Hi,

I am looking for some pointers. We are improving an electrochemical cell for metal ion generation and we want to include a polarity reversal circuit in our rectifier (power supply). Currently our power supplies are 1,000-10,000A, regulated 0-12VDC and we manually (physically) reverse electrode polarity (so that both electrodes get used up).

The problem we have encountered is that there are no suitable electronic components for building an h-bridge at such high currents. The highest rated transistors are capable of handling around 200A. Besides, they are very expensive.

So my questions are:

1)Is it a good idea to make parallel h-bridge modules at those current levels?
2)Is there a more suitable architecture than the h-bridge? (We also want to pulse the current, which is easy with the h-bridge).
3)Is there a way to reverse the polarity before going into the transformer so that we work with higher voltage (220-240) and thus, less current? (components are available for higher voltage and lower current)
4)Would using a switching power supply architecture be better suited, instead of a linear transformer one? (we really don't have much of an issue with ripple so it could make sense).

Any help would be very appreciated

 

Sounds like some industrial relays might be up to the job. It seems like a fairly simple ladder logic control could be developed. Would it matter if the current was completely shutoff between reversals.

 

Thanks for the fast reply Papabravo.

Logic is not a problem, nor is switching off the power before reversals. It is something we already do in smaller scale rectifiers (we have used both mechanical relays and solid-state). But we have not found suitable industrial relays, they also only go up to around 200A (omron has some 400A solid state solutions). But maybe you know another source I am unaware of

We are controlling these power supplies with our own micro-controllers.

I forgot to mention, these are 3-phase rectifiers.

 

1) I think is not a good idea to parallel two h bridges. are not intented for this.
2) The h-bridge is a good choice. you can use SCRs same way like a diodes and control 2 scr for one direction and the other 2 scrs for polarity reversal.
3) I never tried to reverse polarity fron trafo primary but is interesting to investigate.
4) For thay current levels the easy way if via SCRs
5) Like Dream- say: try using 3 phase rectifiers to drop the current per phase.

 

Instead of looking at the usual providers, I just tried a google search for 1000A relays and a few providers appear. That's one step closer, although they are very expensive.

SCRs are a good idea, there are even 3000A ones. Should we use them in h-bridge configuration or just put a pair on each phase and let them do the rectification?

Another idea is to put the rectifier bridge at the mains, then do the polarity reversal, and feed into the transformer. It should work since we still have the AC oscillation to drive the transformer (even if we don't have the zero-crossings), right?

 

It should work since we still have the AC oscillation to drive the transformer (even if we don't have the zero-crossings), right?

You will have pulsed DC which will saturate the transformer core and result in a call-out of the local fire service . Transformers need AC.

 

SCRs could provide the rectification as well as the output voltage reversal but note that SCRs have a higher forward drop then a rectifier so the average power dissipation of each SCR will be over 1000W at 1000A (half-wave through each SCR).

 

SCR's (non GTO types) need a current zero to turn off. Does the "DC" load current have a current zero during the mains AC cycle? Otherwise one then requires a forced commutation scheme.

 

the big plating supplies we us at work use scr's for rectification, mostly water cooled, six for each polarity. in the output, not the input. we have supplies for up to 50,000 amp. all use 3 phase input, some 480 volt and some 12,000 volt.

 

i would try some mechanical arrangement. How often and how fast do you need to do the switchover?
You could make use copper plates and stack them so that they mesh together, then clamp the sides to get good contact.

 

I think the SCR rectification scheme is the better direction. In a 6-phase configuration (from 3-phase mains), we can put 2 SCRs at each phase secondary.

Sounds like this is a "standard" solution, I wonder if there is a source reference for this sort of thing (book).

 

I think the SCR rectification scheme is the better direction. In a 6-phase configuration (from 3-phase mains), we can put 2 SCRs at each phase secondary.

Sounds like this is a "standard" solution, I wonder if there is a source reference for this sort of thing (book).

How does this solve the matter of polarity reversal?

 

I am not sure where your shopping but IGBT and industrial MosFet modules come ins sizes into the 1000's of amps. I have a few IGBT modules at home that are 1000+ amp rated and a few 5 gallon buckets of 600 amp single and dual device modules and they are by no means anywhere near the largest on the market.

 

tcmtech, you know, we had been searching through providers catalogs and octopart. But now that I have been doing a search using just google I have been finding some IGBTs up to 2400A from Powerex and Mitsubishi. You are right! Now, the only thing is that they are super expensive
The MOSFETS are cheaper but only go to 600A (at Powerex site). IXYS has 1500A at a decent price, but apparently they are phased out.

t_n_k, if you place two SCRs at the output of the secondary in inverted parallel configuration, you can activate them to rectify either the top or the bottom part of the AC wave.

 

While electronic switching sounds wonderfully modern for this app. even if parts for the current rating are expensive, I would suggest using DC contactors since it appears switching time is not an issue. We had large DC crane motors being switched using contactors in H bridge config. and they could handle large currents.
They could also be driven by your PLC or whatever for pulsing operation.
Contacts on these contactors are usually replaceable when they wear out. Should be much easier to implement than electronic controls and much cheaper.

 

profbuxton, our polarization cycle is between 30-60 seconds. That's 1440 to 2880 cycles a day, I am afraid the contactors would die before the day is over.

 

t_n_k, if you place two SCRs at the output of the secondary in inverted parallel configuration, you can activate them to rectify either the top or the bottom part of the AC wave.

Understood.

 

profbuxton, our polarization cycle is between 30-60 seconds. That's 1440 to 2880 cycles a day, I am afraid the contactors would die before the day is over.

I have one single pole relay off of a locomotive that I believe is 1000a.
The Loco source may be a viable solution for 2 or 4 pole as they used them for DC traction motor switching before S.S. came in.
I could check with my sources at the RR if interested.
Max.

 

There will be a big increase in contact life when not switched under load. If the contact is capable of carrying the required current without overheating and that you can turn the current of before the changeover happens, then the lifetime should be basically forever with just an occasional cleaning.

BTW what size are your conductors and connections right now?

 

Also when using DC contact switching at that current it is important to get types that incorporate some kind of arc quench at switch off.
There are a couple of methods used.
Max.