- The LaST Upgrade -

PART 6 - ATARI PSU REPAIR

exxos 2012 - Last updated March 21, 2024

QUICK LINKS

 

If your ST/E machine is suffering from random crashes, poor video, random hard drive or floppy access problems, noise in audio, then chances are your PSU is going bad. Considering the age of the PSU's now, it is of no surprise that they are in desperate need of updating.

Please note all of these are UK/230V types! I will also add more PSU and fixes as I come across the various PSU types. All capacitors used are carefully chosen for reasonable cost and very good ESR ratings. I sell kits to update the PSU's listed in my webstore along with new PSU's.

Considering the age of these PSU's now, the ones which still have good regulation are likely to get my recommendation. I like PSU's which have the BU508 transistor as its a good rated device over the smaller TO220 based transistors. Though regulation is the most important quality overall.

I have noticed the scope waveforms are not totally reliable in seeing PSU problems. For example, on some PSU's they appear to show good regulation, though when accessing the floppy drive, the screen brightness dims up to 50%. There seems to be a lack of "on demand" power from a lot of PSU's. I had tried initially adding another 4,700uF on the 5V rail to give it more on demand peak power, but it had little or no change in screen dimming.

If you have any PSU's other than those listed, then please take a clear hi-res image and send it to me. It would also help to know any model/revision numbers on the PSU along with the number on the main switching transistor. Also a list of the main electrolytic capacitors is helpful. If you have any ST/STE PSU's not listed and would like to donate them for inclusion here, then please get in touch.

If anyone has the circuit diagrams for any PSU's listed then please let me know.

NOTE - I failed to mention on the mV readings x10 probe was used, so 50mV is actually 500mV for example. I have corrected the mV results for the real voltages, however the scope readings may confused instead now :)

NOTE2: I keep getting asked about part numbers I use for these kits. Unfortunately I do not list them anymore simply because by the time I have sorted part numbers out and updated my site, the part numbers are no longer valid or discontinued. With several PSU's now on this part all needed various capacitors, it would be a huge undertaking to keep updating this page (time I just do not have unfortunately). Generally I go with the Panasonic FR series as they are really good ESR values. Please don't fit "any old capacitor sold around the Internet" as the cheaper capacitors are not suitable for switchmdoe designs and can easily overheat or explode! I am trying to update my store with new re-cap kits so please keep checking there.

WARNING - PSU's do not operate correctly without a load connected. If you try to measure voltages without the PSU connected to the ST, then the PSU will likely malfunction. If you want to test outside of your ST, then you must load the 5V and 12V lines with a 5-10watt 5V & 12V lamp or resistor. Resistor would be 2.2R - 4.7R 10watt minimum for 5V and 22R 10watt minimum for the 12V.

DISCLAIMER - Mains voltages can kill. Capacitors can explode. I do not accept any responsibility for anyone doing repairs to these PSU's, and the user doing any modifications or repairs do so entirely at their own risk! Only a qualified electronics engineer should attempt such repairs.

** MITSUMI SR98 **

Download schematic

Here is the output of the 5V line. This is actually very bad as it has about 1volt of ripple. This means the power supply is fluctuating between 5volts and 4volts constantly. This will cause random ST crashes and in some cases could totally kill a ST!

 

Here is a list of new parts. Some items on the Atari PSU are 10V but these were replaced with 16V types. The values were also increased to be able to supply larger current demands due to "add ons" and better stability.

Also these are not "cheap" capacitors. These have been carefully selected to give maximum ripple current and low ESR values. Cheaper capacitors may well make the results worse than you could start with, so using good capacitors is always a must.

ECOS2GP680BA or B43305A9686M000 CAPACITOR, 68UF, 400V - Replaces 400V 47uF

KBU1005 BRIDGE RECTIFIER, 10A, 600V - Replaces 2A rectifier

EEUFR1C152L CAPACITOR, RADIAL, 16V, 1500UF - Replaces all 1000uF caps.

