Story of another dead deer and a doppleganger.

[eccerr0r]

I found one of my machines had died and it had an Allied 250W PSU in it...another Deer it looks.

5VSB is dead. Figures.

Anyway as I was searching around for that DR-0183 switch controller chip (DR=Deer? Did they get their own chips?) I found someone had a 400W supply with the same chip. Well, that's not that surprising, TL494/KA7500's are used everywhere.

As I looked closer at the guy's pictures...his 400W PSU looked very uncanny...it was an exact copy of my 250W-rated supply! Same 470uFx2 input caps, placement is identical... Well, almost. Mine has input filters, his has a (possibly fake) PPFC choke. And then looking at it even closer, it seems his just has a bigger TO-247 diode packs instead of the TO-220's in mine.

And his was working apparently, but wanted to mod it for a higher output voltage. I suspect his PSU would probably soon end up like mine soon anyway... dead.

Alas most likely another worthless repair coming right up! Once I figure out why the 2-transistor oscillator is dead...

TBC...

[PeteS in CA]

Re: Story of another dead deer and a doppleganger.

470uF, if within spec, would be good for 250W, but not for 400W. Unless the switch transistors are beefed up, the transistors that were OK for 250W would have probably brief violent lives at 400W. And if the same heatsinks and fan are used, they probably would not be good for 400W. Then there's the output inductors and caps. Using the same PCB for uprated power levels can work, but components have to be uprated.

You probably know this, but check small electrolytics in the Stby primary circuit.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

Yep, the heatsinks are the exact same, a flat piece of aluminum with a comb cut and fanned out of the top. The absolute cheapest heatsink - not even an extruded heatsink with holes. All in all, I agree, this should do 200-250W just fine, but have my doubts of that 400W. However I'm grateful the placement/schematic are the same.

A cursory check with the ESR meter indeed indicates a boatload of bad caps, not only of the 5VSB but also including the primary outputs. Now the question is whether it's worth to repair or diagnose further, I'm hoping the DR0183 isn't fried as collateral damage as I have no spares. Studying the schematic a bit, this circuit is funny, it looks like it was designed so that the same board could take both BJT and MOSFET primary switching transistors which makes it more difficult for me to bypass the 5VSB to see if the rest of the circuit works.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

BTW: the doppelganger is at http://www.webx.dk/oz2cpu/radios/psu-pc400.htm
Attached is my PSU.

Looks like the 5VSB caps are in the upper/center just to the left of the fuse, and the small cap between EEL-19A transformer and the low voltage heatsink is dry. It's not alone, lots of high ESR caps in there...

[PeteS in CA]

Re: Story of another dead deer and a doppleganger.

With those heatsinks I'd be nervous about greater than 200W continuous. Some of those output caps look preggers.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

Yeah, a bunch of caps are bulging. Need to be replaced as well.

I wonder how many watts I can (safely) pump through this if I *just* replaced the heatsinks with some larger ones that I saved from other dead PSUs, and swapped the output rectifier diodes with larger ones...

[stj]

Re: Story of another dead deer and a doppleganger.

a well assembled shit design!
fuse in clips instead of soldered in,
fan using a plug/socket assembly,

but the actual design - i would use it for parts myself!

[kc8adu]

Re: Story of another dead deer and a doppleganger.

and gobs of evil brown glue.

[stj]

Re: Story of another dead deer and a doppleganger.

and hotglue - i hate that wax-based stuff!

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

I thought hot glue was low melting point plastic and not actually glue or wax? It feels waxy however, but not actually melting at finger temperatures like real wax. Feels almost like set RTV.

Are there really any parts worth saving in this? I don't think this even has overvoltage shutdown, or I can't tell if it's included internally within the DR0183.

I was kind of surprised - it even has a full complement of noise filters - full pi filter on output and filter on input, unlike its doppelgänger.

[stj]

Re: Story of another dead deer and a doppleganger.

white spirit will liquify hot glue - whatever that indicates!

