Good day to yall!.
My attempts to design a 400VA (50V to 1KV) adjustable DCDC SMPS has been troublesome. Two versions have been made but are either unstable or < 75% efficient. Push pull using 50KHZ with 2x series outputs was recommended the most and that is my general approach. Oh, I definitely have a more to learn in order to fine tune this, but it seems like a good start is define passive component placement then choose the right specs. "?" on the diagram are a WIP. I have lots of work to do on the 2nd side with feedback. Any and all suggestions very welcome!

 

I do not understand how the two Bridge Rectifiers function properly with the Secondary having a Center Tap ???

 

You only need one Bridge Rectifier across the outer two terminals of the transformer, ignore the centre tap.

 

Oops, my bad! Thanks for your responses. This latest schematic has changes and hopefully the correct bridge layout. Any more advice?

 

No it's still drawn wrong, you put your bridge rectifier across the end terminals of the 80t windings. and link the middle together, this adds the transformer output voltage.

 

No it's still drawn wrong, you put your bridge rectifier across the end terminals of the 80t windings. and link the middle together, this adds the transformer output voltage.

He is doing that because he uses 600V capacitors in a 1KV power supply. He could use a single bridge if the rectifiers are rated high enough.

Trying to regulate over such a wide range with a SMPS is very difficult. You might be able to get 250V to 1KV but that may still be a challenge. Try 500 to 1KV, you might have better success.

 

I would use a TL494 chip, as it will work upto 40V DC, also i would wind a separate 10t winding and use that as the feedback path.

Also your Current Sense isn't connected to the Mosfet.

 

Here is my logic on feedback. V regulation does not have to be perfect. +-10VDC fine. Sag acceptable, so long as it recovers and can maintain 400VA. I was thinking of doing feedback off one of the bulk caps. So that is 25V to 500V feedback. Or use dividers and a low pass filter to buffer the feedback voltage before it gets to the 4430. So make attempts to sacrifice accuracy for stability. What do you think?

btw, the 25V lower supply limit should say 50V. So 50V to 1000V supply

 

Since there are two similar power supplies in series, using the voltage of one for feedback is reasonable, just unusual. It was done a lot years ago to be able to use the cheaper rectifier diodes, and it is still valid today. I have seen one other design of a supply with similar specifications and that one solved a bunch of problems by putting a lot more iron in the transformer stack of laminations. Evidently the transformer design rules don't work for these kinds of supplies and lots more magnetic material is needed than the formulas predict.

 

Also your Current Sense isn't connected to the Mosfet.

Take another look and you will see that the 0.01R is connected to the mosfets
It appears to be a "dual labelled" part - apparently to make the drawing wiring simpler.

 

Oops, my bad! Thanks for your responses. This latest schematic has changes and hopefully the correct bridge layout. Any more advice?

I would make one more change ...
I would relocate the (-) wire that starts at the left side of the Upper Diode Bridge,
I would remove that wire from the left side of the lower choke and connect it to the right side of the lower choke.

I think, each bridge should pass through one choke each, to keep the 2 series connected secondaries "symmetrical".

 

Thx for the response! Can you clarify your change? Each bridge goes through a choke. Why move the lower bridge output to the right side of the choke? Or maybe this is not the wire your talking about. Also is your "(-) wire" change suggestion purely cosmetic?

 

The 20:1 ratio of output voltage is really hard. (50V - 1kV)

You might get the range, but the efficiency will be horrible.

 

The 20:1 ratio of output voltage is really hard. (50V - 1kV)

You might get the range, but the efficiency will be horrible.

As I study the output portion of the circuit it looks like for the lower voltage range the two bridge rectifier circuits could be switched into a parallel connection, since the sense is only using the grounded half of the output. That will improve both the range at the bottom end and also the current capability would be doubled, almost for free. AND it will also improve the efficiency. at the lower end of the range, and probably also make the regulation better.

 

That is a very interesting approach! Switched capacitor setups were once entertained when starting this project but were highly complicated. They needed some 20 to 30 stages, not isolated and a VERY complex controller.

I have already decided to take FB from the lower bridge/cap section such that the FB range would be 50Vto500V and hope for the best. But, your suggestion would definitely make it more stable and efficient. Maybe my user physical dial could also perform a switchover at a certain point. Two fets maybe, to parallel or series the output. Just have to be careful not parallel differential voltages and burn up fets.

 

That is a very interesting approach! Switched capacitor setups were once entertained when starting this project but were highly complicated. They needed some 20 to 30 stages, not isolated and a VERY complex controller.

I have already decided to take FB from the lower bridge/cap section such that the FB range would be 50Vto500V and hope for the best. But, your suggestion would definitely make it more stable and efficient. Maybe my user physical dial could also perform a switchover at a certain point. Two fets maybe, to parallel or series the output. Just have to be careful not parallel differential voltages and burn up fets.

It makes a lot more sense to use either a switch or a mechanical relay to do the switch over, because of the voltages and the requirements of timing and off state. Consider that the negative terminal of the upper bridge is moving from the high voltage side of the lower bridge to the negative side, which is the common. And likewise, the positive terminal of the upper bridge, along with the output connection, is now being attached to the positive terminal of the lower bridge. Really, that needs to happen with a mechanical switch having lots of time and space between contacts.An added benefit of using a mechanical switch is that it would provide an obvious indication of which voltage range was in use.

 

Yeah, I'm definitely considering DPDT relays with extra contact travel. The only issue with these is arcing messing up my mcu boards with emi. Snubbers and RC across contacts helps a bit... If upon closure/break, if differential voltages are < 100V, emi sparks are ok. Anything higher and trouble sets in. Do you think upon make/break of relay that voltage differences will be high?

 

Yeah, I'm definitely considering DPDT relays with extra contact travel. The only issue with these is arcing messing up my mcu boards with emi. Snubbers and RC across contacts helps a bit... If upon closure/break, if differential voltages are < 100V, emi sparks are ok. Anything higher and trouble sets in. Do you think upon make/break of relay that voltage differences will be high?

I was thinking that this would be a power supply that was set at voltages for different applications, not for ramping up and down under load. That would require a lot more filtering to keep any switching noise away from what is being powered. An earlier post had indicated a number of different applications and so I was not considering the ramping challenge.
AND, I have not seen snubbers on high voltage switching contacts.

 

Do you need one power supply to run over such a wide range?
It may be better to have a few supplies more specialized.

 

I was thinking that this would be a power supply that was set at voltages for different applications, not for ramping up and down under load. That would require a lot more filtering to keep any switching noise away from what is being powered. An earlier post had indicated a number of different applications and so I was not considering the ramping challenge.
AND, I have not seen snubbers on high voltage switching contacts.

Sorry for the confusion, but this supply is not changing under load. When it set at a voltage, it is not changed during any loads. My "differential" question was in regards to the bridges/caps being at different voltages when changing configuration states using the relay. In the perfect world, the caps should be the same voltage assuming balance resistors are doing there job right.