Can you set up stacked output rectifiers in this way?

Thread Starter

SiCEngineer

Joined May 22, 2019
299
I always see stacked rectifiers when working with high voltage flybacks, where multiple identical rectifiers and capacitors are placed across multiple windings then connected in series. They are always set out a certain way on the PCB, where each gets its own dedicated area on the secondary of the converter PCB.

I have been giving a very small amount of thought to whether they could be laid out differently for high power density designs. Could they be stacked physically on top of each other, assuming that necessary spacing was placed between them and the connections from the transformer and to the output capacitors, filters and load were made correctly with wires?

I have attached an image toillustrate what I mean. I am very open to being told why this is a stupid idea and won't work, but I'd like to know why exactly. Any thoughts appreciated. Please understand it is a very rough sketch but I hope I get across the idea correctly..Cheers.

EDIT: Note in the image the connections from the transformer to the three stacked rectifiers in configuration 2 are incorrect. Each transformer winding will connect to the top of each board - and each board is then connected in series at the rightmost side, and then the final voltage is connected out from the stack of rectifiers to the capacitor filter and load.IMG_2656.jpg
 
Last edited:

Dodgydave

Joined Jun 22, 2012
10,035
The only advantage I can see is using a lower voltage rectifier and smoothing caps , I've only seen this method used in stereo amps using two different outputs from the transformer.
 

Thread Starter

SiCEngineer

Joined May 22, 2019
299
The only advantage I can see is using a lower voltage rectifier and smoothing caps , I've only seen this method used in stereo amps using two different outputs from the transformer.
Thanks for your reply. Maybe I am not being clear. Both configurations will have lower voltage rectifiers and smoothing caps - this is the reason these are typically employed in HV supplies. And they work very well. Good for higher power outputs as well as you can spread out the power and heat dissipation.

My question is about the way to construct these rectifiers physically. In the first embodiment, the rectifiers are placed side by side and connected in series through the PCB, as normal.

In my probably incorrect embodiment, the rectifiers are "literally" stacked on top of each other. The PCBs will be identical but phsyically seperate, placed on top of each other, and connected in series from the bottom up. The final output voltage at the top of the stack is passed down back to the bottom of the PCB through a cable where the final voltage will connect to the rectifiers, the filters and the load.

Hope that clarifies things a little.

Best.
 

ronsimpson

Joined Oct 7, 2019
1,319
It is not clear to me how the bottom picture works. I think it does not work. It looks like you have one long winding with taps.

I have made many high voltage supplies where there are many low voltage supplies stacked up.
When operating at high frequencies the capacitance inside the transformer becomes a big problem. With high voltage, keeping the spacing is a problem. Not at 60hz but at 100khz to 1mhz switching supplies I often wind one layer of wire in the transformer and them pull out the wires to diodes. Add tape and the next layer of wire.

A 1000 turn secondary has a very low resonant frequency. But 4X250 turns will run about 16x faster.
 

Thread Starter

SiCEngineer

Joined May 22, 2019
299
It is not clear to me how the bottom picture works. I think it does not work. It looks like you have one long winding with taps.

I have made many high voltage supplies where there are many low voltage supplies stacked up.
When operating at high frequencies the capacitance inside the transformer becomes a big problem. With high voltage, keeping the spacing is a problem. Not at 60hz but at 100khz to 1mhz switching supplies I often wind one layer of wire in the transformer and them pull out the wires to diodes. Add tape and the next layer of wire.

A 1000 turn secondary has a very low resonant frequency. But 4X250 turns will run about 16x faster.
Hi Ron. Yes, I made a mistake there. If you see my EDIT note.
Basically, there will be the exact same number of transformer windings. Exact same number of rectifier diodes and smoothing capacitors.

The only difference is that instead of having 1 discrete PCB containing all three rectifiers, they are physically stacked upon one each other and the connections to the PCBs are made through some kind of connector from the PCB below. So rather than the windings be spread out horizontally, they are spread out vertically one on top of each other.

The only issue I envisage is to do with creepage/clearance, considering the high voltage. But if the spacing between the rectifiersis good enough, it should be fine - additional insulation between the physical stacks could also increase the insulation between them.
 

Dodgydave

Joined Jun 22, 2012
10,035
Thanks for your reply. Maybe I am not being clear. Both configurations will have lower voltage rectifiers and smoothing caps - this is the reason these are typically employed in HV supplies. And they work very well. Good for higher power outputs as well as you can spread out the power and heat dissipation.

My question is about the way to construct these rectifiers physically. In the first embodiment, the rectifiers are placed side by side and connected in series through the PCB, as normal.

In my probably incorrect embodiment, the rectifiers are "literally" stacked on top of each other. The PCBs will be identical but phsyically seperate, placed on top of each other, and connected in series from the bottom up. The final output voltage at the top of the stack is passed down back to the bottom of the PCB through a cable where the final voltage will connect to the rectifiers, the filters and the load.

Hope that clarifies things a little.

Best.
So you're saying you will have 3 pcbs stacked Vertically with it's own output, and all three outputs will be in series to provide one Larger output?
 

Thread Starter

SiCEngineer

Joined May 22, 2019
299
So you're saying you will have 3 pcbs stacked Vertically with it's own output, and all three outputs will be in series to provide one Larger output?
Yes, that is the idea. The fundamental behaviour of the circuit and the components don't change, just how they are physically layed out. I had this idea randomly last night. For high density systems it might be useful because you can have the multiple rectifiers but safe in size and area of the PCB by instead stacking them upon each other, meaning a more power dense design.
 

Dodgydave

Joined Jun 22, 2012
10,035
Current will be lower because you're going through more silicon devices and will create more heat waste, rather than one rectifier and one output. It's ok for something like a Tesla coil or HV fly swatter..
 

Thread Starter

SiCEngineer

Joined May 22, 2019
299
Current will be lower because you're going through more silicon devices and will create more heat waste, rather than one rectifier and one output. It's ok for something like a Tesla coil or HV fly swatter..
I think maybe I am still being unclear. The number of silicon devices in approach number one is exactly the same as the number of silicon devices in approach number two.
The number of rectifiers, capacitors etc is exactly the same. For example, to generate 750V, we could have 3 rectifiers and filters which are each 250V.
In approach number one, all the rectifiers are connected horizontally on the single PCB. In approach number two, there are three smaller but identical PC Boards which each contains a set of rectifiers and capacitors. These are then connected to each other vertically, through some kind of connector from the bottom most PCB to the top most PCB. From the top most PCB, a wire extends downwards and back into the PCB which contains the final output capacitor, filter, and the high voltage load.
I hope I can get across the fact that the actual function of the circuit and the number of components does not change - only the way in which the rectifiers boards are connected to one another.
 

Thread Starter

SiCEngineer

Joined May 22, 2019
299
Something like this - but with the rectifiers. Do you see how the boards slot into the holes, allowing a higher power density supply, as the boards are utilizing the z-dimension as well as the x and y direction like in normal converters.

I am unsure how the connetion is made in these - maybe the copper plates on the daughterboards are soldered into the motherboard whihc has matching shapes of copper?

Do you think this could work for a rectifier consisting of multiple identical rectifeir daughterboards?

1602503887061.png
 
Top