High range input voltage isolated topology

Thread Starter

kobi209

Joined May 28, 2017
19
Hi All,
I need to design a DC/DC isolated converter, with a high range of input DC voltage: 50-950V and output 12Vdc, ~30W.
The first topology I thought about was Flyback, which is cheap and easy, however has some drawbacks.
Have somebody had the chance to design high voltage range power supply?
I'd love to hear your suggestions.
Thanks.
 

Irving

Joined Jan 30, 2016
3,843
950v is exceptionally high! Why so big a range?

Its going to be difficult to find a solution to that breadth of range. You'll need some form of pre-regulator stage. I'd look at a resonant LLC converter, which can handle a wider input range than Flyback. Even so, probably not all in one go.
 

ronsimpson

Joined Oct 7, 2019
2,988
Part of the problem is you need high voltage silicon and high current silicon.
1000V x 30mA = 30 watts AND 50V x 600mA = 30 watts
non isolated simple buck wide range I remember this part only good to 1000V. Good read if you are thinking about preregulator.

Power Integrations has a application note I can't find now for wide range. Most of their parts are only good to 850 to 1000V but there a application note on going to 1500V to 2kv. Will look 8 hours later. I did a design this way years ago. Flyback type. If you go looking for it "power supply for power metering".
 

DickCappels

Joined Aug 21, 2008
10,152
The pre-regulator would be a buck/boost (the term is frequently misused. By this I mean the converter can be used in either mode and adjusts its operation as needed to achieve the desired regulation). Otherwise, you can (?) switch ranges, by driving different windings or taps.

Sanyo has a pretty good collection of medium current 1500 volt transistors. The lower current dynamic focus transistors are included here because when working with voltages in that range (over 900V) they might come in very handy.

1592916952516.png
1592917051143.png
 

Deleted member 115935

Joined Dec 31, 1969
0
Hi All,
I need to design a DC/DC isolated converter, with a high range of input DC voltage: 50-950V and output 12Vdc, ~30W.
The first topology I thought about was Flyback, which is cheap and easy, however has some drawbacks.
Have somebody had the chance to design high voltage range power supply?
I'd love to hear your suggestions.
Thanks.

Whats your experience in Power supply designs, especially at high voltages ?
 

SiCEngineer

Joined May 22, 2019
442
Do you have any other requirements, such as transient response performance, does it need to be fixed frequency, etc. As much as you can say will help with suggestions more.
 

Thread Starter

kobi209

Joined May 28, 2017
19
Hi All,
I need to design a DC/DC isolated converter, with a high range of input DC voltage: 50-950V and output 12Vdc, ~30W.
The first topology I thought about was Flyback, which is cheap and easy, however has some drawbacks.
Have somebody had the chance to design high voltage range power supply?
I'd love to hear your suggestions.
Thanks.
950v is exceptionally high! Why so big a range?

Its going to be difficult to find a solution to that breadth of range. You'll need some form of pre-regulator stage. I'd look at a resonant LLC converter, which can handle a wider input range than Flyback. Even so, probably not all in one go.
The big range is due to the solar panel string voltage range can be implemented. This is anauxiliary psu for dc/ac 3-ph inverter. The nominal inverter input voltage is 750V, and this is the 30w working point of the dc/dc. 950V is for the largest string, and the low voltages are for a case of off grid mode, when the system is expected to wake up with minimal panels voltage (i.e at early morning).
 

Irving

Joined Jan 30, 2016
3,843
Would be a damn sight simpler, cheaper and easier to have a 12v 30Ah LiFe battery pack charged off one of the AC phases. It'll physically last >15y under these conditions, and will deliver 30W for 12 hours on a single charge!
 

Thread Starter

kobi209

Joined May 28, 2017
19
Would be a damn sight simpler, cheaper and easier to have a 12v 30Ah LiFe battery pack charged off one of the AC phases. It'll physically last >15y under these conditions, and will deliver 30W for 12 hours on a single charge!
There are some major disadvantages with your solution. A battery pack is not so reliable, it's too big in PCB area and it requires additional electronics for BMS etc. and most important - this solution is very expansive comparing a BOM of 9-10$ of a flyback.
 

ronsimpson

Joined Oct 7, 2019
2,988
Wide range supply This supply does not meet your requirement. It does show the trick I did years ago. The IC U1 can handle 750V or something. Don't remember. Q1 can be another 750 to1000V. When I built the supply high voltage transistors were not in production because CRT monitors and TVs are out. MOSFETs did not go higher than 1kv. I needed 1200V to 1500V for the switch. By standing Q1 on top of U1 you can to high voltage input. This trick works on other power ICs.
1592950227618.png
I will look later to see if I can find the files.
 

RPLaJeunesse

Joined Jul 29, 2018
252
Just a thought - consider a non-isolated buck regulator to knock the input down to say 250 volts max. Make sure it will pass inputs below 250V through, at least down to 50V. Follow that with an isolated converter having a 250V to 50V (5:1) input range (not complicated) to get the 12V output. But that 1KV range buck is not going to be a trivial design.

Since the second stage is isolated it's possible to make the first stage switch be in the negative side, opposite what most bucks do. This gets more switch options available, with 1200V n-channels parts as low as $2, and the regulator stage could run on less than 50V if need be. Output voltage level shifting could be problematic, but doable using a bipolar transistor trick or two.
 
Top