Triplen Harmonics, 5th Harmonic, Drive Isolation Xfmr, Generator

[W@ttson]

Hello All,

Question about effects of 3rd and other triplen harmonics. I have a situation where large SCR based drives are running motors. The existing system had them operational with the previous generator. The generator was replaced in kind (as in KVA) and even with a larger starting KVA capacity. A PMG type exciting system is used on the new generator. Now when the generator tries to run the SCR drives and run the motors it begins surging and then the control system stops (probably due to sensing the surges and dips in the power system). A drive isolation transformer was placed in line on the output of the generator to see if it helps the situation. It in fact did allow the generator to operate without incident.

Detail on the ATS: The ATS is an overlapping neutral type.

My questions is the following. Typically SCR based drive systems produce large 5th and 11th harmonics rather than 3rd and triplen harmonics. To my understanding delta wye transformers (what the drive isolation transformer is) do a good job at trapping the 3rd and triplen harmonics. That being said, my question is if the major source of harmonic distortion is from 5th and 11th harmonics which are negative sequence harmonics, then why when the transformer is connected and filters the triplen harmonics does the installation work? Could it be that the little bit of triplen harmonics found were putting the generator over the edge?

Furthermore, is there any possibility that the generator is not wired correctly, i.e. not having its neutral conductor bonded to ground since it is a separately derived system due to the overlapping neutral ATS? If that is the case could it be that by installing the transformer they are actually providing the neutral connection to ground and this is what is actually eliminating the harmonic destructive results? Could the installation possibly work if it is found out that the neutral on the generator is not grounded since it is mis wired and simply a neutral to ground electrode connection is made?

The reason why I am concerned about the neutral is because I know that the zero sequence triplen harmonics live in the neutral. And due to the complication with the overlapping neutral conductor and possibility for mis-wiring, I hope that not having the neutral connected to ground is somehow exacerbating the situation.

 

[ron]

Is it a fair statement that the SCR based drives are 3 wire feeders only, and no neutral is brought to them?

What is the load power factor when the surging occurs? Is it leading or lagging?

Why is there a suspicion that the neutral of the overlapping nature of the ATS is involved with the problem? Why is there a neutral brought out to the ATS at all? Are there 4 wire loads downstream?

Is the neutral and ground bonded at the generator and again at the utility since you have a 4 pole ATS?

 

[Besoeker]

My questions is the following. Typically SCR based drive systems produce large 5th and 11th harmonics rather than 3rd and triplen harmonics. To my understanding delta wye transformers (what the drive isolation transformer is) do a good job at trapping the 3rd and triplen harmonics. That being said, my question is if the major source of harmonic distortion is from 5th and 11th harmonics which are negative sequence harmonics, then why when the transformer is connected and filters the triplen harmonics does the installation work? Could it be that the little bit of triplen harmonics found were putting the generator over the edge?

The SCR drives are most likely six-pulse units. That is six SCRs in a conventional bridge arrangement. Harmonics are 6n±1 so you get 5th, 7th, 11th, 13th, 17th, 19th and so on.
To a first approximation their magnitudes are the reciprocal of their number.
There should be no 3rd order harmonics. If their are, it would suggest that the drives may be faulty or badly set up causing unbalance.

Can you look at the current feedback with an oscilloscope? You should see six evenly balanced current pulses per cycle of the supply. If not, you have a problem.

 

[W@ttson]

The load power factor should be lagging. We have not verified during surging though. The suspicion is that the transformer solves the issue and the transformer only stops triplen harmonics which are located on the neutral since they are zero sequence. Neutral is from the generator. The generator is a separately derived system due to the overlapping neutral.

I still have not verified if the generator is properly bonded but wanted to see if the performance as stated could be an indicator that it isn't. I guess I am figuring that if there is a solid path to ground from the neutral that the harmonic content can get stabilized.

Is it a fair statement that the SCR based drives are 3 wire feeders only, and no neutral is brought to them?

What is the load power factor when the surging occurs? Is it leading or lagging?

