Delta 3 Phase and VFD problems 1

[EdDanzer]

We have had problems with VFD’s and spindle drives with our 240V Delta three phase. The new Baldor drive required Y three phase only. Searching only provided thread237-106308
Our incoming power has a wild leg that is 210V to ground.
Some systems work, Fanuc, and Haas, but not Allen Bradley, or Benshaw.
Any Ideas?

 

[DickDV]

Floating delta sourced power will always be a problem for VFD's because, to properly protect the motor and motor leads from ground faults, the drive has to create a virtual ground if an actual ground isn't provided thru the source power. As soon as it does this, power supply imbalance to ground and phase-to-ground leakage currents create conflicts which result in unacceptable nuisance faults, usually ground faults.

The solution is either to provide an isolation transformer with grounded wye secondary (preferred) or turn off all of the drive protections for ground and short-circuit faults and then hope the leakage currents aren't larger than the drive input fuses. If they are larger than the fuses, your only remaining choice is the transformer.

Aren't floating delta supplies just wonderful!! Just so you can keep running with a ground fault! I will oppose those systems as long as I am able, if for no other reason than the hazards they present to operating and maintenance personnel.

 

[jraef]

"We have had problems..."
What kind of problems?

"240V Delta three phase..."
What are your voltage measurements? This denotes only the "nominal" voltage.

"The new Baldor drive required Y three phase only..."
What series of Baldor drive? I see that the H2 Series recommends having a balanced voltage with respect to ground, but does not technically "require" it. If they did, they would sell very few 240V VFDs in the US, because 240V Wye does not even exist here. Did you actually contact Baldor and ask them? Sometimes they put this stuff into manuals just to cover the rare problems, not the general use. BTW, I have never seen this issue with Allen Bradley nor Benshaw either.

You may have some other pertinent problems with your incoming line, such as a severe voltage imbalance, but just being a Delta system is not the sole cause. 240V "red-leg" Delta as it appears you have, is a low cost way of having 3 phase main power with a small amount of 120V for lighting and such by connecting to a center tap on one winding of the transformer. That is why you read 210V to ground on one leg, it is the one NOT common to that center tap. The down side with this arrangement is that those 1 phase 120V loads represent an immediate imbalance in current, which creates an imbalance in voltage. The original intent, that of a "small amount" of 1 phase loads tapped off, is often lost on the 2nd, 3rd or 4th owner/tenant and they keep adding on to the system without realizing the consequences. One day, your 3 phase voltage imbalance is so bad that your 3 phase equipment starts giving you trouble.

Bottom line, more investigation is warranted.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[EdDanzer]

Jraef,
We burnt up 2 Allen Bradley spindle drives and one Benshaw VFD This has been very expensive for a small company.
The incoming power is from our power company in rural western Washington.
The between leg voltage is 240V and we do use the other 2 legs for 120V and 240V single phase. There is a neutral provide for the 120V and 240V single phase. The power company has monitored the system as have we and there are no imbalances that should cause problems.

We did install an isolation transformer on the new Baldor drive, but the Fanuc and Haas do not have an isolation transformer do not seem affected. We may have had a similar problem with another VMC that the dealer and factory could not make work correctly, so they took the machine back.
Are there internal design differences that can make one system work and the other not?

 

[ScottyUK]

EdDanzer,

Consider what happens to the voltage of the DC bus as each rectifier diode (or thyristor - I guess diode for this application) conducts: when the device conducts, that pole of the DC bus is connected to the line voltage less a diode forward volt-drop. On a rectifier fed from a balanced Y system the DC bus voltage is roughly split in half relative to ground, and this value doesn't vary enormously as one device commutates to the next. With a grounded delta the bus voltage relative to ground jumps from near zero to full line voltage as the rectifiers commutate, causing a high dv/dt between the DC bus and ground.

