Hello Guest it is October 17, 2019, 07:30:29 PM

Author Topic: UC300 (USB) to Mach3 erratic connection losses  (Read 4093 times)

0 Members and 1 Guest are viewing this topic.

Offline Filou

*
  •  42 42
    • View Profile
UC300 (USB) to Mach3 erratic connection losses
« on: February 17, 2017, 05:45:38 PM »
I observe erratic  >:( connection losses between a UC300 and a Win 7  laptop box running Mach 3. No power RF sources (like plasma or torch) around, simply 3 axis actuated by NEMA 24 steppers.
Motor cables shielded with braid and separated from limit / home cabling which run in their own shielded braid. Single point grounding to avoid ground loops.
Connection losses do not seem to be influenced by machine activity, or the lack thereof.
I could be in IDLE state for an hour, or work for 30 minutes without any connection loss or be idle for 2 minutes and experience 5 connection losses in a row.
I do not seem to be able to find any pattern.  Hard to say if it is Mach 3 / the Win 7 box / or the UC300 motion driver. Tested several USB cables and double checked for ground loops.
Tried to up-date the driver of the UC300 - failed. Power management of laptop: never sleeps.

Anyone facing similar issue ?

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #1 on: February 18, 2017, 06:58:53 AM »
Lack of pattern forced me to proceed step by step. Monitored supply voltages ( 5V dc for UC300 and BoB + higher for drivers ) with oscilloscope and could correlate voltage glitches (drops) in 5V supply with connection losses.  Not easy to say which of V drop or connection loss occurs first so that swapping for another 5V supply is the easiest; the situation improves.

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #2 on: February 18, 2017, 04:57:50 PM »
the situation has improved, but the main issue remains unsolved:

the number of occurrences for connection losses has reduced (since the 5 V power supply to the USB unit is now stable AND drops are no longer registered when a connection loss occurs) but the erratic character remains (so does my mood !).

A LONG run (4H) with a reduced chain (Win 7 running Mach3 / UC300 USB motion controller / BoB / nothing else attached)  can, on occasions, succeed. However, sometimes there is a connection loss after an hour. Same sort of statistics when Mach 3 runs but is idle.

When more equipment (motor drivers and power supply) is attached and powered, there is an increase in the frequency of the USB connection failures. Thus, noise could also play a role.
I don't know how I can improve/modify this wiring: (BoB to motor drivers) lines are shielded and (motor drivers to motor) lines are separated and shielded as well. The first shield is connected to the metallic enclosure shielding the (UC300 + BoB)  and the second shield is connected to the ground of the mains (220 V one phase, powering the motor drivers).

I'm reluctant to connect these two shields, since the first one is shared by the 0V plane of the BoB printed circuit and the second to Earth.

Anyone with a fancy idea for Sunday + Coffee ? Wish you all a nice WE.

 

Offline Davek0974

*
  •  2,578 2,578
    • View Profile
    • DD Metal Products Ltd
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #3 on: February 19, 2017, 03:39:51 AM »
sounds like a possible ground loop somewhere.

I would first make sure (with a test meter) that the USB cable shield and therefore the UC100 is NOT connected to the machine ground at all, this should only be grounded by the PC - do this with all other wiring disconnected.

Now connect the UC to the BOB - the same condition must exist.

Now designate a point on your chassis as ground - this will be you star grounding point. The shields to the motors will connect here but NOT at the motor ends.

Connect the shields (if any) on the limits cables here too, but not at the switches.

connect the 5v supply to the BOB (power off), the USB shield must still be isolated.

At this point you have TWO systems of grounding - the control side (PC,USB,UC300) and the power side (BOB, MOTOR CONTROLLERS, MOTORS)

Imagine it as two circles - control and power, now each circle can be connected to supply ground and no problem as no loop is formed, but let the two circles touch (say by connecting the BOB to ground) and you now have a full loop and issues will arise

That should remove any ground loop effects, of course whether it makes a difference i dont know its how i wire my stuff :)
« Last Edit: February 19, 2017, 03:43:08 AM by Davek0974 »
Bridgeport Mill, Mach3 V062, CSMIO-IP/A controller, AC Servo Drives

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #4 on: February 19, 2017, 05:00:47 AM »
Howdy, thank you Davek0974 for your thoughts.

you say "At this point you have TWO systems of grounding - the control side (PC,USB,UC300) and the power side (BOB, MOTOR CONTROLLERS, MOTORS)"

and I totally agree, this is precisely why I was reluctant (opposed) to connect these two shields, as I tried to describe above.  I fought for many years with ground loop issues on other type of equipment, so I became cautious if not maniac ...

