Power Supply A Guide to 12V PSU

[Thehack]

What exactly needs to be done to get SYNC working?

Can I just connect the 5VSB and PS_ON from the PicoPSU pins to the appropriate connectors on the EPP-400? Or would that be a bad idea?

No that would not work. The simplest route is to use the PS_on micro connector on the meanwell. When connected to ground the psu turns off. So you use a normally closed switch (closed in the off position) and when you manually switch it, the meanwell turns on and outputs 12V. So turning the pc requires two switches: the PS on switch and the regular power button.

The sync feature is something different as it essentially uses a load switch, actually more technical term is solid state relay (I used the term load switch initially since I imagined something that is aftermarket to the plug in board, but since it is now integrated it is essentially a solid state relay). It is for controlling a second board that turns on or off the 12v rail.

 

[NateDawg72]

Is PS_on on the meanwell micro connector fundamentally different than the PS_on signal from the motherboard? That's all I'm hung up on right now, don't understand why that wouldn't work.

 

[Thehack]

Is PS_on on the meanwell micro connector fundamentally different than the PS_on signal from the motherboard? That's all I'm hung up on right now, don't understand why that wouldn't work.

Correct. It gets quite complicated explaining concept of operation of the circuits.

The most basic issue is that the meanwell operate reverse of ATX PS_on signal. Meanwell, when PS_on is connected to ground, it turns off.

ATX PS_on is the opposite. When connected to ground it turns on.

The plug in unit also needs the 12V already on to operate the board logic. The latching PS_on signal on the atx is controlled by the motherboard using the 5V standby provided by the plug in unit.

 

[aquelito]

@Thehack : pretty funny, I have exactly the same conversation with the designer behind the load switch.
He is using the PS_ON and 5VSB from the Meanwell unit to only turn on the 12V Output of the PSU when the PC is turned on... no load switch required.
I will be back with more info asap.

Regarding your distro solution works just fine with a server motherboard Asrock IMB-181-L.
No funny business upon PC shutdown : fans, mouse, etc. remain powered down.

However, I have not tested it with my Asrock Z170 Fatal1ty ITX yet, which did not like the combo G-Unique / Meanwell / custom load switch for both GPUs.
https://smallformfactor.net/forum/t...o-gpus-and-two-240-radiators.7011/#post-95364

I had to do a Clear CMOS as this power solution was apparently messing with the boot sequence after a couple reboot.

 

[Valantar]

He is using the PS_ON and 5VSB from the Meanwell unit to only turn on the 12V Output of the PSU when the PC is turned on... no load switch required.

Sorry for breaking into your discussion here, I'm just trying to keep up. Could you please clarify this a bit? It sounds like you're saying the Meanwell unit provides a 5VSB rail, which ... I've never seen, at least. What units are you discussing? Isn't the 5VSB generated by the DC-ATX unit, running off the 12V rail from the MW?

 

[Thehack]

@Thehack : pretty funny, I have exactly the same conversation with the designer behind the load switch.
He is using the PS_ON and 5VSB from the Meanwell unit to only turn on the 12V Output of the PSU when the PC is turned on... no load switch required.
I will be back with more info asap.

Regarding your distro solution works just fine with a server motherboard Asrock IMB-181-L.
No funny business upon PC shutdown : fans, mouse, etc. remain powered down.

However, I have not tested it with my Asrock Z170 Fatal1ty ITX yet, which did not like the combo G-Unique / Meanwell / custom load switch for both GPUs.
https://smallformfactor.net/forum/t...o-gpus-and-two-240-radiators.7011/#post-95364

I had to do a Clear CMOS as this power solution was apparently messing with the boot sequence after a couple reboot.

I'll be interested to hear you'd accomplish that. You'd need a separate circuit logic since the meanwell operates reverse of atx.

 

[Thehack]

Sorry for breaking into your discussion here, I'm just trying to keep up. Could you please clarify this a bit? It sounds like you're saying the Meanwell unit provides a 5VSB rail, which ... I've never seen, at least. What units are you discussing? Isn't the 5VSB generated by the DC-ATX unit, running off the 12V rail from the MW?

Most psu provides 5VSB. The meanwell does as well.

It provides:
5VSB
PS_on
12v fan

 

[Valantar]

Most psu provides 5VSB. The meanwell does as well.

