Posts by thibalte

Update after ABB technician intervention:

The DSQC 663 unit was dead, swapped for a new one and all smooth sailing.

Tech guy confirmed there was nothing wrong with mains supply nor did we do anything wrong, sadly the dead unit will go back to ABB for repair and enter refurbishing cycle without any chance to know what was the actual failure.

The problem seems to have been bad luck.. maybe manufacturing defect or something, the robot is 107 hours only.

Luckily we were still 1 month under warranty left so that has not been as stingy as expected, ABB replaced the unit at no cost for us.

Ah ok I see now why I was confused, sorry: it's a tripolar circuit breaker so I cannot tell which phase trips it.

I have only access to the location next week and will have an ABB tech coming to look into it, will report on their findings.

Per the IRC5 product specification, the recommended line fusing is 3x25A for 3-phase 400-600V on a 4600, (we are running it at 3x400V).

We've never had any issue whatsoever related to fusing before, only once with differential earth protection on an electrical install with 40mA protection, but most times it's been fine on 40mA. (mind you we run the robot for film productions in studios and on-location so most of the time it's a huge hassle to make the studio/client change their electrical install the day before a film shooting, the IRB4600 is specified as running fine on 40mA if no more than 7m long motor cable is used, above that and a 300mA protection is required). Anyway here it's the fusing that pops so unrelated.

Wouldn't the fuse have been popping way before when we did high-speed high-load movements? Here it's popping even when standing still with motors off

I obviously dont have a robot to test again, but when switching from Auto to Manual, shouldn't the "Motors OFF" state message be seen right after the "manual mode selected" message? There is 18s between the manual mode request and motors OFF state. Is that normal?

From my memory the motors should go off right away no? I cannot remember if they did go OFF when I switched, so I might just be wrong, but I was not pushing on the enabling device so it just seems weird that the motor OFF would be so delayed, am I mistaken?

and a detail of the error from the message log

hi Skooter,

I will have access to the machine only next week again, but this is a brand new IRC5 robot, manufactured in 2021 so I assume "latest" controller revision. What I am confused is that everything was working smoothly for the entire morning, and the only change was going from Auto to Manual, so i doubt that something would have happened right when i switched it. I will go check the incoming power once I'm back there next week to make sure.

When you mean is it the next phase each time, do you mean is the power connector wiring changing or something? We always use the same power connector, same we used many times before without any issue.

Only smoking gun I have is this screenshot of the TP log showing the manual switch, and then few seconds later the error once I tried to jog it. From then on I could not do nothing so I initially assumed some benign safety warning, like on our S4C+ it would give off warnings when the enabling device was just slightly depressed and only one safety channel would trigger too early from the other. But after restarting the controller (just a warm start) then it simply goes out everytime the enabling switch is depressed.

Uff, ok yes i also thought it would be crazy that its against procedure.

Fuse opens as soon as i depress the deadman switch, seems to be before the motors are even on since they dont even release brakes like when you start moving the joystick. From what i gather there was this bleeder resistor circuit error right after, with a warning that the dc link will stay energized. So i suspect that something must be shorted out or something, its pretty much impossible to troubleshoot further since i cannot get any error log as it abruptly goes off. Ive called ABB support, let’s see.

But good to know that there was nothing wrong on my side, i couldnt believe it would just go out on a mode switch..

To specify, its the fuse before the irc5 controller, from our mains supply, not a fuse in the controller that pops. Its not either the ground earth differential protection (fi schutz).

hi all, im quite puzzled by what happened to our 4600 IRC5, 1 year old, about 100hours of light use.

This morning after a project under EGM control, i switched back the robot out of Auto to Manual mode to jog it into a parking position. I turned the mode switch while the robot was still Motors On, and immediately got a fault “34257 Open circuit in bleeder resistor circuit”. I acknowledged and tried to enable motors for jogging and nothing, same error. I restarted the controller, no error but as soon as i enable motors for jogging the 32A fuse of the electrical cabinet pops and the controller instantly shuts down.

So 2 questions:

- there is absolutely nothing in the operating manual mentioning that the robot has to be motors off to do a mode switch, and if it could literally fry anything by simply mode switching during a run there would likely be tons of people sadly discovering it like me, seems to be that the controller should be able to go motors off just like you release suddenly the deadman switch. Still, is it considered proper procedure to always manually go motors off before switching mode?

- did this happen to anybody else? Is it a common thing? For a brand new robot to go dead on me by simply switching mode has made my confidence in it plummet since its meant for film production use and luckily it died at the end of a shoot and not at the start with a whole film crew and angry producer in the studio..

