AT/ATX Power Supply?

[Waltah]

In the breakup of stepdaughter's marriage (don't ask) we acquired an Acer P-II 'parts machine'. Turns out what's wrong is a dead power supply. This supply looks at first like a standard AT supply with +/- 5VDC and +/- 12VDC at the usual sorts of currents and the usual 2-6 pin motherboard connectors. However, instead of connections to a switch it has a another three pin connector for the MB which provides +5VDC standby voltage and it is switched by the motherboard in ATX fashion.

The problem seems to be the +5VDC sb voltage (at least, there is no voltage there when power is connected); in enough hours I can probably figure out and fix the problem. However, I'd rather replace the supply. The question is -- with what? This thing seems to (inexperienced) me to be neither fish nor fowl. Is my best bet to get an ATX supply and change connectors as needed? Or?

The machine came right up when I connected a standard AT supply so we know the problem's just in the supply.

Help!

Walt Hutchens

------------------
Walt

[ranchdog]

Welcome Waltah....

Please check on the Mboard for a J1 that designates
AT/ATX power.

Person may need to use the BIOS or Mboard model # and download
a Mboard manual for some additional help.


Luck.

------------------
....How long is a minute... depends on which side of the Bathroom door you're on. ......
..........

[jahn]

Just out of curiousity when you connected a standard AT supply what did you connect to the 3 pin connector? as an At only has the 2 6-pin connections.

Originally posted by Waltah:
In the breakup of stepdaughter's marriage (don't ask) we acquired an Acer P-II 'parts machine'. Turns out what's wrong is a dead power supply. This supply looks at first like a standard AT supply with +/- 5VDC and +/- 12VDC at the usual sorts of currents and the usual 2-6 pin motherboard connectors. However, instead of connections to a switch it has a another three pin connector for the MB which provides +5VDC standby voltage and it is switched by the motherboard in ATX fashion.

The problem seems to be the +5VDC sb voltage (at least, there is no voltage there when power is connected); in enough hours I can probably figure out and fix the problem. However, I'd rather replace the supply. The question is -- with what? This thing seems to (inexperienced) me to be neither fish nor fowl. Is my best bet to get an ATX supply and change connectors as needed? Or?

The machine came right up when I connected a standard AT supply so we know the problem's just in the supply.

Help!

Walt Hutchens

[Waltah]

John, I didn't connect anything to that 3pin connector. It seems only to control the power supply and since I was doing that with Radio Shack cliplead I just left it unconnected.

Ranchdog, the board is a V65LA, thanks for cluing me on how to research this. A web search turned up plenty of top level motherboard info -- jumper settings and all that but nothing more detailed. The ACER on line info hints that the power supply is considered ATX but the connectors are wrong. The label on the supply says it's a 'Tiger model AI-2200 Rev B' and gives the voltages -- +/-5, +/-12 (usual current ratings) and +5Vsb @1A.

I guess what I hope for is either a more specific nomenclature that will let me search eBay and such places for a supply that will directly plug in, or a little more detail about power control on both the V65LA and ATX supplies so I can be sure of safely mapping a 'real' ATX supply to the connectors used on the ACER board. In particular, the 3-pin connector has red, black, and brown wires. According to the board in the existing (dead) PS, the red is 5VSB, black is GND, and Brown is labled (I think) 'P/O'. 'Power On' perhaps? Does that mean that applying +5 to the brown wire brings up the rest of the supply? Is that how a real ATX supply does it? Could I fry both a new power supply and the MB finding out?

I'm only sure about the answer to one of those questions ... although I may try hooking an external +5 to the brown wire to see if it will bring up the existing supply.

Can someone either point me to a circuit diagram for a typical ATX supply or provide a three sentence tutorial on the design of the low current sections of those supplies? I suppose the safest and cheapest way to go is to fix the supply; I just hate to have to start by tracing the circuit to figure out how the darn thing is supposed to work.

Fuse is okay, +340 volts is present on the input filter caps so it's not totally fried.

------------------
Walt

[bassman]

Hello Waltah,
Sounds to me like you have one of those machines that got caught in the change-over from one technology to another. I have a handful (if thats possible) of machines around here that fit that catagory.

