Jump to content

Astec Power Supply for 8100 - How did I mess this up?

Recommended Posts

I will make a brief summation, since I have a tendency to be verbose.


Simply put: I re-capped a perviously working Astec power supply from my Power Mac 8100, and now it is not reliably functional.


More in depth: all I set out to do was replace the big electrolytics on this thing. There are three boards on this beast, one being a relatively tiny one that has two caps on it, and two other main boards. The smaller board needed to be completely desoldered to replace the caps, as these things are stuffed in there at a 90 degree angle to the other components on the main board. Below is exactly what happened:


1. Took out the supply, disassembled it, desoldered the existing caps, and soldered in new caps with the same ratings as those there previously.

2. Reassembled the supply. Hooked it up. Started the machine. It started up normally. I started up a Game of Marathon, and right as it opened the application, the entire machine just shut down. I tried to start it again, I hear the power supply click, and then off. If I do that a few more times, there isn't even a click.

3. Waited a while. After waiting 15-20 minutes, I was able to turn the machine on again. However, part way through startup, it immediately shut down again. It seems like capacitors bleeding off stored charge somehow helps it be able to start up.

4. Took the supply out, and jump-started it from the trickle-voltage pin to the start signal pin with a resistor. While it was running, I checked voltages. All grounds have no voltage, all +5s read 5, the +12 is 12, and the -12 is -12. So the appropriate voltages appear with it not connected with no load.

5. Got another power supply and tested it with the machine to make sure it was not an issue with the machine otherwise. The machine functions normally with the other supply.

6. Opened up the supply, looked over all my work, desoldered some cap leads and resoldered, made sure they are very clean. Touched up some ground pad areas. Reassembled with the other supply as a reference to make sure things were going where they should. Re-tested....and the same issue persists.


At this point, I'm at a loss to explain it. Though it seems a long shot, I had an issue like this once with the sound circuit on a Quadra 630, where the sound would play initially, but then would fail because there was no path for one of the sound caps to ground. However, this is so complex, I am not sure what goes where, and to my knowledge there is no schematic floating around. It is bothersome because it worked fine before, and I am not sure what I did to cause this.


Roping in @Bolle, @techknight, and @cheesestraws because, well...you guys are always helpful; and have good ideas! I would appreciate any suggestions from anyone for me to try.

Link to post
Share on other sites
  • 68kMLA Supporter

I know this doesn't specifically deal with whatever issue is causing the behaviour on your 8100 PSU, but I have read at least a couple of people mention that these PSUs can become unreliable with age (granted, your issue only started happening after recap so is assumably related to those caps) but have you considered a retrofit of an ATX PSU?

Link to post
Share on other sites

you have a bad connection somewhere in there. 


the thing with power supplies, they either work, or they dont. 


The only time things become intermittent is either with bad connections, or bad capacitors. (which you have already changed). 


If the protection circuitry drifts a bit, it could sense an overcurrent condition and trip it out, then it will cycle, so you cant rule out resistors that have drifted out of tolerance. But if this is random, you have a bad connection. 


Next thing could be a bad optoisolator that is dropping out on you. This is common. Also, was there any cap corrosion? If so, check the nearby transistors and ICs for green crusties. Epoxy isnt perfect, and the corrosion will walk itself up into the chip and into the die causing problems. 

Edited by techknight
Link to post
Share on other sites

Thanks for the insights. So, I'll say: the problem is completely predictable and reproducable. Happens every time exactly the same. I tested yesterday, and it really is an instantaneous issue. +5 and +12v rails are 100% within spec until it shuts right off. Attempting to start again produces a click, and then nothing. The more you try in rapid succession, it will do nothing (no click even). Then, if you leave it alone and unplugged, try a few minutes later, it will start again. It seems like there is some built-up charge or something that causes it that has to dissipate over time.


There was no corrosion or any trace of leaked electrolyte at all. Last night, I went through the entire supply, looked over every milileter of work that I did. Touched up anything that seemed even slightly funny. Fired it back up, and it seemed to work for a while. I thought I was good to go when it cut out..yet again. But it did run for a bit longer I think (this sounds like 4-cycle engine testing with carburetor adjusting more than electronics...).


Lastly, as I mentiooned earlier, this worked 100% fine before the recap.


I will get some photos. I'll find a place to upload them as I am still capped here on the forum as far as attachments.


Edited by LaPorta
Link to post
Share on other sites

I successfully repaired the ASTEC PSU of my PowerMac 6100 three years ago. This process is partially documented in this thread: https://www.badcaps.net/forum/showthread.php?t=62455&highlight=Astec


I highly recommend to skim over the above thread because it contains a lot of useful advices regarding replacement parts for old caps given by professional technicians.


I remember to have replaced all electrolytic caps on the small daughterboard as well as the voltage regulator IC (AS431).

I was told to left the big caps as is.


I also built a device to test the repaired PSU under load before connecting it to the Mac.


Please note that those PSUs need a proper load to get the stable voltages.


I hope you made pictures of the PSU and all its components before desoldering anything.


Can you tell us the part number of your PSU? You can try to search the badcaps forum to see if you could locate a schematic for it.

Edited by powermax
Link to post
Share on other sites
  • 3 weeks later...