Hi,

I am investigating an issue on our brand new IRB 4600 20-2.50. When the arm is extended out forward ("leaning horizontally") we are observing very low rigidity around the axis 1, we can push with 1 finger more than 1 cm in the direction tangent to the axis 1, when pushing multiple times we can make it oscillate back and forth 1-2cm and with an audible clicking noise. Happens with servos ON and also with brakes. We had an ABB service tech coming, tried to adjust the servo on axis 1 by loosening the bolts and the backlash was still there, dismounted the servo 1 and we could see that the gearbox input gear could be moved loosely forward/backward a couple millimeters around the shaft. The robot is gone to ABB for repair but they are now saying that this is "in tolerances".

I honestly have a hard time believing it, the gear felt almost loose. The stated backlash on these gearboxes is approx 1 arc.min on the output shaft, can it really translate to looking like a loose input gear? The service tech said it was clearly abnormal but after 2 weeks of testing at ABB they are considering it "normal". The robot is brand new, 20 hours

If anybody with experience of "normal" backlash on axis 1 of the 4600 serie robot can chime in I'd be very happy to hear!

Thbault

Thank you very much for the precious help! I will study all the options we have then.

Ok, perfect. So that means besides replacing the computer nothing else needs to be upgraded?

So the current computer is a DSQC 639, apparently a DSQC 1000 won't fit in the bay as replacement? Is the whole IRC5 internal structure organized differently for RW6 systems?

That would be great thanks! I have found some Release Notes specifically on RW6.0.7 (just googling) and the Controller Hardware Support lists quite a lot of modules as DSQC10xx, like for IOs, Profibus, DeviceNet, RS-232.. not sure what version is in the IRC5 as I'm currently quoting for purchasing it so no direct access..

Alright, that's what I feared.. I have forwarded the robot key, waiting for a reply from ABB directly.

Hi,

Am am looking for confirmation that it's possible to simply swap out the main computer to a newer one like DSQC 1000 in an IRC5 controller, as I must upgrade towards RobotWare 6 but the current computer can only run RW5. Is it a simple swap? Is there anything else to upgrade regarding hardware?

Thanks!

Thibault

Saved myself the big bunch of money we're a tiny design studio working on creative robotics (aatb.ch) so fixing is preferred rather than costly replacement.. plus figuring/learning out how it's built is also our favorite approach so that we can best implement our various projects since we do everything ourselves and got out of art school!

Thanks! Was a great learning experience My Mc Gyver-esque fix should do for now..

So after much troubleshooting I got the robot back in operation.

Followed all procedures initially to check for any connector/cabling issue, insulation tested all motor phases, insulation tested all cabling up till servo drive connector, without any obvious fault, swapped motors/resolvers 4-5-6 and issue was always on axis 4 which lead me to strongly suspect servo drive damage.

Then I went slightly overboard and got the scope and meters out to study the servo drive unit, reverse-engineered all circuitry, checked pulse trains on the gate drive transformers, all looked fine. swapped the control board between the 2 drives units (the DSP-KORT), same fault, which led me to focus on the power/igbt section and by manually shorting the control pins for the half-bridge gate-driver while applying 24V to the DC rail, where the bus-bar usually goes, I could see 24V on each phases of the 2 other axises, but nothing coming up on the axis 4 section. Checked the board itself, could see all clear voltages on each driver, but no response to input signals, and by inspecting the power section which seems to be supplying the floating supply, fed to the 3 of them I finally found this little blown resistor. Popped it out, frankensteined an equivalent resistor with some through-hole parts, patched it in, booted up and working again like a charm.

Have to say I went slightly deeper the rabbit hole than I wished, but I got a much better understanding of the control circuitry running those beasts now!

Hello,

I have a IRB2400/16 with S4C+ bay that has been working flawlessly for couple of years, gone into storage for a year (clean, dry) and now when turning back on gives an error 38653, Drive unit curr controller saturated" on joint 4 as soon as we try to jog it. All other joints move fine.

I have checked cabling, measured 6.5R between UVW phases of the joint 4 up until the connector going into the Drive Unit 1, DSQC 346E. As motor 4 and 6 are same model I even swapped the UVW plugs between to see if the joint 6 would fail instead but it's still joint 4 failing which seems to indicate the issue would be more on the Drive Unit side.

The robot config was working fine before so I have big doubts the motor config is an issue. The Controller was fully disconnected while in storage so I am quite flabbergasted that the Drive Unit would die out of the blue without any fault or crash? Anybody hit the same problem, any advice?