 So here's a thing. This, if I can try and get this without the wires going over the screen, is a 470k resistor. It's quite near one of the capacitors that I replaced. Which measures, as you can see, that's showing kilons at about 150 to 200. That is R336 and according to the circuit diagram, that should be 470k. Now it could be that there's other circuitry on the board upsetting it. But I am suspicious, so let's smear some flux on and reflow all these, because that was near that capacitor that I think may have been leaking. Because remember that we saw marks in the... But that's a lot of flux. That looks a bit more reasonable. Because we saw marks in the tracks near where it was. It's possible that it's damaged that capacitor. I found another couple of 470s nearby and they seem to all be fine. So I'm just going to melt the solder in each one of these components. Do you see they all look kind of crusty? I said melt. I changed the tips to a needle-nosed one. Interesting. Don't like the look of that. That's not actually melting. So what do we get now for this one? That was the good one. 470. That's what we expect to see. This is the suspect one. It's better. It's still quite low. Okay, I am actually going to try and take this off. This may be a terrible idea. Yeah, that resistor is fine. It may have been a problem soldering it on, a problem of the joints. But now we have to get it back on again. I found a thing. So if you look at the schematic, there's a 3900 picofarad capacitor going from signal to ground on each channel. These are both identical. However, if I actually measure the resistance, let's turn the meter back on, this one here, C323, is the left channel. You don't normally measure the resistance of capacitors and that shows up at about 250 kilo ohms. But if I do the same thing with C333 for the right channel, and this is the faulty one, it settles down to about 3 kilo ohms. So this is different for the two channels. Now, I found this because I found somebody online who had a very similar problem to mine, and it turned out to be that this capacitor here had died, probably due to damage from this capacitor here leaking. So I think it's quite plausible that something similar has happened to me. So what I'm going to do is take this off and try and measure it off the board. I have also replaced the soldering iron tip with a different new one, and this one is now a firm fit. So hopefully this will work this time. Yeah, that's melted. That's so much quicker. So we want to heat the whole thing up. Well, that came off and then it went somewhere. I think it landed on the PCB. These are the two pads. They don't look great to be honest. Okay, so the first thing we're going to do is measure the resistance across the capacitor, which I have found, and that is registering zero, which is pretty much what I would expect, which is interesting. What about the resistance across these two pads? Don't want to lose that capacitor as I don't have a replacement. 100 ohms-ish. Interesting. Did I measure the resistance correctly? Because things are not consistent. Okay, well, that's odd. However, there is one focus. That's better. There is something else we can try, which is if I get the schematic out, we should be able to track one end of this to pin 5 on the op amp. 2, 3, 4, 5, this one. And we get zero ohms. Let me try that again, but on camera. Which means that one thing I thought might be a problem is damage caused by leaking capacitor to the tracks, but those tracks appear to be fine. I also need to test the resistance to audio ground. I happen to know that the other end of 3, 2, 3 over here is also connected to audio ground, so this should be zero. That's not zero. I've got the right end. Okay, that would seem to be the right end. That is zero. Therefore, this end of 3, 2, 3, which is the other capacitor, should be connected to pin 3 of the op amp. Which is here. Yes, that's a zero. Okay, so the tracks to and from this seem to be okay, which is pointing at something to do with the cap, which is peculiar. Well, I built this shoddy test framework by putting the cap on to a piece of strip board so I can attach some wires to it and when I fire up my component tester, it thinks it's a capacitor of about the right value. So that looks fine. That still doesn't explain why we were seeing odd things here. Possibly these joints were more crusty than I thought. Well, the thing to do now is to fire the Amiga up without the capacitor on it. This means that we won't have a capacitor to ground in the filter. It should be okay. We should be able to tell from the output voltage whether anything weird is happening. It may not work right because the cap is missing, but it's worth a try and that's a big no. So it doesn't appear to be the capacitors fault, but we have found definite signs of something odd going on. So the question now is do we put the capacitor back onto the board or do we investigate further? Well, looking at the circuit diagram, I was testing the resistance across the capacitors. So one end is attached to ground. The other end is connected to ground through a resistor here, R336, to another capacitor to ground. On this side, I'm sorry, there's a MOSFET in the way. So I think that's a between those two pins, I believe that acts as a diode. Yeah, it's very difficult to