Wiring Harness/Robot is ghost!

[Robotic Metallizing]

IRB 2400 S4C Controller

I was wondering if anyone has ever had an issue with run away axes?

I will move the robot in joint, linear or reorient mode and I will release the joystick after being held in one direction. Upon the release, an I mean complete removal of my hand from the stick,  an axis or axes (particularly the wrist of the robot, 5/6) will rotate on its own and throw an error of Non optimal movement which will result in the controller continually repeating the error. The only temporary remedy I have found to clear the error or stop it from repeating and beeping is to power down the controller and reboot. I should note that if I move any axis in joint in either the + or - direction and release the joystick quickly the joints 1,2 or 3 will act as if they impacted something and jerk back in the opposite direction of my joystick commanded move. This only happens when moving the joystick in one direction for the aforementioned axes but it could be the = or - direction (I will post and state the direction for each as I do not have my notes in front of me at the moment.)

In addition to this problem when I rotate axis 4 to its 40/-40 (+/- 15 degrees) position, I will trip another code of the payload being too high.... there is not anything mounted on the flange! (I removed all equipment to test the robot and have in the meantime substituted a separate system.

ABB seems to think it is the wiring harness, which I have ordered a megger to test. But is it possible to get these type of result from a failing main board or drive controller board located vertically in the back of the controller cabinet? The process that the robot is used for emits high intensity UV light and may have broken down the wiring insulation...and then again it is an old robot.

To remove the wiring harness I am assuming I would need to remove the 9 pins from the connector blocks in the upper part of the arm, how do you get the pins out? Are they Molex .062" pins? I've purchased numerous pin install and removal tools but have yet to find the correct one to fit in the block but around the pin.

Thanks for the help and I know there's a lot here to swallow. I can break down into separate topics if need be but I wanted to get this up ASAP otherwise I am looking at a ~200-300 program generation programming spree and screaming customers.

Power to the People!

-Zack

[TygerDawg]

"a lot to swallow" indeed, need more description.  I'd start with the basics:  (1) pull, clean, & re-seat all boards  & amps;  (2)  swap the pendant to see if the problem is there;  (3)  pull encoder connectors, clean, re-seat;  (5)  examine closely all cables (that you can see) for wear & damage;  (6) then go to the wiring harness.  Old robots are just that, and they don't last forever.

200-300 programs?  I might have a solution for that, check it out.

[prnuk2003]

Also check that all the brakes are releasing fully (sometimes you can hear a noise when jooging an axis as the brake rubs).
I have seen when one brake is not quite releasing fully it has strange results on how the jogging behaves. Try a test program using AbsJ moves one axis ata time putting in +90 then -90  then 0 for axis 4 then 5 then 6. Then more AbsJ with +20 -20 0 for axis 1 then 2 then 3 (depends on how far you can move each axis).
This may find out if one of the axis is causing an issue or it may give another error (joint collision or torque related).

It could be the teachpendant but normally if a joystick has a problem the robot will keep moving without error until the dead mans handle is released.

Make sure also that the com offsets are all 1.570800 but that problem was more common on older M94a robots.

[jimvw57]

Try replacing the SMB cable from the robot to the controller.

[Robotic Metallizing]

I've got my work cut out for me. Megger should be in tomorrow. I need to get the booth with the substituted robot running with a production job before I can start diving into figuring out the root causes. I will post my progress so please check back momentarily.

TygerDawg - when you say clean the boards and connections and re-seat what would you recommend using to clean the connections and how would you go about doing so?

Prnuk2003 - com offsets meaning having to do with the communications, I'm not sure I'm following you here, please explain further. I did at first start messing with the absj commands to limit axes movement but I didn't think of listening for brake noise (it kind of escapes your thinking when the ambient shop noise is about 110 dB all day long. I'll have to look into that after hours)

I have dealt with joystick issues before... somehow the contact foils became crunched inside of the housing... I think one of the operators damaged it somehow but long story short I was able to disassemble the pendant and open up the joystick to flex the contacts just enough that it resolved the issue. I will take a look at all of the wires and re-seat them.

I will also try a different SMB cable as well just to rule that out.

Anyone think it could have anything to do with main cpu boards or drive controllers, or is that equivalent to pressing fast forward??

[TygerDawg]

"pull, clean, re-seat" :

The way I was taught to do it:

• Get a personal grounding strap & use it.  Properly.  Always.
• Power off, and go have a cup of coffee to give the machine time to discharge any stored energy or capacitance charges in the connectors.
• Pull boards carefully.
• Wipe with cloth to burnish the contact surfaces.  Optional to use a cleaning agent like isopropyl alcohol or similar that won't leave residue after drying.  In cases where the contacts have tarnish on them, using a "rough textured" cloth may be indicated, maybe even something like paper wipes would work for this.  Worst case option is to gently  use something like Scotchbrite pads.  Verify all particles and lint are removed.
• Dry, replace gently.
• In the case of cables with typical D-Sub connector pins, it's more difficult to clean off tarnish if any exists.  Examine with a light and magnifying glass is necesary.  I've used different  connectors on & off several times trying to get some gentle mechanical action on the pins or sockets.  Undetermined if this really works, but it makes me feel good.
• VACUUM  , never BLOW  the cabinet interior to remove all dirt, lint, debris.  You should do this regularly anyway.
  

