Investigating a full linear PSU for the XXHE PC and NOS1

[PeterSt]

So ...

Anthony, since you are as far as you are with all the elementary requirements, why wouldn't this one be made for the job ?

http://www.velleman.eu/downloads/2/ps230210gbnlfresd.pdf

500-600 euros or so.
Does all we need and more. And major advantage : should be safe.
Will boot as well. Haha.

Is this now a bad idea for us hobbyists ?


PS: Side note : I can get these for a lower price. Not sure how that works out net with shipping and all.

[PeterSt]

And to keep things going somewhat ...
Pedal, we have the parts for your PC so we could just as well try to "implement" this one with the great help of Anthony's schematics. Unless someone sees something wrong with it.

So I think I am serious on this Lab PSU and now I need the help of you out there seeing something wrong with it.

Peter

[acg]

The main issues that I have with a lab supply like that for a final solution Peter are noise (the bigger output lab linears are all fan cooled as far as I know) and knobs (and how kids love to tweak them).  Plus, I think their implementation will not be as good as something we build especially for it.

I could not read the manual for that supply Peter...wrong language for me.

[PeterSt]

Hmm ... First part of that manual seems English to me ?

But there's another downside so far : No 115V as it seems.

About the fan(s), yes, you could be right.

[acg]

Based on Peters forward thinking regarding the capacitor bank and the ATX psu to start up and shut down the pc, following are my ideas on a clean implementation.

As mentioned earlier, I am sure that I can use a $2 chip to make a small circuit board to control the linear power rails and motherboard to full ATX power specification.  Peter has said that he has some decent linear regulators used in the NOS1 that can go to 3A output (is this maximum output Peter, or continuous output?).  There are two rails that we will have an issue with needing more than 3A of current:  the 12v P8 socket which will need at least 5A, preferably more during startup and shutdown; and the 5V rail into the P24 socket which will need about 5A at shutdown.  The remainder of the rails can be serviced by a 3A or smaller supply.

So, the only time that we should need more current is during startup and shutdown...which I should be able to identify by the voltage on the 'PWR_ON' leg of the aforementioned $2 chip (each state has a seperate voltage range).  So I propose one power switch and one toggle...the toggle to tell us that we want the linear ATX PSU to go into "Extreme Low Power Mode" because the XXHE PC is optimised and ready to roll in low voltage mode...the power switch to simply turn on and off the pc.

When the toggle is NOT set for "Extreme Low Power Mode" my little circuit board will boot on the pc using the normal voltage rails and a special low spec, low cost 12V 5A+ and 5V 5A+ as well as the normal high performance rails for everything else.  I can use cheap monolithic 7.5A or 10A LT regulators and use a dual wound transformer for thes 'high power' rails.

When the computer restarts with the toggle in "Extreme Low Power Mode" the circuit board will still start the XXHE PC as mentioned in the previous paragraph, but once it is booted it will use the capacitor bank and turn off the high amp/low spec rails and turn on the low amp/high performance 5V and 12V rails.  Likewise, when a shutdown is triggered, the mobo will let the $2 chip know what is going on and it will use the capacitor bank to turn off the two low amperage rails and turn on the high amperage rails for shutdown.

I think I can implement this.

Anthony

[acg]

The lt series is only good for a few amperes of regulated current. That will do a HDD or a fan best. There are other more capable monolithic solutions. The datasheets also tell you how to boost the current with extra power BJTs. They are very low drop and available in the needed voltages.

Thanks Coen.  The LT1083 will do 7.5A with 5V and 12V.  They cost about AU$20 each over here.  That may be a decent solution for the two high amp/low spec. rails I mentioned in my previous post.  I would love to hear your other suggestions though...I only had a look this afternoon and these were the best I could find.

I feel heat is still an issue. I know that you can get away with other constraints ie a lower ambient temperature, but as Peter said, component life expectancy progressively reduces with increasing remperature.
You also have to realise that most regulating solutions do NOT regulate transients or hf components. Actually the regulator may have an inductive output feeding oscillations.