tell because there's a lot of other circuitry. There is a route to ground here, which goes through this resistor through another capacitor, through another pair of resistors, and then we're at ground and the same applies on this side. So I'm not sure that's particularly relevant. So I removed both caps and well, across these two pads, I see about 15 to 20 ohms, it seems to vary a little. Across these two pads, this is the right channel one that's broken. This is the left channel one that works. We see 235k. So that is definitely pointing at a dead component somewhere on the right side. That is shorting pin 5 to ground somewhere. So if you find a random ground and go to pin 5, we do in CEC about the same resistance. So I now need to find out where this might be. This probably sounds like a bridged connection, possibly. All the components that I've been able to test, these various passives, the capacitors, etc, have been absolutely fine. So the remaining options are damage to the board, possibly caused by this capacitor leaking in the past, that rather this capacitor's predecessor leaking, or damage to the MOSFET itself. Trying to diagnose this stuff in situ is going to be hard because all the components are connected together and they all interact. So like just measuring the capacitance of this capacitor, you'll get a meaningless value because it's connected to various other capacitors, etc. So what I'm going to do is the drastic measure. I'm going to take the MOSFET off the board, plus all the surface mount components that connect to this little network of tracks here. This should isolate this piece of the PCB tracks. That will let me test those in isolation. I can also look at the op-amp disconnected from the board and see whether it looks all right or not. I really don't want to take this thing off the board, but I think I'm going to have to, and that means I'm going to need to get the hot air gun out. So this is going to be work. I have attached alfoil and captan tape to protect the other components. I've got the hot air gun out. So we fire it up, let it heat up, and let's get started. You can see I've also glopped flux over the component to help make the old crusty solder melt. So now let's take a roll to the chip and we start testing it. The flux melts and we just keep playing the hot air over the chip. That was remarkably painless. I was expecting that to take way longer. Oh well. Okay, well immediately we get some results. If I do the resistance test, let's get that wire out of the way, again over this capacitor for the right-hand channel, I get 50 odd ohms. If I try it for the left channel, the one that works, mega ohms to 250k. So that suggests that the problem is not due to the off-amp. So let us start removing some more components. I've got the hot air gun out. I might as well use it. But I'm going to take off, looking at the circuit diagram. I'm going to take off that cap. Very easy. I should have actually used this before, but it's a bit loud. So I have taken off the MOSFET again and C3 C3 again. This is now completely isolating the section of circuit that runs from the top end of C3 C3 to the MOSFET, I think that pin. It also goes to pin five of the off-amp. So I should be able to test this and you see you get a resistance of zero indicating a actual circuit. This one, not that one, this one, zero. Okay, it's that pin. And also I have removed R333 on the underneath of the board. So these pins are not connected to anything. So why then? If I take the resistance here, are we getting a non-zero resistance? These should not be connected to anything. I believe that this is a reasonable sign, just picking a different ground, a reasonable sign that we have board damage and a track somewhere is shorting through to something else. How to identify where that short is is going to be a slightly trickier problem. Because it could be anywhere. Now chances are it's related to the damage you can see here on this component on these tracks caused by the old leaky C303. So maybe something around here. But I've never actually seen this kind of damage from a capacitor before. Well, this is interesting. I removed R349 here. I was intending to remove several of these so I can see the traces underneath. And it's rather more charcoal-y under there than I was really expecting. Plus if now that the resistor is removed, if I measure the resistance across these two pads, we get a very high number, which is kind of what I was, which is what there should be. So is there some damage here? The resistor itself, by the way, R349, tests out fine. So I'm not quite sure what's going on there. I will clean this up and see what there is to see. Bad news, these three pads are connected together. You can actually see the tracks connecting them. And if I measure the resistance between them, they show up as zero, which is correct. Why is this bad news? Because according to my PCB schematic, which is for the same revision of PCB as the one I've got here, a revision 1.5 model, C333, which is here, and R349, which is here, and R336, which is here, are not connected together. R336 and 349 are through this wire, but C333 and R336 are on either sides of this MOSFET, which I've removed. So I no longer know what's going on. I don't even know whether I've successfully isolated this piece of