Why is there a suspicion that the neutral of the overlapping nature of the ATS is involved with the problem? Why is there a neutral brought out to the ATS at all? Are there 4 wire loads downstream?

Is the neutral and ground bonded at the generator and again at the utility since you have a 4 pole ATS?

 

[W@ttson]

That's what I figured. That there should be minimal 3rd harmonics. But then why does the drive isolation transformer fix the issue?

The SCR drives are most likely six-pulse units. That is six SCRs in a conventional bridge arrangement. Harmonics are 6n±1 so you get 5th, 7th, 11th, 13th, 17th, 19th and so on.
To a first approximation their magnitudes are the reciprocal of their number.
There should be no 3rd order harmonics. If their are, it would suggest that the drives may be faulty or badly set up causing unbalance.

Can you look at the current feedback with an oscilloscope? You should see six evenly balanced current pulses per cycle of the supply. If not, you have a problem.

 

[Besoeker]

That's what I figured. That there should be minimal 3rd harmonics. But then why does the drive isolation transformer fix the issue?

I don't know. Drive isolation transformers are not common here. The only thing I can think of is that it adds a bit of impedance to the drive input in which case input line reactors would have done the job.

 

[Jraef]

The problem I have observed with old thyrisor DC drives is not harmonics, it's common mode noise. In generators, that common mode noise gets into the AVR system and the voltage and frequency drift coming from the generator starts to affect the SCR firing, which makes everything worse. If it's and old GTO based AC Current Source Inverter, it may also be line notching, which then causes sever CM noise.

So most likely what your drive isolation transformer did was to stop that CM noise from getting back to the generator and it's AVR. Most of the drive isolation transformers are electrostatically shielded specifically for this reason.

 

[W@ttson]

Jraef,

I think you hit the nail on the head. It is an old (1980s or so) thyrister drive. I was considering the common mode noise issue as an alternative but could not formulate it as eloquently as you did. What do you think could be the reasons as to why the old (from the 1980s) generator worked but the new generator does not? Could the new AVR control schemes be that sensitive? Is there a possibility that the gen manufacturer can fine tune that?

The problem I have observed with old thyrisor DC drives is not harmonics, it's common mode noise. In generators, that common mode noise gets into the AVR system and the voltage and frequency drift coming from the generator starts to affect the SCR firing, which makes everything worse. If it's and old GTO based AC Current Source Inverter, it may also be line notching, which then causes sever CM noise.
most likely what your drive isolation transformer did was to stop that CM noise from getting back to the generator and it's AVR. Most of the drive isolation transformers are electrostatically shielded specifically for this reason.

 

[Jraef]

In the 80s when solid state drives and soft starters started to really proliferate to where they ended up in places that were on generators, this problem was rampant. Caterpillar was notorious for it and there were a LOT of Cat gensets out there, especially in the fishing industry in Seattle where I was working at the time. So Cat came up with a retrofit kit for their AVR, basically just a filter for their sensing circuit. Worked great and I installed a bunch of them, then eventually they built it into all of their new gensets and the problem went away.

Now what I think might be happening is that a lot of low cost generators are hitting the market that are knock-offs built in China from expired patent documents, so the designs they are knocking off are pre-fix. But at the same time, thyristor drives are now so rare that the issue doesn't show up enough for anyone to take notice. The reason I suspect this, is that the same thing happened with soft starters, but as I said, the problem disappeared in the 90s and 2000s because of the fix. But lately I've been seeing it show up again on new soft starter installations for sewage lift stations, and every time I go out to investigate, it's been a cheap Chinese genset that looks almost EXACTLY like the old Cats I used to see in the 80s, just not yellow. I checked with Cat to see if I could get their retrofit kit any more, nobody there had ever even heard of it, they were likely all too young... Drive isolation transformers do the trick too, but it's a much more expensive solution.

