ATX to Tanzania / 4400/7220 PSU adapter

[Franklinstein]

LPX is apparently the standard from which ATX was developed, at least in terms of standardized power supply dimensions and most voltages/signals. So the LPX and ATX soft power signals are the same: there's a 5V standby, a PWR ON pin, and a PWR GOOD pin. The 4400 has 5V SB and PWR ON (and a GND) in that little 3-pin connector while PWR GOOD is on one of the main power plugs as you've noted previously. 

On that -5V issue, it's possible the 4400 doesn't use it, as mentioned above. Try omitting it to see if the machine boots and runs stably; if it doesn't then obviously it's required and you'll need to provide it somehow, but temporarily not having it isn't going to blow anything up so there's no harm in trying.

If there are other wires that aren't used on the new power supply, such as the extra 4 pins on the main connector or the 12V CPU aux connector, you can safely ignore them; just because they're provided doesn't mean they need to be utilized.

Try opening both power supplies. Most of the good ones will have the voltages/functions listed on the component side near where the wires go into the board. I haven't done it yet but I was planning to completely desolder the harness from a modern 20+4-pin ATX power supply and install a harness from my Umax S900's defunct power supply in its place. The Umax's PS is a weird non-standard thing but it's electrically similar to ATX so there are corresponding points in the replacement ATX unit for all of the required wires. By doing this there are no splices or vestigial connectors hanging around (such as the CPU 12v connector, SATA connectors, etc) and everything is still the correct length.

 

[jessenator]

Also, I won't need to scav extra +3.3V wires as I can just remove them from the SATA power connections I'm not going to use. That will make an even 6. The grounds I can take care of as well.

planning to completely desolder the harness from a modern 20+4-pin ATX power supply

I was just thinking about if that was viable. That might be the best way to go. The AXT output board is clearly marked on the Rosewill, from what I can see (I can add as many of those 3.3V and grounds as I need.

The only mild concern I have is a disaprity between the total max watts on the +5V: it's 20W on the original PSU and only 14W on the Rosewill. I'm thinking since the draw from that bitty CF-IDE adapter is so small it won't matter. The CD-ROM shouldn't take over 10W I'm thinking.

edit: uegh, I don't think the Rosewill works. 

Also it fits like a glove, save the pointless metal pieces that slot in on the original (I'm going to secure it with screws anyway)

 

[jessenator]

The Dell PSU's standby light on the back is on, I get the 5V from the VSB pin. I wonder if there's a way to trick the PSU into thinking the computer is "on"

I mean, it was fully operational before I pulled it to sell the motherboard to a friend across the country.

 

[NJRoadfan]

The reason for the lower rating on the +5v rail is that modern machines use +12v to power the VRMs on the board for the CPU. Older machines directly fed everything with +5v. Honestly all that machine needs is a cheap ATX to AT adapter and some wiring of the 3-pin power switch harness. The only "work" would be wiring the +3.3V harness.

 

[IlikeTech]

I wonder if there's a way to trick the PSU into thinking the computer is "on"

To test an ATX PSU without a computer, jumper the Green to any Black wire

 

[jessenator]

So last Saturday I got cold feet after attempting to desolder output connections from one of the dell PSUs and the rosewill and failing— tl;dr my iron was going haywire, which may be a tip or wand issue. It's an inherited Weller WES51. I think the wand has a small break in the wiring somewhere as the heat indicator LED intermittently shuts off. I'm also having a devil of a time just keeping the screwdriver tip (weller brand) from oxidizing and therefore not heating to temperature at the very tip. Do manufacturers coat their boards with an inhibitor or something to prevent desoldering? something that would foul up a soldering tip? The actual ruining part came in the shape of damaging a diode on the Dell, and some other component on the Rosewill (after reading though, the particular model I have is very much loathed by the PC community, and has very poor voltage regulation, so good riddance…)

Anyway, I had one remaining, working PSU which is all prepared to just splice, which I can do all day, everyday with nary an issue. (I swear I'm not a knucklehead), And then I thought, well why the heck am I getting worked up over a $5 purchase that might fix my PSU … so I bought some fuses which arrived today! If I can conclusively prove the original PSU is dead, I'll splice away.

