AB ultra3000i drives going bad

[Robofrog]

Hello all,
I am brand new here . I'm limited in my controls knowledge but always trying to learn more.

I recently started working at a new manufacturing facility and one of the assembly lines I have aquired is having issues with AB ultra3000i servo drives failing. Apparently these drives are failing multiple times per year and are expensive to replace.

The drives are on a machine that has light curtains that are arranged in a manner that they are accidentally being interrupted during servo motions. The narrative I've heard is the light curtains are being interrupted so often therefore the drives fail prematurely.

Some other things I have noticed:
1. The enclosure for the drive is fairly warm. There are vents on the enclosure with a fan that appears to no longer work.
2. There are 9 bowl feeder controllers crammed in the same panel. I'm assuming this could create some noise?
3. I don't see ac noise filters. Don't know how critical these are.
4. There was some suggestion that the light curtains are cutting the 460V instead of using 24V to signal stop. Im trying to read the documentation to understand how it is typically wired. I assumed there was only one way to do this.

Can anyone point me in a direction or things I can look at?

 

[robertmee]

If the light curtains are killing the line side 480V to the entire servo, it could prematurely reduce the life. Had a customer doing the same with powerflex 40s, and they were losing them all the time.

The Ultra has two sets of power inputs. L1, L2, and L3 are the only ones that should lose power on safety. L1Aux and L2/NAux should remain powered.

 

[Robofrog]

If the light curtains are killing the line side 480V to the servo, it will prematurely reduce the life. Had a customer doing the same with powerflex 40s, and they were losing them all the time.

The Ultra has two sets of power inputs. L1, L2, and L3 are the only ones that should lose power on safety. L1Aux and L2/NAux should remain powered.

I believe this may be the case. What is the alternative to killing the line side?

I'll check that the other inputs are still powered tomorrow.

 

[robertmee]

I believe this may be the case. What is the alternative to killing the line side?

I'll check that the other inputs are still powered tomorrow.

Not sure....the Ultra3000s are pretty old, before the electronic safe off features of modern drives. Given the drives age, the charge capacitors in the drive are likely stressed from repeated power cycles, and probably the culprit of failure. You need to investigate the actual contactor to the line side, see what power its dropping....im not certain if bringing the drive to a safe stop before releasing the contactor is possible in your scenario but it may help. However likely not enough to redo the setup. It's been awhile since I messed with Ultras. Hopefully someone with more direct experience will chime in.

 

[Robofrog]

Not sure....the Ultra3000s are pretty old, before the electronic safe off features of modern drives. Given the drives age, the charge capacitors in the drive are likely stressed from repeated power cycles, and probably the culprit of failure. You need to investigate the actual contactor to the line side, see what power its dropping....im not certain if bringing the drive to a safe stop before releasing the contactor is possible in your scenario but it may help. However likely not enough to redo the setup. It's been awhile since I messed with Ultras. Hopefully someone with more direct experience will chime in.

Confirmed we drop the 480V to L1/2/3 but the aux remains powered. I guess this line used to be enclosed and was consolidated and light curtains added. Feedback from AB was the drives are being powered on and off too often.

 

[FilthyTech]

When you say that they are fairly warm, does warm have a number? Heat is the killer of electronics. We have many of these same drives, but all of our enclosures have an A/C unit to keep the temps in check. When they fail or fault do you get an error code?

 

[joseph_e2]

Frequent power cycles will kill a drive prematurely. I've seen it with Lenze servo drives as well. The safety stop was dropping line power and they gave the operator ability to use 2-hand controls to "bump" the machine with the guards open. That meant that the drive was powering up every time they tapped the 2-hand buttons and dropping when they released them. With about 10 in service, we were replacing 2-3 a year. The post mortem reports always referred to blown bus caps and inrush resistors.

It may be time to replace the drives with ones that have STO capability built in. That will save the input stages. Newer drives, and the engineering to migrate to them, are expensive but less costly in the long run than replacing the old ones repeatedly.

 

[Robofrog]

When you say that they are fairly warm, does warm have a number? Heat is the killer of electronics. We have many of these same drives, but all of our enclosures have an A/C unit to keep the temps in check. When they fail or fault do you get an error code?

Don't have a number at this point. The surface of the drive is warm to the touch. I'll take some readings today. The error we have gotten in the past is no power DC bus.

 

[Robofrog]

Frequent power cycles will kill a drive prematurely. I've seen it with Lenze servo drives as well. The safety stop was dropping line power and they gave the operator ability to use 2-hand controls to "bump" the machine with the guards open. That meant that the drive was powering up every time they tapped the 2-hand buttons and dropping when they released them. With about 10 in service, we were replacing 2-3 a year. The post mortem reports always referred to blown bus caps and inrush resistors.

