To repair or to replace? Atari PSU

[cbmeeks]

I have six Atari ST's.  Out of the six, only one has a power supply.  That power supply is internal to the computer.

I haven't powered it in years.  So as I was restoring it, I noticed that one of the caps is slightly bulging.  Nothing appears to have leaked.  And, the bulge is SLIGHT.  But, you can feel it.

The bulging capacitor has a big red sticker on it saying "Warning shock hazard. Heat sink is LIVE..."  It's the only cap with that sticker.  It's a 47uF 250V cap.  I assume it's on the AC side of the PSU.

Now, I'm about a 5 out of 10 on soldering.  I've repaired many computers, soldered projects, etc.  But, I've never been on the AC side of the PSU.  I've built many DC power supplies.

Anyway, my gut tells me I can spend the $20 in parts to recap this PSU.  My brain says spend $75 and get a new one (well, a refurb one with warranty from official Atari store).

What do you guys think?

I can send pictures if you want.  The model is:

ASP34-1

117V 50W DC  O/P 5V 3A 12V 1.6A

Thanks for any suggestions.

[Delta]

Go with your gut, mate.   

I too treat the mains with a huge amount of respect, but soldering is soldering.  Use plenty of flux and get the joint nice and hot before applying the solder.

[linux-works]

I just did a house move and found my atari 520st in a forgotton box.  want it?  I have both monitors (bw and color) and 2 drives plus some original PSUs.

PM me.  its just collecting dust here, and I'd like to see it go to someone who could appreciate it or cares about it.

[cbmeeks]

I just did a house move and found my atari 520st in a forgotton box.  want it?  I have both monitors (bw and color) and 2 drives plus some original PSUs.

PM me.  its just collecting dust here, and I'd like to see it go to someone who could appreciate it or cares about it.

Yes please!

PM sent.

Thanks!

[cbmeeks]

I've decided to try and repair the PSU myself.  I will just be super careful.  I ordered the caps.

After that, I'm going to design an S-Video adapter for it like I did for my Amiga.  Then, some sweat Atari ST action!

[Delta]

I've decided to try and repair the PSU myself.  I will just be super careful.  I ordered the caps.

After that, I'm going to design an S-Video adapter for it like I did for my Amiga.  Then, some sweat Atari ST action!

Pfffft!  You with your fancy-Dan 16bit computers!  Pah!  I'll stick with my C64 thank you very much! 

[cbmeeks]

I've decided to try and repair the PSU myself.  I will just be super careful.  I ordered the caps.

After that, I'm going to design an S-Video adapter for it like I did for my Amiga.  Then, some sweat Atari ST action!

Pfffft!  You with your fancy-Dan 16bit computers!  Pah!  I'll stick with my C64 thank you very much! 

LOL.  Not to brag...but I bet I got ya beat on vintage computers.  I started out with a TI-994/a and now I own six of them.  But the C64 is my main squeeze.  You can find me on Lemon64.  :-D

But, seriously, I have many vintage computers but for some reason, I've never gotten any of my ST's to work.  Which is frustrating since I also have six of those!  However, in my defense, I've only got the one PSU for them.  And, three of them are motherboards only.  So I felt like I was neglecting the ST and not giving it the proper vintage love it deserves.  :-)

Oh, especially since I just finished some restores on my Amiga's.

[retiredcaps]

But, I've never been on the AC side of the PSU.  I've built many DC power supplies.

Discharge the 47uF 250V capacitor first and then verify it reads close to 0V with a multimeter.  Then solder.

[cbmeeks]

Discharge the 47uF 250V capacitor first and then verify it reads close to 0V with a multimeter.  Then solder.

Excellent suggestion.  However, it hasn't seen a power source in about 5 years so it's probably discharged.  LOL

But I get it...best to be safe.

[retiredcaps]

However, it hasn't seen a power source in about 5 years so it's probably discharged.  LOL

But I get it...best to be safe.

Even though it is 5 years, you can't assume anything.  See my experience with a PVR that was not plugged into for 4.5 weeks.

http://www.badcaps.net/forum/showthread.php?t=41319

[bitseeker]

I've decided to try and repair the PSU myself.  I will just be super careful.  I ordered the caps.

Good call. At least give it a go. If it doesn't work out, you can always buy a replacement supply. One thing on my to-do list is to build a central power supply for all my vintage computers, rather than continuing with the gaggle of wall warts.

Not to brag...but I bet I got ya beat on vintage computers.  I started out with a TI-994/a and now I own six of them.  But the C64 is my main squeeze.  You can find me on Lemon64.  :-D

Cool. I learned to program on a TI-99/4. Then, got a first-gen CoCo. Later a few Atari 800XL and 130XE. I have a C64 around somewhere, too. Never had an ST, though from time to time I've contemplated getting one or a TT.