Overnight I tried to acquire some statistic for the failures, with a reduced chain: (PC / UC300 / BoB connected and running).
Occurrences are random, or at least I cannot distinguish any pattern.  The number of failures remains too large for any practical purposes.

Thus, I moved to another battle field, my old enemy, the PC ...
I realized that, for some reasons, TWO plugins for the UC300 were listed in the Plugin Config menu of Mach3. This state of affair may be due to my earlier attempts to up-date from a 2015 driver from Balazs, which failed.

Trashed the oldest plugin file, reboot, restart Mach3, reconfig,  and just started with long term tests using the reduced chain without the second "circle" as you nicely describe it.
I'll inform about the result soon.
Enjoy you Sunday and thank you for the input. Philippe


I

Offline Davek0974

*
  •  2,578 2,578
    • View Profile
    • DD Metal Products Ltd
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #5 on: February 19, 2017, 05:46:27 AM »
Thanks for the update , two drivers would certainly be a bad thing.

Hope the test goes well.
Bridgeport Mill, Mach3 V062, CSMIO-IP/A controller, AC Servo Drives

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #6 on: February 19, 2017, 09:13:43 AM »
much better with only ONE driver, but the issue is NOT GONE entirely. Typically one failure / 3000 G-code steps .... still no pattern

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #7 on: February 19, 2017, 01:49:08 PM »
failure rate increases at dusk !  Beer time.  Beside any joke it is hard to understand a moderate failure rate and a subtle increase as time goes by without changing anything in the configuration.
If failure is still caused by noise or interference related to the mains I could understand on a week-day, but on a Sunday the load variations are moderate (moreover we are not in a city).  Current connection loss rate higher than 1 / 100 lines of G-code ...

Offline Davek0974

*
  •  2,578 2,578
    • View Profile
    • DD Metal Products Ltd
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #8 on: February 19, 2017, 03:37:40 PM »
If yu have just the UC300 and BOB connected does it disconnect then?
Bridgeport Mill, Mach3 V062, CSMIO-IP/A controller, AC Servo Drives

Offline Filou

*
  •  42 42
    • View Profile
Re: UC300 (USB) to Mach3 erratic connection losses
« Reply #9 on: February 19, 2017, 05:15:55 PM »
Good evening,

yes, but less frequently as under Mach 3 operation,  but it does clearly happen. Right now for instance, as I moved the mouse to wake-up the screen. A long job was running flawlessly (UC300 + BoB only). It could have been the mouse event, but it could be an artifact. I remember having used the mouse for looking at some G-code with an editor while Mach was running without observing this behavior.

I'm still hoping to find some sort of pattern, correlation / anti correlation.

I observe other things that I dislike:

- during startup, Mach3 shows alternatively two screens, a red one with a Mach 3 logo and a version number X.X (literally) and the other one (more recent graphical style, I'd say)  with blue letters and clearly labelled version 2.0.

- in the Mach3 folder I observe the creation and up-date (dates match with most recent run) of a set of files named Test.extension. The extensions range from .aab, .aac, .crv, .d2c, .job,. to .xfa.
What a heck are these files ?

In the Config Plugin menu I tried to fire the "video Ver-1.0" as well as another "Flash-Flash screen something" since I do not know what they should do and I want to obtain the simplest possible configuration.  But my choices keep coming again after reboot, and I don't know if they are deactivated for sure if I only uncheck them.  Maybe we cannot fire these. Do you happen to know ?

I seem to recall the PC alone disconnecting and reconnecting itself to "something" that could be a USB (UC300) device. Same sound as when the PC looses its link to UC300 or if I remove a USB stick.

I'm not good at windows; is there a clean way to monitor the actual traffic on the USB bus ? 

If I leave the PC and Mach 3 running but IDLE I can also loose the connection.

BTW: PC power management is "never sleeps"