Amiga 2000 Power supply suddenly stopped working

[roadrash]

Cant believe my luck recently as everything I seem to touch goes wrong and this is the latest. I just spent the last few days making some other small repairs to the computer like reseating chips etc due to system instabilities and got it all working perfect. Turned it off, built it all back up and then when I turn it on the Power supply just when "Tick Tick Tick" and the fan jerked. I took it apart and nothing looked bad inside so I plugged in a PSU tester and powered it up again to see where the ticking sound seems to be coming from a Philips BUW13A (Silicon diffused power transistor) and I also accidentally touched its heatsink and it shocked me which I am sure its not meat to do either. So is it possible this is what has failed.

some pics attached.

[stj]

Re: Amiga 2000 Power supply suddently stopped working

live heatsinks arent unusual.
your tick sounds like a startup capacitor or an output short.
given it's lasted atleast 20 years, i'd replace the caps anyway.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

That's interesting. Don't think its a short as I've tested it out of the Amiga and it even ticks with nothing connected unless there is a short somewhere in the PSU itself caused by a failed component or is that what you meant. Are the two big caps in the picture the startup caps? Yes it might be worth replacing all the caps now after all these years right enough. I've got another Amiga 2000 but its PSU failed years ago too and its a older model 2000 anyway. Had this computer from new and its got all the kit like PC-XT bridgeboard 5.25" & 2 x 3.5" floppies plus Amiga hard drive and PC-XT Hard card. So really want this one fixed back up again its kinda special to me..

[stj]

Re: Amiga 2000 Power supply suddently stopped working

i did mean a short in the psu - usually an output diode.

startup caps are small ones - often 10-47uf.
you find them near the controller chip - in this case, ic1

[kc8adu]

Re: Amiga 2000 Power supply suddently stopped working

change the small lytic near the orange tag.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

change the small lytic near the orange tag.

Ok thanks will give it a try..

[PeteS in CA]

Re: Amiga 2000 Power supply suddently stopped working

If I read the date code on the LM339 correctly, that PSU was built in late 1988 or so. You should regard all the electrolytic caps, possibly excepting the two big Rubycon I/P caps, as suspect. The O/P caps are "Nicon" brand, which I've not heard of.

Usually I'd expect a PSU with a single switch transistor like that to be a discontinuous flyback, but I'm pretty sure it is not. First, there are two large O/P inductors. In a flyback the transformer primary serves as the inductor. Second, discontinuous flyback designs of that era were usually self-oscillating, and that PSU has a UC3524A PWM IC. I've not seen a 3524-type PWM since the late 80s. The SG3524 was, I think, the very first PWM IC, introduced, I think, in 1976. The "A" is an improved version. The TL494 and clones have surpassed it in low-end voltage mode switching power supplies. Anyway, the UC3524A has an internal oscillator. The vendor, Unitrode, was bought by TI in the 1990s, IIRC.

Re your immediate problem, that cap kc8 pointed out is a key suspect. I don't think that PSU will start up if its capacitance is way low or its ESR is way high. Checking for shorts in the output circuit, as stj suggested, would also be a good idea. It may be a slim chance, but Phihong has an office in Fremont, CA. You might try emailing them, usasales at phihongusa.com. Or maybe their Netherlands office, sales at phihongeu.com. I do know they have a design and a tech support group at their Fremont office.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

change the small lytic near the orange tag.

That Cap is a 100uf 35v. It looks ok but I know that means nothing

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

If I read the date code on the LM339 correctly, that PSU was built in late 1988 or so. You should regard all the electrolytic caps, possibly excepting the two big Rubycon I/P caps, as suspect. The O/P caps are "Nicon" brand, which I've not heard of.

Usually I'd expect a PSU with a single switch transistor like that to be a discontinuous flyback, but I'm pretty sure it is not. First, there are two large O/P inductors. In a flyback the transformer primary serves as the inductor. Second, discontinuous flyback designs of that era were usually self-oscillating, and that PSU has a UC3524A PWM IC. I've not seen a 3524-type PWM since the late 80s. The SG3524 was, I think, the very first PWM IC, introduced, I think, in 1976. The "A" is an improved version. The TL494 and clones have surpassed it in low-end voltage mode switching power supplies. Anyway, the UC3524A has an internal oscillator. The vendor, Unitrode, was bought by TI in the 1990s, IIRC.

