Recapping a IIci: Question about derating capacitors.

[CptSkyfox]

Sorry if this is the wrong section for a general question. I didn't know where else to put it.

So, this is my first post, and it's about my first shot at repairing anything. I'm working up to doing my first recap and repair, on a IIci. It's... Pretty much the bog standard issue.

I've already cleaned this up a bit, and did some prodding for continuity. Haven't been able to find any breaks so far. So it's not as bad as it looks. Neither is the damage in the audio/serial section. Maybe I'll find a problem later on, but it looks like this is going to be straightforward so far.

That's the context. So on to the question:

I'm building a BOM for the repair, and I ran into an issue... Its an issue I'm sure everyone is running into, but I can't find anyone talking about it.

How are people derating 47u capacitors on a 12v rail? Because I'm struggling to find a 25v that fits.

My original plan was just to buy a bunch of 25v solid caps and put them on all the 47u, just for simplicity's sake... But that went right out of the window when I went looking for them.

I'm looking on Mouser, and as far as I can see... Tantalums in a C package cap out at 20v, Polymer-tantalums in the same package caps out at 10v. You can get 25 volts in a D package easily, but they'd overshoot the pads. Meaning, at best, you have to solder them on the underside, and they'll be floating above the board.

Aluminum Organic Polymer caps seemed ideal for a moment... They come in the same kind of cans as the originals, and are apparently pretty good. ...But apparently not for an audio circuit. The ESR is too low, and they have too high voltage leakage. This is a problem, because +/-12v is on the audio.

I then saw someone talking about hybrid caps. Aluminum Organic Polymers with a small amount of juice in them. Apparently, they have far lower DC leakage. But the ESR is still kinda low, and they can still leak.

So I guess what I'm asking, is... What is everyone using to derate? Are they cramming the D size caps in there, or are they using 20v tantalums? Or did I just miss something?

 

[bibilit]

Hi, not sure to understand, those are 16v capacitors, so your best option is to buy 16 volts caps, if you can find 25 volts one, sure doesn’t hurt, but 16 ones are totally fine.

 

[Byrd]

Can confirm 'D' size tantalum caps fit quite neatly onto the pads if you pre-solder the pads, flux and position the larger cap on top (pre-tinned).

 

[Phipli]

Hi, not sure to understand, those are 16v capacitors, so your best option is to buy 16 volts caps, if you can find 25 volts one, sure doesn’t hurt, but 16 ones are totally fine.

The spec sheets for tantalum capacitors recommend that you over rate the voltage by 100% - i.e. double it.

It is true too - as a community we have seen 16v tantalums explode on a 12v rail, including ones fitted by Apple originally, back before the advice was updated to say double the rating.

Yes it begs the question why call a part "16v" if it is only good for 8v, but I guess it is for continuity.

 

[jmacz]

Can confirm 'D' size tantalum caps fit quite neatly onto the pads if you pre-solder the pads, flux and position the larger cap on top (pre-tinned).

I have used D size for the ones that are on 12V rails. It’s a little tricky but they fit as @Byrd mentioned.

 

[CptSkyfox]

So, with some more research on capacitor types, and you guys confirming you can fit size D on this footprint, I've narrowed things down to two options.

1. 25v everywhere, using Aluminum Organic Polymer Capacitors on most of the board, but Aluminum Hybrid Polymer Capacitors on the audio circuit.

2. 20v c size tantalums on everything but the 12v rails, with 24v size D polymer on the 12v.

I'm leaning towards option 1. Having cans that match the original footprint well is very appealing.

I heard the hybrid caps don't leak. Either voltage *or* cap juice... Though they are kinda low ESR. I suspect they'd be fine on the audio circuit, since I see people using low ESR tantalums for that. I do wonder whether the low ESR could be a problem elsewhere, but I suspect not. Maybe someone else knows better though.

I also have to figure out what explicitly is on audio and 12v. I'm not good at reading schematics yet, but I've already figured out which the digital filtering caps are, and I know one cap is on -12 because it's in reverse polarity.

