Power Supply PicoPSU schematic

alexs

Chassis Packer
Original poster
Nov 20, 2018
17
30
It looks like schematics for PDCBs are not easily available, at least I was unable to find one to look at when I wanted to. So I ended up tracing a spare board I've got. If you're wandering how these things work, what can and cannot be done, maybe thinking of doing something custom, well, here it is.



Dual-channel buck converter built around U1 provides 5V and 3.3V. This part is pretty much the reference design from ISL6440 datasheet. The 5V channel is always enabled, its output goes to ATX 5VSB. The second channel, 3.3V, shuts down in OFF state by the means of T3. The two switches on the right, Q5 and Q6, cut off ATX 12V and 5V when the board is OFF. ATX 5V is just 5VSB through a switch. ATX 12V is just input voltage through a switch. There's no switch on the 3.3V line, instead T3 shuts down the whole left side of the converter.

ATX -12V is made from 3.3V. I have no idea what's going on there but I guess it's got to be some sort of charge pump or something. Anyway, no 3.3V means no -12V output.

The second chip, U2, monitors ATX PS_ON input, provides ATX PG (power good) output, and controls the cut-off for 12V, 5V and 3.3V. FPOB is the control line that shuts down all non-5VSB ATX outputs.

Nearly half the parts on the back side of the board are there only to implement the OFF state. If PSON is wired externally to shut down the 12V supply, and there's an external source of 5VSB, the board can be reduced to basically ISL6440-based converter wired according to datasheet. ISL6440 is capable of producing PG output itself, doesn't need a supervisor chip.


Note all Ts are three-pin SOT-23s with no useful markings on them. T3, T5, T6 are just some small transistors, probably. Most likely not JFETs, that's just a convenient symbol. T2 is probably a dual Schottky diode like T4, probably to protect U2 from reverse polarity input. T4 sits right where there datasheet calls for a pair of diodes, that's how I know what it is.


Datasheet for ISL6440: https://www.renesas.com/eu/en/www/doc/datasheet/isl6440.pdf
Datasheet for WT751002: http://datalinker.com.hk/uploads/spec/WT751002S_v1.00.pdf







 
Last edited:

Thehack

Spatial Philosopher
Creator
Mar 6, 2016
2,800
3,650
J-hackcompany.com

Thanks :D

I had actually taken a look at this when he posted. It's pretty simple, just a 5v VRM, 3.3V VRM, and a SSR essentially.

The only issue about using 5v standby to power it is that some motherboards the 5v standby also powers the USB ports. That and the Meanwell units have very weak 5v sb and you'd need switch over from the 5v sb to the main 5v rail. At least that is my educated guess.
 
Last edited:

Thehack

Spatial Philosopher
Creator
Mar 6, 2016
2,800
3,650
J-hackcompany.com
It looks like schematics for PDCBs are not easily available, at least I was unable to find one to look at when I wanted to. So I ended up tracing a spare board I've got. If you're wandering how these things work, what can and cannot be done, maybe thinking of doing something custom, well, here it is.



Dual-channel buck converter built around U1 provides 5V and 3.3V. This part is pretty much the reference design from ISL6440 datasheet. The 5V channel is always enabled, its output goes to ATX 5VSB. The second channel, 3.3V, shuts down in OFF state by the means of T3. The two switches on the right, Q5 and Q6, cut off ATX 12V and 5V when the board is OFF. ATX 5V is just 5VSB through a switch. ATX 12V is just input voltage through a switch. There's no switch on the 3.3V line, instead T3 shuts down the whole left side of the converter.

ATX -12V is made from 3.3V. I have no idea what's going on there but I guess it's got to be some sort of charge pump or something. Anyway, no 3.3V means no -12V output.

The second chip, U2, monitors ATX PS_ON input, provides ATX PG (power good) output, and controls the cut-off for 12V, 5V and 3.3V. FPOB is the control line that shuts down all non-5VSB ATX outputs.

Nearly half the parts on the back side of the board are there only to implement the OFF state. If PSON is wired externally to shut down the 12V supply, and there's an external source of 5VSB, the board can be reduced to basically ISL6440-based converter wired according to datasheet. ISL6440 is capable of producing PG output itself, doesn't need a supervisor chip.


Note all Ts are three-pin SOT-23s with no useful markings on them. T3, T5, T6 are just some small transistors, probably. Most likely not JFETs, that's just a convenient symbol. T2 is probably a dual Schottky diode like T4, probably to protect U2 from reverse polarity input. T4 sits right where there datasheet calls for a pair of diodes, that's how I know what it is.


Datasheet for ISL6440: https://www.renesas.com/eu/en/www/doc/datasheet/isl6440.pdf
Datasheet for WT751002: http://datalinker.com.hk/uploads/spec/WT751002S_v1.00.pdf








Do you have any data on the back of the board?
 

alexs

Chassis Packer
Original poster
Nov 20, 2018
17
30
What kind of data?
It's marked "Rev 1.1" near the 5 pin connector, and there's nothing else particularly useful there.
 

adalbert

What's an ITX?
New User
Jun 20, 2021
1
0
Hello, I know that this is old topic, but I wonder if OP or anyone else tried to mod this PSU to provide high current single +5V output. What I mean is connecting the two phases together (+3.3V output and +5V output) and jumpering feedback resistor pins, so both phases use only +5V feedback resistors. Datasheet specifies that ISL6440 runs 180 out of phase, but it has no reference design for single voltage output. So I wonder if such configuration would be possible (some other PWM chips offer this). Any opinions on that?

If this mod would work, then two Picos could be paralleled - first would have 2-phase 3.3V and second would have 2-phase 5V output. Some jumper wires would probably also be needed to be ran between WT751002 chips.

That would help running very old PCs with AGP graphic cards and +5V intensive CPUs. Currently it's pretty difficult to find DC-DC PSUs with strong +5V / 3.3V lines as modern PCs are +12V intensive.