D'oh. Looks like I just missed out on linux-works ST clearance sale.

[stj]

before you do anything,
can you actually see the metal top of the cap?

a lot of high voltage caps have a plastic disc on them that can bulge wihout anything being wrong with the actual cap.

[cbmeeks]

Success!

can you actually see the metal top of the cap?

Yes.  I may have been paranoid but I replaced it anyway.  The old cap was tenting slightly.  So I replaced it with a new one (Nichicon). 

And viola!  The Atari ST lives!

I bought other replacement caps too.  I figured if one was old and going out the others may be too.  I will probably recap them sometime but I'm not sure if I need to immediately.  Computer seems to be running fine.  Even with floppy disk access.

[CJay]

Success!

can you actually see the metal top of the cap?

Yes.  I may have been paranoid but I replaced it anyway.  The old cap was tenting slightly.  So I replaced it with a new one (Nichicon). 

And viola!  The Atari ST lives!

I bought other replacement caps too.  I figured if one was old and going out the others may be too.  I will probably recap them sometime but I'm not sure if I need to immediately.  Computer seems to be running fine.  Even with floppy disk access.

I bought and repaired a job lot (200 ish) of Atari STs many years ago a month or two before Christmas, most common problem was the PSU, IIRC it's a bent metal bracket with a small board mounted on it, the problem with the TL494 based PSUs was the startup supply capacitor and dropper resistors (IIRC 470K and 47uF)

Replacing those got a good 90% of the PSU faiulres fixed, one or two had explosively disassembled and required a TL494, FET, bridge diodes and fuse.

Others had shorted output rectifiers.

As someone else pointed out, the plastic cap on those caps is usualy what bulges, not the cap itself. It's there to give an extra level of insulation.

Happy days even though I was an Amiga fan.

Don't remember replacing anything else but I could probably find my records from way back when if I could find a 5.25" floppy drive.

[bitseeker]

And viola!  The Atari ST lives!

I bought other replacement caps too.  I figured if one was old and going out the others may be too.  I will probably recap them sometime but I'm not sure if I need to immediately.  Computer seems to be running fine.  Even with floppy disk access.

Excellent! Glad you gave it a go. Congrats on a happy patient.

[analogix]

I realize this is an old thread.... but I just came across it as I have a couple of Atari STs which probably need their power supplies recapped, given their old age.

Besides checking each capacitor's value, voltage rating, temperature rating, height/diameter and pin spacing I'm at loss for what kind of electrolytic capacitors to get as there are highly conflicting information floating around the net. Some say to get low-ESR type caps while others say the opposite as the former could drastically change the characteristics of the PSU (which was designed years before such low-ESR capacitors were available).

Obviously I want to do this right even if it'll cost a little, so should I look into (Japanese quality branded) "normal" capacitors or the "low-ESR" types? 

[stj]

the high voltage cap on the input should be a regular but 105' rated one - rubycon make the best.

the other caps should be low-esr

[analogix]

I agree about the temperature rating. Also I'm trying to find the same long hour rating (10 000) for all capacitors so there's no second guessing about which capacitors need replacing first, the next time around.

In other forums I've been told that with older switched mode power supplies you should be wary of replacing the existing caps with low-ESR ones as the original design didn't take the much lower impedance into account and this may ultimately cause problems and instability such as oscillation of the feedback loop (don't quote me on the exact explanation, and also I don't know what that means). Is there cause for concern?

There's an explanation on recapping my particular Mega STe PSU, which lists different capacitor values than the ones originally used. Panasonic FR capacitors are listed at the very top of the page. Furthermore it says that new, higher values have been chosen to deliver a larger supply current for better stability and to allow various computer hardware add-ons.
Finally it explains that with lower ESR values and higher capacitance, extra strain will be put on the rectifiers when power is applied, so a higher current rating for the diode(s) should be used. I'm not sure if that was referring to a different Atari power supply or in general (the page lists numerous types) but though it's worth mentioning since the capacitor values have been changed for most of the caps.

I'm not sure if I feel confident about replacing caps in a high voltage PSU with something other than their original values, but then again replacing them with low-ESR types could affect their performance in the rest of the circuit, so I'm wondering if using different values is to compensate for that, ultimately making things right, or if using low-ESR caps with their original values would make things wrong because of their performance in the circuit.... complicated matters 

[bitseeker]

It's true that in switched mode power supplies, the ESR of the caps influences the behavior of the circuit. Too little ESR and there could be oscillations that were originally suppressed by the built-in resistor (i.e., the ESR).

You could try low-ESR caps and look at the results with a scope to check for oscillations. If that's not something you want to do, though, just stick with caps of similar ESR to the originals.

