HV supply load regulation issues

[Opcom]

They can be stacked without filters but the filters isolate the power transformers behavior from each other. There is no real difference. It's a choice.

[KA2DZT]

They can be stacked without filters but the filters isolate the power transformers behavior from each other. There is no real difference. It's a choice.

The xfmrs are isolated by the diode rectifiers.  I have done this with a lot of different xfmr combinations and never had any problems.

[AMLOVER]

Opcom,

I also prefer to connect separate bridges and filter caps to individual supply and then connect them in series.
As far I don't have access to any supply simulator, could you please simulate for me the attached 3-phase supply with individual single phase transformers in order to get the rates for chokes 1,2 and 3 for a reasonable ripple 3-5%...?
There are 2 conditions, one is the 1.8A steady carrier and the other is 3.6A mod peaks for 50% duty cycle.
Thank you in advance.

[AMLOVER]

There is of course this set-up which needs only one choke but I have never used it before and I don't know if it is efficient.

[KA2DZT]

There is of course this set-up which needs only one choke but I have never used it before and I don't know if it is efficient.

Yes, this is the correct way to do it.  I've done this many times using many different xfmr combinations and it works perfect.  The PIV rating of each FWB diodes only have to be rated for the xfmr they're connected to.  You can lower the voltage by simple turning off any of the xfmrs leaving everything else connected  You can step-start by turning on one xfmr at a time.  Doesn't matter which xfmr is turned on first.  Only thing is the voltage insulation of the secondary winding has to able to handle the higher voltage above ground.  This is for the xfmrs higher up the series.  All the xfmrs have to be able to handle the current load.

The secondary voltages of the xfmrs do not have to be the same,  any voltage combination will work. The highest voltage xfmr should be at the top of the series.  Use as much input inductance you have to improve regulation.

Do not put any chokes or capacitors between the FWB rectifiers,  just connect the FWB diode output directly to the next higher FWB negative terminal.

Fred

[Opcom]

They can be stacked without filters but the filters isolate the power transformers behavior from each other. There is no real difference. It's a choice.

The xfmrs are isolated by the diode rectifiers.  I have done this with a lot of different xfmr combinations and never had any problems.

What might be the effect of one transformer having a higher resistance and it being in series through the diodes with a lower resistance transformer? The currents must be equal in all parts of the circuit, right? 50 Ohms secondary resistance is not out of line for a 500V 1A transformer. I concede that 10 Ohms shown is very low for a HV unit but I think these are worth considering if no filters between them. Not so much a problem as a performance detail.

[Opcom]

Opcom,

I also prefer to connect separate bridges and filter caps to individual supply and then connect them in series.
As far I don't have access to any supply simulator, could you please simulate for me the attached 3-phase supply with individual single phase transformers in order to get the rates for chokes 1,2 and 3 for a reasonable ripple 3-5%...?
There are 2 conditions, one is the 1.8A steady carrier and the other is 3.6A mod peaks for 50% duty cycle.
Thank you in advance.

Yes you can download PSUDII from duncanamps.com. It does not do 3 phase but it is free. Be sure to also download the many additional rectifiers file that users have made. It's a text file called rectifiers.txt. The "Perfect" rectifier I used is something I just rewrote. You will see how to do it. Here is my rectifiers.txt file  including that but it may not have the latest additions by others.

[AB2EZ]

I'm a little puzzled by the fact that AMLOVER is proposing to feed each transformer's primary with a different phase of a 3-phase service. Therefore the various bridge diode paths will not be opening and closing at the same times. I've never simulated something like this, and I can't model it in my head.

Stu

[KA2DZT]

I'm a little puzzled by the fact that AMLOVER is proposing to feed each transformer's primary with a different phase of a 3-phase service. Therefore the various bridge diode paths will not be opening and closing at the same times. I've never simulated something like this, and I can't model it in my head.

Stu

I also gave that some thought but haven't concluded anything about it yet.

Fred

[KA2DZT]

Adding xfmrs in series with just rectifiers works fine provided they're in phase.  Three xfmrs on three different phases probably will not add together to the full peak voltages.  When one xfmr is at its peak another is only at its peak (x sin60) and I think the third is also at its peak (x sin60).

So I think the voltage sum would be the peak voltage + (peak voltage+peak voltage)(sin60)

Stu, maybe you can look at this and figure it out, my brain is smoking.

Fred

[SM6OID]

Stefano, please send me a PM with your e mail adress.

