Low cost DC-DC converter/SMPS design idea and suitable topologies

[Md Mubdiul Hasan]

Hi there,

I am looking for a customized design where good power solution can be applicable.

My aim is to develop a DC-DC sutable toloplogy ( flyback, buck, isolated ,..so on) converter with following parameter, with the idea https://www.smps.us/topologies.html

INPUT: 250 to 650 VDC

OUTPUT: 24V/5A , 5V/3A, 12V/2A.

Lets think its saparate system for all output. Offcourse this design is challenging since it has wide range of input.

 Its for battery charger or EV application.


Right at this moment I am thinking about quasi resonant flyback converter topology.
Take a look here what Infineon expart has made a note,  https://www.mouser.com/pdfdocs/2-9.pdf


What do you think ?


 

[T3sl4co1l]

Front end PFC solves a lot of issues, and is often compulsory (i.e. to meet CE).  Maybe with a SEPIC topology, if you don't mind the somewhat higher losses -- this avoids the high DC bus of a boost PFC.  (Which still isn't a problem with respect to 900V+ transistors being readily available, but gate drivers not so much. There are a few bootstrap drivers that can do that, but most likely you'll need a transformer or fully isolated solution.)  With PFC, LLC is reasonable for the iso stage, and probably not a bad idea for the power level and if it needs to be compact.  Otherwise, sure a quasi resonant flyback isn't bad, or 2-switch forward/flyback.

Have seen industrial Meanwell and other supplies with 240-480VAC input (or even lower, I forget?), and I think either plain flyback, or 2-switch forward (no PFC).  Your power level isn't crazy, a little high for flyback but understandable given the range.

Tim

[NiHaoMike]

PFC is irrelevant since it's DC input. I would say go for a half bridge design, generating the 24V directly and then buck converters for the 12V and 5V.

[Md Mubdiul Hasan]

PFC is irrelevant since it's DC input. I would say go for a half bridge design, generating the 24V directly and then buck converters for the 12V and 5V.

You are right. PFC is not acceptable for DC-DC application.
Right at this moment, TI suggest a very easy flyback type, take a look, just need to remove bridge diode, https://www.ti.com/lit/ds/symlink/ucc28740.pdf?ts=1637161141066&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FUCC28740%253FkeyMatch%253DUCC28740%2526tisearch%253Dsearch-everything%2526usecase%253DGPN
Here may be transformer turn can be modify.

Seconly, Infineon suggest a good design, https://www.infineon.com/dgdl/Infineon-ICE2QS02G-DS-v02_00-en.pdf?fileId=db3a30431a5c32f2011abef4f7fd3bc8
12V.5V and 24V can be set by 950V or 900v mosfet, https://www.infineon.com/dgdl/Infineon-IPD95R450P7-DS-v02_01-EN.pdf?fileId=5546d462636cc8fb01643b1b778a55d4 or
https://www.infineon.com/dgdl/Infineon-IPW90R120C3-DS-v01_00-en.pdf?fileId=db3a3043183a955501185000e1d254f2

Kindly review those datasheet and make a decision.

[T3sl4co1l]

For some reason I keep forgetting it's DC. Yes, obviously PFC is N/A.  So, unless you're putting in a prereg, LLC is probably out.  The rest are fine.

Tim

[TimNJ]

Quasi-resonant flyback is probably the most straightforward approach. It can handle wide-input + wide-output range without too much effort.

The transformer will probably need to be roughly PTS40/20, PTS40/25, or PQ40/30 size...It's rather clunky, but it's do-able.

http://www.chinadmegc.com/en/upload/201305/88d52b8ce984e6490dfc8bd774136b8e.pdf
https://www.tdk-electronics.tdk.com/inf/80/db/fer/pq_40_30.pdf

The main thing you need to consider for wide output voltage range is how to maintain a (relatively) constant Vcc for the primary-side controller. Not terribly difficult, just need two auxiliary windings instead of one. Some new controllers (those specially designed for USB PD, like TEA19363T) handle this neatly. Otherwise, you can just use a "normal" QR flyback controller and use an external NPN/zener regulator (or two).

Beyond 120W, you'd probably be better off with LLC, but you'd likely need a tracking buck/boost pre-regulator to track the output voltage setting. LLC really only works well with a fixed conversion-ratio...meaning: The transformer turns ratio should match the actual Vin:Vout ratio. There would need to be some sort of communication between LLC controller/user input and the pre-regulator to make sure Vin is adjusted to match the requested Vout setting. Maybe timing and other logic involved.

[TimNJ]

BTW, did I misinterpret this?

OUTPUT: 24V/5A , 5V/3A, 12V/2A

Are there 3 outputs? Or 1 output and 3 options?

If you need all outputs simultaneously (at full load), then that's 159W. If that's the case, I'd probably design a 24V/165W LLC stage, and just throw two buck-regulators on the output for the lower power 5V and 12V rails. You'd still need a roughly 700V boost rail though.

[Md Mubdiul Hasan]

BTW, did I misinterpret this?

OUTPUT: 24V/5A , 5V/3A, 12V/2A

Are there 3 outputs? Or 1 output and 3 options?

If you need all outputs simultaneously (at full load), then that's 159W. If that's the case, I'd probably design a 24V/165W LLC stage, and just throw two buck-regulators on the output for the lower power 5V and 12V rails. You'd still need a roughly 700V boost rail though.

Nice to get back your reply.
Lets consider 24V/5A as a single system.
Other 2 option can consider at multiple output.

[Md Mubdiul Hasan]

Quasi-resonant flyback is probably the most straightforward approach. It can handle wide-input + wide-output range without too much effort.

The transformer will probably need to be roughly PTS40/20, PTS40/25, or PQ40/30 size...It's rather clunky, but it's do-able.

http://www.chinadmegc.com/en/upload/201305/88d52b8ce984e6490dfc8bd774136b8e.pdf
https://www.tdk-electronics.tdk.com/inf/80/db/fer/pq_40_30.pdf

The main thing you need to consider for wide output voltage range is how to maintain a (relatively) constant Vcc for the primary-side controller. Not terribly difficult, just need two auxiliary windings instead of one. Some new controllers (those specially designed for USB PD, like TEA19363T) handle this neatly. Otherwise, you can just use a "normal" QR flyback controller and use an external NPN/zener regulator (or two).


Beyond 120W, you'd probably be better off with LLC, but you'd likely need a tracking buck/boost pre-regulator to track the output voltage setting. LLC really only works well with a fixed conversion-ratio...meaning: The transformer turns ratio should match the actual Vin:Vout ratio. There would need to be some sort of communication between LLC controller/user input and the pre-regulator to make sure Vin is adjusted to match the requested Vout setting. Maybe timing and other logic involved.

Initially I was thinking Quasi-resonant flyback, but its cost relatively high compared to flyback one. Infineon suggest a good pair, https://community.infineon.com/t5/Power-MOSFET/QR-Flyback-Converter-for-DC-DC-buck-action/m-p/324704

I do understand the benifits of LLC, but I didnt find any suitable controle IC for wide range of input. What I have find an expensive one.https://www.wolfspeed.com/products/power/reference-designs/crd-06600dd065n

Now TI gave me a solution which is cheaper and easy, I am trying to make it, kindly take a look here, and suggest possibilities of 24V/5A or 120W, https://www.ti.com/lit/ds/symlink/ucc28740.pdf?ts=1637161141066&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FUCC28740%253FkeyMatch%253DUCC28740%2526tisearch%253Dsearch-everything%2526usecase%253DGPN

Dont you think, the transformers turns ratios, output inductor, protection circuit and input can be modified ?