Wolf in sheep's clothing - Enermax in a Deer casing

[Dan81]

As the title says, I got a "AO" (no, not AOpen , their units are FSP) PSU that sits doing nothing and I have a complete PCB from a Enermax EG-365P/VE sitting here. It's a "P4 ready" variant according to the old casing. (which I ditched because of way too much rust)

Here are my goals with them:

- remove whatever PCB the Deer has (and probably reuse it when I need a solid 200/250 watt unit for say a Pentium MMX - Pentium 3 range machine)
- remove the auxiliary AT- style plug from the Enermax and the -5v rail as well since they have no use anymore
- rig up an extended 4 pin adapter for the plug using the old plug from the Deer as well as reuse the SATA plugs as well
- recap the Enermax
- replace the fan with either an clean Xinruilian LP/EP or a newer Globe Fan (which ironically is made by Jamicon)
- upgrade the 12v rectifier if necessary

Now several questions:

- When powered up, the Enermax does an insanely high pitched sound that keeps on going higher. What could it be? There's very little glue left on the board.

-Given I do the upgrades, what can I power with it? I could identify the following rectifiers - 2 of them that might be for 3.3 and 5v are S30D40C but the one in the middle which should be for 12v (I guess) is a S40SC4S, which I couldn't find a lot of info about.

[Dan81]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

Okay, so far I replaced the K3067 5vsb transistor with a C3866 from a gutted Hipro of the same vintage and the K2607 switching transistor with a K2654 (which is just one amp less, so not that much of a change) and now I don't get the high pitch anymore, but neither do I get 5vsb or anything, and the parts were good (the Hipro wasn't even used that much, and it was some kind of SFX PSU by the way, about 150W or so)

Any ideas what to do next? I removed most of the glue and replaced disc capacitors (because somebody at Enermax decided it was a GREAT idea to use fake Y caps. Shame on you Enermax.) with the ones from the Hipro, but no change. Should I check the little JPCE-TUR buggers with small values?

[kaboom]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

Okay, so far I replaced the K3067(MOSFET) 5vsb transistor with a C3866 (BJT)

You replaced a MOSFET with a BJT in that standby section. Put the MOSFET back; what will properly bias one will not work for the other.

Replace the 2SC/KSC945 across the source shunt resistor- don't bother testing. Any shorted zeners, or any that look to have gotten hot?

[Dan81]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

You replaced a MOSFET with a BJT in that standby section. Put the MOSFET back; what will properly bias one will not work for the other.

Replace the 2SC/KSC945 across the source shunt resistor- don't bother testing. Any shorted zeners, or any that look to have gotten hot?

None.

Noted, will put back the K3067. Is it located near any of the switching transistors?

[kaboom]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

There should be a resistor from the source to B-/common, something from half an ohm to perhaps 4.7 ohms, depending on how hard they drive the transformer. The high side of this resistor drives the base (maybe thru a small resistor) of this shunt transistor.

They're TO-92 parts, and not always C945s. Can also be C1845 or C1815. Not all of these supplies have that shunt transistor. But if present, it's a TO-92 in the vicinity of the aux/5VSB switcher. Not near the main switcher(s).

Check/replace the UF4007 (or equiv) from the top of the line-side aux winding on the standby transformer to the aux-supply cap.

Did you have anything on 5VSB when it was squealing?
If you had nothing, change that 5VSB rectifier- squealing in that case would be it stuck in limiting into an overload (the bad rectifier).

[Dan81]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

I had 5vsb when it was squealing and fan would even spin if green and black wires were shorted.

[kaboom]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

- When powered up, the Enermax does an insanely high pitched sound that keeps on going higher. What could it be? There's very little glue left on the board.

I had 5vsb when it was squealing and fan would even spin if green and black wires were shorted.

Sorry, I missed this.


Recap the entire aux section, both line and load side. A new 5VSB post-rectifier cap may (should?) stop it.
Does it calm down after a "preheat" with a heater/hair dryer/heat gun blowing on the PCB? If so, it's certainly those caps.

[Dan81]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

I'll likely just go on replacing most of the 5vsb parts with new working parts, but I was suspecting the small caps since they are in some quite hot places.

[kaboom]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

Change the caps. Rule them out first before condemning anything else, as they'd be my prime suspects for a unit of this age.

Again, does it quiet down after a "preheat" from a heatgun, hair dryer, or heatsoak outdoors? If so, it's the caps.
You can test this possibility w/o changing any parts.

[momaka]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

Change the caps. Rule them out first before condemning anything else, as they'd be my prime suspects for a unit of this age.

+10

Squealing 5VSB is almost always a result of bad output caps on the 5VSB. If the 5VSB is 2-transistor design with a "critical" cap, sometimes that cap could be the issue as well. Last but not least, I've seen one very old Morex PSU that used feedback both from the 5VSB and the 5VSB secondary-side aux. rail. Although such designs are probably very very rare, it's worth replacing the filter cap for the 5VSB secondary-side aux. rail as well.

But in short, replace ALL caps that are in the vicinity of the 5VSB section. No point in playing the "guessing game" and changing other parts at this stage until you do a recap.

Also do NOT power On the PSU anymore until you recap the 5VSB. Otherwise, you risk of killing your PWM controller and/or the secondary-side BJT driver circuit.

[kaboom]

Re: Wolf in sheep's clothing - Enermax in a Deer casing

+10

Squealing 5VSB is almost always a result of bad output caps on the 5VSB. If the 5VSB is 2-transistor design with a "critical" cap, sometimes that cap could be the issue as well.

The once-common Panasonic VCR SMPS had a 1u 50v cap (C21), that when leaky, shunted the feedback signal in such a way the outputs never fully came up. It overbiases the LED in the optocoupler. This supply, and the Samsung and Goldstar copies would squeal just like an ATX PSU when the caps started going; incidentally, the GP caps Panasonic used in the early 80's would sometimes even last 5+ years. Compare that to modern, fake "low-esr" junk that's bad after a year or two.

Also do NOT power On the PSU anymore until you recap the 5VSB. Otherwise, you risk of killing your PWM controller and/or the secondary-side BJT driver circuit.

Also in those Panasonic SMPSes, extreme cap failure causes the zeners across the outputs to short (energy from the primary has nowhere to go), hopefully protecting the VCR. The earliest designs, some w/o the zeners, supposedly could run away to the point of blowing all the ICs in the machine up! The switching frequency was said to coincide with the leakage L and parasitic C of the transformer- ie, ferroresonance! In a little SMPS transformer!

I mention that supply because it's the same topology as a typical aux supply- a blocking oscillator flyback thing. Complete with "energy recovering" snubber, in 1982.