[AMLOVER]

Stu,

The supply I have attached in my prior post is working 2 years now under this attachment's set-up. It off course sags in mod peaks and after our conversation about how to improve a bit the drop I thought that adding some inductance and larger capacitance would improve it's performance.
The voltage with no load is 10850Vdc rectified from 3 transformers, one in each different phase, with a total of 7700Vac.
I soft start it through 3x10ohm/200w resistors switching on/off one transformer after the other by 50A ac switches. Never something strange happened till now except the sagging that is not that much like the smaller one phase supply for which I opened the topic.
I don't know if that sag comes from the difference of phases by 60 degrees but the DC voltage is there it should be 10850Vdc.
I only thought to save 2 chokes, to series connect only the bridges, to use only one very well isolated choke on the highest voltage point and put all smoothing caps in series on the end of the supply.

Fred,

I have never tried before to series connect only the bridges, even to lower voltage supplies and I am a bit suspicious on how it would behave.
Saving 2 chokes though is enough to give it a try.
I must now find the inductance I need.

Opcom,

Thank you for the simulator link. I will check it this evening.

SM6OID,

I will send you my email by PM as soon I find the way to send PMs here.

[AB2EZ]

The supply attached to your last post allows each transformer and each bridge rectifier to operate completely independently of the other two transformers and the other two bridges.

Each transformer and its associated bridge rectifier can deliver charge to its associated output capacitor, at times and in quantities that are independent of the other two.

Therefore the phase relationships of the three primary AC feeds don't affect the charging process for each of the three output capacitors.

The D.C. output voltages add, and the residual ripple voltages add with relative phase offsets that provide a beneficial reduction (vs. having all three primary supplies in the same phase) in the total residual ripple at the output.

Also, the peak primary current is lower than it would be if all of the primary supplies were in phase (and driven by a single phase of the AC service), and this improves the regulation.

If you force the currents from each bridge rectifier into a single series current flow (by using the alternate design (i.e the single choke design), I think the phase offsets of the primary supplies will cause significant problems.

Stu

[KA2DZT]

Putting xfmrs in series with just the rectifiers works fine provided the xfmrs are all on a single phase.

Three xfmrs on three different phases, I would agree with Stu, seems there would be too many issues.  Using separate filters as you pictured would be needed.

Fred

[AMLOVER]

Opcom,

I have added a rectifier 15kv as you can see in the PSUD2's rectifiers database.
Which way I can add it to my supply simulation?
The program offers a small rectifier's database to choose from.
I couldn't add any rectifier from the user's database in the program choices...
I was worried if these chokes must be gapped or not as far their inductance is small.

Stu and Fred,

I will keep then my set-up (total supplies in series) improving it by 200uf extra capacitance and if possible 3 smoothing chokes with small inductance.

[Opcom]

Putting xfmrs in series with just the rectifiers works fine provided the xfmrs are all on a single phase.

Fred

Not disagreeing with you about filter inclusion, by the way, because I know it works fine without filters; but always wanting to learn more I usually try to dig into the smallest details when different circuits are put forth.

[Opcom]

Opcom,

I have added a rectifier 15kv as you can see in the PSUD2's rectifiers database.
Which way I can add it to my supply simulation?
The program offers a small rectifier's database to choose from.
I couldn't add any rectifier from the user's database in the program choices...
I was worried if these chokes must be gapped or not as far their inductance is small.

I see that you added it online. Now, you have to download the file and put it in place of the rectifiers.txt that comes with the program.

BUT - I notice that you have a huge voltage drop as though it is a vacuum tube rectifier. That is not going to work and is going to mess people up. That unit should be removed from the library of the site. Please use the forum and tell them it was entered in error.

The short answer is that you can use the existing 575A vacuum tube model. It's a 15KV 1.5A mercury rectifier.

Otherwise you can use the file that comes with the program and add a 15KV rectifier to it. Carefully edit, then save, the rectifiers.txt file with windows notepad. You should first copy/back up the original rectifier.txt file as rectifiers-old.txt or something.

For a near-perfect diode rectifier, add this line as the first rectifier:
__perfect, SS, 0.01, 20, 9.353, 100000, 1000, 1000

For a stack of fifteen 1KV 10A diodes, use this line:
_1KV10Ax15, SS, 6, 5, 9.353, 15000, 100, 10
The "__" is not required but I used it to show to myself that it is one of mine whenever I open the file for editing.



===

As to the choke, it does not matter about the gap as long as the inductance is at the rated/desired current of 1.1A or whatever DC current you specify. It should be designed for rectifier filter use with appropriate voltage ratings. If the choke is gapped it may swing the inductance less as the current through it varies and be less likely to saturate. The point is that it should not saturate at the peak current because that is where its effect is wanted.

[AMLOVER]

Opcom,

I have tried hard to find a connection or to erase my wrong data rectifier addition but no result. If you have a communication connection or link to the Duncan's webpage let me know.
Thank you.