It provides:
5VSB
PS_on
12v fan

Thanks for clearing that up

After having a look at the datasheets for the MW EPP-200-12 and 400-12, it seems like only the higher power models have 5VSB; it's listed under "function" for the 400-12 and has its own output, while on the 200 there's nothing. The same goes for PS_ON. I take it that means there's no way for this kind of setup to work on the lower power units, and they'll remain "on" no matter what?

 

[Thehack]

Thanks for clearing that up

After having a look at the datasheets for the MW EPP-200-12 and 400-12, it seems like only the higher power models have 5VSB; it's listed under "function" for the 400-12 and has its own output, while on the 200 there's nothing. The same goes for PS_ON. I take it that means there's no way for this kind of setup to work on the lower power units, and they'll remain "on" no matter what?

Yup. Which is fine. These things have low no load power consumption.

 

[aquelito]

I'll be interested to hear you'd accomplish that. You'd need a separate circuit logic since the meanwell operates reverse of atx.

Great minds think alike

I asked him indeed to design a small circuit that allows to allow to short and maintain shorted the PS_ON wire of the Dell DA-2 brick or Liteon AIO PSU , with the computer Power Button !

Since these kind of PSUs have a PS_ON function, you don't need a load switch if you can "start" them with the power button.

I think he tested this circuit with the Meanwell PS_ON first but his message was very short and I may be wrong.

I am waiting for his confirmation.

 

[Thehack]

Great minds think alike

I asked him indeed to design a small circuit that allows to allow to short and maintain shorted the PS_ON wire of the Dell DA-2 brick or Liteon AIO PSU , with the computer Power Button !

Since these kind of PSUs have a PS_ON function, you don't need a load switch if you can "start" them with the power button.

I think he tested this circuit with the Meanwell PS_ON first but his message was very short and I am waiting for his confirmation.

The cheapest way I think is to do a momentarily and latching switch that is normally closed.

Concept of operation:
- the normally closed switch keeps the psu in the off state by shorting PS_on and ground
- when pressed, the button operates both as a latching and momentarily switch, the latching function switches the psu to on by opening the latching circuit, and once the 12V is applied and the board logic begins, the PC turns on when it reads the momentarily switch to turn on.
- the user then let go of the button and now the computer turns on.

This can be done completely mechanically and will work 90% of time as long as you don't use a deep hibernate function that requires the power button pressed to turn the pc back on. I can imagine some fringe cases where you're messing with it.

Assuming this is a simple electromechanical switch, can we have have a dual stage switch as well, where when half pressed it operates as momentarily switch.

A small simple circuit can also be programed where when fully pressed, it operates a time delayed momentarily switch that automatically turn the pc on.

 

[aquelito]

Does it allow to maintain the brick PS_ON wire shorted ?

Some people already tried but they had to use a complicated circuit involving an optocoupler to maintain the PS_ON shorted as long as the computer is running :

1. Uses the power-on button to start the Dell&PicoPSU simultaneously
2. Uses the power-LED pins from the motherboard and an optocoupler to keep the Dell on

http://www.silentpcreview.com/forums/viewtopic.php?f=6&t=65652

The designer I'm working with found a much smaller solution but I have yet to wait for his explanations (I'm no EE).

 

[Thehack]

Does it allow to maintain the brick PS_ON wire shorted ?

Some people already tried but they had to use a complicated circuit involving an optocoupler to maintain the PS_ON shorted as long as the computer is running :



http://www.silentpcreview.com/forums/viewtopic.php?f=6&t=65652

The designer I'm working with found a much smaller solution but I have yet to wait for his explanations (I'm no EE).

For a simple electromechanical switch this is how I would have it. It may require two presses for it to work, but it is two function switch based on two stages. You can also have two separate switches, one a toggle and one a power button and it'll work all the same.

For something fancier this is what I imagine:

The long press function should be P1 and P2 change stage between closed and open.

You can also take second input from the power LED that whenever it reads a state change from high to low it reconnects (closes) P1 and P2.

 

[Valantar]

With a mechanical two-stage button setup like that, how would it "know" when to turn off the MW PSU after the system has been shut down without requiring pressing the power button again?

 

[Thehack]

With a mechanical two-stage button setup like that, how would it "know" when to turn off the MW PSU after the system has been shut down without requiring pressing the power button again?