EEUFR1C472 CAPACITOR, RADIAL, 16V, 4700UF - Replaces all 2200uF caps.

Just out of interest the old capacitor values were measured.

47uF 400V = 42uF
2200uF 10v = 1912uF
2200uF 16v = 1868uF
2200uF 10v = 1951uF
1000uF 16v = 887uF
1000uF 10v = 975uF

The main switcher transistor is 2SC2979 800V 3A 40W.

 

Rectifier fitting example submitted by Jonathan Whiteside.

The upgraded PSU. As this one lasted nearon 30 years, lets see how long this one lasts ;)

Also worth nothing the Rectifier legs were slightly larger than the holes in the PCB. These were carefully drilled out using a 1mm drill bit. No damage was caused to the pads as I guess the holes were already 0.8mm.

Also note as the capacitors are lower ESR values, and higher capacitance, this will put extra strain on the rectifier on turn on. 10A is a overkill, but I would suggest no less than 5A 600V type be used.

 

The new 5V line, Here we can clearly see the huge advantage in updating the PSU. There is about 400mV (0.40V) of noise on the 5V rail now, which is about as good as things get!

Overall the stock PSU is in desperate need of capacitor replacements. Once changed it has very good regulation and peak power demand isn't to shabby. There is slight dim on display during floppy access, but a lot better "peak power regulation" than some other PSU's I have seen.

 

SR98 PEAK CURRENT UPDATE November 3, 2015

I have found the SR98's peak current demand isn't as good as it could be. This is generally seen as a dim on the video when the floppy drive is being accessed. So I started to look at this issue.

At first I changed the opto coupler for a faster one (actually same one as used on the ASP PSU) thinking it would give a more rapid response time, but it changed nothing from what I could see. I later changed it back (after doing all the tests below). The faster one might be slightly better in regulation, but it is so slight its probably not worth changing.

I then started to follow the resistor networks on the PCB and found that there is a capacitor which is pretty much directly across the opto's emitter diode. What this would mean is there is a small delay when the voltage is applied to the diode as it has to charge up a 1uF capacitor. It would then mean the 1uF would delay in turning off also. It is difficult to exactly work out whats going on with the diagram but I did notice when the 5V was turned up to 5.5 - 5.7V things improved a lot.

For this test I used a STE. I always had some odd ghosting on it and wavy lines on the screen. When the voltage was turned up (way to high in fact for 5V) the video noise problems and ghosting pretty much went away. Though I did notice some odd juddering on the edges of objects afterwards. But the most interesting thing was when the floppy drive was turned on, the video did not appear to dim as it did before.

So I found 2 resistors R201, R202 which seemed to be a voltage divider which also goes via the preset pot. This divider then is feed to a small transistor which in turn charges the 1uF capacitor to the opto diode. R203 is in the mix aswell, seems to be another resistor in the 0V line to the transistor. In anycase, I added 10K resistor across the top of R201, R202 and now the video problems went away at 5.25V. Also the dimming problem is not there. Any lower voltage then all screwy video problems come back. So clearly there is a biasing problem. I have also replaced the 1uF with a ceramic to rule out aged capacitor issue, Though I am not totally sure it made a difference.

Overall how it looks is when the PSU gets under load rapidly (when floppy motor turns on) It takes to long to charge the 1uF capacitor and the voltage drops (actually only about 0.02V) but this is visible to a quick eye that the video dims. So the idea is to increase the current though the circuit to give a faster charge rate of the capacitor, which will turn the opto diode on faster, and this does seem to be drastically improving stability.

I have also tried lower resistor values down to 2k2, but this did not appear to change anything. I upped the 1uF to 2uF. Its hard to say but this might have improved some very slight video issues. This particular issue only seems visible on the "drag bar" on GEM windows where the horizontal lines have vertical noise issues about every 1cm. Though without this new PSU mods, those lines start flickering pretty bad. This STE seems to run happiest at 5.25V. Any lower and the video noise issues start coming back. Overall the peak power can be improved with just a couple resistors and a new capacitor :)

So the final round up.