 

[Besoeker]

In the 80s when solid state drives and soft starters started to really proliferate to where they ended up in places that were on generators, this problem was rampant. Caterpillar was notorious for it and there were a LOT of Cat gensets out there, especially in the fishing industry in Seattle where I was working at the time. So Cat came up with a retrofit kit for their AVR, basically just a filter for their sensing circuit. Worked great and I installed a bunch of them, then eventually they built it into all of their new gensets and the problem went away.

Now what I think might be happening is that a lot of low cost generators are hitting the market that are knock-offs built in China from expired patent documents, so the designs they are knocking off are pre-fix. But at the same time, thyristor drives are now so rare that the issue doesn't show up enough for anyone to take notice. The reason I suspect this, is that the same thing happened with soft starters, but as I said, the problem disappeared in the 90s and 2000s because of the fix. But lately I've been seeing it show up again on new soft starter installations for sewage lift stations, and every time I go out to investigate, it's been a cheap Chinese genset that looks almost EXACTLY like the old Cats I used to see in the 80s, just not yellow. I checked with Cat to see if I could get their retrofit kit any more, nobody there had ever even heard of it, they were likely all too young... Drive isolation transformers do the trick too, but it's a much more expensive solution.

Odd.
I've been involved with DC thyristor drives, thyristor front ends on inverters, and Kramer recovery bridges since the early seventies. Not a problem I have ever encountered.

 

[W@ttson]

Ah so the PMG for the AVR isn't quite helping with the sensing its just providing a steady power source for the AVR. I thought the PMG was the Ace in the hole for situations like this. I guess the AVR needs a filter to be able to sense better.

We always get PMG generators but I guess I over estimated their abilities.

In the 80s when solid state drives and soft starters started to really proliferate to where they ended up in places that were on generators, this problem was rampant. Caterpillar was notorious for it and there were a LOT of Cat gensets out there, especially in the fishing industry in Seattle where I was working at the time. So Cat came up with a retrofit kit for their AVR, basically just a filter for their sensing circuit. Worked great and I installed a bunch of them, then eventually they built it into all of their new gensets and the problem went away.

Now what I think might be happening is that a lot of low cost generators are hitting the market that are knock-offs built in China from expired patent documents, so the designs they are knocking off are pre-fix. But at the same time, thyristor drives are now so rare that the issue doesn't show up enough for anyone to take notice. The reason I suspect this, is that the same thing happened with soft starters, but as I said, the problem disappeared in the 90s and 2000s because of the fix. But lately I've been seeing it show up again on new soft starter installations for sewage lift stations, and every time I go out to investigate, it's been a cheap Chinese genset that looks almost EXACTLY like the old Cats I used to see in the 80s, just not yellow. I checked with Cat to see if I could get their retrofit kit any more, nobody there had ever even heard of it, they were likely all too young... Drive isolation transformers do the trick too, but it's a much more expensive solution.

 

[Phil Corso]

Odd.
I've been involved with DC thyristor drives, thyristor front ends on inverters, and Kramer recovery bridges since the early seventies. Not a problem I have ever encountered.

How many on Generators?

Phil

 

[Besoeker]

How many on Generators?

Phil

All of them.

 

[W@ttson]

2 - 30HP each or so motors on the drives. (that is 2 drives total, one 30HP motor on each)

Ah so the PMG for the AVR isn't quite helping with the sensing its just providing a steady power source for the AVR. I thought the PMG was the Ace in the hole for situations like this. I guess the AVR needs a filter to be able to sense better.

We always get PMG generators but I guess I over estimated their abilities.

 

[Jraef]

Odd.
I've been involved with DC thyristor drives, thyristor front ends on inverters, and Kramer recovery bridges since the early seventies. Not a problem I have ever encountered.

Interesting. Well if you had ever witnessed it, you wouldn't forget it. The generator engine would start acting funny, the soft starters or drives would then start going haywire and motors would stall, then if they didn't trip fast enough, the generator would. On a crab boat in the Gulf of Alaska, that was a major catastrophe...

Maybe it was a peculiarity with the AVRs that Cat was using? It wasn't just Cat though, it's just that they were so ubiquitous. It might be that it was the AVR that was the issue, and other people here were using the same AVR that Cat was using.