VFD Troubleshooting - Low DC bus voltage and motor not as smooth as possible

Hi guys,

Question. I've been trying to diagnose a problem with the VFD of my Okuma LS lathe.

• Original 1973 7,5 Kw motor, over here running on 400V 3 Phase supply (motor in delta)
• VFD replaced old star/delta starter
• VFD is a SEW Eurodrive Movitrac unit rated for 7,5kw worst case.
• VFD fitted in original cabinet with ventilation.
• Most wiring replaced.

What happened. Didn't use the lathe for a month, it was simply too cold in the shed. Went to use it again, motor not running, only jerking and banging the gears of the headstock. Within short time overcurrent fault triggered.
Tried a spare (old) 4 KW motor, same effect, with original parameters and with current etc set for the other motor.
Checked IGBT's and diodes for continuity, all checked out fine with my Fluke meter. Power supply good too, all 3 phases 400V.

Checked the parameters. Some seemed off. Changed them back to factory setting, motor runs good! Checked motor resistance good. Motor phases same voltage. Maybe slightly uneven but hardly noticeable.

Now the strange things. Works fine on Vector field control mode. On VF mode the lights start to dim/flicker at high rpm and some high pitched humming noise is heard from the VFD.
And, the DC bus is only 570 Volt and decreasing with higher motor rpm. Normally a 3 phase 400V grid has a DC peak value of 690V. DC bus voltage just doesnt get over 570V.

Maybe bad filter cap?

The issue here is that this VFD has a form factor that works good with the original cabinet. So not easy to just buy the cheapest VFD on my local Craigslist and use that one.



Mains swith (4 pole) - circuit breaker - relay - VFD.

Thanks in advance!

400 VAC RMS input will give you 565.6 VAC P-P which should be what you're reading on the bus once filtered. If I've done my math correctly... As load increases the bus the voltage will begin to sag. With a 480 VAC RMS input you should see close to 700 VDC on the bus....

I think you have a parameter or two that needs to be looked at. First of all I would check to see if there is a parameter to select input voltage and change to match your input volts - 400. Or, you may have a parameter that sets the minimum bus voltage.

What is the mfg of the VFD? EDIT - nevermind.... I see that now.

Also, check the motor parameters to be sure you've specified the proper motor voltage and amperage. Voltage may/may not be selectable.

The input filter capacitors do not like cold temps. Might have partialy frozen..

Appling power after having been frozen.. Might have partialy opened/shorted them...

Also - run the motor at half speed and verify that the applied voltage reads half the nameplate voltage - if it doesn't match, adjust the motor cos phi or power factor or whatever that drive calls it to bring the voltage into compliance with the nameplate, this will ensure that you are not saturating the windings with current and thus pulling the bus voltage down at rated motor speed.

Thnx guys.

Yeas, (input) capacitors dont like cold. Maybe it sounded extreme, but temps hovered around and just below freezing for a while and that was it. Lots of condensation though. I'm going to disassemble the VFD and check the capacitors.

As far as I understand s phase systems, the root of 3 is used (1.73), not the root of 2 (1.41).
I asked at work, where some peoiple are developing power electronics and a suggestion was that the VFD might have PF compensation and used triacs in the rectifier to limit inrush etc. But the DC bus is 560 volt and only drops inversely with rpm.

Just missed out on a brand new one Ebay. Went for more money that i wanted to spend so. Have to figure out why this one is misbehaving.

The lathe is converted to VFD for maybe 6 months and has only seen light use (only time the motor is strained is from braking or turning on, the chuck is heavy and ramp up/down is ca 2 seconds). Maybe the motor is shorting inside and making the VFD adjust for that? Of course not related to the DC bus.

motion guru said:

Also - run the motor at half speed and verify that the applied voltage reads half the nameplate voltage - if it doesn't match, adjust the motor cos phi or power factor or whatever that drive calls it to bring the voltage into compliance with the nameplate, this will ensure that you are not saturating the windings with current and thus pulling the bus voltage down at rated motor speed.