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

Gasoline and other hydrocarbons dissolve polystyrene, which isn't a wax.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

Okay...This sucks.
Multiple failure time.

I got 5VSB to work again but PSU is still broken. I want you guys to take the wildest guess what the problem was with the 5VSB. I couldn't believe it myself but when all the usual suspects are ruled out, the only thing that remains must be the problem.

Useless hints:
transformers are all good
V+ (the voltage across the main caps) is good, unlike that other PSU I fixed.
The burned area across the 1K resistors have nothing to do with the 5VSB failure.
I did replace a part.

Take a wild guess of why the 5VSB was dead, while I figure out why the main supply is going into shutdown right away... So it looks like quite possibly the DR0183 does have over/under voltage protection...

---

Oh well, guess wasn't enough time as I figured out the shutdown. And the supply has been completely rootcaused: the mystery part and bad capacitors on +5 (main) and +3.3 lines. Supply is deemed fixed as I stuck a capacitor in parallel with the bad ones to test my theory. I just need to solder new capacitors in...

[stj]

Re: Story of another dead deer and a doppleganger.

wire broken loose from the board??

[momaka]

Re: Story of another dead deer and a doppleganger.

I got 5VSB to work again but PSU is still broken. I want you guys to take the wildest guess what the problem was with the 5VSB. I couldn't believe it myself but when all the usual suspects are ruled out, the only thing that remains must be the problem.

Useless hints:
transformers are all good
V+ (the voltage across the main caps) is good, unlike that other PSU I fixed.
The burned area across the 1K resistors have nothing to do with the 5VSB failure.
I did replace a part.

Well, if I remember correctly, those 1K resistors go to the Vcc of the DR0183 PWM chip. When they burn, that's a sign that the secondary-side auxiliary rail from the 5VSB has been going over-voltage, typically either due to a bad critical cap (22 uF, 50V, I believe) or the output caps going bad. Thus, I'm putting my money on either bad critical cap for the 5VSB, bad output caps on the 5VSB, or possibly the 5VSB FET/BJT died to save the rest of the 5VSB circuit. Yes/no???

Yep, the heatsinks are the exact same, a flat piece of aluminum with a comb cut and fanned out of the top. The absolute cheapest heatsink - not even an extruded heatsink with holes. All in all, I agree, this should do 200-250W just fine, but have my doubts of that 400W.

Yup, don't expect more than 200-250 Watts, depending on your room temperature.

A cursory check with the ESR meter indeed indicates a boatload of bad caps, not only of the 5VSB but also including the primary outputs.

JEE, what else did you expect!

Now the question is whether it's worth to repair or diagnose further, I'm hoping the DR0183 isn't fried as collateral damage as I have no spares.

It's not that bad of a PSU, actually. Once recapped, these older Deers and L&Cs are pretty dependable. But they are extremely basic PSUs - UVP and OVP trip points are quite low/high, and I think the 12V doesn't actually have either OVP or UVP (forgot which). Also, the efficiency of this platform is just sad nowadays - maybe 75% on a good day, if even that.

But like I said, they are dependable PSUs, once recapped fully. I've been using mine (a Deer DR-250ATX) for 5 years now (after recapping it, of course). Granted I don't have it running 24/7... heck, not even 1/7 - I use it maybe for one to three weeks every year when at my grandmother's house. After that, I turn off the PC, disconnect it from the wall, and store away.

Studying the schematic a bit, this circuit is funny, it looks like it was designed so that the same board could take both BJT and MOSFET primary switching transistors which makes it more difficult for me to bypass the 5VSB to see if the rest of the circuit works.

That's shouldn't be the case. These old Deers are based on the classic half-bridge design with BJTs. Only the 5VSB may use either a MOSFET or a BJT, but that's a different story.

Anyway as I was searching around for that DR-0183 switch controller chip (DR=Deer? Did they get their own chips?)

