Voltage spike on rotary phase converter? 2

[chipwitch]

Hello all. I have a 25 h.p. rotary phase converter that was purchased new about 3 years ago. Since then, we've burned up the idler motor twice. Both were new Reliance induction motors. The phase converter is of the transformerless variety.

I'd like to know, is there anyway one could wire up a 3 phase motor to 240V, single phase, using only capacitors, that could cause a voltage spike sufficient to burn out the stator winding? I don't think it's inrush current burning up the motors and that's probably the obvious suspect. I want to consider the voltage as the cause.

The converter is never started under load, electrical or mechanical. Once running, the load is balanced within 5%. All machines operate without one iota of problems. Most are manual machine tools under 6 hp. One machine is a CNC mill with a 10 hp spindle. The mill is loaded with built-in electronics, onboard computer, etc. The mill does have a vector drive and it's been suggested that harmonics could be creating a problem. While I'm not disputing that, I am reluctant to commit to it as the cause. In fact, I believe this problem is happening at start up. Both times the motor burned out, was shortly after startup. The last time, I hadn't even turned on a load.

Here's the part that bothers me. A varistor circuit protects the CNC mill. They've never faulted, except for one time. I had inadvertently thrown the mill knife switch to on before turning on the RPC. When the mill didn't power up, I quickly realized my error and switched on the RPC, forgetting to turn the mill off beforehand. The varistors blew immediately. Since the varistors can only blow during an over-voltage condition, I'm wondering what could have caused it? There IS a 3 phase buck boost transformer in the mill. I don't know if that matters.

I did discover some wiring errors in the RPC control box, but I'll withhold that information for now. First, I want to see what you think. Is there anyway to create a momentary high voltage with just a motor and some caps? Sufficient to burn up the insulation on 18 gage windings? Assume nothing wrong with the wiring from panel to RPC.

 

[Skogsgurra]

Is there any smell from the converter during operation? Does it run hot? Does the insulation look "burned" or did it just fail?

In the former case, I would suspect current form factor. Most drives, vector or not, do not have any PFC, so the current drawn from the three-phase has a rather bad shape. Too much third harmonics (zero sequence) in the current means that the speed of the converter is reduced and the idler needs to work more than it should. Compare to DC braking during operation.

Check if the load current is OK, you need a scope and a current clamp for that. Ideally, it should be a nice sine. Also check if the converter's speed is OK. If below rated speed, it indicates harmonics braking.

The varistors - are they connected line-line or line-ground? Do you have impedance ground or solid grounding?

I find it unlikely that overvoltage is the problem. Most latent failures "self-indicate" during start when the current is high. A problem that has developed slowly over time will eventually show when extra stress is put on the system during the next start. It is then easy to conclude that the start caused the problem. But it isn't always so. Might be, but more often not.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[waross]

This is not an absolute indication but a burn-out due to over current will often result in the entire winding being overheated. A burnout due to over voltage will often show local arcing damage in the area of the initial insulation failure.
Delta connected motors are not common in north America but there are some sold. The effects that Gunnar mentions may be much worse if the motor is connected delta and you won,t be able to measure the extra current easily as it will be circulating internally.
Have you investigated the possibility of running the vector drive on single phase?
In North America, on 240 Volts, a wye connected nine lead motor motor will have two leads connected to each incoming line and three leads connected together and insulated. A delta connected nine lead motor will have three leads connected to each incoming line.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[chipwitch]

I forgot to mention it's an open frame motor. No smell, no external signs of heat whatsoever. I took the motor to a shop for an estimate. He wrote up the failure as "winding insulation failure". Clearly visible at one location once the armature was removed (blackened). I assume that is the insulation to which you're referring. All feeder insulation looks new.

Perhaps I should have had a tach on it, but the rpm's "sounded" about right. It was always smooth and just a quiet hum, like one would expect.

Right now, testing isn't an option. The control box is going back to the manufacturer today. They're going to go through it, repair what needs repaired and check it against a new motor. I'll make sure they scope it.