It's manual you do it yourself. Supposed to be cheap to make since latching circuits are common.

 

[Valantar]

It's manual you do it yourself. Supposed to be cheap to make since latching circuits are common.

So, you're bringing the AT PSU spec back? "It is now safe to turn off your computer"

Not that it's a bad solution per se (this is after all intended for enthusiasts who know what they're getting) - quite the opposite, really - but I have to wonder how this will affect the PC. If the momentary switching action (ATX-ON) is before the latching one (MW PS-ON) in the switch, won't that mean momentarily turning your PC back on after shutting it down when you're trying to shut off the PSU? And can't that, potentially, lead the BIOS to reset (if it thinks it's failing to POST repeatedly, or even failing memory training or other pre-POST steps) or similar issues?

 

[Thehack]

So, you're bringing the AT PSU spec back? "It is now safe to turn off your computer"

Not that it's a bad solution per se (this is after all intended for enthusiasts who know what they're getting) - quite the opposite, really - but I have to wonder how this will affect the PC. If the momentary switching action (ATX-ON) is before the latching one (MW PS-ON) in the switch, won't that mean momentarily turning your PC back on after shutting it down when you're trying to shut off the PSU? And can't that, potentially, lead the BIOS to reset (if it thinks it's failing to POST repeatedly, or even failing memory training or other pre-POST steps) or similar issues?

Yes that is a potential issue. The solution that aquelito linked is an between the manual solution and the fancy solution.

However generally most PCs register on the release of the switch.

The most correct answer to work this is for the plug in unit to use +5v for its microcontroller, or as a secondary function. The microcontroller will operate completely on +5VSB, and then provides separate PS_on functions to the AC-DC with configurable NC or NO modes.

However this requires a complete design of the plug in unit. All plug in units use the main voltage only by design.

 

[NateDawg72]

I'm loving this thread even though the recent discussion is a bit beyond me very interesting to read

Suppose the development/sale of new direct plug units wasn't blocked, is this something that could be resolved if a new PDCB was designed with meanwell ps_on/5VSB in mind? (ignoring the very small market for a PDCB like that, just hypothetically) Even if it added 2 more wires for that functionality it seems like that would be optimal for simplicity.

I'm looking forward to what aquelito can share with us

 

[Thehack]

I'm loving this thread even though the recent discussion is a bit beyond me very interesting to read

Suppose the development of new direct plug units wasn't blocked, is this something that could be resolved if a new PDCB was designed with meanwell ps_on/5VSB in mind? Even if it added 2 more wires for that functionality it seems like that would be optimal for simplicity.

I'm looking forward to what aquelito can share with us

I have recently found out that I can continue with the development of a Plugin unit.

I can certainly add request the +5VSB and PS_on for a meanwell, but that means it's a circuitry to support a single type of PSU. It presents the following issues:

1. Additional cost and complexity to the board.
2. Requires additional wiring. It is not just two more wires, it requires a different set of connectors as well.
3. It only really cover Meanwell PSU. Even though they are the most preferred 12V PSU, they are still very very niche.

So those issues will prevent me from wanting to invest in it. A new PDCB is not out of the question, something that is much more commonly available than G-Unique but it does present a few issues that I have not though of a solution for yet.

1. G-Unique uses soldered pads instead of connectors. This means all of his stuff is built to order. If I use connectors, it greatly increases the amount of board space for it. If I use solder pads, or soldered wiring, it means that all off the cables are permanent, a solution I'm not found of.

2. Support the PDCB with high power AC-DC units. I will not be modding any devices, but getting power to the PDCB is a bit of a challenge. So far I'm integrate Meanwell units, but it's hard to keep compatibility across the board. It likely means I'll have to stock Meanwell AC-DC units and provide them as a full kit.

At the end of the day, I'm just a hobbyist, not a full fledged business owner so my time is limited and I'm only scaling up what I can do.

 

[Soul_Est]

Thank you @Thehack for the wonderful guide and everyone who contributed to the discussion. It made for a worthwhile read.

I've been trying to figure out the power solution for my next build that runs under 400W. Run one EPP-400-12 through a PDCB or run one EPP-200-12 through a PDCB and one EPP-200-12 to the video card? Handling power to the power supplies won't be a problem from the AC side, it's more a space savings measure.