I think those chips are either an early version of the 2003/2005 "chip-of-the-year" ICs or perhaps a close variant of the AT2005B or ATX2005. In any case, the difference between those and the TL494/KA7500 ICs is that TL494/KA7500 only have two comparators inside them that can be used for the protections, whereas the earlier mentioned chips like the DR-0183 pretty much take the 3.3V/5V/12V output directly through a resistor to handle SCP/OVP/UVP.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

I guess it's gone long enough, and dang you momaka, you took my JEE joke for the solution to this puzzle

OK. At first due to those 1K'ers frying, I did think either the 5VSB output caps (1000uF C18/C19) and the critical 22uF C9 caps were bad. They were indeed bad but I had a bad feeling that it wouldn't fix the problem. Since their ESRs were awful, I swapped them anyway.

I powered up...

Still dead. Bummer.

So I went through finding other parts that were bad, testing a few out. After pulling a lead of R3, a FREAKING 1Mohm resistor, I found that this resistor was OPEN.

This would totally explain why 5VSB was dead. So I replaced it, and 5VSB came right up. So the solution to the puzzle was the replacement of R3 (or R5, since the two are in series) was open.

So I went ahead and shorted GREEN to GND with a load... a pulse of life...and dead again. Waited a few seconds and tried again, same. => Yes this thing is shutting down due to OVP or UVP or something.

For the heck of it I scoped the outputs and sure enough, the 3V3 line was going way overboard.

Now the joke that was taken away from me: JEE, I wonder why. 1 ohm of ESR.

And that was the end of the story.

[momaka]

Re: Story of another dead deer and a doppleganger.

and dang you momaka, you took my JEE joke for the solution to this puzzle

Oh. Sorry!

So I went through finding other parts that were bad, testing a few out. After pulling a lead of R3, a FREAKING 1Mohm resistor, I found that this resistor was OPEN.

Crap, that would have been my 4th guess!
I've seen this happen a few times, but it seemed like the failures were all random (no sign of the resistor overheating to become open-circuited).

For the heck of it I scoped the outputs and sure enough, the 3V3 line was going way overboard.

Now the joke that was taken away from me: JEE, I wonder why. 1 ohm of ESR.

Well, joke stolen or not, I still smiled at that one.

JEE is one of those names that is just asking for someone to poke fun at it, just like Canicon, PowMax, and Powork. Couldn't help it!

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

Yeah I was dumbfounded when I found that dead resistor, how could it be? 1M has no way of dissipating much power with the voltages we're dealing with here, even with 340VDC. Plus no burn marks. I have to chalk it down to perhaps bad resistor manufacturing.

So the bad units were JEE IM 2200uF/10V, Jun Fu HK 1000uF/10V, CapXon KM 10uF/50V (ESR was bad compared other 10uF caps I had but I think this one was in a "don't care too much about ESR position", and Fuhjyyu TM 22uF/50V.

Lest we not forget Fuhjyyu

[PeteS in CA]

Re: Story of another dead deer and a doppleganger.

Yeah I was dumbfounded when I found that dead resistor, how could it be? 1M has no way of dissipating much power with the voltages we're dealing with here, even with 340VDC. Plus no burn marks. I have to chalk it down to perhaps bad resistor manufacturing.
...

Really high value resistors can sometimes break down internally and open up. Another possibility is that resistors do have a maximum rated voltage. It usually is way higher than the circuit voltages, but on the primary side or in snubber circuits it can be relevant and a cause of failure.

[eccerr0r]

Re: Story of another dead deer and a doppleganger.

The other thing is that there were two of these 1M resistors in series, so as a voltage divider, only half would appear on either one. Still very unexpected for it to fail.

It is, however, on the primary side. They're virtually but not completely across the voltage doubled mains. The bleeder resistors flow more current (2x220K vs 2x1M+47K).

The 1M'ers were standard 1/4 W carbon composite resistors, and I thought most of them were rated about 500V, mostly due to arcover from terminal to terminal. Then again at 500V, the 1/4W 1M resistor would be dissipating 1/4W...