If anyone is curious, this is how the board cleaned up so far:


This is with just iso and a really good cleaning brush. As is probably visible, a lot of it was just surface level corrosion and the dried salts from the caps.

Continuity is testing good on everything I've tried. Even the loose chip is testing in diode mode like it might be good. And yes, that bare trace going between that serial chip and that capacitor, it's connected fine.

That trace was originally covered in black corrosion, and I scraped it off with titanium tweezers to see what the copper under was like. It's completely intact. I got a proper fiber pen on the way and might do the vinegar -> distilled water -> iso thing if it seems neccessary.

I still gotta figure out how I'm going to remove the caps. Like I said, first time doing this. I do have equipment, but right now I only have one iron, and I'm waiting on a response from Yihua over the defective handle on the hot air station I bought. So I'm really reticent to start, especially because the prospects for this board look so good and I don't want to screw it up.

Another question... Or maybe more of a verification: The PM 7100 power supply is backwards compatible with the IIci, correct? I don't know the state of the IIci's power supply, as the symptoms the original owner listed are ambiguous. I do however have a known working PM 7100.

I also have a Centris 650, but I'm not completely sure of the state of the PSU in that. It works, but it doesn't click when the computer turns on. Which worries me there's a problem with the relay.

 

[CptSkyfox]

I think I found my first break in the traces. I think this should be connected. It's not connected. So. Difficult questions time:


Does anyone know what pin 6 on UE13 does? I think I've learned about half a year's worth of electrical engineering in a day trying to figure this out.

I think this pad is the end of a series of NAND gate logic designed to send what I imagine is the on signal up through pin 9 on the PSU connector, when you hit a power button. ...But I'm still trying to decode the schematic. I just spent a couple hours learning how these NAND gates and flip flop chips work, so I could be wrong.

Either way, assuming my diagnosis isn't bollocks and I'm not about to send 5v somewhere it shouldn't be... Any advice on how to add a bodge in a spot that's going to be under a surface mount chip?

 

[Melkhior]

It is true too - as a community we have seen 16v tantalums explode on a 12v rail, including ones fitted by Apple originally, back before the advice was updated to say double the rating.

Tantalums are not all created equal; modern tantalums can be safe if you buy the right ones:
James Lewis - They're JUST Capacitors

The entire video is a must-watch before picking capacitors IMHO.

 

[CptSkyfox]

Tantalums are not all created equal; modern tantalums can be safe if you buy the right ones:
James Lewis - They're JUST Capacitors

The entire video is a must-watch before picking capacitors IMHO.

Oh, he's showing the inside of the caps and going into the material science. Thank you, this is exactly the kind of thing I was looking for. I'll watch this before I order.

 

[Melkhior]

I was weary of tantalums myself (had one blow up and catch fire when turning on a Sun 3/80 motherboard...), but for bulk capacitance I still ended up with them on the IIsiA7 Mini; polymer tantalums are much much more expensive but compared to the cost of the other components like the FPGA itself it doesn't really matter [and regular tantalums are dirt cheap on JLCPCB, so even polymer tantalums aren't expensive in absolute terms]. I'm using T520 series Kemet.

 

[CptSkyfox]

I was weary of tantalums myself (had one blow up and catch fire when turning on a Sun 3/80 motherboard...), but for bulk capacitance I still ended up with them on the IIsiA7 Mini; polymer tantalums are much much more expensive but compared to the cost of the other components like the FPGA itself it doesn't really matter [and regular tantalums are dirt cheap on JLCPCB, so even polymer tantalums aren't expensive in absolute terms]. I'm using T520 series Kemet.

Yeah, I read that polymer tantalums had a safer failure mode and required lower derating. That was actually going to be my first choice. The problem is they seem to cap out at 10v in a C footprint.

So you either cram a motherboard full of oversized caps, or you use something else. Maybe I'll try using D size caps, but I need to get a caliper to takes some better measurements first.