R202 2.4 k red yellow red + 1.8k = 1.029k Ohms
R201 2.7k red purple red + 2.2k = 1.212k Ohms

Replace R202 with 1K
Replace R201 with 1.2K

The trimmer pot may need to be adjusted back to 5.00volts on the 5V rail.

 

 

DIODE UPDATE

The output diode in the LV side is a HRW34. Not a bad spec but gets pretty hot on the heatsink plate. So I looked around and found a better one a STPS10H100CFP. This one is much faster switching and offers a fair bit less voltage drop.

After I recapped a SR98, I took some new readings on the 5V rail..

Where we have about 500mV (0.5V) of noise. I measured about 49c temp wise on the heatsink after about 10mins run time. This is "burn ya finger" kinda hot ;)

After the diode change I got this..

Basically the "noise" had reduced by a drastic amount!

I zoomed in a bit more..

Here I now have about 200mV noise (0.2V) from about 500mV (0.5V) of noise before! Was actually amazed the regulation was improved by such a drastic amount! Temperature wise I got about 39C after about 10mins run time.

So the lower voltage drop diode (along with the insulated package) has reduced the heat output, where the efficiency of the PSU will be a fraction better. Though more importantly, Its heating up the capacitors next to the heatsink a whole lot less which will greatly improve their life. Over all, its a win-win modification :)

 

 

*** DVE DSP-508A ***

Download schematic

The ripple may look really bad, but this was taken on 50mV range. So this PSU would seem to have a lot better regulation than the SR98 PSU which appears to have around 400mV of ripple. Yellow is 5V rail. Blue is 12V rail.

After one of my SR98's died, I found the main switching transistor had fallen short circuit. This is a 2SC2979 which is a little wimpy.

On the 508A PSU, it just so happens to have a BU508A transistor..

Voltage and current ratings of the BU508 are much better than the 2SC2979. Due to this I am inclined to favour the 508A based PSU at this time. Considering the age of the PSU and still very good ripple figures and better switching transistor, It is probably one of the best PSU's I have seen so far.

 

I measured the capacitor values just for the hell of it. They seem pretty good overall, but a meter measurement does not necessarily mean the values will remain constant under load. I am sure I was told many years ago that capacitors giving a higher value than expect can be a indication they are failing also. Though going by the age of the capacitors they need changing anyway, so its what I did :)

330uF 35V = 378uF (replacement 680uF 25V EEUFM1E681 )
330uF 35V = 355uF (replacement 680uF 25V EEUFM1E681 )
330uF 35V = 376uF (replacement 680uF 25V EEUFM1E681 )
2200uF 16V = 2330uF (replacement 16V 4700UF EEUFR1C472 )
2200uF 16V = 2310uF (replacement 16V 4700UF EEUFR1C472 )
47uF 250V = 48.2uF (replacement 250V 47UF EEUEE2E470 )
47uF 250V = 48.6uF (replacement 250V 47UF EEUEE2E470 )

KBU1005 BRIDGE RECTIFIER, 10A, 600V - Replaces 2A rectifier

It is also worth while re-soldering the BU508 legs. Anything which gets hot in fact should be re-soldered. Anything which looks suspect should also be re-soldered also of course!

I took some better images on how to fit the rectifier. The good thing is the holes did not need drilling which is another plus point of this PSU! Click on any of the 3 images below for a higher resolution shot!

A small improvement which means the new capacitors are doing their job :) The difference is not huge, but that was to be expected considering the PSU was in pretty good condition to start with. Regulation is below 200mV which is pretty good and no noticable dim on video during floppy access, so peak power is very good also.

ASP34-2

Download Schematic

Main transistor is the BU508.

