A1200 rev 1B notorious for going bad?

[TomCrazy]

Hey all!

I'm currently in the middle of repairing an A1200 rev 1B motherboard. What I first thought would be a quick and easy repair (like e.g. replacing a capacitor and that's it) turned out to be a lot more complicated and time consuming when I started to notice that component after component had failed bad.
I have so far found around 10 ceramic capacitors (yes, ceramic capacitors, not electrolytic ones) that have shorted out, most of them being decoupling caps hooked up between Vcc and GND.

I first started to suspect that, since all bad capacitors that I've found are hooked up between the 5V rail and ground, maybe it was a bad power supply that killed a lot of these components on the motherboard. A PSU with a lot of voltage spikes on the 5V rail, perhaps...?

But now I have found even more ceramic capacitors that are partially shorted and that are not hooked up between the 5V rail and ground. So my theory about someone having used a bad PSU with this motherboard might not hold true anymore.

As many of you already know, A4000 is notorious for having had bad electrolytic capacitors on the motherboard out of the factory.
What I'm wondering here is if anyone have heard or experienced anything similar about the A1200 rev 1B, but with the ceramic capacitors instead of the electrolytics? Did Commodore use a bad batch of ceramic capacitors when manufacturing rev 1B of the A1200?
If so, can these bad capacitors be traced to a certain value of capacitors? Like for example, all of the 220nF capacitors being prone to short out, or something like that?
I've read that on the A4000 motherboard, the factory mounted 22µF electrolytics seem to be the ones that fail the most, so I was thinking that it might be the same with the A1200 rev 1B, but for a certain value of ceramics instead?

I would rather not have to do a complete ceramic recap of the whole motherboard, because there are so many of these caps on the board. That's why I'm wondering if there is any specific value of capacitors on these rev 1B motherboards that are particularly prone to go bad.

I've read a post from someone on an Amiga forum a while back that the rev 1B motherboards are notorious for going bad in general, but it was never mentioned in the post exactly what components usually fail bad on this particular revision of the A1200.

[tesla]

I find it hard to believe that so many ceramic caps went bad, I don't think I've ever seen a bad one that wasn't obviously physically damaged.

Just so we are on the same page, you, (at the very least) de-soldered one lead of the suspected bad ceramic cap, and measured it with your DMM, and you have a direct short? if you didn't check the cap this way, (or completely remove the cap to check it), you may be measuring something else that is shorted in-circuit.

I read several years ago, that some Chinese manufacturers were putting out defective electrolytic caps, I heard Dell was bitten bad, others were also, I believe that this is the main problem that has been discussed here about electrolytic cap problems. Just as a reference, most of us have owned equipment using electrolytic caps that are in some cases are decades old, and still working just fine.

Going back to the ceramic caps, if they are genuinely bad, I suspect faulty manufacturing, perhaps going back to the the same problems with the electrolytics.

BTW, if the ceramic caps are bad, they should show a low resistance reading on the lowest resistance setting on your meter. One, maybe two bad, but more?

If you find you have few or no bad ceramic caps, I suspect something shorting out the voltage rails, since the ceramic caps are normally used to decouple chips, and are installed across the voltage rails.

[demolition]

Ceramic capacitors don't generally short out, so if that is indeed the case for several in one machine, then they must have seen some seriously high voltages on the supply.
Question is if you have tested them after desoldering as tesla writes, and if that is the case, then how were they desoldered? If not done right, they could have been damaged during desoldering, causing you to think that they were always bad. It should be done using two irons or a suitable dual tip.

When fixing up one of my 600s I came across one ceramic cap which had shorted out. It was the decoupling cap for the 2.5V audio ref voltage and the location meant I could test it in-circuit and it was indeed close to 0 ohm, pulling the Aref voltage to 0V. There was no physical indications of damage on it, and it had not been resoldered by anyone (I broke the factory seal on the Amiga to fix it). After replacing it, the correct voltage level was restored, so they can go bad.

[Loedown]

Two things happened with capacitors.

1. A recipe for capacitor electrolyte was stolen but it wasn't a complete recipe so when these Chinese / Asian knock off parts surfaced they had the so called capacitor plague. Check out Capacitor Plague on Wikipedia or read Badcaps forum.

http://en.wikipedia.org/wiki/Capacitor_plague

http://www.badcaps.net/

2. Some capacitor parts were being sold as Nichicons for example and they were found to have an inferior capacitor brand substituted in the shell of the part.

http://makezineblog.files.wordpress....pg?w=500&h=311

[TomCrazy]

Originally Posted by tesla

I find it hard to believe that so many ceramic caps went bad, I don't think I've ever seen a bad one that wasn't obviously physically damaged.

