A600 Capacitor Replacement

[AmmoJammo]

Originally Posted by kipper2k

ok, heres where i am at. I removed all the caps except for the audio caps. All caps except for 3 were in the 5% tolerance limit

C306 = 9.2uF (should be 10uF)
C612 = 8.0uF (should be 10uF)
C409 = 85.8uF (should be 100uF)

electrolytic capacitors are 20% tolerance

having said that, even if a cap measures within range, doesn't mean its working properly

[alenppc]

Thanks kipper as always for starting this thread and for the excellent part list. I have recently bought an SMD rework station and will be practicing with old PC crap before trying to recap my Amigas.

I have gone through the digikey part list you mentioned, and some components are no longer available but there are suitable RoHS alternatives, I assume there will be no difference in using them other than they are a lot more expensive than non-RoHS counterparts.

Specifically, PCE3009CT-ND is now replaced by PCE4197CT-ND, and PCE3002CT-ND is now replaced by PCE4155CT-ND.

I also found this youtube video specifically dealing with A600 caps: [ Show youtube player ]

It may as well have been posted here somewhere, but I think it's useful in this thread.

[jimbob]

Originally Posted by Stedy

Two questions to ask yourselves

1) Does the Amiga audio output go negative?
2) Do you really need 35V parts for a 1V audio signal?

I can answer question #2, no you do not need 35V parts. Any 6.3V or above part should do as line level audio is only 1V. You are primarily concerned with the voltage across the capacitor, 1V + any DC offset which will be small, allowing for s short circuit to ground.

To answer question #1 it would be necessary to measure the outputs of the LF357/TL084 op-amps with an oscilloscope. I suspect the signal will be +/- 0.5V around ground.

Ian

A while ago I know but I have often read and myself repeated that the output coupling caps should not be polarised. But I don't think it is true. Please correct me if I'm wrong but before you do make the measurement Stedy suggests yourself.

On A600 at least, there is a 2.5V DC bias on the output pins 1 and 7 of U15. I measured the audio signal at max 3v peak to peak or +\- 1.5V. Therefore audio output, (- side of cap), cannot be positive relative to op amp output, (+ side of cap).

Anyway, there are good enough other reasons not to use electrolytics.

[Loedown]

The caps that need to be changed from polarized to NP are as follows.

C 324 & C 334

That applies for ALL Amiga models except CD32


A4000 needs two additional caps replaced with NP, C 317 & C 318. The other 22uF 16V is on the -5 rail and doesn't suffer polarity issues. People do have all sorts of ideas relating to caps but from experience in repairing A4000 main boards it's the 22uF caps that leak the most and cause the most grief on the boards in the vast majority of cases if not all. The other one that usually gives the most problem is the 10uF in the video section ( A4000 ).

I will always err on the side of caution because if you're going to fix your hardware why do a half arsed job and need to do it again down the track prematurely?

[jimbob]

I know the 4000 has a different audio output circuit and I'm not familiar with it so can't say if there is a problem with it.

But the reason given for changing c324 and c334 to non polarised on other models is that the audio signal swings positive and negative due to the bipolar supply of the op amp. This isn't the case, the op amp outputs are not 0V +/- 1.5v, they are 2.5v +/-1.5V. Even if the +side of the caps was centred around zero, the plus side doesn't sit at zero while the the output goes to -1.5V. The - side follows the signal on the +side and could only fall below it by a tiny amount caused by the esr and this is prevented by the positive bias.

I've only done about 10 A600 boards but the worst corrosion I see is around the power inlet where there are a lot of caps not at the audio output where they are supposedly being abused. Apart from failing to boot, dodgy audio would be the most obvious symptom of failing caps but just because its the thing you notice first, doesn't mean those caps are actually failing first.

I do agree it's better to use superior caps anyway but I just don't think there is this problem with the design.

[Loedown]

Originally Posted by jimbob

I know the 4000 has a different audio output circuit and I'm not familiar with it so can't say if there is a problem with it.