Capacitors are
250V 33/47uF x 2 - Replaced 250V 47/68uF
35V 330uF x 3 - Replaced 470/820uF 35V
16V 4700 x 1 - Replaced 4700uF 16V
16V 2200 x 1 - Replaced 4700uF 16V

 

Regulation on the ASP PSU seems to be struggling a bit on first glance, though probably down to aged capacitors. The PSU make a fairly loud ticking sound, probably 4 or 5 ticks a second which is a little unusual. This PSU dims on floppy access much like the SR PSU, though it is only slight.

After updating the capacitors the results were interesting. The screen dim on floppy access was almost not noticeable, so great improvement there and peak power seems very good. Probably the best I have ever seen. The overall regulation seems to have gone a fraction worse at 500mV and slighty worse than the SR98.

 

OTHER ASP REVISIONS

(note some of these images are shown after being upgraded)

 

I can't really see much difference other than the REV5&6 have C3 missing and a small diode on the bottom right. C3 & C2 seem to do the same job, just a coupling capacitor to the chassis. The diode seems to be in series with the 12V line. Not sure why, could be protection on early machines or the 12V was to high. In anycase, the diode is hardwired over on the PCB itself in later revisions.

I did noticed on 2 REV6 PSU's that some resistors near the BU508 were different, only by package. Though also Q1 which is normally a small package C1384 transistor is a TO220 packaged C2331. This is the first PSU I have seen, its possible it could have been repaired previously or the manufactures decided to use different parts during the REV6 run. I will keep a eye out for more PSU's to see if this is a common thing done.

I re-tested the REV3 Vs REV6 and did not see any difference. The diode on the REV3 12V is a little lower at 11.52 Vs 11.55V on the REV6, but that could be PSU tolerances. The 5V regulation doesn't seem any different either. The oddball V6 Variation doesn't seem to operate much different. Its hard to say, but the regulation may be a fraction worse but again it could just be PSU tolerances.

 
TOKIN 4501E1

Main transistor is 2SC3460 800V/6A . I had to look up the specs on that one as I had not seen it before.

Overall its a good device, good voltage and current rating. Not as good as the 508 but comes in a close second place.

Capacitors are
400V 100uF x 1
10V 2200uF x 2
25V 470 x 2
10V 470uF x 1
25V 220uf x1

 

What I noticed about this PSU is that when the floppy drive was first accessed it caused screen to dim by about 50% as the drive motor powers up. I had seen this before though assumed it was due to the capacitors not delivering the peak power needed. So I added a good ESR capacitor 4,700uF onto the 5V rail and it did not improve matters. So while this PSU appears to have reasonable good regulation, at around 300mV! its peak power capability seems very poor though.

So far this PSU is probably the worst PSU I have come across so far. It would be possible to update the PSU, though as the capacitors are so squashed up with no margin for error, then I will probably not do any fixes for this PSU. I have not seen many of this type so I do not think it is common anyway.

FALCON PSU

Thanks to qq1975b for submitting the Falcon PSU image.

Main switching transistor 2SK1507 - N-chan mosfet 600V 9A. One thing which is a little odd about that mosfet is it has 0.850R resistance which is about as high as I have ever seen. While these PSU's are 20 years old , I would have expected something a little lower. Though there is nothing wrong with that in itself as it is mostly heatloss on that figure. The really odd thing thing is the turn on time which is around 100uS , I frown on 100ns parts nevermind 100uS. So not really sure why the turn on times are incredibly slow. Without knowing the ins and outs of the design there is no way to tell. The transformer itself says input 50/60hz, with all the delays of the mosfet, it probably makes around 3,000Hz switching cycle. So possible the slow mosfet is done deliberately to better match the transformer. I can't help but feel this shouldn't be how switchmodes should be operating.

The other heatsink as a dual diode attached. As this gets warm also I had a quick look at its specs and found it has a pretty low voltage drop of 0.55V. Looking at modern parts, this is pretty hard to beat!