Re your immediate problem, that cap kc8 pointed out is a key suspect. I don't think that PSU will start up if its capacitance is way low or its ESR is way high. Checking for shorts in the output circuit, as stj suggested, would also be a good idea. It may be a slim chance, but Phihong has an office in Fremont, CA. You might try emailing them, usasales at phihongusa.com. Or maybe their Netherlands office, sales at phihongeu.com. I do know they have a design and a tech support group at their Fremont office.

Wow I wish I had all your knowledge about PSU's. Motorcycles are my speciality but some PSU repair knowledge would come in handy too keeping all my vintage computers working. I've attached another couple of images for you to see now I have it out of its enclosure. I have another Amiga like this one but it has a different PSU and it went a few years back so will be fixing that one next and you will be able to see what they did with that earlier version. Apart from the 100uf 35v cap near the label is there anything else I should check. Is there some voltage readings etc at certain parts I can take that will lead us to the area of the fault.

[stj]

Re: Amiga 2000 Power supply suddently stopped working

to be blunt, given the age i would be replacing every cap other than the mains filters before i even test it.

if you want to learn something, download the datasheet for the controller chip - they usually have example circuits and descriptions of how it all works.

infact a lot of simple psu's and amplifiers are just clones of the example in the datasheet.

[PeteS in CA]

Re: Amiga 2000 Power supply suddently stopped working

The ancient 1976 Silicon General datasheet for the SG3524 was very informative, and it did have a sample design. TI's datasheet for the UC3524 looks pretty helpful, a bit more so than for the UC3524A.

That being built in 1988 or 1989, I'd suggest using older series parts from Nichicon or Nippon (or United) Chemi-Con - PS or PM series from Nichicon or LXV series from NCC/UCC for the output caps. Modern parts can be "too good" and cause stability problems. The series I named are similar to what was available when that PSU was designed and built.

[redwire]

Re: Amiga 2000 Power supply suddently stopped working

I see a dent on the top of 4700uF 16V cap and a couple TO-220 parts near it looked baked like they've run hot. You might be drawing too much from this PSU with all those addons. I would look there first, for a shorted semi.

It's very close to an IBM-PC power supply, with a special "TICK" signal which is probably AC line pulses. Electrotanya has motherboard schematics but no PSU schematic.

When working on this, be careful you don't bump parts into touching each other. There are many parts close to each other with no insulation.

[stj]

Re: Amiga 2000 Power supply suddently stopped working

you dont need the "tick" signal,
there is a jumper on the mobo to select between mains clock and using the video-circuits to get a frame clock.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

Just ordered a completely new set of electro caps from farnels today so will see if that sorts it. Like was said they are all very old now so its best replace them all now anyway. Will post bak how it goes when i have them fitted.

[stj]

Re: Amiga 2000 Power supply suddently stopped working

btw, have you put a usb floppy emulator in it yet?
those things are pretty nice.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

btw, have you put a usb floppy emulator in it yet?
those things are pretty nice.

I haven't yet no but will be trying one sometime as these old Amiga drives were pretty unreliable things always getting read errors, something that never often happened with the 720k msdos formatted ones. Even the HD 1.44mb msdos disks were more reliable than the Amiga's. I have a large amount of Amiga software and hardware I built up over the years. I even have one of the first ever Amiga A1000 computer systems I bought at its London launch back in the 1980's, after seeing the amazing things it was doing like the bouncing ball.

[stj]

Re: Amiga 2000 Power supply suddently stopped working

i saw that ball demo at the closed showing.
why they did it in secret inside a show-booth i dont now.

[roadrash]

Re: Amiga 2000 Power supply suddently stopped working

I am having great problems finding the 2 Large Electro caps (68uf 200v). They are 26mm diameter x 50mm high but the nearest I have found from RS & Farnells are 16mm x 25mm. How important is it the caps are the right size? These are half the size of the originals and surely not suitable. Where can I get these big size caps from?

[stj]

Re: Amiga 2000 Power supply suddently stopped working

forget the original size,
all you need to know is the maximum size that will fit,
tell me that info & i'll find what you need.

btw, those are probably still o.k.
they are decent rubycon ones.

[PeteS in CA]

Re: Amiga 2000 Power supply suddently stopped working

The 3 critical factors are that the replacements be 68uF or higher, 200V (or 250V), and the same lead spacing. Shorter is OK, 22mm diameter is OK if the leads fit. 100uF would just give you amazing hold-up time.

Like stj said, your originals are a good brand, and input 'lytics are less stressed. So they may be OK. What's the voltage across them when you apply power?