I don't blame you for almost not believing me, but I can ensure you that what I have written in my original post is very much true. If it wasn't, I would have no reason to bring up this topic.
And let me assure you, I know perfectly what I'm doing.

I don't know about your history as a repair tech, but if you have never encountered a bad ceramic cap that hasn't been physically damaged, then I would probably say that it's more of a coincidence that the machines you've worked on so far haven't had any bad ceramic capacitors on them. I don't know if it's only because of that or if it might also have to do with that you haven't done that many repairs, but I'm quite sure that you will find ceramic caps that have gone bad in electronic equipment sooner or later if you are doing repairs on a regular basis. Because like Demolition says, they can go bad. I have this A1200 rev 1B board here to prove it more than enough.
I have also encountered shorted ceramic capacitors on other motherboards as well, but not to such an extent as this A1200 rev 1B. This is a pretty rare and extreme case, I think.

While it's true that ceramic capacitors are more sensitive to physical stress and are more prone to crack than other types of capacitors, they can very much go bad without having any physical damage to them.
In my own experience, the most common failure mode of a ceramic capacitor is a short between the two terminals.
When I say "short" here, I am not talking about a "dead short". I know that the correct definition of a "short circuit" is when two points in a circuit have very low or almost no resistance at all between them in a case where such a low resistance is not desired and/or expected.
However, a capacitor should have an infinite DC resistance between the two terminals. If the DMM gives a steady resistance reading over the terminals even when the capacitor is out of circuit, the capacitor has failed bad.
So just to clear one thing up; When I say "shorted capacitor", I refer to a capacitor that doesn't necessarily have a dead short between the terminals, but has a certain dedicated DC resistance between them. Which is, of course, something it shouldn't have.

The shorted ceramic capacitors I've encountered so far have all shorted out not completely, but partially. They haven't given me a reading close to 0 Ohm, but rather anything between 15 Ohm and 2 kOhm, depending on the capacitor. All measured out of circuit, of course.

Originally Posted by tesla

Just so we are on the same page, you, (at the very least) de-soldered one lead of the suspected bad ceramic cap, and measured it with your DMM, and you have a direct short?

There are other ways to spot a shorted capacitor, but yes, I did of course desolder them to verify that they are in fact shorted. (Or "partially shorted", as I explained earlier.)

Originally Posted by tesla

if you didn't check the cap this way, (or completely remove the cap to check it), you may be measuring something else that is shorted in-circuit.

This is not necessarily true.
In the case when a capacitor is hooked up in parallel with other components, which is most often the case with decoupling capacitors for example, then it is not possible to check its DC resistance without desoldering it first. Same thing goes for other components in parallel, like resistors. And since there are not only a lot of decoupling caps on a motherboard, but also many other components hooked up between the 5V rail and ground, there is no point at all in trying to check the capacitor in circuit.
However, there are cases when a capacitor doesn't have any other components in parallel. One such example would be an AC coupling capacitor, which is coupling an audio signal or video signal or whatever between two different parts of a circuit. If these two parts of the circuit are isolated from each other (apart from the capacitor), it is very much possible to check the DC resistance and even the capacitance of that capacitor while it's still in circuit. You may get some stray capacitance included in your reading which is coming from the parasitic capacitance of the traces of the PCB, but that is so incredibly small in comparison to the capacitor so it might just as well be neglected for troubleshooting purposes.

Originally Posted by tesla

I read several years ago, that some Chinese manufacturers were putting out defective electrolytic caps...

Originally Posted by Loedown

Two things happened with capacitors....

This whole electrolytic scandal that happened in 2002 can make up for some interesting reading for anyone who doesn't know about it yet. However, as interesting as this story is, it has very little relation to the situation I'm dealing with right now, because ceramic multilayer capacitors doesn't contain any electrolyte at all and are therefore not affected by the electrolyte scandal. And even if they did contain electrolyte, they would still not have affected any Amigas since 2002 is long after all the classic Amigas were manufactured.

Originally Posted by tesla

Going back to the ceramic caps, if they are genuinely bad, I suspect faulty manufacturing, perhaps going back to the the same problems with the electrolytics.