But the reason given for changing c324 and c334 to non polarised on other models is that the audio signal swings positive and negative due to the bipolar supply of the op amp. This isn't the case, the op amp outputs are not 0V +/- 1.5v, they are 2.5v +/-1.5V. Even if the +side of the caps was centred around zero, the plus side doesn't sit at zero while the the output goes to -1.5V. The - side follows the signal on the +side and could only fall below it by a tiny amount caused by the esr and this is prevented by the positive bias.

I've only done about 10 A600 boards but the worst corrosion I see is around the power inlet where there are a lot of caps not at the audio output where they are supposedly being abused. Apart from failing to boot, dodgy audio would be the most obvious symptom of failing caps but just because its the thing you notice first, doesn't mean those caps are actually failing first.

I do agree it's better to use superior caps anyway but I just don't think there is this problem with the design.

Agreed the worst corrosion on A600 boards is around the video caps but as it uses virtually the same audio design as all the others it suffers in the same way. I agree that the audio caps are usually the first symptom people notice but by that stage many of the others have given up their guts and the audio stage can have quite subtle faults prior.

[demolition]

Originally Posted by jimbob

But the reason given for changing c324 and c334 to non polarised on other models is that the audio signal swings positive and negative due to the bipolar supply of the op amp. This isn't the case, the op amp outputs are not 0V +/- 1.5v, they are 2.5v +/-1.5V. Even if the +side of the caps was centred around zero, the plus side doesn't sit at zero while the the output goes to -1.5V. The - side follows the signal on the +side and could only fall below it by a tiny amount caused by the esr and this is prevented by the positive bias.

True, during normal operation those caps will not be reverse polarised at any time, so that is not a reason to switch them to non polarised. The main reason to change them is that it is possible to fit ceramics instead, which will last longer and not start leaking.

[Loedown]

Originally Posted by demolition

True, during normal operation those caps will not be reverse polarised at any time, so that is not a reason to switch them to non polarised. The main reason to change them is that it is possible to fit ceramics instead, which will last longer and not start leaking.

Ceramics still present problems, more information to follow soon.

[GadgetUK]

Regards the audio caps (22uF bipolar), i've just ordered some in the UK from this seller, he's still got more left and they are in packs of 5 so I should have 1 left after doing my 1200 and 500.

They look reasonably sized, have the correct orientation (radial).

http://www.ebay.co.uk/itm/22UF-35V-B...item1c2c3184fa

[demolition]

They look quite heavy which means that could likely rip the pads off the PCB if you're not careful. To avoid that, they should be on the side glued to the board, but I'm not sure there's room for that, at least in the 1200.

[Loedown]

Originally Posted by demolition

They look quite heavy which means that could likely rip the pads off the PCB if you're not careful. To avoid that, they should be on the side glued to the board, but I'm not sure there's room for that, at least in the 1200.

They don't need hot glue to hold them in place, you simply fold the capacitor to sit on top of some of the surface mount resistors / capacitors. C 334 folds to the right if you are looking at the board from the front and C 324 to the left. This stops the parts from fouling the A1200 keyboard connector, it's a tight fit but it works. The two reasons not to use hot glue are;

1. If you need to fault find when you break the seal of the hot glue you can rip the SMT pad off the board.

2. The computer isn't being moved around much and the parts won't be flopping around on the board, they are well held in place by the SMT pads. Any flex in the parts from heat or vibration is taken up by the way the component is bent.

[Stedy]

Hi,

Attached is a snapshot of the Amiga audio signal on my A600 either side of the AC coupling capacitor. According to the measurements on my oscilloscope there was a 150mV drop across the capacitor.

As with any AC coupled signal, without bias at the receiver, it goes negative, nothing to get concerned about.

The two biggest killers of electrolytic capacitors are too much heat (the 1000 hour rating is at 85/105C) and lack of use. If you use the equipment and it does not get too hot, you will not have problems. Your mileage may vary.