Click to expand...

Will do, excellent suggestion!

Last week I adjusted the 'boost' parameter down to the normal/stock value and that helped. Might have a similar effect.

Usually, a noisy inverter is an "unhappy" inverter. Noises may be current limit acting, etc. They normally have a "retry rate" that is in the audible range, and so they often "squeal".

While you are at it, look for a monitor parameter that shows output current on the display. Make sure that is in-range with the motor data plate. (I'd say check with meter, but many meters do not like the pulses from a VFD, and may give strange readings.)

When you changed the motor, did you re-tune the VFD to the new motor? For Vector Control to work properly, it must know the motor equivalent circuit, more than just the Voltage and FLA values. So you can enter those other values manually, but since it's uncommon to know them, most Vector drives will do an "auto-tune" routine to learn them empirically. If you didn't re-do that for the new motor, it doesn't get things right.

Didn't he say that vector mode was fine, but V/F mode was messed up? Seems that would suggest the vector is happy enough and has "solved the motor" OK. If it hasn't it usually does something that alerts you all may not be well.

quote: " Works fine on Vector field control mode. "

Maybe what is being called "vector" here is actually not????

JST said:

Usually, a noisy inverter is an "unhappy" inverter. Noises may be current limit acting, etc. They normally have a "retry rate" that is in the audible range, and so they often "squeal".

While you are at it, look for a monitor parameter that shows output current on the display. Make sure that is in-range with the motor data plate. (I'd say check with meter, but many meters do not like the pulses from a VFD, and may give strange readings.)

Click to expand...

Indeed, thats my feeling to. Noisy = bad.
Regarding the output current, no load current of the lathe running was 5 Ampere. The separate test motor has a no load current of 3 Ampere.
I am thinking of cobbling up some kind of 3-phase current transformer setup to check the output current(s). Might take a while.
When the problems arose, current value shot up from the usual 5 (or 3) to trigger the 'over current' fault code. Max amp was adjusted to 12 for the lathe's original 7,5Kw motor and a little less for the other motor.
Might be useful to test a like new 100 Watt 3-phase 400V motor that I have but I had the feeling a bit more load would be good to stress-test the drive. But I have no idea how good this test motor is (junkyard), which runs OK directly on line but still. VFD's with their fast switching IGBT's could trigger all kinds of effects with old (ab)used overheated etc motors..

Jraef said:

When you changed the motor, did you re-tune the VFD to the new motor? For Vector Control to work properly, it must know the motor equivalent circuit, more than just the Voltage and FLA values. So you can enter those other values manually, but since it's uncommon to know them, most Vector drives will do an "auto-tune" routine to learn them empirically. If you didn't re-do that for the new motor, it doesn't get things right.

Click to expand...

Yes I changed the parameters to the ones for the new motor, as far as I could find them on the nameplate. The test-motor is a Leroy Somer which has the usual data volt/amp/hz/Cos-phi/rpm on the plate. This VFD does the autotune after changing parameters yeah, AFAIK one can't skip the auto-tune with this VFD. Takes about 2 seconds and I can hear the coils in the motor do something.

JST said:

Didn't he say that vector mode was fine, but V/F mode was messed up? Seems that would suggest the vector is happy enough and has "solved the motor" OK. If it hasn't it usually does something that alerts you all may not be well.

quote: " Works fine on Vector field control mode. "

Maybe what is being called "vector" here is actually not????

Click to expand...

This controller has 'Voltage-controlled vector control VFC': SEW-EURODRIVE Products: MOVITRAC(R) MC7B
The auto-tune function could change things after a while? Could this drive be checking the motor every time it powers up?

I wonder if a bad motor (with some kind of coil-coil-phase-phase etc short) could wreak havoc with the VFD. Sure it is protected but still.

Sadly no time for testing this weekend, will report next week what my findings are!
Many thanks for the replies, really helpful and its great for me to test the suggestions!