They are bad, trust me on this one.
Exactly, faulty manufacturing is what I suspect too.
The reason I started this thread was to ask you guys if you think this is because of faulty manufacturing (or rather, a bad batch of ceramic capacitors being used during manufacturing of A1200 rev 1B boards), or if it is more likely to be because of a bad PSU that has blown the crap out these capacitors.
I will give you some more info about this. Here is what I know so far:

1. I have so far found at least 7 ceramic decoupling capacitors (hooked up between the 5V rail and ground) that have shorted out, all with different resistance readings between the terminals when measured out of circuit.
The lowest was around 30 Ohms and the highest around 1.2 kOhm, if I remember it correctly.

2. I have found two shorted capacitors that were not hooked up between the 5V rail and ground. These two were AC coupling caps.
This indicates that maybe it isn't a PSU that has destroyed all these caps after all, because these AC coupling caps are not even connected to Vcc or GND, they are just coupling some signals in the system.
Both chips that these two AC coupling caps were coupling between are fine. No damage to the chips, but these two caps were partially shorted.

3. No chip on the board seems to be damaged, and the motherboard boots up fine, except for a few minor things that had to do with the bad AC coupling capacitors.
In other words, the ceramic capacitors are the only components that seem to be affected.

4. No electrolytic capacitor seems to have gone bad according to my tests.

So we're dealing with 7 or 8 ceramic decoupling caps and 2 AC coupling caps that have gone bad so far, all on this rev 1B board. I know, It sounds crazy, but I'm not joking here.
Why else would I be writing and asking you guys what kind of experience you have with faulty A1200 rev 1B boards?

And, you know what the funny thing is? I'm pretty sure it doesn't stop here. I still have to try to find the rest of the shorted decoupling capacitors on the motherboard. I know there are more of them, because I still get a lower DC resistance reading between the 5V rail and GND than normal, even with all the shorted caps removed from the board. It's totally crazy!

Originally Posted by tesla

If you find you have few or no bad ceramic caps, I suspect something shorting out the voltage rails, since the ceramic caps are normally used to decouple chips, and are installed across the voltage rails.

The ceramic capacitors are what is shorting out the 5V rail and ground. A steady reading of around 19 Ohms between the 5V rail and ground along with the motherboard drawing almost 3 Amps on the 5V rail during idle (no expansions connected) made me start looking for shorted components. I first thought it was just one component that had shorted out, so imagine my surprise when I found capacitor after capacitor that was shorted.
After removing one of the shorted capacitors, the resistance between the 5V rail and ground went up a bit, but just a little. After removing the next bad cap I found, the resistance went up a bit more. And so on...
After having removed 7 or 8 shorted decoupling caps, I'm now up to around 80 or 90 Ohms (steady reading) between the 5V rail and ground. This is still too low and is not normal. I have never gotten such a reading on a working Amiga board before. There is at least one more shorted component on the board, that's for sure.

Originally Posted by demolition

Ceramic capacitors don't generally short out, so if that is indeed the case for several in one machine, then they must have seen some seriously high voltages on the supply.

Yep, exactly my point. They can go bad, but they generally don't if they are not
1) of extremely bad quality.
2) they are subject to conditions of which they are not specified to handle.

What is the cause here on the board I'm working on? Is it cause nr 1 or 2? That's what I'm trying to figure out.
Why does it matter? Well, because if they have gone bad because of bad quality, then it might be worth considering doing a complete ceramic recap on this motherboard, or at least a recap of the decoupling caps. Because if so many of them have already failed, then how can we know that not a few more will fail after just a few months time? Simply put, if they are of bad quality, I can not trust them at all and I think I better replace them.

But if a bad PSU is the culprit here (which then keeps me wondering why all chips are still working fine), then there is no reason to replace all of the ceramic caps, I think. Only replacing the shorted ones will be just fine.

Originally Posted by demolition

Question is if you have tested them after desoldering as tesla writes, and if that is the case, then how were they desoldered?

Like I wrote earlier in this post, I know perfectly what I'm doing so you don't have to worry about that.
I always use a temperature controlled iron. And I might also add that I have actually never managed to damage a multilayer ceramic capacitor when soldering it, or any other component for that matter as far as I can remember. While it's true that surface mounted multilayer capacitors are quite sensitive devices as far as temperature goes, they will not die at the very instant when you touch them with your iron. They can usually survive correctly performed soldering procedures without any problems at all. Sure, if you go berserk and crank the temp up on your iron like crazy and solder with a dwell time that is beyond any common sense, then you might run into problems.
But when using reasonable soldering techniques... No, you won't kill a ceramic capacitor that easily.

