3 phase autotransformer neutral 4

[Comaqc]

Our company sells and installs low power (max (12 kW induction heating systems to heat various elements in the industrial environment. For instance, there are more than 50000 of our power supplies reliably heating paper machines rolls.
Our systems run on 3 phase power. Normal voltage is 208 Volts. A new customer in Europe is using a 400 to 220 volts auto transformer to power our systems. In his case, the transformer feeds six 8 kW power supplies. The reason for this introduction is that our power supplies regularly destroy themselves when power is interrupted, either by throwing the main disconnect or in one case, by a power failure in the plant.
Because a large number of identical units are running reliably when using a delta/Wye secondary transformer, we suspect the auto transformer is generating abnormally high voltages on its secondary, either at turn ON or turn OFF, The neutral of the auto transformer is not used as a common nor is it grounded. The front end of all our power supplies is a rectifier bridge with 120 µfd capacitor bank on the DC rail. Does someone have expertise with auto transformer (or isolated secondary XFMR) causing large voltage transients? Should the neutral point of the auto transformer be earthed?

 

[itsmoked]

Add to this tale, what exactly is damaged during these failures.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

With a floating neutral the voltage on each leg will be inversely affected by the load on each leg. (Affected but not exactly proportional.) The voltage will rise on the lightly loaded legs.
There are some important differences between a neutral and a ground. The autotransformer wye point should be connected to neutral stabilize the voltages. The autotransformer wye point should be grounded to limit voltages to ground and to trip the protection in the event of a ground fault.
An open delta autotransformer connection will give stable voltages without a neutral connection, but the power supply voltages will be elevated abouve ground.


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

 

[jjmoab]

Hi Comaqc

There is something that sounds familiar to me. Here in Spain, our plant is a paper mill, we have a CALCOIL system used for heating paper rolls.

We have experienced several problems with these actuators after the annual paper mill suthdown this year.

i will check our configuration, I know there is a XFMR for adjusting the tension level.

 

[roydm]

Is it possible one phase is being interrupted slightly ahead of the others.
Are you sure the Auto trany is connected in Star & not Delta (you say "The neutral of the auto transformer is not used as a common nor is it grounded")
Is any part or your inductance heater circuit grounded?
Can you post a drawing?
I agree with the others the star point should be grounded or connected to the grounded neutral.
Roy

 

[Comaqc]

In all occasions, the dual IGBT used as a high frequency switch blows. It is a 200 Amps, 600 Volts device. This happened 3 times now on a group of 6 heaters.
The transformer primary disconnect is manual; I presume one leg could disconnect before another, but for every short time. Also, each induction power supply has its own individual circuit breaker.
The problems occurs only when the main power is interrupted or turned back ON (we do not know which situation initiates the defects. If all the individual power supplies circuit breakers are switched OFF when transformer primary turns ON or OFF, no problems occur when we later turn ON each power supply in sequence.
JJMOAB; I am the inventor of the Calcoil system. The 50,000 power supplies I referred to in my initial message are mostly Calcoil power modules. The application we are having problems with are in plastics industry, where we heat injection machine parts. The major difference between your Calcoil system and the one installed in the paper industry is in the main transformer. Calcoil systems use isolated Delta to Wye windings. The plastic industry systems installed in Europe use auto transformer for cost and size reasons. We are not having similar problems in plastics systems installed in North America where isolated transformer are being used.

My main question to this board is whether the connections to the neutral of the transformer could be causing the problems we are seeing?

Thanks to all for helping out
I am attaching a rough sketch of the system to this reply. hope I did this right!

 

[itsmoked]

Compaqc; I'm giving you a star because you're attentive and are providing clear questions and answers - fairly rare - and great!

The four things that will typically toast IGBTs are:
1) Ringing. I discount this in your case as that would have nothing to do with switch ON/OFF of the mains and would be endemic, powered anywhere.

2) Overheating. Same answer.

3) Shoot thru. Very deadly to IGBTs. This is having them both ON briefly. Can this be happening because your control circuit goes insane briefly during power down. Another words your controller's boundary condition not being properly constrained? For instance the processor going into reset as the power fails and pulling all outputs to some state that turns ON both halves of the IGBT bridge.