Varistors are connected leg to leg. I don't know what "impedance grounding" is. Just googled a quick description. Everything is grounded normally as per typical grounding methods. There are no added devices in the grounding, unless there is something in the mill enclosure?

Harmonics was our first plausible focus. It wasn't until it occurred to me that the varistors had blown during start up. Before thinking about the varistors, I would agreed with your point on latency manifesting during start up. But varistors? Could that same latency somehow damage the varistors causing them to blow? They won't blow without voltage exceeding their rating as far as I'm aware. I'm just looking for clues.

 

[chipwitch]

Bill, this is a delta wound motor. All were. The two that fried and the one I have for re-installation. I have considered running the CNC (therefore, vector drive) on single phase. I've spoken to a company that sells parts for and maintains my line of machines. I can get a transformer replacement for single phase. It's supposed to be a one to one swap. The only part that confuses me about that is the nameplate on the vector drive says 3 phase input. If I replace the transformer with single phase, how will that affect the drive?

The burnout is clearly focused. Windings all look new otherwise. Of course, you can see where the plastic sleeves melted after insulation failed and the short caused overcurrent, but only in that winding. I counted 48 slots in the stator, so that would mean their are minimum 8 coils per phase? Point being the coils are rather short looped and are thus relatively isolated from the rest of the stator.

 

[crshears]

FWIW, we supply a [three-phase output] VFD from a 240VAC single-phase supply, which the VFD sees as an open corner delta connection; the VFD works very well, no issues at all.

The supplier was made aware that we only had a single-phase power supply available, and sent us a unit normally rated for 60 HP on a three-phase supply but de-rated to 40 HP since one leg would be idle.

Since the unit had not beeen modified in any way, however, the nameplate on the unit had not been altered, which at hook-up time generated a couple of phone calls, first to the supplier, then to the manufacturer, in order for us to learn how to correctly wire the unit to the power supply.

CR

 

[waross]

The AC is rectified to DC and the DC is inverted back to PWM AC.
Some VFDs are capable of charging the DC bus from a single phase supply. All are, but there may be issues with ripple, overloading, single phase protection. If the manufacturer says that the VFD may be supplied from a single phase source, don't worry.
An advantage of the delta motor for a phase convertor is that it develops a well balanced three phase output. The disadvantage is that it will suffer circulating currents if multiples of the third harmonic are present. The VFD harmonics may be taking out your motor.
Running the VFD on single phase may take the strain off the phase convertor and allow a smaller phase convertor to be used.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[chipwitch]

Thanks Bill. It's an option we're considering. We'll wait to hear what the manufacturer has to say when they inspect the box. It shipped to them this morning.

Do you think a line reactor could sufficiently manage the harmonics?

I mentioned a wiring problem in the RPC controller. There are 9 324 MFD caps in parallel on the start circuit. Problem is, 4 of them were only connected by a single lead. Meaning, 5 caps were all that were connected. 3 of those were leaking. Attached is a schematic I drew, minus the low voltage control wiring for clarity. Notice how a lead from the bank of 4 caps and a lead of the 5 caps both go to separate poles of the contactor. And from the "load" side of the contactor they join electrically. Since only one lead was connected on the bank of 4, the joining on the load side of the contactor is of no consequence, but it seems the unit was built at 3:30 of a Friday. What role could the start caps play in damaging the motor, if any?

 

[waross]

There are others here who are better with harmonic suppression. Let's wait for them to answer the reactor question.
personally I would go single phase on the VFD and get the harmonics out of the RPC circuit.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

Now we are up and smiling on our side of the Pond!

Yes, line reactors are there to reduce current harmonics. A four percent reactor will usually reduce harmonics enough to avoid the more severe problems. But I think that you may need more.

The first problem with that approach is that you MAY get a problem reaching full speed AND full load at the same time. So, if you need to run the spindle at highest speed and maximum torque, then you should check if there are any margins in the drive. If you seem to have a 20% margin both in speed and torque, then no probs.