The peak voltage should be 565 at 400Vac. More importantly the voltage on each capacitor should be over 489Vdc. Overlap of 3phase is at 120deg or 565V X Sin120 =489V.

Is there a difference in mains input current from leg to leg?

Yes, a motor fault, especially an intermittent one, might really foul up the measurements, and make the VFD do odd things.

You mentioned running rough, which often is from one phase having a problem compared to the other ones

DNS said:

The peak voltage should be 565 at 400Vac. More importantly the voltage on each capacitor should be over 489Vdc. Overlap of 3phase is at 120deg or 565V X Sin120 =489V.

Is there a difference in mains input current from leg to leg?

Click to expand...

+1 Sounds as if:

A)'leaky' cap. Easy to find with an ignorant VOM.

B) Short in the motor. Not hard. either, just not so easy with a vanilla VOM.

The VFD in-hand, if itself 'healthy' as-in still runs OTHER motors properly, IS your 'test equipment'.

Check the caps.

First. Or "again".

Bill

I should add that the voltage on the caps should be closer to 565Vdc. I would think that if the voltage drop by more than 5%(536V) @ FLA the the drive couldn’t reconstruct the peaks for 400Vac.

DNS said:

Is there a difference in mains input current from leg to leg?

Click to expand...

I worded that incorrectly. Is there a difference in current on L1, L2 and L3. If you measure the current across 2 legs you will short the 2 legs. Use a current clamp to measure.

As for leaky capacitors.... When the thing is hooked up to mains voltage, a leaky capacitor is not gonna do jack to the bus voltage. They just cannot leak that much without more problems.

If one was in fact THAT leaky, the next thing it would do is to launch out of there, because there would be a LOT of heat dissipation. It would boil the electrolyte, arc, and probably blow the case off. Seen it happen.

FAR more likely to be a problem with either the load or the VFD. If the VFD runs another motor of roughly the same size*, it's likely to be OK, so it's time to look at the load. Motors with a shorted turn in the winding will draw excess current, probably will be lossy enough to foul up the vector motor test, will heat up, and will eventually fail hard when the insulation fails from being charred by the heat.

Problems can also occur if wires intermittently touch, etc. So it's worth checking wires if nothing else turns up.

*
I say roughly the same size, because I HAVE seen cases where the current sensor got messed up on one leg, causing it to sense a lower current as being an overload. So it would run a smaller motor OK, but went into limit when a more nearly matched-power motor was connected.

If the voltage is low, the peaks will not be correct, but the VFD might just widen the pulses to correct it, if it is measuring output RMS. Vector drives can calculate the number so they don't need to measure, assuming the motor can be characterized. A bad motor won't act right, and the calculated number will be wrong..

Some VFDs have what they call "3rd harmonic injection". This can be thought of as just widening the pulses and creating what is pretty much like a "clipped" sine wave. It has higher RMS, but lower peak voltage.

JST said:

boil the electrolyte, arc, and probably blow the case off. Seen it happen.

Click to expand...

FELT it happen.

1938 Sparks-Whitington art-deco 'waterfall front' console receiver.

Dumb kid (ich) had seen the plates of a 5Y3 glow white-hot, upgraded to a handy 5U4, still bright red, got a rush of brains (NOT!) to the anatomy and hit up his mate's Dad (who owned the local 5KW AM broadcast station) for a used-but good 5R4 that only glowed dull red at the plates... and then.. the stench of molten beeswax and brown shite dripping out of the transformer just as an old Borax, water, and tinseled papier mache (well NOW it was..) cap-ass-i-mate-er made itself master of that side of the room as if to protect B-29's from RADAR...

That said ... modern polyfilm caps are near-as-dammit 'dry' are they not?

Bill

Monarchist said:

...
That said ... modern polyfilm caps are near-as-dammit 'dry' are they not?

Bill

Click to expand...

Electrolytic are electrolytic, still. Yes, mostly just "moist paper" wet, but they still take off like a mortar round if abused. Happened on the production line every so often.