4) Overvoltage. 208V rectified gives 293V which is below your 600V by a comfortable margin. This leaves seeing a larger voltage due to the transformer center not being tied to a neutral. Or inductive kicks caused by interrupting the transformer inductance. Or the result of not having all three contacts opening simultaneously.

T'were me and no one here gives you a solid answer on [4) above], I'd set up a test in my shop using just the rectifiers and caps and an isolated scope. I would try every nasty mode I could think of. Single contact opening two contact opening and three. Find the culprit that creates more than 600V.




Keith Cress
kcress -

http://www.flaminsystems.com

 

[Comaqc]

Itsmoked,

Thanks for the useful info.

Your point 3 is quite valid about IGBT shoot thru failures. However, because some of our induction systems simultaneously turn ON more than three hundred 6 k Watts power supplies every 20 to 30 minutes and OFF for about 10 minutes ( the time for a paper reel change), we have tried extremely hard to design our control board to be immune to problems at start up and shut down. While we could be getting some weird effect from the auto transformer, as we power our control board from a single phase, I believe that your suggestion to test the autotransformer is the one I have most confidence in.
This afternoon, we have been able to locate an auto transformer with the same ratings as the ones used in Europe. We will test it in our lab this week. I'll definitely keep this forum posted on our results.

 

[itsmoked]

We will watch with great interest.

I hope you have a scope.

You can probably just look at the DC with various connections initially.

Q: Have you considered using the 1200V IGBTs and skipping the hassle of the transformers for your 400V, 460V customers?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[roydm]

"a large number of identical units are running reliably when using a delta/Wye secondary transformer"
Since the Auto Transformer is Star connected , same as Delta/Wye you would expect it to act in a similar manner on loss of power except that it's not isolated. I am wondering if you loose one or two phases before the third if the voltage generated by the collapsing field is adding on to the voltage of the remaining phase causing an excessive voltage to ground. Grounding the Star point (to neutral) should eliminate that possibility
Is ii possible that the spike in voltage to ground is causing both IGBTs to turn on together?
You power the control board from a single phase, is it transformer isolated?
I assume you have considered all the surge protection options.
Roy

 

[waross]

You can't compare a delta wye connection to an autotransformer.
In the delta wye each delta winding gets the same applied voltage and each leg of the wye develops the same voltage.
To get an idea of the action of the autotransformer with a floating neutral, consider a 120:240 single phase circuit where for some reason the neutral si disconnected from the supply transformer and starts floating. If the loads are unbalanced, the voltages will be unbalanced.
A three phase autotransformer will have a similar action. Look at your circuit and see what happens if one phase goes high or low. What if the phase with the supply transformer goes high or low? What may happen if the phase with the control transformer opens early or late?
The other consideration when one phase opens early is the phase shifts.
If A:B phase opens early, the loads connected from B:C and C:A will go in series across A:B phase.
The A:B voltages will drop from (normal volts / 1.73) to (normal volts / 2.0) (eg; 120V to 104V)
The combined voltages across A:B and C:A will equal the voltage across A:B (eg; 120V to 104V / 2 = 52V)
All the phase angles will be either in line with A:B or 180 degrees away.
You may wish to evaluate the possible reaction of your circuits to varying voltages and phase angles.
Remember also, if the reaction results in unequal phase loading, the voltages may become unbalanced also. Higher than normal voltages are possible.

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

 

[Comaqc]

Thanks for all your pertinent responses. Here are some comments:
1-The power for the control board comes from the same auto transformer. The 2 micro controllers on the board manage the power up and down functions. Power drawn is under 700 ma. Also, remember we have a whole lots of identical power supplies running perfectly in North America.

