though is that if there is noise or some temporary short circuits (circuit defect) on the BOB or on a drive's input which sucks the voltage level down on the driving UC's output, that will cause the measurement to drift while not the driving UC has the problem. So, I see a possibility for measurement results misslead.

[Derek]

Installed a new NIC and cable with fresh drivers. Still loosing position.

Works fine with USB
Not so good with ETH.

Derek

[cncdrive]

Hi Derek,

I was thinking about this, but I don't know yet what could be the problem.
One thing I've noticed is you using the M44 motherboard and I think we probably don't run final step/dir measurement tests with that motherboard, however there are common codes, but maybe there is somekind error about that motherboard's own code part... but this is yet just my thinking ... speculation, based on that most users running the -5LPT and we did not receive this kind of problem reports at all.
So I think we should run a test with it to see if we see any lost steps.
As said my collegue is currently out of the country for his brother's wedding who leaves in another country and I think we can check this out properly when he will be back.

Meanwhile if you could post the profile file would be great and if you could post a g-code file with which you sure experienced this issue.
And if you have any other computers lying around to test with to see if the issue is reproducable on another computer too would be also very helpful.

[CT63]

One thing you might check is take a look at the communication statistics graph. The button is in the profiles page. I don't see how the communications can fail only in one stepper direction, but its worth a shot.

[Derek]

At this point I'll wait until you guys can test it. I moved everything back from my lathe to the mill and I'm back up and running.

I've decided to transition away from CNC4PC boards at this point as they have become time vampires for me. I stopped counting the hours and now counting the days I've wasted trying to get things to work. The only C62 that is reliable for me is revision 2. All the others have given me headaches. The C32/UC-300USB in the Alliant has been flawless for years.

Balazs. Do you run multiple HDBB boards in your builds? If not what are you using? VFD control, relays, inputs???? I'll happily buy whatever is working for you.

You can have the best motion controller in the world but if you can't get reliable IO to the machine what good is it? You seriously need to put together something.

A very frustrated Derek.

[cncdrive]

Hi Derek,

Yes, we running multiply HDBB2 boards in our installations where many I/Os are required.
The HDBB2 has relay drivers (NPN outputs) on the upper 4 outputs, so they can externally connected 5V, 12V or 24V power relays directly on the upper 4 output pins.
And so the same relay drivers can directly drive VFD input pins, because they can be interfaced to these higher voltage levels also.

And the HDBB2 just passing the step dir signals through, I mean it does not do anything with those signals, just a stright wire through,
so you could even try (if you have some time to do so) a very simple test to connect a -5LPT motherboard to the UC300ETH and connect the step dirs of the drives directly to these pins just for a short test.
It would be even enough I think to connect up only the axis which you have the problem with, that is basicly 3 wires only (step, dir and gnd) to see if it looses steps or not this way.
If not then we could go on from there and then using the HDBB2 will also not loose steps, because as said it just passes through these signals just like a direct wire connection in the test I adviced now.

And I agree with George (CT63) that you should also check the latency test on the Config/Profiles/Statistics window, because maybe the issue is with the computer's LAN network,
but the latency test can quickly show if that is the issue. It is not good if the signal time goes to the very top to 20msec.

And I'm sorry for the time wasted, I wish I could help more and better.

[dezsoe]

Hi Derek,

just an idea, but in the past I saw a very similar issue. Finally, the problem was the DIR signal latency. When the axis had to turn back, the DIR signal arrived a little bit late, so the first step went to the wrong direction. This latency occured only in one direction, so the same happened as on you machine. Maybe, you could test this with 50 or 25 kHz kernel frequency. If it works well on lower frequency then maybe your problem is the same.

[Derek]

Hi Derek,

just an idea, but in the past I saw a very similar issue. Finally, the problem was the DIR signal latency. When the axis had to turn back, the DIR signal arrived a little bit late, so the first step went to the wrong direction. This latency occured only in one direction, so the same happened as on you machine. Maybe, you could test this with 50 or 25 kHz kernel frequency. If it works well on lower frequency then maybe your problem is the same.

Thanks for the info. I just ordered up a HDBB2 and the assorted other bits and will install the ETH in my lathe for testing. The mill is going to be tied up for a few weeks on a job.

[cncdrive]

Hi Derek,

Sounds good. Did you receive the board yet?
If you will need any advice installing it then just e-mail me or post it here please.