A Compaq 486 that, from the back, you would swear was an ATX machine with onboard vid, parallel, serial, and sound, but is in fact a 486 with the mobo mounted backwords in the case. ODDBALL!?

With an ATX power supply, you will also have 3V. leads, so I would recommend staying with an AT and modifying the case for a typical AT power switch. You may have a heck of a time finding the exact right PSU for this machine.

Hang in here for a bit as someone here may have experience with this setup and be able to help you better.

Good luck http://www.PCGuide.com/ubb/cool.gif

------------------
Waiting patiently for the future to arrive Frank's Place (http://www.members.aol.com/frankscompsrvice)

[jahn]

Why not just use the AT supply then?

I'm not sure but the signals you describe seem like an LPX supply not and ATX. An ATX has 2 rows of pins with some 3.3 volt lines.

I have an LPX mainboard and I used a standard at supply to power it and it works fine. It has the 2 standard 6 pin connectors and the same signals. P8 and P9 traditionally known as.

Just wondered why you couldn't just use the AT supply.

[jahn]

I have posted 2 replies that didn't get through so here is try #3.

Why don't you just use the AT supply?

It sounds like the signals you describe are actuall from an LPX and not ATX. I think ATX has 2 rows of pins and about 20 signals from the PS. Including 3.3 volts.

I have an LPX mainboard that has the standard AT connectors. P8 and P9 as traditionally known and it works fine on this board.

Since you got it to come up why not use the AT supply?

[Waltah]

Jahn said:
> Since you got it to come up why not use the AT supply?

No technical reason, and that's probably the sensible answer. Basically I'm just too obsessive to do it that way right now!

I like being able to walk away after telling the OS to shut down. There could be some minor mess with Win 95 thinking that the machine can be shut down under software control but probably that could be solved. Maybe you can turn that off in the CMOS? The real for true reason is psychological.

Of course if I fry something I'm gonna wish I did it the sensible way. I am going to see about putting an external +5 on the sb line, so stay tuned ...




------------------
Walt

[Paul Komski]

I bought a pc with very similar "hybrid" wiring on which the PSU blew. My main problem was that the PSU was only 3"x3" and I couldn't get any replacement to fit inside the case. My "fix" was to use an ATX PSU and cut the wires near the MB connector. I kept the PSU outside the case (but earthed them together) and ran the wires into the pc. I then cut the wires off the old PSU and connected the two sets of wires together; colour for colour etc, etc. I just blanked of the 3.5v wires and made sure that the other 3-wire (standby component) had its three wires connected to the correct wires coming from the new PSU (I did a lot of checking and rechecking here since the wiring was of different colours). There was one additional wire that went into Pin-1 I think and ran back into the ATX that didn't exist on the AT. This concerned me; but I learned that this is just some sort of feedback/safety device and doesn't affect functioning.

This "temporary" fix is still working fine 6months on. But I now know that you can simply buy an ATX/AT conversion cable that has two 6-wire connectors for the M/B plus the 3-wire connector for the Standby; which I would suggest is the safest and neatest way to go. If I can find the link I will post it here - but it was 6months ago. http://www.PCGuide.com/ubb/wink.gif

------------------
Take nice care of yourselves - Paul

Here it is http://www.pcpowercooling.com/products/alarmandaccesories/accessories/popups/indexatxtoatcable.htm

[This message has been edited by Paul Komski (edited 01-30-2002).]

[Paul Komski]

That pc was a 2nd-hand slimline made by Tatung; if that's of any interest.
I thought I'd put this in instead of a stupid double-post that I made. http://www.PCGuide.com/ubb/mad.gif

[This message has been edited by Paul Komski (edited 01-30-2002).]

[Waltah]

Jahn, I looked up the LPX form factor; that supply is pretty much AT in that it has a hard wired 120 volt switch controlling the power supply. My supply has internal power switching controlled by a signal from the MB; the front panel switch is just a momentary that toggles circuitry in the MB.