I agree with DickDV's comments regarding the G/F detection and the possibility of the detection circuitry misbehaving as a result.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[rbulsara]

I am not a VFD design expert, but it would seem if the VFD's rectifier is desinged to be fed by a balannced Y system, that is equal voltage to ground across diodes(scrs), but connected to a delta system with one center-tap would cause problems. Because the three phases to ground would read 120V,120V and 208V in a 240V/120V delta system. While the ground reference point of rectifier is looking for a balanced 120V or so input. See the attached diagram:

http://services.eng.uts.edu.au/~venkat/pe_html/ch05s1/ch05s1p1.htm

One of solutions would be to install a isolation tranforerm with Wye secondary as suggested before. I would suspect an ungrounded delta system would work too, but you need the ground the center tap for 120V.

 

[rbulsara]

May be the diodes/scrs used in the VFDs that have no problems are rated for peak voltages of 240V nominal input and those they don't are rated for peak voltages of 120V nominal.

Does this make sense??

 

[EdDanzer]

What I was trying to find out if there are different designs that allow one drive to work and the other fail.

Rbulsara,
The link shows one design that would not work with delta power. (thank you)

Is it possible to connect all the DC together and solve the imbalance?

 

[davidbeach]

EdDanzer, the biggest difference might be the ratings of the diodes/SCRs. Grounded wye would allow lower voltage ratings than would be necessary on other types of systems.

 

[rbulsara]

eddanser:

I am afraid not. It is most likey what davidbeach and I are suspecting. You need to talk to the manufacurers to give you their reasoning.

 

[rbulsara]

In fact all DC are tied together, it is the PIV (peak inverse voltage) that diode and SCRs will see.

 

[jraef]

Huh, I for one am learning something new here. I agree that the only conceivable difference would be in the ratings of the power devices in the bridge. I know this is done by some manufacturers of Medium Voltage solid state equipment because it makes a fairly substantial difference in component cost and count, but for 230V VFDs? I had no idea that anyone was using Diodes or SCRs that were rated only for Y systems, it makes no sense other than to MAYBE save a few cents on the component costs. That would mean they are finding diodes rated for only 300PIV? Wow, now that is getting cheap!

Then again, to some bean counters those few cents may mean the world. The strange thing is that this just shifts the costs to the end user, in fact INCREASES cost substantially if they have Delta system, and as I mentioned, I have never heard of a 240V Y system in the US! 208/120 Y for certain, but not 240/138V Y. Ludicrous if you ask me.

Unless the issue is the potential voltage IMBALANCE that can occur in red leg delta system. Then again, how does one explain the fact that most 240V VFDs 3HP and under can operate on 1 phase input? You can't get any more of an imbalance that having one phase completely missing.

I do agree with DickDVs thoughts on the GF issue however. I have seen this a lot. We have quite a few 480V ungrounded Delta systems in the US, this has always been problematic with respect to GF tripping.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[rbulsara]

jraef:

GF tripping would not burn up drives...or would it? A higher than rated voltage would, I think. Trust me, I do not pretend to be an expert in VFD manufacturing.

As for bin counters...lets not get started...

 

[DickDV]

jraef, I do almost no 230V 3 ph work so I may have been making some useless assumptions on the type of source.



Are you saying that 230V 3ph power in North America is almost always delta source, ie, floating with no neutral?

I'm learning some things in this post too.

 

[rbulsara]

jraef:

Loss of a phase, for correctly sized devices in a rectifier, will not cause them to get damaged either. They may or may not work but will not burn up.

 

[ccjersey]

Most common 230-240 V 3 phase is 4 wire in the US. Much of this is supplied by an open, delta secondary transformer bank. (In our area it is open wye/open delta supplied by transmission lines with 2 phase conductors and a neutral).

There may be some corner grounded 230V delta and some ungrounded 230V delta somewhere, but not common like the 230V 3phase 4 wire center tapped delta. Many power companies are replacing these as transmission lines are upgraded and all 3 phases are available for a wye transformer bank. Our REA power company recently upgraded the line adding the 3rd phase, but they just moved all the residential and single phase on to the new phase. Ok by me, I'm not ready for a shakeup, voltage wise.