The rectifier, well, I have not known any of these to fail, but it is only a 1amp bridge which seems a little odd as its got a 2.5A fuse on board. Saying that, its peak current is 30amps so probably more than good enough. I've never personally been a fan of 1Amp rectifiers on mains, so I would have liked to have seen at least a 2amp part in there. But of course that is my own personal preference.

So lets get the scope out and see whats going on :) Tests were done on a stock machine with only the floppy drive connected.

 

The overall regulation on initial glance looks good. On closer look, it has about 100mV of overall regulation which is pretty good. Though I did manage to capture some 600mV (0.6V) spikes which could be higher or lower than the overall regulated voltage. This gives about 1.2V of spiky pulses. Pretty bad, but we must think this PSU has been used for years and has aged. There does not seem any apparent voltage drift even when the floppy drive turns on. So regulation and peak power is very good.

So time to upgrade and see how things turn out afterwards..

 

What can I say other than simple "wow". Those voltage spikes have totally vanished and regulation is a amazing 20mV! Clearly the aged capacitors were not delivering peak power which they now do thanks to Panasonic's ultra low ESR capacitor range.

In conclusion, while this PSU has a bit of a WTF on parts used, clearly whatever the manufacture did works, and works very well indeed. For a simple switchmode design, this PSU really does punch out some seriously good regulation.

FURTHER TESTINGS

I did some more experiments trying to work out the max output current of the PSU.

So checked the Falcon's current pull from the PSU..

5V @ 1.81A
12V @ 0.13A

Machine is a stock machine. Floppy drive and 14MB ST RAM only.

So its not really far off a stock STFM really. Depending what voltage the PSU is on the ST's 5V rail (IE 4.8 - 5V) its about 1.8A - 2A.

The Falcon transformer itself has 5V 7A, 12V 0.2A stamped on the side. So in theory it could push that output current. Though that is assuming 7amps is RMS current.

12v x 0.2 = 2.4watts
5 x 7 = 35watts..
230v x 0.45 = 103watts.

So really, round up the wattages, to say 40watts total output with 100watts input.. thats 60watts loss across the transformer going by those figures. I start to frown at 2watts heatloss.. Though I'm not surprised at that, basically 40% efficient which is probably about right.

Overall, the PSU is probably capable of 7amps on the 5V rail, but I doubt it could sustain that without overheating badly. My next tests is get my rheostat resistor and load it at 2amps and see how hot the switcher gets temp wise. Regardless of what the PSU is "capable" of. I would say 50-60C is as much as I would be happy at it running at. In hotter parts of the world, its probably 50C here, is 80C+ in hotter areas of the world.

Ambient temp was 18degC

2 amps load for 30mins 48.5C
4amps load for 30mins 66.2C

So almost 18C from 2amps to 4 amps. I would guess 6amps would push around 100C.

So I think the PSU could actually push 7amps without to much trouble, BUT, its not possible as the thing would simply run to hot. If it had a fan on top yes.. probably..

One thing which makes the tests not fair exactly, is that feking metal PSU cover. Over time heat builds up and could probably push the temperatures up by 10-20C easily.

Also it was a cold day here, 18deg ambient. so for 2amps, its like 48.5-18 = 30.5deg temp rise at 2amps. If it was 30C ambient, then it would make 60.6C temps. Add that under the metal cover and you probably would hit 70-80C. IMHO I'm not a fan of stuff running hotter than 50C. Though most semis these days are run at 80-120C..

So conclusions ? eye of the beholder I guess... I would suggest if anyone is running more than a stock machine to remove the metal PSU cover, it will keep temps a lot lower. Overall, I would say 2-3amps is about which is realistically usable on the PSU.

FALCON PSU UPGADE KIT

Parts I supply in my upgrade kit as sold in my STORE are as follows.