Despite some advantages to using dry film caps, they are still roughly 3x the cost to VFD mfrs. In small drives that added cost is offset by the ability to make the VFD much much smaller, which is what the market wants, plus other advances in technology have kept the overall cost down. But when you get above 10A or so, that added cost of the dry film caps looms larger. From a standpoint of cost increase per unit of amps, the two most expensive parts of drives from that point on are the caps and heat sinks. The cost of transistors and diodes stay relatively flat with current increase. Control electronics are virtually the same regardless of size, so as current rating goes up, the percentage of cost that represents is actually going down. But as current goes up, the cost of caps and heat sink increases faster than the increase in capacity. So anything they can do to mitigate that to stay competitive is employed. Electrolytic caps are therefore still the norm in many cases.

Something else not mentioned earlier, but would have a similar effect, might be an input phase loss. In many drives, because of their inherent ability to run on single phase, there is no input phase liss protection. But if you do lose a phase, the DC bus ripple increases a lot and it gets worse under load. That's why you (at least) double the size of the VFD of intending to run on a single phase input. Double check your input circuit. You might have a blown fuse or a breaker that has high resistance on one pole. In that latter case, check the voltage drop across the breaker under those higher load conditions, because simply checking continuity or voltage at no load may not show it.

The film cap VFDs are generally ONLY usable on 3 phase input, and have ZERO ability to operate on single phase except perhaps in units having an input power factor controller. This is because the film capacitors are so small in capacitance. A 50 uF 400V part is a little under the volume of a cigarette pack, and will handle around 20A ripple current. I've used them. With 3 phase input, you may need relatively little capacitance, only enough for simple "bypassing", because 3 phase rectifies to 86% of peak at minimum.

The same volume of electrolytic may be 5x to 10x the capacitance and 1/4 the cost, although the maximum ripple current is less. Film are liked because they have longer life, are small for the ripple current, and are sufficiently high capacitance for a 3 phase input. For military uses, they stand cold very well, while an electrolytic is barely a capacitor still at -40C or -50C.

Despite advantages of film in stability and response to very cold temperatures, they still have little capacitance compared to electrolytic, and if used on single phase, the voltage drop under load is so much that the units are virtually unusable. A sufficient amount of capacitance in film types is very expensive, and takes a lot of volume.

Normally capacitors are designed to vent upon failure. A capacitor would not normally blow up, but i"m sure they do.

My reasoning for a input DC buss failure:

Botje said:

Now the strange things. Works fine on Vector field control mode. On VF mode the lights start to dim/flicker at high rpm and some high pitched humming noise is heard from the VFD.

Click to expand...

Vector control or Field Oriented Control(FOC) reduces the required voltage on the motors windings for the condition of operation. Therefore a reduced buss voltage would operate the motor normally when unloaded.

To give you a better understanding of FOC. Your motor is not only a motor, it is also a generator. The current that is being generated is called BEMF(Back electromotive force). The BEMF produces a signal that will give an approximate rotor position while spinning. If the actual rotor angle is lagging behind the electrical field more voltage will be applied. Yet still PWM sinusoidal.

The Reason they call it Vector control is because they use sum of vectors to determine the rotors position. The Vector calculations are determined by inductance of the motor and the signal current from 2 or 3 current sensors. If one of these current sensors failed the motor could not spin up to speed. A motor will start in the VF mode and the switch to FOC. At very low speeds there wouldn't be enough BEMF to measure.

I will also mention that if there is not enough buss voltage the VFD will compensate with more current due to the lag in rotor angle.

Botje said:

Tried a spare (old) 4 KW motor, same effect, with original parameters and with current etc set for the other motor.

Click to expand...

I gather that the 4Kw motor was unloaded since it would be a different frame size from the 7.5Kw.

I believe a connection, cap or one of the DC buss diodes failed . As I mentioned before measuring the current on the input mains can determine this. Use a current clamp.