The shuntreg is a worst case class A circuit, that is that is has to provide the peak current all the time, so if you need say at peak 2.5 Amps, the currentsource and sink have to both dissipate at least this 2.5 Amps. For a 12v currents sink alone this is 30Watts. Adding the source, you easily end up with a 40-50 Watt dissipation. You should save those for small signal audiocircuitry or ultra energy efficient PCs.

As they say on the silentpcreview site: a little air goes a long way. This is very true for the heatsinking of my cpu. When called for duty very little airflow can keep my megahalems cool compared to a hot touch  when absent. If you decide to go the powerdissipation route, you might consider previous generation cpu coolers as heatsink (to save some money). The are designed to get 100+ watts out of the CPU and keep it cool with a little air. As can be seen on the heatsink datasheets forced air can get small heatsinks to archieve 0.8deg/watt or lower.

To be honest I have not looked into heat too much at all.  An aluminium case with plenty of mass and decent heat-sinking can be VERY expensive to get made (unless there are suitable mass-produced amp cases on the market - now that's an idea) but I would love to make the entire XXHE PC passively cooled, but that may not be feasible if we want the components to last.

Thank again Coen.

Anthony

[PeterSt]

Peter has said that he has some decent linear regulators used in the NOS1 that can go to 3A output

Small correction : Peter said that another option was to have separate LPSU's for each rail and that one of those is in the NOS1's left leg. So this is no regulator but a whole PSU.
(and yes, that 3A should be for continuous use, though I never tried that)

Peter

[acg]

Peter has said that he has some decent linear regulators used in the NOS1 that can go to 3A output

Small correction : Peter said that another option was to have separate LPSU's for each rail and that one of those is in the NOS1's left leg. So this is no regulator but a whole PSU.
(and yes, that 3A should be for continuous use, though I never tried that)

Peter

Even better Peter.  I was of the opinion that best practice would probably be a separate psu for each rail.  Sounds like it may save some development time also.

Anthony

[CoenP]

I found the nice LDO regulators in my archive. The typenumber(s) is:

Micrel MC29750 to 29753.

 "Only" 7.5 amps but very nice specs. A few of these should do a music PC. Only0.5V dropout so you can keep input voltage close to the desired voltage (under minimum volt line condition).
There is also a nice LDO design section in the datasheets that covers most implementation aspects of safe operation.

I once thought of these to reduce the ripple of a normal switching supply. That is: readjust the ATX voltages to a minimum 0.6v overvoltage (under load) and kill the ripple with these LDO babies.

Intriguingly the Velleman labsupply has an "operation" of 8 hours....
Undoubtably this is about the heat building up.

Regards, coen

[PeterSt]

Intriguingly the Velleman labsupply has an "operation" of 8 hours....
Undoubtably this is about the heat building up.

Haha, I saw that too. But no, that would be non-sense. I merely read it in the realm of "with this you can finish your working day" (no breaks ).

Peter


PS: But of course no heat builds up "more" after the 8th hour of operation.

[CoenP]

Intriguingly the Velleman labsupply has an "operation" of 8 hours....
Undoubtably this is about the heat building up.

Haha, I saw that too. But no, that would be non-sense. I merely read it in the realm of "with this you can finish your working day" (no breaks ).

Peter


PS: But of course no heat builds up "more" after the 8th hour of operation.

Ok, you're right  no heat capaticance that keeps charging for 8 hours in such a small casing. But is remains an odd statement to me and will probably have to do with the lifespan/guarantee at serious lab loading (hence heat).

Regards,  Coen

[acg]

I found the nice LDO regulators in my archive. The typenumber(s) is:

Micrel MC29750 to 29753.

 "Only" 7.5 amps but very nice specs. A few of these should do a music PC. Only0.5V dropout so you can keep input voltage close to the desired voltage (under minimum volt line condition).
There is also a nice LDO design section in the datasheets that covers most implementation aspects of safe operation.