3x 1,000uf replaced with 3x 1500uF
1x 33uF 400V replaced with 47uF 400V
1x 330uf replaced with 680uF
2x 47uF & 1x 22uF replaced with 47uF (ripple filter capacitors)
1Amp rectifier replaced with 2Amp

I only supply quality low ESR Panasonic capacitors in my upgrade kits. Please do not source them from dubious sources around the Internet as there are to many fakes about. Only use reputable supplies such as Farnell. Do not use generic or low cost capacitors and your PSU will likely become damaged or causes damage to you precious Falcon!

I mostly supply SIP type rectifiers but this turned out to be a small nightmare as there isn't physically room to fit it properly so I opted for a smaller round type. unfortunately the Falcon PSU rectifier has a nonstandard pinout so fitting a modern one is a little tricky.

Please do not attempt this upgrade unless you are reasonably competent at soldering and understand basic electronics and also understand the risks of working on mains voltages equipment.

The capacitors are pretty much self-explanatory in fitting, only the rectifier is a little more tricky as the pinout is different on standard rectifiers..

 

TAKE GREAT CARE

Make sure you get the + and - markings on the rectifier match the markings on the PCB.

Not the rightmost pin on the rectifier as down on image bends under and over to the bottom left hole on the PCB. Take great care none of these legs are touching ANYTHING else otherwise the PSU will likely explode!

TRIPLE CHECK YOU HAVE THE RECTIFIER INSTALLED CORRECTLY BEFORE POWER ON

 

VOLTAGE ADJUSTMENT

My PSU always had low voltage and there isn't any easy way to alter the voltage in any meaningful way. so I designed this little add-on board which will allow you to more significantly change the voltage higher or lower by about half a volt.

Two resistors below get removed ( R18,R19) and the patch board built up and fitted as illustrated below.

 

 

 

LITON POWER PE-2350-2

A new PSU to pass my workshop today :) Not seen one of these before so a rare treat :)

This one has the C4233 switching transistor. A quick look shows its a 3A NPN 60W. So Almost the same as the 2SC2979 (3A 40W) as used on other PSU's.

I powered up a stock STFM like in my other tests and took a scope image, and I must say I am impressed already.

 

Considering the age of the capacitors and the ripple being only 200mV that is probably the best I have seen on a Atari PSU! Another fantastic thing is the floppy drive "dim" on the white background I can't notice a thing there. So the pulse power of this PSU is awesome.

The main output capacitors are 2x 220uF 10V and 2x 220uF 25V. Awesome again as they are even 105deg caps! I swapped them for my favorite FR series 4700uF & 680uF and re-tested. Still the same regulation! Not surprised really as it was 200mV to start with.

Overall, this is one awesome PSU. Even the solder joints still look good. I seriously doubt any PSU would better this, other than the PSU I designed that is ;)

PHIHONG PSM 2512-2 REV A4

Main switcher 2SC3150 Similar to the SR98's 2979 only slightly higher rated.

Peek current on floppy access isn't to shabby at all. The white dim is barely noticable. So good marks there.

Main capacitors are
400V 47uF - Replaced 400V 47uF
2x 2200uF 10V - Replaced 16V 4700uF
1x 1000uF 25V - Replaced 25V 2200uF
1x 460uF 25V - Replaced 680uF 25V

So onto the regulation..

 

Pretty much as other PSU's. Around 480mV to 680mV.

So after the recap..

 

Now we have about 160mV of noise which is pretty good and actually beats the DSP508 in regulation. The dimming is next to nothing as overall its a pretty good PSU once updated.

MSTE PSU PHIHONG PSM-5341

(Thanks to Rui for loaning me this PSU)

12V pulls 0.11A (excluding hard drive) 5V pulls 2.90A when connected to a MSTE.