The second problem is that it may be voltage transients and not harmonics that kill the motor. The easiest way to find out is to simply look at the current's wavform. If it contains lots of third, fifth etcetera components, then you know what the problem is. If the current looks nice and clean, then you shouldn't add any reactors at all. But look for other problems. A current clamp and a scope is what you need. May or may not exist somewhere in the neighbourhood. Ask around.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[chipwitch]

Welcome to the thread Skogsgurra!

The manufacturer suggested the line reactors. I've been reticent because it seems too much like a guess. If it isn't harmonics, I could be replacing another motor in a month. First off, 25HP motor on the converter, 10HP spindle, CNC mill has an LED bar indicating torque in terms of percent. Either it's broke or I just don't work it hard because it sits on about 30%. I rarely take the speed above about 70%.

I own both a clamping ammeter (clamp independent of meter, attached by leads) and a scope. I barely know how to use it and have read just enough to scare the hell out of me regarding working on mains power. As a former electrician, I have an excellent working knowledge of electrical wiring and am pretty sure the reason people warn against using scopes on mains power is the fact that the probe is attached to ground and many scopes can't have a maximum voltage input of less than mains. BUT, because I don't know all there is to know about scopes, I'm reluctant to use it on mains. Also, lots of people on electronics boards are afraid of mains and think if one is a novice with electronics they aren't familiar with the dangers inherent to working on mains. I've had my bare hands inside live panels on numerous occasions and THAT part doesn't bother me. One thing I DON'T know is if there IS a harmonics situation, would they be high enough to destroy my scope?

I realize you're talking about current. And I'm assuming you're suggesting I probe the leads on the clamp to read the current wave? Can't imagine there'd be any danger there.

This converter is in a large residential area of town. Any chance the power company could be doing something to disrupt the line and damage equipment? (No, I'm not one of those conspiracy theorists). Poisoning the well, so to speak.

 

[Skogsgurra]

Oh, that is probably Fluke that have scared you with all the pictures of guys in hard hats, gloves, protective goggles and all that stuff that makes any practical work impossible. If you are a seasoned eletrician, you shouldn't have any problems checking the current waveform with a current clamp and a scope. Do this:

1. Bring out scope and fire it up.
2. Bring out current clamp.
3. Connect current clamp to scope vertical input (usually marked CH1, CHA, Vert or Y).
4. Put clamp around phase wire. Make sure it closes (the metal surfaces shall contact each other).
5. Run CNC.
6. Press Aut. (usually green button) on scope. Adjust V/div and ms/div if necessary.
7. Look at screen. Does it look like a sine? Or more like a series of humps going up and down with more or less straight horizontal parts between them?

If nice sine, then you can forget about harmonics.

If humps. You should do something about it.

If there is an FFT option on the scope, you can make a frequency analysis. You should change ms/div to 100 - 500 ms/div and then press FFT. A spectrum will show. Look for peaks at 150, 250 etc Hz. They shall be at least 10 dB below the 50 Hz peak, that corresponds to roughly 30% current harmonics and usually doesn't harm. If you are 15 or 20 dB below 50 Hz - then no sweat at all. Of course, the frequencies will be 180, 300 and 60 Hz if the grid is 60 Hz.

The prospectice short-circuit current doesn't seem to be very high in your installation. So you can forget about hard hat and goggles. And I don't think that you need gloves either.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[chipwitch]

I don't see an FFT button. My scope is a Tektronix 2230. I don't have an actual clamp probe designed for use with a scope. What I have is one that plugs into my portable DMM. It's range is to 400A. Can it just be connected directly to the scope? Or does it require some kind of external power source?

In any case, I won't be able to test it until the control box comes back from the manufacturer. But, I'm looking forward to testing the line once it's reinstalled. Your detailed instructions are the best diagnostic procedure I've encountered. With it, we should be able to actually determine the cause rather than guessing, installing line reactors and hoping for the best.

Thank you

 

[Skogsgurra]

The 2230 is digital. Rather complex, I would say. There *should* be an AUTO button somewhere. At least an Autotrig.