2-We are under pressure from our European customer to develop a 400 VAC input model of induction power supply. Our excuses for not doing it at this time are higher IGBT switching losses, the need to replace most power, control and output connectors to comply with CSA and CE requirements and naturally, higher equipment cost.
3-Because we are feeding 3 phase rectifiers with the transformer secondary, we assume that phase loading will be always balanced. At start up, the only load on the transformer comes from the bank of 120 µfd capacitors on each power supply. The only slight unbalance is on the phase which also feed the control board, about 0.7 amps.
4-In relation to the above, we were wondering if the temporary short circuit resulting from charging the capacitor from zero to maximum voltage was not causing the auto transformer to also temporarily present close to the level of the input voltage.
5-We are getting a 30 k VA, 600 V to 208 auto transformer to-morrow. We will measure the effects of floating, common and grounded connections. I will post the waveforms and results.

 

[waross]

If your supply has a neutral, connect the wye point of the autotransformer to the neutral.
I don't think that depending on the load to maintain balanced voltages is a good idea. Any component failure may result in unequal voltages.
Grounding is a different issue. It will not help voltage excursions unless the supply system also has a grounded neutral. Then you will have effectively connected the autotransformer to the system neutral through ground. This is not good practice and may results in compromized ground fault protection. Better to directly connect the neutrals together.
If you have a system neutral to connect to, it will provide the system ground. If the system is ungrounded or high impedance grounded, a ground on the autotransformer neutral will certainly cause voltage excursions in the event of a ground fault.
You mentioned 400V to 220V amd now you are suggesting 600V to 208V. What is your system voltage and does the system have a neutral point?



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

 

[roydm]

Waross, I know the auto tranny is not quite the same as the Wye of a Delta/Wye but assuming the load is balanced the normal voltages are the same.
I was trying to visualize what happens to the voltage to ground at the moment of power interruption.
Delta/ Wye - the DC voltage to ground should remain the same
I suspect the with the Auto Transformer DC voltage to ground will be all over the map as the transformer field collapses especially if one phase is still connected to the supply.
I think we agree the key is to connect the star point to neutral.
Roy

 

[waross]

You may have switching issues due to the phase shift when one phase opens slightly ahead of the others. This may lead to voltage shifts.
On a wye/delta bank with a floating neutral, (NOT delta/wye) you may encounter energization voltage surges on one phase. This effect is not present when the neutral is connected.
If the autotransformer uses a three legged core it may act as if it has a phantom delta winding. This will tend to reduce or eliminate voltage swings.
Given that the units run well when they are running, I suspect that your problem may be more related to phase shifts than to voltage surges.
If one pole of the disconnect switch opens early the phase shift on two phases may result in IGBTs being on at the wrong times. There may also be issues with inductive kick from the load if some IGBTs are off at the wrong time relative to the phase shift.
Connecting the autotransformer neutral to the system neutral will correct both possible voltage unbalances and phase shifts.



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

 

[Comaqc]

Thanks again for all comments.

Some answers to posts:
-The problematic auto transformer powered systems are situated in Germany. They run from 400 VAC primary to 220 Volts. Our power supplies accept 200 to 250 VAC 3 phase input. Here in Canada, the transformer primary voltage will be 550-600 volts to 208 volts secondary. This is as close as we will ever get in duplicating the auto transformers used in Germany.
-Our power supplies do not use a neutral connection. We use the rectified 200 to 250 volts (about 300 V DC) to feed a series resonant, 20 kHz IGBT inverter. When using isolated transformer, we ground the Wye secondary neutral point to earth as close as possible to the transformer location. The chassis of our power supplies are also grounded in their cabinet. This has worked without problems for years. Now, because we do not use the neutral for our power supplies, isn't our only option to ground the transformer center point?
-We cannot confirm for sure, but the engineer installing the system in Germany believes that the failures must have occurred during power shut down rather than when re-applying power. His reasoning is that he was working close to the power supplies cabinet when it was re-energized and did not hear any noise from 40 amps circuit breakers or internal power supply fuse popping.

 

[waross]

This has worked without problems for years. Now, because we do not use the neutral for our power supplies, isn't our only option to ground the transformer center point?

