appropriate replacement for a double capacitor?

[Logg]

While working on a Compaq Portable II (1986), I found this double capacitor rated for 1700 uF 20v but only measuring about 38 uF. Seems obvious enough; this is probably why the machine won't turn on twice in a row.

I'm thinking I probably want to buy 2 of these 1800 uF capacitors

In case you're not sure what I mean by "double capacitor", here's a picture... it has 2 positive leads and one negative lead.  Is it kosher to solder the two capacitor's negative leads together, and then solder the positive leads to the two points on the power supply board where the original cap went to? and is a 1800 uF capacitor (with 20% tolerance) an acceptable replacement for a 1700 uF capacitor? I think it is.

[davy peleman]

never heard of double caps??? isn't it just one 1700uf cap with two leads. you should check the schematic and see if theres 2x 1700uf in the power supply. in any case there's no problem in joining the negative leads together. parallel caps are common practice. only when you series them up they should have the same ratings, which is also the case if you'd replace it by 2 1800uf caps rated for the same voltage. so don't worry you will not cause damage doing this. and yes they will be good replacements as long as the rating is close to the original, you should be fine. also most of the time a little more is better than less. so check the schematic that will make you wiser i think. my guess is that it's a single cap with 2 leads. if it is a double cap you can replace it with those 2 1800uf caps and solder the leads together no problem.

[Logg]

I think it's two capacitors in one package? I measured the resistance between the two positive leads as "open loop". Not sure that I can find the schematic for the power supply.

[davy peleman]

ok than it seems it's a double cap indeed. so just do what you wanted to do. it will not cause damage even if it's not the case. have you already checked that the cap isn't shorted, 38uf is very low for that cap, so it has a severe issue.so there could be some other damage. check the power supply. also check the cap with an ohmsmeter. the resistance should climb steadily until it reaches OL. surely check the output voltage of the power supply and check the fuse if there is any. if that's alright change the caps.

[Logg]

there was also a blown (looks like it poofed, totally black) 15uf tantalum cap on the board. I'm only guessing it used to be 15uf based on other tantalum caps on the board. Completely turned to carbon, no schematics anywhere online or in the service manual. I just replaced it with 2 electrolytic caps, a 10 and a 5. Might get a proper 15uf tantalum one since I have to make a minimum order anyway.

this was an eBay find claimed "working", I get it and there's lithium ion battery acid turned into this white powder all through the machine... could probably fill a drinking glass with all the acid powder. The way it was stored it ate a hole through the case metal, not the board. Even with this apparently faulty power supply, I still managed to get it to boot from a floppy once. It just needed about 7 minutes once the power turns off before it'll turn back on successfully again... Hoping replacing this cap will fix it.

When the machine is on, the 12v line reads 12v and the 5v line reads 5v.

about measuring ohms on the double cap, it behaves exactly like you said, it's measuring 232 kOhms right now and still slowly climbing.

[davy peleman]

don't replace tantalums with electrolitics. electrolitics are polarized + and - whereas tantalums are not. you should replace it with a proper tantalum otherwise you're in for trouble. the electrolitics are going to be fried if they're rereceiving ac voltage. also ceramics are way to unstable when getting hot. so you need tantalums they hardly move in value when temperature rises. sorry to say that but it's like that. seems that your power supply is working. the fried tantalum cap is an indicator that there is a problem elsewhere.

 so electrolitics are polarized, ceramics not but they move alot with temperature it will cause drift in the circuit. pretty sure the bad battery caused the issues. you should replace it. sorry to say but you will have to resolve these issues. changing caps alone will not suffice. the problem will reoccur and fry the new caps. also clean the entire circuit board with a soft brush. the acid probably made a short somewhere when the machine was on. check the traces on the board they could be damaged by the acid. sorry to say but this will not be an easy fix.

[davy peleman]

if i'd only knew the function of that tantalum cap. was it a decoupling cap or an ac coupling cap. probably a decoupling cap.then you could replace it with electrolitics, but if you're not shure it's not easy. it could be a cap in the clock circuit and in that case you need the precision of a tantalum cap. i'm no expert to but that's my knowledge so far. they don't use these tantalums if they don't have to since there more expensive. so always replace a component with the same one.

[davy peleman]

when measuring ohms on a cap you should look like it this way. in the beginning the cap is empty so it passes current for a short period until it is charged by the meter then no current can pass through and it should read OL. the bigger the cap the longer the charging off course. When you use the continuity check. the meter will beep for a short period then stops because current can no longer pass when it is charged by the meter. that's another indication for a good cap

[Logg]

Yeah, when I saw the acid everywhere I wasn't happy. But all the traces I tested are fine and I already cleaned the board with ipa, replaced the battery. Will do on the tantalum cap.

[davy peleman]

ok then when you replaced the tantalum cap, you should give it a go. i hope that will fix the issue. the tantalum is probably going to be a decoupling cap since it's 15uf. but should be replaced with another, because of the mentioned heat issues. electrolitics are not happy near heat and ceramics will move alot in value.  good luck then! hope you get it working again.

[cdev]

NO NO NO..

Tantalums are polarized and what's more not all of them are clearly marked, and so they present a danger, under a number of conditions, especially if you put them in backwards by accident and don't derate them suitably they can literally explode into flame.

don't replace tantalums with electrolitics. electrolitics are polarized + and - whereas tantalums are not. you should replace it with a proper tantalum otherwise you're in for trouble. the electrolitics are going to be fried if they're rereceiving ac voltage. also ceramics are way to unstable when getting hot. so you need tantalums they hardly move in value when temperature rises. sorry to say that but it's like that. seems that your power supply is working. the fried tantalum cap is an indicator that there is a problem elsewhere.

 so electrolitics are polarized, ceramics not but they move alot with temperature it will cause drift in the circuit. pretty sure the bad battery caused the issues. you should replace it. sorry to say but you will have to resolve these issues. changing caps alone will not suffice. the problem will reoccur and fry the new caps. also clean the entire circuit board with a soft brush. the acid probably made a short somewhere when the machine was on. check the traces on the board they could be damaged by the acid. sorry to say but this will not be an easy fix.