As for the capacitance values, larger won't hurt if it's not overdone. Go too far and higher or longer-duration inrush currents may be a concern. I hadn't thought about higher inrush with low-ESR caps, but that seems plausible, though I'm not sure that it'd be a huge difference. That's probably why it's recommended to use regular caps on the high-voltage side.

[analogix]

What exactly are oscillations in a switched mode power supply, and what kind of problems can it cause?
I do have a small 200KHz range scope -will it suffice, and what should I look for when testing the outputs?

Regarding the changed values, as described here, which you indicate is OK.... Here's a list of the original values and their replacement values:
 
C6 120uF/400V (150uF/400V replacement)
C7 10uF/400V (same replacement)
C15 1000uF/16V (1500uF/16V replacement)
C16 1000uF/16V (1500uF/16V replacement)
C14 2200uF/16V (4700uF/16V replacement)
C13 4700uF/16V (same replacement)
C17 680uF/16V (1500uF/16V replacement)
C18 680uF/16V (1500uF/16V replacement)
C19 100uF/10V (same replacement)
C23 22uF/25V (10uF/50V replacement -not a typo!)
C20 3.3uF/50V (10uF/50V replacement)
C26 4.7uF/50V (10uF/50V replacement)
C8 100uF/25V (220uF/50V replacement)
C24 220uF/25V (same replacement, but 50V)

The article also lists a SSP6N60 (600V 6A 1.8R N-MOSFET) which I'm guessing means that Q1 (MTP6N60) should be replaced by it.
What might the reasoning be behind all of these component changes, and is it safe?

[stj]

just use FR, if the guy who made that schematic did and it worked then that's settled.

i know what pete said at badcaps, but he always says that.

[analogix]

So you've been following the badcaps thread as well 
I downloaded the schematic for my PSU from somewhere online (I forgot where), then added the red/blue notes myself before uploading it here (and elsewhere), based on the recapping info found here.

I was hoping someone with the right knowledge could take a look at the schematic and verify if using the different valued caps and/or low-ESR caps would work out or not, as for me this is pure guesswork. I'm handy with a soldering iron and a bit more, but with high voltage stuff (and something that could damage my vintage computers if done wrong) I want to get it right the first time without any guesswork if you know what I mean.

[stj]

like i said, if it works for others.

[bitseeker]

What exactly are oscillations in a switched mode power supply, and what kind of problems can it cause?
I do have a small 200KHz range scope -will it suffice, and what should I look for when testing the outputs?

Oscillations in a switched-mode power supply (SMPS) are instability in the circuit typically manifested by an unstable output. There's a lot more to it with regard to phase shifts caused by capacitance and inductance, and their effect on the gain of the control loop, etc., etc. If you take a look at Figure 4 on page 7 of this paper from Kemet about the effects of low ESR, the first scope capture is an example of oscillation where the output is constantly changing between approximately 4.8 and 5.0 volts. The two scope captures below it illustrate what happens when adding some resistance (effectively restoring higher ESR) or capacitance to stabilize the control loop. The output is finally stable at approximately 4.9 volts.

The article also lists a SSP6N60 (600V 6A 1.8R N-MOSFET) which I'm guessing means that Q1 (MTP6N60) should be replaced by it.

Yes, that's what that means. I wasn't able to find the MTP6N60 on Mouser or DigiKey, so perhaps it's no longer available, hence the suggested replacement if yours is dead.

What might the reasoning be behind all of these component changes, and is it safe?

From what you wrote previously, most of the capacitor changes are to take advantage of modern low-ESR technology wherever the circuit will tolerate it in order to get better efficiency, more transient power delivery capability (via higher capacitance), longer life (less self-heating, more voltage headroom, lower relative ripple), etc.

The nitty gritty of all the goings on in an SMPS circuit are quite involved and there are folks here who are much more adept at that level of detail than I am, if you really want to dive into how a particular power supply will behave with particular changes.

Otherwise, you'd need to rely on the real-world outcome documented by those who have already tried it. If they achieve stable operation with the changes as stated, then you'll likely have the same outcome. If you don't, you'll certainly see it with either a scope or a DMM set to AC or Hz mode. Large oscillations would even be apparent in DC mode (output constantly going over and under the desired value). Then, you could opt to change the components back to factory-original values.

I'm handy with a soldering iron and a bit more, but with high voltage stuff (and something that could damage my vintage computers if done wrong) I want to get it right the first time without any guesswork if you know what I mean.

Yes. So, be sure to do plenty of verification with a DMM before connecting the supply to anything you value. If you have an electronic DC load or power resistors, it would be good to verify operation under load as well prior to putting the supply into normal operation.

[stj]

if you want to get technical,
you just have to try it.

to understand the circuit you need to know the frequency it runs at, and the inductance of the output coils.

then you can try to calculate the optimum esr for the pie filters.
i really wouldnt go there - even people designing supply's dont like that stuff.