Ian

[jimbob]

Not certain if I'm reading your trace correctly. Is the green arrow on the left the 0v level for both traces. If so that's pretty much exactly what I measured 2.5Vdc, 3Vp-p signal. I'm not sure about the 150mV drop you describe, is the green trace smaller amplitude than the yellow? It isn't easy to see.

[Stedy]

Hi Jimbob,

Yes both 0V references were aligned (the green arrow). This shows the DC offset internally of the audio signal.

When I used the PC software to grab the oscilloscope screen, it did not capture the measurements I had on screen (Vpp for each channel), that is where the 150mV difference was observed.

Ian

[alenppc]

Hi, I need some troubleshooting ideas.

I replaced the capacitors using the listed digikey parts on an A1200 and everything works well. Later, I used the same parts to replace capacitors on an A600 which was working, but which also had quite a major capacitor leakage (a small pool had gathered underneath the TV modulator, which I later removed).

After replacing the capacitors, it now refuses to boot.

When powered on, it displays a dark grey (black?) screen on RGB, and no signal on composite.
I have double checked and triple checked all the connections, everything is OK. The new capacitors are installed correctly, with correct polarity and correct values, making electrical contact with their relative traces.

The keyboard LED stays lit permanently, regardless of what kickstart chip is installed (I tried using different ones).

Thinking that it might have something to do with capacitor leakage, I decided to replace capacitors on another (working) A600 board, paying extreme attention to everything I was doing and to not causing any damage. I can safely say that I removed all the capacitors without damaging the PCB. I carefully replaced them using the same stock as described above.
Lo and behold - the same symptoms as described above, occurred again, for this 2nd A600 board.

I am at a loss as to what might be causing this. Again, I rechecked all the connections, everything is there and working. Checked at least 10 times all the new capacitor polarities and values: everything is fine. Measured the tiny SMD resistors with a multimeter to ensure they didn't get damaged by hot air during desoldering. They all show correct values. Tried applying small amounts of tin to the custom chip pins, in case some of them got detached from the motherboard during desoldering, still no change.

I don't have a capacitance meter, so I can't meausre the tiny SMD caps. This is the last idea I have as to what might have gotten damaged by heat, but as I read, these things are supposed to be quite resistant, so that's unlikely.

I am not sure what else I can do. I am thinking that the A600 might simply be "incompatible" with digikey capacitors, as weird as that might sound. I am thinking of trying to put back the original ones and see if that makes a difference (sigh).

I am grateful for any help...

[john1979]

Can't help you with the problem but one tip I saw is:

When you replace the caps try to boot the machine after replacing each one. Then when it fails to boot you know where the problem is.

[Loedown]

Originally Posted by alenppc

Hi, I need some troubleshooting ideas.

I replaced the capacitors using the listed digikey parts on an A1200 and everything works well. Later, I used the same parts to replace capacitors on an A600 which was working, but which also had quite a major capacitor leakage (a small pool had gathered underneath the TV modulator, which I later removed).

After replacing the capacitors, it now refuses to boot.

When powered on, it displays a dark grey (black?) screen on RGB, and no signal on composite.
I have double checked and triple checked all the connections, everything is OK. The new capacitors are installed correctly, with correct polarity and correct values, making electrical contact with their relative traces.

The keyboard LED stays lit permanently, regardless of what kickstart chip is installed (I tried using different ones).

Thinking that it might have something to do with capacitor leakage, I decided to replace capacitors on another (working) A600 board, paying extreme attention to everything I was doing and to not causing any damage. I can safely say that I removed all the capacitors without damaging the PCB. I carefully replaced them using the same stock as described above.
Lo and behold - the same symptoms as described above, occurred again, for this 2nd A600 board.