Tantalums are safe if you make sure to always get the polarity right, and use them being careful that they don't exceed their rated voltage even for an instant.

 I will typically use a part that is rated at several times (4 or more times) the voltage I expect there. 

[helius]

No matter how much you derate them, they will self destruct if powered with reverse polarity. The solid MnO₂ type will ignite, whereas the solid polymer and the sulfuric acid electrolyte types will only short out. In general, tantalums behave similar to aluminum electrolytics, but they usually have lower ESR and longer service life if correctly used, since they lack the drying-out and dielectric erosion effects that cause failure in aluminum electrolytics.

[vk6zgo]

never heard of double caps??? isn't it just one 1700uf cap with two leads. you should check the schematic and see if theres 2x 1700uf in the power supply. in any case there's no problem in joining the negative leads together. parallel caps are common practice. only when you series them up they should have the same ratings, which is also the case if you'd replace it by 2 1800uf caps rated for the same voltage. so don't worry you will not cause damage doing this. and yes they will be good replacements as long as the rating is close to the original, you should be fine. also most of the time a little more is better than less. so check the schematic that will make you wiser i think. my guess is that it's a single cap with 2 leads. if it is a double cap you can replace it with those 2 1800uf caps and solder the leads together no problem.

Multiple caps in the one package are quite common.
They allow more convenient placing of PCB tracks, & are easy to mount rigidly.
Yes, you can replace it with two individual capacitors, but the mounting may be more difficult.

[james_s]

I've seen a lot of multi-section caps in old tube/valve gear but I don't recall seeing one in anything modern. I usually gut them and stuff a couple of modern electrolytics inside the can.

[dhaillant]

Hello everyone,

I'm having same trouble trying to identify and understand how to replace the same kind of 3 leaded capacitors in a slightly newer SLT 386s/20 Compaq Laptop Power Supply:





I found a Datasheet where it is said "NC" (as in "No Charge potential") for the third leg : https://www.vishay.com/docs/42055/678d.pdf

And on the PCB, there's no apparent copper trace connected to that third leg. But, if I check for capacitance between the 3rd leg and one of the two others, I read a 8-9 µF value. The reading is not stable at all. If I read one value, the second changes, ramping up... So it's obviously internally connected in some way.

Are those capacitors Radial caps with a 3rd "mechanical" leg that can be ignored?

Thank you,
David.

[helius]

Good work finding the datasheet. The picture in the OP is this datasheet: https://www.vishay.com/docs/42054/672d.pdf

Basically, the design is a bridge between radial and axial capacitor construction. One reason that axial caps were favored in quality equipment, and still available (at high cost) for military/aerospace, is that they are much more soundly mounted by their leads compared to radial caps. The cap lying flat against the board with leads soldered at both ends is less subject to vibration and flex. But the industry moved to radial caps since they could be made with higher CV factors and smaller sizes. So the idea was, why don't you take a radial cap (- and + are both at the same end) and put a dummy lead on the opposite side so it could be mounted axially? You would get the benefit of both better structural rigidity of axial mounting, and better CV of radial construction.

Are those capacitors Radial caps with a 3rd "mechanical" leg that can be ignored?

That's exactly what they are. Extra mechanical legs can also be found on the same side in some caps. The dummy leg is connected to the case, and there is some capacitance between the case and the foils, but not a lot and the measurement is easily disturbed by external fields as you noted.

The advice in the thread above is all nonsense: the cap in the OP, and the second picture in post #15, are not double capacitors at all, and the reason that the OP reads such a low value with his DMM is that he had shorted out the - and + legs of the cap. The cap is a red herring and the failure is elsewhere (a dead rectifier or mosfet would be my guess).

[davy peleman]

sorry my bad i'm not giving advice anymore since my knowledge clearly sucks. really sorry. just wanted to help. for now i'll be only asking questions myself. i'm a bit embaressed  about all this.

[vk6zgo]

Good work finding the datasheet. The picture in the OP is this datasheet: https://www.vishay.com/docs/42054/672d.pdf

Basically, the design is a bridge between radial and axial capacitor construction. One reason that axial caps were favored in quality equipment, and still available (at high cost) for military/aerospace, is that they are much more soundly mounted by their leads compared to radial caps. The cap lying flat against the board with leads soldered at both ends is less subject to vibration and flex. But the industry moved to radial caps since they could be made with higher CV factors and smaller sizes. So the idea was, why don't you take a radial cap (- and + are both at the same end) and put a dummy lead on the opposite side so it could be mounted axially? You would get the benefit of both better structural rigidity of axial mounting, and better CV of radial construction.

Are those capacitors Radial caps with a 3rd "mechanical" leg that can be ignored?

That's exactly what they are. Extra mechanical legs can also be found on the same side in some caps. The dummy leg is connected to the case, and there is some capacitance between the case and the foils, but not a lot and the measurement is easily disturbed by external fields as you noted.

The advice in the thread above is all nonsense: the cap in the OP, and the second picture in post #15, are not double capacitors at all, and the reason that the OP reads such a low value with his DMM is that he had shorted out the - and + legs of the cap. The cap is a red herring and the failure is elsewhere (a dead rectifier or mosfet would be my guess).

Unless I'm really losing it, there was no photo prior to the OP editing his posting.
From the description, it definitely sounded like a conventional dual cap.