And in order to emphasize it even more that the capacitors were actually bad in the first place, and not damaged by soldering:

1. If there were no problem with the motherboard, and the motherboard was working perfectly fine in the first place, then why would I even have to start working on the motherboard?
I of course started to work on the motherboard because there was a problem with it. I quickly discovered a very low resistance between the 5V rail and ground.

2. Like I mentioned, after removing the shorted capacitors one by one, the resistance between the 5V rail and ground started to go up more and more respectively.

3. There are troubleshooting techniques that can be used to spot a shorted (or partially shorted) capacitor without even touching it with the soldering iron. I won't go too much into it all, but one way is to simply feel the capacitor with your fingertip, or use any kind of temperature measuring device for a more accurate reading. If it is a decoupling capacitor in a 1206 package, it will usually get very hot if it has started to act as a resistor more than a capacitor. If there is 5V on one side and 0V (GND) on the other side (since it's a decoupling cap), there will be a constant current flowing through the capacitor. The amount of current will be determined by how much the capacitor is shorted out. That's Ohm's Law.
The 7 shorted decoupling capacitors I have found so far got very hot. I even burned my fingertip when touching some of them.
Once desoldered, it's possible to measure the DC resistance of the capacitors. By using Ohm's Law again, it's easy to calculate how many watts of heat a certain capacitor was dissipating while it was in circuit. The result of my calculations made me realize that it wasn't strange that some of the capacitors got incredibly hot, especially considered that all this heat was being dissipated from a little 1206 package.

In the case of the shorted decoupling capacitor that Demolition found in his A600, it's a completely different story since the audio Vref is derrived from a voltage divider consisting of two 1 kOhm resistors. A completely shorted ceramic decoupling cap for Vref will mean that one of these 1 kOhm resistors will be completely bypassed, but there is still another 1 kOhm resistor limiting the current going through the capacitor, so this capacitor will not even get warm to the touch.
But it's a bit different with capacitors that are hooked up directly between the 5V rail and ground. Since there is no other resistive element present, the current going through the capacitor will be directly determined by how much the capacitor has shorted out, in an inverse proportional manner.

So in all... Nope, it was certainly not my soldering technique that killed the capacitors.

Just to clarify, I'm not asking for help about how to repair this motherboard. I already know how to repair it.
What I'm asking for is your opinion on whether I should replace only the bad ceramic capacitors I find, or if I should suspect that Commodore used a bad batch of ceramic capacitors when manufacturing this motherboard, and therefore replace all of the decoupling caps with good-branded ones?
What would you guys have done in a situation like this?

Also, if anyone has any experience working with A1200 rev 1B boards in particular, please let me know. Did you find multiple components that were bad when working on the board?
Or rather, have you ever come across anything like this before on any revision of any Amiga motherboard? Which motherboard was it? Do you remember the revision?
(We don't have to take the A4000 into account here I think, because we already know that it's notorious for having bad electrolytic caps mounted at factory.)

Let me know your experiences. Thanks.

[demolition]

That's one long post..

I think I'd just replace the whole bunch. They are not particularly expensive and it takes too long to figure out which ones are bad or not, and you've found several so far. As you also write, not all of them necessarily gets hot so they can be tricky to find, and with the right equipment they are quickly replaced anyway.

[Loedown]

That's a nice scroll bomb

I only included links to electrolytics because they do have some relevance to this thread. I've got an inverter here that wasn't charging the batteries and it had a violently blown fuse and a polyester capacitor with part of the side missing. 67R is what it was reading.

I've not come across many capacitors apart from electrolytics that have gone bad in anything, this inverter is only the second item behind an old tuner with a shorted ceramic and never seen a bad one in an Amiga.

It's easy to say that something has over driven the rail but the reality is that the chips would go first followed by the electrolytics and then the ceramics. If the ceramics were all the same value then I would say it was a manufacturing fault, several bad ones on a reel in manufacture ( SMD ) otherwise if they are all different values then I would only suggest a sudden shock ( physical ) would be the only other reason.

I think it's safe to say that no one revision of Amiga 1200 is worse or better than another. Clearly later boards have later revisions because Commodore fixed up circuitry or made modifications to allow for more functionality.