Oversize Line reactor

YES! Today, after I put the 160 series drive in was the first time it faulted while I had the DC Bus volts on the display.

I had just set AUTO RETRIES = 3, DELAY = 3.0 . . .

So I got to see the DC Bus go from 740-somthin' back down to 630 over a period of about 3 seconds.

The other two times it faulted while I was staring at it (with the PF40 yesterday) it did fault with no run command (it was enabled).

Same problem, new drive, but I saw what looked like serious regeneration.

It was the meat that was missing.

When I tried to make it fault with a rapid decel ramp, I wasn't putting enough meat on the belt. When the inspection operator runs the system in full auto, there is probably 500 more pounds of meat on this belt! Doh!

So I wish I had thought of it while I was at work, but first thing in the a.m.:

1) Try coast stop.
2) If stop distance is insufficient, experiment with DC injection braking.
3) If #2 proves successful, monitor Heatsink temp and motor temp.
4) Order DB and 1 Pole OL relay.

Yes, I am thinking I'll put a DB on it no matter what, and I gotta get us a working scope. I could have ordered the solution yesterday if I had the right tools.

And two more drives are going in this MCC soon. (2 x 7.5HP) and they all need devicenet, so I can chart DC Bus easily and identify regeneration.

Thanks again...and I'll try to update what happens tomorrow.

We had a bus overvoltage issue on PF70's. we finally realized the line reactors were put on CE compliant drives. Removing the reactor fixed what was thought to be a resonance problem.

Not sure if this is relevant or not, but we recently had an issue with a 1336 that was tripping on over-current randomly, but often enought to be consistent.
This machine is sounds similar to yours, it goes from 3Hz to 85Hz, then down again, very fast. There is a braking unit or the drive, and this does it job.
The problem was, that the machine also had a 400V solenoid that triggered a clutch. Both legs of the 400V were being switched, whcih in turn (we think, in certain parts of the Sin cycle) would cause a spike (we think) on the AC line (the drive & solenoid are on the same AC line).
We changed it so only 1-leg switching, and the problem went away. This did have me scratching my head for a bit, but we got a solution in the end.

Not sure if this could be a similar scenario, but maybe there is something else interfering?

Interesting...

I put my meter on DC Peak Hold on the DC BUs terminals. Went away for an hour to work on three other things, and came back to find the max reading OL. So, my DC Bus is peaking above 1000vdc, and I have seen it fault while not commanded to run so I still suspect have transient on the 3 phase input.

I have not seen any strong signs of regeneration today. Even when heavily loaded, the DC bus barely breaks through the 650 volt level.

These drives have no onboard filtering. So, I am back to thinking I need input filtering and less sure about the need for a dynamic brake. What I saw yesterday that looked like regeneration was very possibly a rare fluke occurrence, and I could not see the machine from where I was looking at the drive in a remote MCC room, so I have not idea what the physical conditions were when it happened. By the time I got out there, the belt had run and emptied itself.

It's always a good idea to use Line Reactors for the drives, especially if a drive is located close to an MCC. I always install line reactors on drive's output if motor wiring length exceeds 300 ft.
Isolating transformer on the input side may help too. You can always ask your electrical power supplier to perform a power quality inspection at your location. They normally install a power quality monitor for a couple of days. Most of the time the service is free.
But it may not be what's causing your overvoltage problem. Short infrequent voltage spikes should not cause an Overvoltage problem. CD bus capacitors will filter it out.
Can you be a little more specific on the drive application. How long are the motor leads between the drives and the motors? Are they all in the same conduit?
It looks like you do frequent start/stop cycles. Can you disable dynamic braking and use coast to stop feature instead of ramp to stop?
You definitely need external braking resistors if there is any regeneration (ramp to stop) involved. Running a drive in regenerative mode will cause an overvoltage problem (DC BUS).
We had problems with PF70s when they sometime tripped (overvoltage) during frequent start/stop cycles. Braking resistors were installed properly. We ended up enabling DC Brake hold feature for 0.5 seconds to dump some energy into the motor and help reduce DC bus voltage.

OkiePC, please take a thoughtful moment to sort out what you are hearing on this problem.

First, motor lead length, motor lead reactors, and regeneration cannot be a factor if your faults occur when the drive is not running.

Second, if your meter saw voltages higher than the DC bus voltage faulting level (between 810 and 850Vdc), it also cannot be output side issues since the drive releases the output instantly when the fault occurs.

Third, this leaves only input supply issues since the bus voltage clearly continues to rise above the fault detection level.

Fourth, the cheapest solution to input side transients is to install line reactors but, unfortunately, line reactors drop the same % of voltage at rated current as their impedance %. So, if you have a 460V supply and a 5% line reactor rated 3amps, at 3amps the line reactor will drop 5% of the 460V supply which is 23V. Now you only have 437V left to run the drive which is not enough. With your supply running under 460V, you can't afford to loose any more voltage by adding reactors.

Fifth, the suggestion to use a drive isolation transformer is a good one. It filters better than a line reactor and, if you get one with taps, you could even tap up your voltage a bit to 470 or 480V on the drive which is much better for the motor at full load and full speed. The problem with a transformer is its cost and about 3% energy losses thru it.

Sixth, when I recommended you try braking resistors, I wasn't implying that your problem was regeneration. Brake choppers operate whenever the DC bus rises above their trigger level regardless of why. So, they will start to clamp the bus voltage on input surges as well as regeneration. Since they are cheap and don't waste continuous energy, I suggested you try these first.