It may be time to replace the drives with ones that have STO capability built in. That will save the input stages. Newer drives, and the engineering to migrate to them, are expensive but less costly in the long run than replacing the old ones repeatedly.

This is in line with the failure rate were experiencing here. I've started the task of justifying the cost of upgrading. Seems like the only option at this point.

 

[alan_505]

The replacement for the U3000's is the K5100's, they have hard wired Safe Torque Off.

 

[Robofrog]

The replacement for the U3000's is the K5100's, they have hard wired Safe Torque Off.

Thanks for saving me some time!

 

[kolyur]

Our standard practice with servo drives or VFD's that lack electronic STO is to wire the line side directly to mains power, with proper fusing of course, and put a contactor on the load side of the drive--between the drive and motor. A safety relay with off-delay contacts is used to cut the power to the contactor coil a short time after dropping the enable signal to the drive. The off-delay function is critical, as it ensures that no current exists on the load side when the contactor opens.

The advantage of this method is that the frequency of E-stops or light curtain trips becomes moot; the drive isn't affected by a load side disconnect as long as it isn't trying to source current. We have used this method for 20+ years on hundreds of drives and have experienced very few failures.

 

[Robofrog]

Our standard practice with servo drives or VFD's that lack electronic STO is to wire the line side directly to mains power, with proper fusing of course, and put a contactor on the load side of the drive--between the drive and motor. A safety relay with off-delay contacts is used to cut the power to the contactor coil a short time after dropping the enable signal to the drive. The off-delay function is critical, as it ensures that no current exists on the load side when the contactor opens.

The advantage of this method is that the frequency of E-stops or light curtain trips becomes moot; the drive isn't affected by a load side disconnect as long as it isn't trying to source current. We have used this method for 20+ years on hundreds of drives and have experienced very few failures

Wow this is a great idea. Would the delay increase the stopping time / safe distance?

If the servo was in motion and light curtain is blocked would this not have the same effect?

Apologies if the answer is obvious. I'm trying to learn and it's a lot all at once.

Edit: I might understand ... Even if the drive was in motion the off delay would ensure there is no current when the contactor opens?

 

[kolyur]

The method I suggested would not increase stopping time; in fact one could argue that if done properly this is the quickest way to stop a drive safely.

One problem with opening the line power as a safety stop, besides the longevity issue, is that the large capacitors on the drive can allow it to keep functioning for some time after line power is cut. I recall seeing a machine years ago where the Estop opened contactors that cut the line power to a series of VFDs. When you would hit the Estop, the machine would keep running for 1-2 seconds because it took that long for the capacitors to drain.

The behavior of the load side contactor is dependent on the function of the enable input. With the Ultra 3000, I believe that if the motor is moving when the enable is dropped, it will ramp down the motion before cutting the power (this may be configurable in Ultraware). With other brands of drives, the hardware enable simply cuts power immediately and they leave it up to the user to decide how the motion should be stopped. In either case, you would need to calculate or measure how long it takes for current to shut off on U,V,W and set the off-delay accordingly.

Determining the best way to put a machine into a safe state is not a trivial task. The traditional method of simply dumping everything the instant a safety interlock is tripped is sometimes not the best route as far as personnel safety is concerned. With your servo application, you'll have to determine the safest action when the light curtain trips. Can it coast to a stop, or is the hazard close enough to the curtain that someone could get pinched? In this case the servo would need to ramp down, or it could be that a light curtain is not an acceptable solution. This is where risk assessments come into play.

 

[OkiePC]

Wow this is a great idea. Would the delay increase the stopping time / safe distance?

If the servo was in motion and light curtain is blocked would this not have the same effect?

Apologies if the answer is obvious. I'm trying to learn and it's a lot all at once.

Edit: I might understand ... Even if the drive was in motion the off delay would ensure there is no current when the contactor opens?

There are two ways I have seen this implemented, but basically you want to command the servo to stop immediately when the safety relay is deactivated and after the time delay, drop out the contactor.

If the normal (untimed) contact breaks the ENABLE circuit to the servo/vfd, it will "let go" of the load which can be bad or very bad for some applications, but might be fine for others.

If the normal contact breaks the RUN circuit or otherwise, triggers a rapid stop of the servo/vfd, it will try to stop the load as fast as it can before the contactor opens and the load is disconnected.

In both cases, you want to give the drive time (even if it is just a fraction of a second) to stop sourcing current to the motor before breaking that current with the contactor.

Either of these methods will require thorough testing and a thorough review or repeat of a safety analysis.
I support this method of control because in many situations, it is the best solution and can improve safety as well as the reliability of the equipment.