I once thought of these to reduce the ripple of a normal switching supply. That is: readjust the ATX voltages to a minimum 0.6v overvoltage (under load) and kill the ripple with these LDO babies.

Thanks Coen.  I looked these up but their noise was rather high at about 260uV and no PSRR figures.  I am reading the specs differently to you...but that 25 page spec. sheet very helpful.

Anthony

[Arjan]

Hi,
I am reading this interesting discussion. I am not an expert!! but here are my ideas:

I will use 2 simple adjustable supplies (PEAKTECH-6080) for the audio grade usb3 card 5V and HDD (via esata) 5V and 12V. I already ordered this.

Will try later a picoPSU 160 XT with following fixed output supply, EA Elektro-Automatik EA-PS 2012-15 (12V 15A).

Question will this picoPSU work on the XXHighend PC? If not, possible solution: Leave the standard ATX supply on the 24pins and use the picoPSU to feed only the CPU and fans?

regards, Arjan

[acg]

Hi,
I am reading this interesting discussion. I am not an expert!! but here are my ideas:

I will use 2 simple adjustable supplies (PEAKTECH-6080) for the audio grade usb3 card 5V and HDD (via esata) 5V and 12V. I already ordered this.

Will try later a picoPSU 160 XT with following fixed output supply, EA Elektro-Automatik EA-PS 2012-15 (12V 15A).

Question will this picoPSU work on the XXHighend PC? If not, possible solution: Leave the standard ATX supply on the 24pins and use the picoPSU to feed only the CPU and fans?

regards, Arjan

Hi Arjan,

One of the things that has not yet been explicitly stated in this thread, but I suppose I am doing it now, is that one of the design goals is to have the computer managed by a single well designed power source.  As things are looking now this power source will actually be a number of smaller linear power supplies within the one 'box' that can be managed as a single unit.  This may improve grounding issues and it will certainly improve usability and computer life.  With a 'one box' solution we don't have multiple power supplies of possibly dubious construction feeding hash back into the mains (just one...haha) and the computer can be managed as it is used to being managed and hence its lifespan will be extended. 

I would not be surprised if much of the gains made by this psu were because of the reduction of hash feeding back into the mains to impact other components.  I am sure there will be gains by improving the USB card performance and we will have to wait and see on the other rails.

I don't know if a 160w pico will start up the standard XXHE PC that Peter makes, but there is no need for the picoPSU now because we can build our own version of it without the noisy switching supplies onboard, which should give a better result. 

Cheers,

Anthony

[CoenP]

I found the nice LDO regulators in my archive. The typenumber(s) is:

Micrel MC29750 to 29753.

 "Only" 7.5 amps but very nice specs. A few of these should do a music PC. Only0.5V dropout so you can keep input voltage close to the desired voltage (under minimum volt line condition).
There is also a nice LDO design section in the datasheets that covers most implementation aspects of safe operation.

I once thought of these to reduce the ripple of a normal switching supply. That is: readjust the ATX voltages to a minimum 0.6v overvoltage (under load) and kill the ripple with these LDO babies.

Thanks Coen.  I looked these up but their noise was rather high at about 260uV and no PSRR figures.  I am reading the specs differently to you...but that 25 page spec. sheet very helpful.

Anthony

Arjan,

Which picopsu do you intend to use? I wonder if any of them is fit for being a xxhe supply.

As Anthony posted before a pc supply has to conform to the ATX standard for a board to start and function properly. This is about having voltages within the specs before supplying them to the mobo. You get yourself into trouble if you use two (or more) supplies. Who is doing the powermanagement?

So what may work is one supply for the mobo 24 and 4/6 pin connectors and one more for the periferals (fans hdds etc.). You have to orchestrate the power up and make shure the sata hdds etc have power to be recognised by the BIOS.

Regards, coen