In the event I will just splice wires with the AT-connectors, which wire goes to the board's "Power_good (+5V)" pin? The wire on the OE PSU is orange, which doesn't correspond to anything on the ATX supply (it's not +3.3V according to the pinout). Is that the "PWR_OK" gray wire, or the "PS_ON#" green wire? 
View attachment 26752

Every other vacancy I've filled with +3.3V wires and ground leads from the respective, unused SATA wiring. Soft power isn't a priority at this point for me.

Cheers

Edit, the plus side of the Rosewill PSU going is now I have a replacement fan for the OE PSU in the event that it works. The original was seized up, and probably louder when it was new than this one.

 

[jessenator]

Plugged it in, (and the battery), got a click, tried the power button again and got another click then nothing.  fuse still has continuity ¯\_(ツ)_/¯ something else in there must be the culprit, and I don't have the acumen to tackle it.

Like a dingus I didn't just search the p/n on the PSU itself.... it's still available refurbished, for $$ https://www.ebay.com/itm/283428794270 but that's to be expected for a niche item.

And a someone in Japan was doing something along these lines as well: https://translate.google.com/translate?sl=auto&tl=en&u=http%3A%2F%2Fwww.wind.sannet.ne.jp%2Fpenpen%2Fpm4400%2Fdengen%2Findex.html

 

[jessenator]

Also, on my Dell PSU there's a medium-Brown wire that's conjoined to pin-13, which from a cursory search is a voltage sense wire. Is it necessary to wire it in on a +3.3V line somewhere?

 

[jessenator]

Alright, So as mentioned in my restoration thread, I got the PSU spliced to the original cabling and to "work"… the machine powers on, but not much else. No booting, etc.

The next thing I need to do is handle the wiring with the "PS_ON#" to keep the PSU on. Funny story though, I though everything was completely dead on the board because absolutely nothing happened (even though the standby light was on in back of the Dell PSU. I wanted to test it, so I looked up that one only need short the PS_ON# wire to ground and the PSU should act like the power is on. I imagine though that there is a more elegant solution, but …I don't know what it is.

Is there a way I could use the front panel button to operate like that? or is it a momentary switch?

Also if anyone who's replied/followed this thread help me out on my restore thread, I'd appreciate any insights you have on what to test/further actions to take on the 4400 itself.

Cheers.

 

[jessenator]

Ok, I think I have it...

the soft power is totallyt doable, and I just wasn't seeing it! According to that link from the Japanese site, I just need the +5V VSB, PS_ON#, and a ground. Time to splice!

 

[jessenator]

View attachment 27096

 

[jessenator]

Cross post=====

Started re-crimping an ATX* + aux PSU for use in one of my 4400s. Everything was going peachy right up until I got to the soft power connector. I tried to extricate the crimped wires from the housing (to reuse it as I had done with the board power housings) and ended up cutting the connector on the sharp metal of the crimped portion. Old, fragile and might've broken at some point anyway. Good news is that Molex still makes that part so I was able to get a replacement from Mouser.

Tricky part was with the P1 connector. There's the +5V "Power Good" (PWR_OK lead on an ATX PSU), connected with a 10nF capacitor to a common ground. U was able to salvage the insulating tubing, which wasn't heat shrunk I guess.

But, since the PSU won't really do anything without the soft power connection, I can't really test to see if it all works. But unless the PSU is just bad, it should be fine.

*I managed to get a decently-priced, relatively unused ATX supply which had both a decent amp rating on the 3.3V rail as well as one that had an older "aux power" connector in the ancient AT/LPX/NLX Molex connector style. The keying wasn't useful for the Tanzania power board connector, but at least I didn't have to re-crimp all of those wires. Just stole another 3.3V connector from the +4 side of the ATX power. I've got everything I needed from the 20+4 connector, including the unnecessary(?) -5V, and the aux power, so I didn't have to go aping off or destroying any of the other connectors. Also there was floppy power (for my CF-IDE adapter) right off the leads, so no adapter needed there.

In case anyone is interested and doesn't want to hunt for the crimp connectors, here they are. Sorry for the mixed suppliers... some have one part but not the other, or make you buy bulk.
wire connectors

P1 P2 P10 P11 wire housings