I am at a loss as to what might be causing this. Again, I rechecked all the connections, everything is there and working. Checked at least 10 times all the new capacitor polarities and values: everything is fine. Measured the tiny SMD resistors with a multimeter to ensure they didn't get damaged by hot air during desoldering. They all show correct values. Tried applying small amounts of tin to the custom chip pins, in case some of them got detached from the motherboard during desoldering, still no change.

I don't have a capacitance meter, so I can't meausre the tiny SMD caps. This is the last idea I have as to what might have gotten damaged by heat, but as I read, these things are supposed to be quite resistant, so that's unlikely.

I am not sure what else I can do. I am thinking that the A600 might simply be "incompatible" with digikey capacitors, as weird as that might sound. I am thinking of trying to put back the original ones and see if that makes a difference (sigh).

I am grateful for any help...

Don't put the old caps back it will make no difference. You will not have incompatibilities with Digikey caps either. If you have a working A600 board of the same revision then take resistive measurements between +5 / Ground, -5 / Ground, +12 and -12 to Ground, if any of those read really low then you have a short somewhere. When I recap any boards I always perform this check before turning it back on in case I have made a horrible mistake.

Stop assuming complex faults, the boards worked before you started repairing them so chances are it's something simple. Vias are the most likely point of breakage as you have heated the board using a hot air gun I assume and they are the most delicate and also most likely to have been damaged by the capacitor leakage.

Keyboard led staying permanently lit sounds like a reset fault. I found on a board this exact fault, long switch on time. R611 goes open circuit or desolders. Look around the NE555 ( U14 ) circuitry. Ground pin 3 temporarily and see if keyboard light goes off

R514 could be open circuit or no +5, also R511D the same, Q511 faulty or broken wire. C612 open circuit to pin 6 of U14 which is most likely. Look for change of state on Q511 pin 3, ground initially then +5V after some seconds for boot ( reset signal )

[alenppc]

Thanks for the help Loedown. So far, I have been utterly unsuccessful in my efforts to revive the 2 boards.

I have recapped a 3rd A600 board, also rev 1.5. This one did NOT present the problems of the first 2, it works fine after recapping. Although for this one I have switched to 35V 10 uF caps (which is what Commodore had originally used) instead of 25V. I have tried changing the 25V caps to 35V on one of the dead boards but this did not help.

Still talking about the dead board, I have now socketed the CIAs and the keyboard MPU. I have desoldered the keyboard connector and changed the audio caps behind it (could not reach them otherwise). I have then proceeded to replacing several NPN 3904 transistors, in case they were damaged by heat from the heat gun. I also have spare 3906s but did not change any yet.

None of this helped. I have made in-circuit transistor measurements using the diode setting on the multimeter and have observed the exact same values as on the working board.

I have measured nearly all of the diodes and resistors and observed the same values as on the working board.

The only difference between the working and non-working board that I have observed is a different resistance measure around the video caps, and I cannot explain that, because all of the point-to-point connections appear to be correct and working, but there could be something dodgy going on. Does the A600 motherboard have a middle inaccessible layer?

Now I am waiting for an ESR meter to arrive to check the in-circuit caps. Failing that, the only other assumption that I can make is that some IC failed, but I can't find out which one without an oscilloscope.

[Loedown]

Have you even looked at the reset circuit which is most likely the cause of all your problems? You could check it using a multimeter and Google NE555 circuits. An ESR meter isn't going to tell you anything other than the capacitors will be ok. You could always shoot a PM off to Kipper2K who is in your neck of the woods for a possible price of repair.

[alenppc]

Hi, I have checked and even replaced the Ne555 with a new one. Anyway... I got one of the boards to work, and it seems all it needed was for it to be washed!

I guess there was residual solder flux left which must have been conductive. I had cleaned it with alcohol, but clearly not well enough. So I soaked the board in the sink with some dish-washing detergent and rinsed it under the tap. It was drying for the last few days. Today I powered it up and it works! Never thought that would be the cause. Now, I have washed the second one and I am waiting for it to dry. Hopefully that cures that as well.

Crazy stuff.

Thanks for all the help!!