CAPACITORS
400V 120uF - Replacement 150uF 400V
400V 10uF - Replacement 10uF 400V
16V 1,000uF x 2 - Replacement 16V 1500uF
10V 2,200uF - Replacement 16V 4700uF
16V 4700uF - Replacement 16V 4700uF
10V 680uF x2 - Replacement 16V 1500uF
10V 100uF - Replacement 10V 100uF
22uF 25V - Replacement 10uF 50V (not a typo!) (2019 NOTE - cap is now supplied 22uF C23)
3.3uF 50V - Replacement 10uF 50V
4.7uF 50V - Replacement 10uF 50V
25V 100uF - Replacement 220uF 50V (needs to be mounted higher due to resistor in way)
25V 220uF - Replacement 220uF 50V

MAIN SWITCHER SSP6N60 600V 6A 1.8R N-MOSFET

SCHEMATIC

 

Above is the 5V rail. We are on x10 so noise is 176mV (0.127V) P-P. Its not to shabby considering the age of the PSU and the fact its under almost a 3amp load..

 

12V is also powering a hard drive (not sure on current consumpion, maybe 500mA or so) and the results are somewhat worse than the 5V rail. We have 380mV of noise P-P (0.38V) So this is a little worrying that! So its time for the PSU to be serviced and see what happens...

 

The 5V rail showed a fair improvement from 176mV down to 116mV noise. Looks like the ripple filter struggles overall, dispute the values of the capacitors being increased. It is overall a lot less "spikey" so its a good as things are likely to get with the 5V rail.

 

The 12V still looks spikey, though likely due to the hard drive motor causing this noise. Here we dropped from 384mV to 328mV. A fair drop down but would have been nice if the PSU showed better results.

Overall, the MSTE PSU I had for testing had fairly good regulation before the recap. Though the usage of the PSU was unknown. I suspect the machine had not been used all that much as everything inside the MSTE looked almost like new.

Clearly there is a improvement with the re-cap kits and likely machines which more use will show better improvement. As the MSTE is a rarer machine I think its worth while re-capping with quality capacitors to ensure the motherboard gets a clean as possible power rails.

 

MSTE PSU PHIHONG PSM-8012-2

SCHEMATIC

*on to do list* to investigate.

 
 
 
 
 
 
 
MITSUMI SR151 240V

(Thanks to Andrej for loaning me this PSU)

This PSU I assume is a later version of the SR98. It has a SR98 "feel" about it aswell. Most noticeably is the horizontal transformer and mains filter. A pleasant surprise that the main switcher is a 2SC3459 800V 4.5A 90watts. So a bit more "meat" in it over the previous SR98 transistors.

 

Regulation seems pretty good at 80mV. Though there was some larger spikes which I was unable to capture, I would guess around 200mV. I didn't capture the 12V but it was similar. I did notice the 12V line increased in voltage a little when the floppy drive was active, but the 5V did not. So this PSU has some minor regulation issues. But nothing to worry about.

I also noticed a lot of "dim" on the white when accessing the floppy drive. This is a shame as it was also similar on the SR98. I would have hoped with a larger transistor that it would have been solved, but seems not.

So time for a re-cap...

CAPAICTORS:
100uF 400V - Replacement 150uF 400V
3x 3300uF 10V - Replacement 3900uF 10V
1x 2200uF 16V - Replacement 4700uF 16V
1x 1000uF 16V. - Replacement 1500uF 16V

The rectifier needed the PCB holes drilling (1.2mm should be about right)

 

I did not take a image as the overall regulation was the same as before. Though the large spikes which I had seen before have now completely gone.

The PSU seems to have good regulation overall. Whats worrying is the huge spikes (look at the Falcon PSU for a example) though due to the age of the capacitors its not surprising. Even though the regulation did not change after the re-cap, I still recommend a re-cap due to the spiking issues.

 

BAD SOLDER JOINT EXAMPLE

A classic example of a bad solder joint on a PSU. One of the main failures of PSU's is bad joints and should be resoldered even if they look good.

... and if you can't spot the bad solder joint, then you should probably not be working on it...

Click the image to reveal the bad solder joints..

 

 

HOME