reactor vs drive isolation transformer

[don_resqcapt19]

I have seen both drive isolation transformers and line side reactors specified for the line side of larger VFDs. What determines which device should be used, or are they even necessary with modern drives?

 

[petersonra]

I have seen both drive isolation transformers and line side reactors specified for the line side of larger VFDs. What determines which device should be used, or are they even necessary with modern drives?

It is a good question and one that does not have a stock answer.

VFDs are somewhat application driven.

My guess (and it is just a guess) is that the vast majority of applications do not appreciably benefit from either.

It used to be that they were more common because the drives themselves could not take the available short circuit current, but that is much less of an issue these days.

Sometimes you have capacitors on the line side that can create havoc and an isolation xfmr or line reactor can help with that.

There are also cases where harmonics are a factor and both can help with that issue.

Not as common as it once was, but some drives are only designed to work with a solidly bonded supply. if you only have delta power, you have to make a solidly bonded supply with an isolation xfmr.

 

[Jraef]

Bob covered it well, I'll just give you more of the "whys".

Adding inductance to the front end of a VFD can help with protecting the rectifier from transients on the line side, and to a small extent, mitigate some of the harmonics. I don't get too excited about the harmonic mitigation capabilities of just a reactor, and that can be done just as effectively with a DC bus choke, maybe even better, if that's all you want. But the protection from line transients is important. The rule of thumb I use it that if the source kVA capacity is more that 10X the VFD kVA rating, then line transients can cause damage to the diodes or SCRs on the front end. Adding the inductor ahead of the VFD slows down the rise time of those transients and the potential damaging effects of current spikes they can cause. Either a reactor or a drive isolation transformer will provide roughly the same benefit in that regard.

But most drives now are designed to work on a solidly grounded Wye supply, where the line to ground potential is 58% of the line to line potential. Protection components such as MOVs and Common Mode Chokes are almost always referenced to ground, so if you have a Delta system or a HRG Wye system, those components can be damaged during a fault when attempt to become a lower resistance Wye point for the entire grid ahead of them, very very briefly, then they are no longer there. You can remove the ground reference on them (some more easily than others), but then you no longer have that full protection. So the better option is to feed the VFD with a Delta-Wye isolation transformer and create a smaller "Wye grid" just for the VFD(s) which has a solidly ground reference. Then if you are going to use that, there is no need for a reactor too.

If you are looking at older SCR based DC drives, the isolation transformer was also necessary for common mode noise mitigation from firing the thyristors. Most new DC drives are now PWM however and therefore have a front end similar to a VFD, so that's becoming less of an issue. Older six-step and Current Source Inverter AC drives used a similar thyristor based voltage control front end so they needed the isolation transformer as well, but those technologies are no longer on the market for LV drives. That's where a lot of the older specs came from however.

 

[don_resqcapt19]

Thanks guys!
The system is 480 solidly grounded wye fed from a 1000 or 1500 kVA 4.16k/480 volt delta wye transformer. The drive will be for a 75 hp motor and will likely be a Yaskawa A1000 drive.

 

[Jraef]

Thanks guys!
The system is 480 solidly grounded wye fed from a 1000 or 1500 kVA 4.16k/480 volt delta wye transformer. The drive will be for a 75 hp motor and will likely be a Yaskawa A1000 drive.

So you're well over the 10X rule, if it were me, I would put in a line reactor. Cheap insurance. No need for an isolation transformer however.

 

[Besoeker]

Bob covered it well, I'll just give you more of the "whys".

Adding inductance to the front end of a VFD can help with protecting the rectifier from transients on the line side, and to a small extent, mitigate some of the harmonics. I don't get too excited about the harmonic mitigation capabilities of just a reactor, and that can be done just as effectively with a DC bus choke, maybe even better, if that's all you want. But the protection from line transients is important. The rule of thumb I use it that if the source kVA capacity is more that 10X the VFD kVA rating, then line transients can cause damage to the diodes or SCRs on the front end. Adding the inductor ahead of the VFD slows down the rise time of those transients and the potential damaging effects of current spikes they can cause. Either a reactor or a drive isolation transformer will provide roughly the same benefit in that regard.

But most drives now are designed to work on a solidly grounded Wye supply, where the line to ground potential is 58% of the line to line potential. Protection components such as MOVs and Common Mode Chokes are almost always referenced to ground, so if you have a Delta system or a HRG Wye system, those components can be damaged during a fault when attempt to become a lower resistance Wye point for the entire grid ahead of them, very very briefly, then they are no longer there. You can remove the ground reference on them (some more easily than others), but then you no longer have that full protection. So the better option is to feed the VFD with a Delta-Wye isolation transformer and create a smaller "Wye grid" just for the VFD(s) which has a solidly ground reference. Then if you are going to use that, there is no need for a reactor too.

If you are looking at older SCR based DC drives, the isolation transformer was also necessary for common mode noise mitigation from firing the thyristors. Most new DC drives are now PWM however and therefore have a front end similar to a VFD, so that's becoming less of an issue. Older six-step and Current Source Inverter AC drives used a similar thyristor based voltage control front end so they needed the isolation transformer as well, but those technologies are no longer on the market for LV drives. That's where a lot of the older specs came from however.

I've dealt with variable speed drives all my working life - the better part of fifty years.
All manner of systems. Ward-Leonard, Carbon Pile, Differential Regulators, thyratron, DC Chopper drives, Slip recovery systems, SCR, IGBT so I suppose I can reasonably claim to have a bit of experience.

As a general rule, we have not used isolation transformers.
For sure, we have used transformers but this is usually to step down from incoming supply voltage and/or to get the right winding configuration for multi-pulse (12, 24) systems.

On our SCR based DC drives, we used input reactors, typically 50uH air cored units* to limit di/dt and to work in conjunction with the individual RC snubbers to limit dv/dt.
In the early days, SCRs were quite sensitive to that.

*Aluminium Inductors for the most part.