As I tried to point out, grounding the center point is not the same as connecting the center point to the neutral.
In residential and commercial and light industrial systems in North America the neutral is usually required to be grounded by code and many people who work in the field on these systems use the terms ground and neutral almost interchangeably. When you start working on ungrounded systems and impedance grounded systems the difference becomes important.
If your problem stems from a floating neutral, grounding it may or may not help.
1> If the supply system is grounded, grounding the neutral may avoid your problems, but would be against the code in North America and possibly in Germany.
2> If the supply system is ungrounded or impedance grounded, then grounding the neutral may have no effect or may cause system ground detection issues.

If the supply system has a neutral, bite the bullet and run another wire to connect your neutral to the system neutral.
If the supply is delta, (No neutral point) say so, and we can go from there.
Be aware that when the primary of a wye delta bank with a floating neutral is energized there are switching surges that may cause overvoltages on some phases. As I remember you may expect 173% voltage surges as a regular event.
These surges are the result of the core of one phase saturating and drawing excess current which drives up the other phase voltage(s). The same issues will probably be present with a wye autotransformer.
Be aware that if the supply is ungrounded then grounding the neutral will not affect these over voltages.
Are you using three individual transformers or a three phase auto transformer?
If you are using a three phase auto transformer, is it a three legged or a 5 legged core?


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

 

[Comaqc]

Bill,

I apologize, but I have a mental block concerning the discussion around neutral and ground. To help me understand how and where we should hook up connections, let me ask the following question. If our 3 phase Wye auto transformer was connected to a 3 phase electric motor, where should we connect the neutral point of the transformer?

Thanks,

René

 

[waross]

By the way, I use an open delta autotransformer connection for adjusting voltages on motors. Usually 460V to 550V or 550V to 460V. It uses two single phase transformers and does not require a neutral. The calculations are simple.
You would connect the wye point of the three phase autotransformer to the wye point of the supply transformer. The accepted way is to run a conductor from the wye point of the autotransformer to the neutral bus in the supply panel.
The wye point of the supply transformer may be grounded. Connecting the wye point of the autotransformer to ground may compromise the ground detection system of the supply system.

The supply system may be impedance grounded (impedance grounding includes resistance grounding). Grounding the autotransformer wye point may compromise the impedance grounding system.
The purpose of a neutral is to carry unbalanced currents.
The purpose of a ground is to create equipotential zones to eliminate touch and step potentials in the event of a ground fault. The ground system also provides a path for ground fault currents so a ground may be detected and action taken. The simplest action will be a solidly grounded system where the ground path passes enough current to trip the circuit breaker or blow the fuse. A more complex, impedance grounded system may shunt trip the circuit breaker but limit the ground fault current. Some systems alarm on a ground fault but do not trip the breaker in the event of a ground fault.
Some systems have been designed where a contactor opens the grounding impedance and leaves the system floating, ungrounded until repairs are made.
Typically, the neutral conductors are connected to the neutral bus in the panel or to the neutral connection at the supply transformer.
The grounding system and the grounding grid is a separate system from the neutral system. Typically there is only one connection between the neutral and the ground system. This connection may be a jumper at the main panel or at the transformer. Ground faults may be detected by monitoring the current through this jumper or the current may trip the breaker on the faulted circuit.
In an impedance grounded system, the jumper is replaced with the impedance grounding device. Grounds may be detected by either voltage or current.
Your autotransformer wye point should not be grounded. It may be connected to the system neutral at the panel.
In an impedance grounded system, there will be a voltage difference between the neutral and the ground in the event of a ground fault. When a fault is causing a voltage difference between the ground system and the neutral bus,you want your autotransformer wye point to be connected to the neutral potential, not the ground potential.
If you are unsure ask again and I'll try to explain a different way. No problem. This is a little more challenging than a face to face explanation.

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

 

[NAZ55]

This may be a case of ferro-resonance folks. As you mentioned, since this is a floating wye, when one phase opens up, since there's no path to ground, there is a series connection between the transformer coil (inductor) and the capacitor. and if the auto is lightly loaded there's no Resistance in LC circuit causing a natural frequency. where jwL = 1 / jwC (i.e. impedances cancel themselves out essentially a short circuit)and you will have abnormally high voltages and currents.