0
\$\begingroup\$

I have a project cominig up where it would be benicical to have the large input decoupling cap's top surface be a touch contact point for Gnd for the circuit.

I know that in the Old Days of yester year (tube era) the aluminium cans of large caps were one of their primary conductors (Gnd)

I'm not considering standard electrolytics (with srink wrap) that have the X,K,Y relief cuts in the top.

What I'm planning on using is the all metal (japanese) style caps as seen on performance Motherboards, Videocards, etc. As this style of cap has its value data printed on the top so I may have to sand/scuff to top surface to be a better contact.

The space I have to work with is so small and oldly shaped it would be very advantageous to do this. Also the currents are small in the tens of mA.

So my question is; is this something that is done, can be done, or really bad? Would standard electrolytic s be better?

\$\endgroup\$
5
  • 3
    \$\begingroup\$ (1) Are you sure the case is grounded? I've never seen them grounded at the PCB pin and I suspect that if they are grounded it might be a weak connection for noise reduction. (2) The can is aluminium and will have a layer of oxide. This will make it difficult to contact properly and impossible to solder to. \$\endgroup\$
    – Transistor
    Commented May 14, 2016 at 17:41
  • \$\begingroup\$ The old aluminium ones were. I'm not sure on the new Japanese style ones. If there not I might still consider connecting the case to Gnd mechanically, over other options. So either way, the question is still valid. \$\endgroup\$
    – user1488660
    Commented May 14, 2016 at 18:19
  • \$\begingroup\$ @transistor axial leaded capacitors usually have the case as one of the terminals \$\endgroup\$
    – Jasen Слава Україні
    Commented May 14, 2016 at 23:28
  • \$\begingroup\$ IF the capacitor actually has a connection to it's case, I can't see a few mA causing any major problems (seeing as those polymer caps are rated in AMPS of ripple current anyway). But if you've got a few spare caps, try it, see what happens, see if you can measure any low resistance between one of the terminals and the case with a multimeter. \$\endgroup\$
    – Sam
    Commented May 15, 2016 at 0:20
  • \$\begingroup\$ these are a new type of electrolytic capacitor en.wikipedia.org/wiki/Polymer_capacitor \$\endgroup\$
    – Jasen Слава Україні
    Commented May 15, 2016 at 3:11

2 Answers 2

Reset to default
1
\$\begingroup\$

This would be a bad thing to do. Most likely you won't get the rated capacitance, and the ESR will be worse. Also, even if it seems to work when its new, as noted in the comment, the aluminium can will easily oxidise, and your connection won't be reliable.

If you need an axial capacitor (contact at each end), that is what you ought to use. Modern components are more optimised than old ones, and are not as easy to abuse.

\$\endgroup\$
3
  • \$\begingroup\$ An axial cap would only expose the one lead to the contact, This is not how a decoupling cap is conected. The circut, and the cap lead must connect to the contact. \$\endgroup\$
    – user1488660
    Commented May 14, 2016 at 18:23
  • \$\begingroup\$ Also the case of the cap is used as a contact for the complete circut. So the ESR of the cap wpuld be unaffected, AlO on the surface would only add to the incoming wire resistance. Aluminium Oxide forma \$\endgroup\$
    – user1488660
    Commented May 14, 2016 at 18:28
  • \$\begingroup\$ Seems you already know the answer to your question. \$\endgroup\$
    – Sean Houlihane
    Commented May 14, 2016 at 23:04
1
\$\begingroup\$

The case is not connected to either terminal, so to pass Gnd through it you would have to make two connections. Most modern 'all metal' capacitors have a lacquer coating which you would have to sand or file off to expose the bare metal.

Aluminium forms a thin oxide layer almost instantly on exposure to air, and the oxide is a very good insulator. To get a reliable connection you need a firm airtight contact, not just 'touching'. In most circuits having a good ground connection is critical, so adding two dodgy contact points is just asking for trouble.

Connecting the case to ground probably won't affect the capacitor's performance, but unless the manufacturer permits it you can't be sure. Your idea might be OK for a hack, but is not good design policy.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.