network, which is going to cause problems. The one thing I do know for sure is that if I measure this resistance, it now shows as a very high number. The meter is actually kind of unhappy. I can't quite tell whether it's connected or not. And that is a change in behavior. Since I removed this and cleaned up the char underneath, it is now behaving better. So I think that what happened is that this thing oozed just a little, just enough to slightly damage these tracks here and these here, and caused gunk to accumulate under this component, which caused partial conductivity between presumably this pad and this track here, or possibly this one down here. I can't tell what these do because I don't have a accurate layout chart for this PCB. So I'm willing to bet that if I were to just put everything back on, the problem would be fixed simply due to having cleaned under this component here. So I put those two components back. What do we get for this resistance? Nothing. That's what we like to see. It's a slight solder blob on top of the top pad, so I can't get a good contact. Okay, so I suspect that was the problem. So I suppose I now need to stick the other components on, which I have carefully taped down here so that I don't lose them. Well, I may have spoken too soon because now that the components are put back on, if I measure the resistance across this one, this is the correct channel, 200-ish k ohms. If I measure the resistance across here for meg ohms, they are still different, but now the resistance is too high rather than too low. It may be that there is leakage somewhere under here and that the resistance is high enough that it doesn't actually matter, or maybe not. I think that I'm going to have to put the op amp back on and these two capacitors and try it. Annoyingly, I actually missed a trick with these. When the Amiga was made, you could only get to 20 microfarad capacitors in electrolytic, which is what these are. These days you can get them in ceramics. So I could have used one of these, or rather the slightly larger version of one of these here and here, but you give me loads more room and generally been easier to work with. Anyway, let's try and put this thing back on again. There's two ways to do this, the right way and the way I know how. The right way is to solder all the legs at once using lots of flux and solder paste and hot air and stuff. The way I know how is to solder on one leg a bit further up. It's to solder on one leg and then go through and very carefully solder all the others on. You need to make sure the chip is properly aligned. So that leg should hold it roughly in place. I can make fine adjustments by heating this up to get the alignment just right. And then I go through and solder all the other legs, trying very hard not to bridge anything. Moment of truth time, let's see if it works. Let's apply power and see if anything explodes. I put that power switch here. Okay, don't know if you saw that, but that was magic smoke escaping. It came from one of these components down here. I'm just going to go back and review the footage and see exactly where it came from. I believe that particular panic turned out to be unnecessary. I examined the footage and I think what's happening is that when I cleaned the board with IPA some got sucked underneath a couple of resistors by capillary action. And then when I turned the power on and the resistors heated up marginally, it evaporated and squirted out, producing that smoky effect. In fact, it turned out there were two resistors that did this. And I didn't touch one of them, so I don't think it's my fault. So let's give this another go. So watch carefully. Right, that was a red dot and a pop. My original hypothesis, I think, is in fact wrong. Glorious. Well, R301 here is now dead. See, the reason why I thought it was IPA is I could actually see splatter coming out and then evaporating. So that resistor is a X resistor. What about this one? This was one of the other smoky ones that is showing a resistance. I think that says 100 for 10 ohms because that is actually what I'm getting here. Well, curses. There's two questions. Firstly, why was it doing that? And secondly, what did I break that made it do that? R301 is that on my sheet? I don't think it is. I figured out what the problem is. I took this off, which was vastly simplified by the fact that it was in fact in several pieces. And this pin here, this is a 10 ohm resistor that goes from plus 12 through the resistor and then into the power supply for the op amp. So I thought maybe I'd managed to bridge plus and minus 12 so that it was shorted across. That would cause a current flow that would fry the resistor, the resistor acting as a fuse. And then I noticed that this dot here is here. This dot should not be here. This dot should be here. I put the op amp on backwards. There is a possibility that it is not in fact broken. I'm going to have to take it off and put it back on the right way round. But oh dear, I suspect it may be fried and I think I rather than go to the hassle of actually taking it off and resoldering it just to see if it doesn't work, I may just wait until a replacement arrives, which is a shame. I also need to replace this, but I can fairly easily do that with a non-surface mount resistor. There is more or less enough room.