Re. the current clamp. It is hard to say. If it does measure DC also, it probably needs an external supply.
You CAN do the neasurement by putting the line through a plain vanilla current transformer. Do not forget to connect a low ohm resistor (I always use 1 ohm, because then I don't have to do any calculation) across the secondary winding's terminals first. Then measure voltage across resistor with the TEK 2230. That will give you the same information that the current clamp would.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[chipwitch]

I'll have to refamiliarize myself with the scope. If FFT and auto are the same thing, then yes, I think it has that.

My understanding of the clamp is that the current flowing through the conductor generates a small voltage. I'm guessing that voltage would be representative of the current signal. We aren't looking for quantitative results here, so it wouldn't have to be linear. I'm guessing it is linear though, and the actual amperage could be calculated from the voltage readings.

I can't say I've ever heard the term "current transformer" before. Looking it up, it sounds pretty much like my clamp probe. Correct me if I'm wrong, but wouldn't that be different from a hall effect sensor? I understand both, if not one and the same, are commonly used in clamp probes. Given that my probe uses no battery and plugs into my DMM's voltage plug, it must be a "current transformer."

 

[Skogsgurra]

Yes, a CT is passive and doesn't need any power. If your current clamp plugs directly into the DMM voltage connector and no other connections needed - then you can use it with the 2230. You do not need the 1 ohm resistor, a suitable resistor is already built into the clamp.

No, FFT and Auto are not the same thing. You will have to eyeball the screen and judge for yourself.

This link shows what you should look for:

http://sound.westhost.com/lamps/pf-f2.gif

The lower set of curves show what it shouldn't look like. If it looks like the top set of curves - you are probably measuring in the wrong place because it is not common to have such a nice current.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[chipwitch]

Great. Thank you for your help. I'll update the thread when I hear from the manufacturer... and then again when I get it back and test it myself. Also, I'm going to suggest he do the test you suggested.

 

[ScottyUK]

The Tek 2230 is an analogue scope with digital control and some text display on the CRT. It's not a DSO and has no FFT capability. It's of similar vintage to my old 2465B and uses a similar user interface. I'm not certain it has an AUTO button either. Set it to trigger off the line and you should get a usable display when working with signals from the AC mains.

 

[Skogsgurra]

Is it THAT old! OK, no TTF then. Scotty's tip on line triggering is good.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[chipwitch]

Update and new question...

I got the control box back from the manufacturer. He did find what would have been responsible for burning up motors under the right conditions, but I couldn't in all honesty confirm that those conditions would have ever existed. This issue he found was in the wiring of the motor contactor. This was one of his company's first units and he corrected the problem before returning the controller to me. The problem was, and I should have caught this myself, the motor contactor was powered by a separate power source from the main motor feed. In my case, we ran a dedicated circuit for the purpose in the same conduit as the single phase feed to the 3-phase converter box. Where this might burn up a motor for example, is in the event the single phase breaker controlling the motor tripped and then was manually reset. The contactor would have remained in the closed position since the dedicated circuit would not have tripped. Of course, if the motor were to be left in this powered state without being started for some length of time, it would degrade the winding insulation. As I said, I couldn't recall a single instance where this might have occurred. We have had mains power failures before during 3 phase operation, but in those events, the motor contactor would have opened and not presented an issue.

So, now I have reinstalled the unit with a used motor lacking a label. I took the motor to a large local motor repair facility where they were kind enough to test it for me. Like the previous motor, it is delta-wound, 1720 rpm. Based on the current and size of the motor, he estimated it to be 25HP, the same as the previous motor. In a no-load test, I found the voltages were way out of whack. L3 is the created leg. L1 and L2 was 244V. L1 and L3 was 288V. And L2 and L3 was around 160V. I called the manufacturer. He said to remove two of the run caps and try again and that the motor was likely smaller than 25HP. I did. Now, I'm reading, 244V, 255V and 272V respectively. Better, but still not very good. When I turn on the Bridgeport mill, a 2 hp single speed motor, I get a drop of about 10V on the high leg.

Again, with no-load, I read the current at the motor with a clamp. I'm getting readings of 303mV, 415mV and 640mV! I guess it's not really relevant what the mV/A is, the ratio is more than 2:1! Nevertheless, the current equals 15mV/A, so that means 42A on one leg. What might cause this disparity?