[Robotic Metallizing]

Okay. I went through on Friday and cleaned out the cabinet with a vacuum and checked all connections. Pulled the boards while wearing the antistatic wrist band and made sure they were all seated securely and checked the breakers and such. I booted the console again and started loading the software but on the last disc 8(8) the system is hanging and the teach pendant is saying "Teach pendant communication lost". I cannot reinstall robotware.

I talked with ABB and they suggested trying the newer version of the software off of their ftp site which I did and the same thing resulted only this time right after inserting the key disc. I went back to the packaged software that came with the robot and the key disc is as far as I can get at this point.

After talking with ABB again they are saying to try swapping teach pendants which I will try after resolving some other production issues and if that does not resolve the issue replace the main computer board. P/N 3HAB 2241-1 and then the panel board if still no luck P/N 3HAB 7215-1.

Anyone have any other suggestions? I can't really progress with testing the axes for issues without a working controller.... Oh, I should also mention of the other robots I have none of the components are compatible with this controller so I can't do a swap...or so I've been told they are not compatible.

Please help!

[Robotic Metallizing]

So I sent the main computer board out for diagnosis and repair and it turns out that it has passed all associated tests.

Does any one have and idea as to the direction I should go in?

[cobree]

Hi give us a call, Robotservice.se
+46 (0) 70 6710763 in sweden ask for Mikael, he will guide you, and we have all the parts for ABB ships world wide

Br Conny

[King_Kingston]

For some reason, this sounds a little similar to an issue that I had a while back with an IRB140 with an IRC5 controller.  After inspecting the run chain, changing out the teach pendant, changing out the SMB and a few swearing fits, we finally replaced the harness and what do you know?  Success. 

An attempt was made to inspect the old harness and ring out (ohm) whatever we could.  However, it did not yield any positive results.  It may have been a broken wire that would only wig out when the robot was in a certain position.

[Skooter]

Hi Zack,

Your original post sounded like it may be the brakes weren't releasing.  You also mentioned the connectors.  I assume you're talking about the Souriau (used to be FCI, which used to be Burndy) Trim Trio connectors.  I use an old FCI # RX16D11D1 tool that has paid for itself many times over.  You can also use an Amp # 305183-R, it fits a little loose but you can wiggle it and get the pins out.

In your second post, you asked about the comm offsets.  Comm is short for the motor commutation.  At this point, those numbers will be correct once the software is reloaded.

In your third post, you don't mention what software you had and what you're trying to load now.  You also don't mention the errors you get when the software fails to load.  Which Robot CPU and EPROM set do you have?

Skooter

[Robotic Metallizing]

Thank you forum community for your time and responses, I haven't been available to consistently spend time troubleshooting the controller like I would like to but here is where I currently stand.

The main computer board was thought to have an issue, I sent it out for testing an repair and had it returned no work done after it was installed in a system and found to be working correctly.

The panel board was then thought to be the culprit, I swapped this board with a working controller and the issue of the robot ware not fully booting upon completing the installation on the controller persisted, so I will rule the panel board out.

To answer others questions, the brakes not releasing is a very good possibility as to my initial "too heavy of payload" error however I cannot listen closely to each axes as it moves if I can't boot the system so I'll have to remedy this issue first. This is not the first time I've had this mentioned as a possibility either so thanks for thinking to mention it.

As to the robot ware versions I'm working with, the original version the robot was shipped with is 3.0 and the version I downloaded off of ABB's ftp site is 3.2 Rev. 56. I currently have the first version 3.0 fully installed as in I went through all of the disks during the install process only when the controller restarts itself for the initial boot it never fully boots the system. The teach pendant will display "Welcome to the S4 controller WAIT: The system is starting up... Boot= iocbase (some numbers that i cannot make out from the pic I snapped on my phone)" After this is displayed for no more than 10 seconds the teach pendant will go blank and the system will hang. Inside the top half of the cabinet I noticed that one of the led on the right side toward the back wit the letter "EN" above it does not light up during the boot sequence, I'm not sure if this will help trace the problem but figured it was worth mentioning, all other LED's remain solid.

One other thing I noticed is that the cooling fans underneath the bleed resistors are not getting power I know some older controllers have a switch to detect when the cabinet is open and therefore cut power but I'm unfamiliar with something similar with the S4 controller. Could this disconnect in voltage for the fans possibly cause an error with the boot sequence???

[Skooter]

Regarding the fans, I just looked at this circuit last week and the 24V that runs the fans comes from the 24V Brake, they don't come on until the brakes release.

In the boot sequence, ioc stands for I/O Computer.  It resides on the Robot CPU board.  Is your Robot CPU a DSQC373?  If it passes diagnostics but won't load a known good set of software, I would suspect a corrupt flash which is located on the Robot CPU.

Skooter

[Robotic Metallizing]

Thanks for checking that skooter! I checked the board number its DSQC 363. It does seem to pass diagnostics just fine and I have tried two versions of software. I'm wondering if it would be worth tracking down and trying an odd revision in between.

I can send the board out for a free evaluation with the company that I typically deal with and see if they can get it running in their test rig and if not make the necessary repairs. I'll try an get this out today.

[Skooter]

Zack,

Write down the models of the Main CPU, Ext. Memory board, Robot CPU and the part # listed on the Robot CPU EPROMs.  Call ABB and check to make sure your setup is compatible with the 3.2 software.  ABB tech support has a spreadsheet that tells them software to hardware capability for the S4C.  I think your older DSQC325/363 might not work with 3.2, but ABB will know.  Ask them what software you can use with your hardware setup.

If you send out your DSQC363 for testing, ask them to try the same software rev that you are going to load.

Skooter