Finally, to repeat, input line reactors are not going to be acceptable unless your motor never runs at full speed and full load. If you can limit its speed to 90%, you will only need 90% of full voltage, etc etc. Otherwise, the transformer is your second option after the resistors.

I hope I've made this a bit more understandable for you, OkiePC. I'd like to help you avoid making a new problem when solving the old one.

Good luck and advise how this works out.

This is a 2HP drive running a 3/4 HP motor (it was a 1HP drive, but we thought the increased bus capacity of the 2hp drive may help)

The drive in installed in an MCC bucket by us, not made that way. It is fed with a 7amp breaker which plugs into the MCC bus.

The motor leads are probably around 100' long and routes to the machine in metal conduit. There is a disconnect switch near the motor.

The design B TEFC stainless steel washdown motor runs a plastic conveyor belt that is about 12 feet long. I don't have the gearbox specs, but I think it is 40:1. The conveyor travels uphill in the forward direction, but the incline is very slight, maybe 3 degrees.

The operator feeds a stream of large chunks of beef onto the belt for inital inspection. There is a metal detector on the belt that will kick in the reject cycle (stop for three seconds, run reverse for 15 seconds, stop for a minimum of three seconds).

This belt was a continuous rubber-like material and was controlled with a reversing contactor until about 6 months ago when we replaced it with the segmented plastic belt. Upon startup, they were concerned about the hard stops and starts breaking this new style belt, so they asked me to add the drives which are basically intended as soft starters.

Right after installation we saw a couple of occurrences of the High DC Bus faults, but could nt reproduce the fault. I did not know until a few days ago that these drives have been faulting at least once a week since installation, and that it is getting worse for one of them.

Since our incoming power averages just below 460, I am a little concerned about dropping voltage to the motors with line reactors. They are already sized such that under full load, they run at about 95% of the motor FLA setting. Now, we can easily install 1 HP motors to overcome that, but I would like to be able to solve this problem without replacing the motors if possible.

So I am considering installing filters instead of line reactors. Will the 160-LFB1 (LF Filter) cause a voltage drop like the line reactors would?

EDIT: Okay they also sell a 160-RFB-xx module as well which is an RF filter...

EDIT: I wrote this before DickDV post above. The problem drive is set to run at 70Hz, we could go back to 60Hz, but not less than that, so the voltage drop is definitely an issue.

You will have to call your AB rep on the filter. My guess is that it is an L-C Pi configuration per leg with additional caps leg to leg, which will put inductors in the current path anyway. These will have some amount of votage drop, although I don't know if they will be as great as an equivalently sized reactor.

Have you thought of simply sticking MOVs or another sacraficial spike remediation technology into the circuit? I don't thnk they are overly expensive and that would certainly tell you if it is spikes. they also wouldn't drop any voltage.

Keith

Thanks Keith. The 160 drive I am using is a series C. The book says about this revision "Standard input voltage transient protection has been increased to 6 kV and contained in the drive itself. The optional MOV module is no longer required."


I got our old o-scope working and didn't see anything too bad on the incoming lines. I only have a 10:1 probe though, so I can't "zoom out" far enough to look for really high voltage spikes. With that probe, the sine wave goes almost from top to bottom, and has a bit of noise riding on it, but I have seen a lot worse.


Also, my earlier meter DC Peak reading was wrong. When I did not set the manual range (earlier today) I was reading into the 650volt area normally, but when I looked closely I could see that the meter range had defaulted to 600v...not sure why it showed 650 just fine and where the cutoff is before it shows OL, but I left it for awhile, and cmae back to view the peak and it showed OL.

So this last time, I manually set my meter range to 1000V first, then went into peak hold and recorded a peak of 757vdc after about 20 minutes.

So far, there have only been one or two faults today, but I have learned quite a bit, and have a working scope now...

Just a thought, is the transformer that is feeding this drive hard grounded at the star point or a resistive ground?

On a pflex 70 there are jumpers to remove to isolate the incoming voltage from ground. If you don't do this and there is a ground fault somewhere on the same transformer then the dc bus voltages can go snakey.

Found this out the hard way after seeing lots of smoke and smelly capacitor juice.

It's hard grounded. I have the stop mode switched to coast and have only had one fault in the last 24 hours and it was a phase W to ground short! It's running now, and running good. The short was probably our 3rd shift hose-monkeys washing out the disconnect with the drive running. It's standard operating procedure here to shoot water into every nook and cranny every night.

Anyway, the coast mode seems to have taken care of the regen problem (duh!), and I am going to check into isolation transformers today. There is a big (est. 36 x 36 x 12) enclosure labelled "MCR" right above the two buckets that need isolation or filtering, and I figured it was full of running hardware and stuff, but today I opened it up and it's spared out. It has two nice big 2" conduits connecting it the the MCC center, a backplane, and even a bunch of old timing relays on a din rail, like I need more of those in my "drawer b".

Paul

OkiePC said:

Anyway, the coast mode seems to have taken care of the regen problem (duh!), and I am going to check into isolation transformers today.

Click to expand...

I'd stick with a brake resistor. Cheaper than an isolation transformer and it will take care of your regen issues so you can control your stops.

I have some 2HP Pflex4 drives making rapid speed changes. I used to fight over-volt problems when the conveyor was loaded up and had to slow down quickly. Bought the recommended brake resistors and haven't had a problem since. When I was initially testing, I found that with the brake resistor I could put my ramp down time as low as 0.1 secs. with a fully loaded conveyor without faulting. Without the resistor, 0.4 secs. was about as low as I could go and I would still occasionally get a fault.