HI there,


Sorry I missed your reply. Now that I am back to working on my 8100, I can get back onto this. I do have part numbers:


Made by Delta Electronics.


Model: DPS-225AB Rev: A0


Apple Part #: 614-0040


I'll search the site and see. I have replaced each and every electrolytic. I have some photos of it originally. 

Link to post
Share on other sites

So, I have been working with people over at badcaps.net. I replaced a capacitor that one of the members mentioned probably was not in spec vs. the original part, and it is now operating stably. However, I found out something that I had not previously realized:


This PSU is the entire reason that the machine didn't function reliably in the first place!


When I power the 8100 with my borrowed supply, the entire thing works perfectly.


When I use this PSU, it will start, run, etc, but I get crazy random freezes, buss errors on startup, etc.


My conclusion: there must be some sort of weird interference in the power distribution ("noisy" or "dirty" power) that is messing up the digital circuitry just enough to cause errors and freezes on the logic board. Question is: how would I track down what causes the issue to start with? It must be a faulty part, because this issue has plagued the machine since I received it in 2002 (I never used it, just stored it since then).

Link to post
Share on other sites
  • 4 weeks later...
19 hours ago, LaPorta said:

I can’t believe it. I recapped the good power supply. Result? It now doesn’t work either!


I think I am done recapping power supplies. They are just too temperamental. I feel like such an absolute failure.

Oh man. That totally sucks. I hope you get it figured out. I haven’t yet touched any of my power supplies because I’m afraid to make them worse. 

Link to post
Share on other sites

well I haven't actually destroyed anything, thankfully. I do clean as well, and I do remove one by one and recheck. I certainly can't install one, reinstall the supply, test, then take it out again. That would be ideal, but as a busy physician, husband, and father of three, that would literally take me weeks. Not to mention, if I waited another decade until they are older to do this, no doubt some of my boards will have become toast due to electrolyte.


So, I suppose I will have to just order other caps closer to these. It is in this way that I have learned just how sensitive power circuits are. The same thing happened to me with another power supply, and, indeed, just one did not quite match needed specs (even though capacitance and voltage specs were the same), and made the supply shut off after a few minutes. I'll keep everyone updated with progress.


This thing is also one of the most convoluted supplies to work on. It has literally four different boards, one main and three soldered in at right angles to it. Replacement requires desoldering entire boards to work on them, etc...it is mind-boggling.

Link to post
Share on other sites

That really sucks.  I usually try to match the series or at least get caps that are made for switching power supplies. Who knows what the ripple current is doing as I am no engineer.    I think with PSUs or analog boards on Macs that don't support soft power it is probably not as big of a deal but it seems the ones that do have soft power it might be more crucial.  I could be wrong as I am only going by what I have experienced myself.

Link to post
Share on other sites
  • 68kMLA Supporter
On 4/12/2021 at 9:40 AM, LaPorta said:

Well, I suppose now I am going to change my strategy. I will remove all caps, document, THEN look them up by series. I have a feeling these PSUs are very, very finicky with the tolerances of the caps.


It is really frustrating/disappointing when a repair project doesn't work out as planned, even worse when its preventative maintenance. I think many of us have been there. :-/


I guess I've been relatively lucky with the power supplies I've recapped lately.  For what its worth (maybe not much!) this is the approach I take:


With so many options on Mouser my process has been to start filtering under Aluminum Electrolytic Capacitors:


Manufacturer = Nichicon (preference I guess, think I've heard good things in general)

Term Style = Radial

Capacitance = <original value>

Voltage rating = <original value> (understanding that I can later increase if I have to for some reason, e.g. no match in needed diameter)

Diameter = <caliper measures original value> (often a real physical limitation)


That usually get me down to handful of options, I look at the series "Datasheets" since I noticed several Nichicon series explicitly state "For Switching Power Supplies", so far these are my notes, the last two being fall back options:


UTT = Long life UPW  = Miniature Sized, Low Impedance, High Reliability For Switching Power Supplies

UPW = Miniature Sized, Low impedance, High Reliability For Switching Power Supplies


UPA = Miniature Sized, Low impedance, High Reliability For Switching Power Supplies


UPV = Long life UPJ = Miniature Sized, Low Impedance, High Reliability, For Switching Power Supplies

UPJ = Low Impedance, For Switching Power Supplies


UPM = Low Impedance, high Reliability

USV = Long Life Assurance


If there are still options I look for longer "Life" ratings, lower "Lengths" (more room for airflow), and "Bulk" packaging (personal preference).


I honestly haven't tried to find the detailed specs (and/or series) of an old capacitor based on its markings, it seems like that might be difficult for old capacitors but will try to do so next time. 

Edited by Fizzbinn
Link to post
Share on other sites

Fizz, I like your methodology. I am currently looking through United Chemi-Cons stuff, to closely match the series (the series are old, so no longer made). Low impedance as like you said, switchers cycle really fast. I'll bet you are onto something: I will bet that as the current switches back and forth, there is a ton of impedance in one of the parts I have which is why it only lasts a minute or so on before it quits. I'm no electronics person, but I would venture that is a possibility.


I'm ordering from mouser now myself. I really like your chart: thank you for that, huge for future reference!

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...