IBM 330 PSU proprietary switch and stand-by noise

[maferv]

Greetings friends,

I have this 6571 motherboard with its proprietary PSU (nice Rubycon caps), it has no power switch but a proprietary 3-pin connector that plugs into the MB.

This connector has 3 cables, colored black, white and red.

I measured it when the PSU is on, and red is 5v, black GND, but the white cable is not -5v. When the PSU is on, the white cable measures 0.33v, when it's off, it measures full 5v.

If you use a regular PSU the MB gets the right power but it refuses to even POST, no errors though, no beeps.

In the past when I was braver and more foolish, I remember wiring the header of this connector to a regular PSU based on the colors (red 5v, black GND and white -5v) and the MB worked with a regular PSU. But now that I have a little more experience I don't think it's safe to wire the white cable to -5v. I'm ready to do more measurements and tests if you can give me directions and find out what's best way to adapt this MB to use a regular PSU.

Next to this connector's header there are two undocumented jumpers that are open WP JP1 and JP2. Maybe changing them to close would allow me to use a regular PSU?

One more thing about this IBM PSU is that when it's on stand-by mode it makes an audible clicking noise which is quite unsettling since I don't see the fan running. What would that be and why is this PSU so strange?

Merry Christmas and happy new year!

 

[jackb77]

I have two 330 486 dx2 without PSU, so I'm interested if can use a standard AT PSU, any news?

 

[GiGaBiTe]

The white wire is PS_ON, it is NOT -5v. The red wire is +5vsb (NOT +5v) and black is ground. If you short the white wire to ground, the PSU should turn on.

That 3 pin connector is IBM's attempt to make a soft start AT motherboard. A few other manufacturers did something similar.

IBM pc 330 450dx2 PSU questions \ VOGONS

www.vogons.org

 

[maferv]

Well, thank you for that GiGaBiTe. That's helpful!
I indeed was able to use a regular AT PSU.

Now I'm thinking how could I use this IBM PSU with a regular motherboard? Shorting the pins would make it work but how do you turn it off? I know, you can remove that short but as I mentioned before, the PSU emits a clicking sound, a quite unsettling clicking sound when it's off but plugged in. It's not loud, but it's there. So far nobody has commented on the source and nature of this sound. Maybe my unit is defective? I doubt it.

My idea is to cut the blue and brown cables right in the middle and wire them to use a regular AT switch. I'll take the switch from another PSU.

Here you can see its entrails and specs. It has quality components.

 

[GiGaBiTe]

Clicking, ticking and whining is usually an inductor or transformer resonating at a frequency that can be heard.

Most of the time, it's harmless. But it can be an indicator of failing components. I had a Dell PSU years ago that would click and tick, it would fall flat on its face if loaded much above the internal cooling fan. Turned out to be one of the main line caps was shorted.

I would recommend pulling and testing ALL of the capacitors, especially the tiny ones buried deep in between hot components like the mosfets. Just because a capacitor looks fine visually, doesn't mean it is fine. You'll need an ESR meter and a capacitance tester. Those cheap Chinese component testers work well enough if you don't want to spend a bunch on a pro grade meter.

It also looks like you have the bukakke of death. I'd recommend chipping that brown glue off wherever you can, it goes acidic over time and eats into components. It also becomes conductive and can short various parts of the PSU out and cause strange problems.

 

[maferv]

I would recommend pulling and testing ALL of the capacitors, especially the tiny ones buried deep in between hot components like the mosfets. Just because a capacitor looks fine visually, doesn't mean it is fine. You'll need an ESR meter and a capacitance tester. Those cheap Chinese component testers work well enough if you don't want to spend a bunch on a pro grade meter.

I have a device to measure ESR and capacitance. It's a time consuming task to test all capacitors, including the smaller ones, but I'll have to do it to confirm everything is fine. Should I also measure the ceramic and film capacitors?

Clicking, ticking and whining is usually an inductor or transformer resonating at a frequency that can be heard.

Most of the time, it's harmless.

I'm under the impression that the noise is part of the way this particular PSU is designed to work. Maybe someone else with a similar proprietary IBM PSU can can share his/her experience.

It also looks like you have the bukakke of death. I'd recommend chipping that brown glue off wherever you can, it goes acidic over time and eats into components. It also becomes conductive and can short various parts of the PSU out and cause strange problems.

You mean the all the hot glue? It looks grayish to me.

Here's another picture

 

[GiGaBiTe]

I have a device to measure ESR and capacitance. It's a time consuming task to test all capacitors, including the smaller ones, but I'll have to do it to confirm everything is fine. Should I also measure the ceramic and film capacitors?

Ceramic disk capacitors generally don't go bad unless they're physically damaged, in which case, they usually short. Metal film capacitors can fail, but it's also not very common. I would stick with the electrolytics for now.

I'm under the impression that the noise is part of the way this particular PSU is designed to work. Maybe someone else with a similar proprietary IBM PSU can can share his/her experience.

SMPS supplies are not supposed to make audible noise. It's either because of a design problem, or something isn't working properly.

You mean the all the hot glue? It looks grayish to me.

It might be fine if it's just hot glue. Many PSU manufacturers used a hard yellow epoxy to hold down components, and sometimes they just splooged a whole ton of it all over the place. That epoxy degrades with time and heat, and can eventually turn brown or black if it gets cooked enough. It'll become conductive and corrosive and cause problems.

You can try poking test leads into some of the glue/epoxy with a meter set to measure resistance. If you get any reading at all, the stuff needs to be removed.