 

[jraef]

rbulsara,
I didn't think GF would necessarily damage the drive, I was just agreeing that if can be a problem, i.e. nuisance tripping. EdDanzer never said the Baldor drives burned up, in fact he never said what his problems were, just that they didn't work. He did say that AB and Benshaw VFDs burned up though, and that bothers me. I completely agree that if properly sized, the semiconductor devices should be fine under these circumstances, but the point you and ScottyUK raised is that maybe some manufacturers are skimping on components and subjecting users to this kind of risk. I have installed and commissioned a lot of 240V drives on Delta systems before, this has never been a problem. Admittedly though, I can say I have never done so with a Baldor, AB or Benshaw drive, so maybe there is something to this! Still conjecture for all of us at this point however.

DickDV, ccjersey's post answers your question quite well, that is my experience too. I know that overseas there is 380Y220 or 400Y230 or 415Y240, but here we have either 480Y277 or 208Y120. 240V is always delta in one way or another, the most common being the type with a spit winding and a grounded neutral to provide 120V. Here is a transformer connection book. Figures 11 and 15 are the most common connections for 240V 3 phase systems.

EdDanzer,
My experience having lived in Western Washington for 12 years is that there is a lot of the Open Delta version (figure 15) being used in legacy systems. It works fine, until, as I said, the user connects too much 120V load, then it gets severely unbalanced. It could very well be that the day your Utility came out to measure, you were not using as much as other times. It could also be as you suspect, some drives may not be capable of being hooked up without that grounded Y secondary transformer in front of them.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[EdDanzer]

The incoming service was installed in 1977. 2 of the phases will provide 120V to neutral, these two legs will provide 240V single phase. The wild leg is 210-215 VAC to ground or neutral. These 120V legs are fairly well balanced.
We monitored the power for about 5 days several years ago, and adjusted the 120V balance, and continue to monitor it when changing machinery mix or other loads.
Both the AB and Benshaw drives blew diodes and IGBT’s on the input or front side.
I should be getting a call from Benshaw with a more complete explanation later this week.
What I don’t understand is how converting the AC to DC should care about voltage to ground. This may change the DC buss voltage some?
Several years ago I designed and built 3ea. 100 amp DC power supply for some servo amplifiers and did not have voltage imbalance problems.

 

[jraef]

Had it been designed around a 240V nominal voltage level, you are correct, it shouldn't really matter. Why it would "care" is if the PIV (Peak Inverse Voltage) rating, the semiconductor device's ability to block voltage, was selected based upon their reliance on there being a Y system, where the peak voltage that can be seen by the device is the Phase-to-Neutral voltage, not Phase-to-Phase. PIV ratings typically need to be 2.5 times the nominal line voltage. So in a 240 Ph-Ph voltage environment (delta), that is 240 x 2.5 = 600V, a common rating for diodes. If they designed it because they thought it would be a lower voltage to ground reference in a Y configuration, i.e. 120V or 138V, then the diodes they selected may be unable to survive the peaks that could happen on that high leg. To be honest it doesn't make sense for them to do that, but it might explain why they insist on a grounded Y system. Again, just conjecture at this point. I have put in messages to some of my sources in VFD design to get their take on this, including someone I know at LS (formerly LG) in Korea, the people who make the Benshaw inverter.

An unbalanced voltage however is a problem in that the difference between phases can increase the ripple on the DC bus and if not able to be absorbed by the capacitors, can cause stress and possibly damage to the IGBTs. By the way, the balance between the 120V loads is somewhat irrelevant, the culprit would be the ratio of 1 phase to 3 phase loading on the transformers.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[EdDanzer]

One observation is both the AB and Benshaw units do not have very large capacitors.
How would you check the ration between phases?