Paul -- that adapter sounds like it would solve the problem. I will start digging around and see if I can find it, but if you get there first please let me know! The 3.3 volts doesn't seem like a problem; I just won't connect it. Those supplies all seem to have a husky +5 volts as well. And the 3.3 and 5 volts seem to be just separate windings on the same transformer since the power rating always seems to be shared. So no worry about required loads.

I did try an external +5 on the 5sb line; no effect, regardless of how I hooked up the brown wire. I didn't have that connector hooked to the MB so the meaning is that the problem in the supply is not as simple as a lack of voltage there -- could be a diode in series and bias for other supplies is taken from that voltage, or something like that. I am going to look at that supply a little more closely; does anyone know a source for any ATX supply circuit diagram?




------------------
Walt

[Waltah]

One more thing: I found out that a real ATX supply turns on by dropping the control line to 0 volts, off by raising it to +5. (There are details, like tolerances and current to be sources/sunk.) Sort of the reverse of what you expect ...

------------------
Walt

[Paul Komski]

He He You must have missed the link ! http://www.PCGuide.com/ubb/biggrin.gif

Here it is again: http://www.pcpowercooling.com/products/alarmandaccesories/accessories/popups/indexatxtoatcable.htm


[This message has been edited by Paul Komski (edited 01-30-2002).]

[YODA74]

http://pavouk.comp.cz/hw/en_atxps.html
http://www.aapspower.com/stan145sf.html
http://www.geocities.com/nzjrs/frames/hardware/powersupply.html

should beable to get an idea from all these pics & schematic

------------------
PETE'S Computer Information Links (http://www.dreamwater.com/paleopete/computer.htm)

[jahn]

Paul do you know the model number of the tatung board. Or at least can you describe it slightly?

QUOTE]Originally posted by Paul Komski:
That pc was a 2nd-hand slimline made by Tatung; if that's of any interest.
I thought I'd put this in instead of a stupid double-post that I made. http://www.PCGuide.com/ubb/mad.gif

[This message has been edited by Paul Komski (edited 01-30-2002).][/QUOTE]

[jahn]

Paul do you know the model number of the tatung board. Or at least can you describe it slightly?

QUOTE]Originally posted by Paul Komski:
That pc was a 2nd-hand slimline made by Tatung; if that's of any interest.
I thought I'd put this in instead of a stupid double-post that I made. http://www.PCGuide.com/ubb/mad.gif

[This message has been edited by Paul Komski (edited 01-30-2002).][/QUOTE]

[Paul Komski]

jahn I don't have the pc here at the moment - lent to someone who is stuck for one. Tatung are a Taiwanese Electical Company that got into and out of computers - but I did get a manual (1.1MB pdf file) from them before they stopped their support and I have taken the the liberty of eMailing it to you. It has been scanned with both NAV and AVG - but take any precautions you feel necessary.

Let me know if you want any other details because I will be over at my friend's house in the next day or so. Good Luck.

------------------
Take nice care of yourselves - Paul
"I can resist everthing except temptation" (Oscar Wilde, I think; and me too!!)

[Waltah]

Paul, I did indeed miss the link. Blind? Worse? Anyway, thanks for reposting.

The links to the diagram of an ATX and related info are very helpful and I'm on the trail, I think. The diagram of this unit isn't identical of course but it's close enough that the data helps a great deal. Sure enough the (separate) oscillator for the +5Vsb supply isn't running. It is getting voltage, the transistor seems okay on a low voltage test, but there is no forward bias when I apply power. The circuit is pretty simple -- seems like at least a couple more hours of work ought to be worth it.

------------------
Walt

[Waltah]

And the answer is ... a $0.02 470k 1/2w resistor which should have supplied bias to the oscillator for the second supply -- the one that furnishes the +5Vsb. It was open, hence no forward bias, so no oscillation. Not fried -- it looks perfectly normal, just no continuity.

I'd have found it a bunch sooner but I was looking for more common problems like a shorted transistor. With a new resistor the supply works normally, the computer came up and works fine. Now all I have to do is figure out how to get an OS on it ... another question for another topic.

Thanks a bunch, guys!



------------------
Walt