3
\$\begingroup\$

I have a PCB with an RJ45 connector, which naturally requires a 1000 pF/2 kV capacitor connecting chassis GND to system GND.

Unfortunately the PCB was designed with an 0805 package, and I can't find any 2 kV rated 0805 capacitors. I can get one in a 1206 package however, and when I overlaid the 1206 footprint onto the 0805 in EAGLE it seems as though there should be enough overlap to solder the larger capacitor. There are no other signals around the pads, in fact they're both engulfed by same-signal traces.

Will soldering a 1206 high-voltage capacitor onto the 0805 footprint in this manner result in any unintended consequences? I know that this is a bit open-ended, but I'm not even sure where to begin with that line of questioning in terms of what to explore.

For context, this board was a small quantity prototyping run. I just need to prove that an Ethernet signal can pass through this board, and then design the correct footprint size for the manufacturing run. Reordering another prototyping quantity without using these to test and debug would be less than ideal.

\$\endgroup\$
5
  • 3
    \$\begingroup\$ Ya, you'll be fine. I did the exact same thing, same function, on a proto spin. We hand-soldered them. No issues. Go for it. \$\endgroup\$ Commented Jan 4, 2023 at 21:49
  • \$\begingroup\$ It appears to me that the voltage rating of that capacitor is totally insignificant to what you want to achieve with this prototype. Why not use a 0805 with lower voltage rating? \$\endgroup\$
    – feynman
    Commented Jan 4, 2023 at 21:55
  • \$\begingroup\$ @feynman from other answers on this site(re: electronics.stackexchange.com/questions/66119/…) it is my understanding that this capacitor for an Ethernet connector must be 2kV. If my limited functionality needs for this prototype means that assumption is invalid, I would appreciate more elaboration. \$\endgroup\$
    – InBedded16
    Commented Jan 4, 2023 at 22:04
  • \$\begingroup\$ In this specific case it'll be fine aside from the mechanical inconvenience, because parasitic properties of the capacitor aren't particularly critical. Just keep in mind that for certain other cases (e.g. decoupling caps on sensitive high frequency devices, and AC coupling caps in high speed buses) the small part size keeps the parasitic inductance low. \$\endgroup\$
    – Polynomial
    Commented Jan 4, 2023 at 23:38
  • \$\begingroup\$ It's a prototype. Just do whatever works. You could use a lower voltage cap, you could solder it sideways, you could use two in series sticking out of the board and connected at the top, you could use a bodge wire, etc etc etc. Just fix it in the next run and remember any testing related to this particular capacitor isn't valid. \$\endgroup\$ Commented Jan 4, 2023 at 23:44

3 Answers 3

1
\$\begingroup\$

You can try to solder the cap but you can also leave it out for initial testing.

A missing cap between chassis and signal ground likely won't cause any Ethernet problems under normal circumstances.

If it does then you likely have problems with the cap in place too.

You can try a capacitor that fits, with smaller voltage rating. Generally you would not have 2kV between chassis and signal ground anyway, or if you need to withstand more than 50V under normal conditions, then it is an indication of something being wrong already.

Sometimes you might already see that connector chassis and signal ground are already connected together via some other route so the AC coupling cap does very little.

\$\endgroup\$
1
\$\begingroup\$

Should be OK but might need precision hand soldering. As a worse case you can use a leader capacitor, again hand soldered. For mass production you need to get it right.

\$\endgroup\$
1
\$\begingroup\$

Will soldering a 1206 high voltage capacitor onto the 0805 footprint in this manner result in any unintended consequences?

Not really, you'll most likely have to hand solder it, so if you are doing this yourself and have already manufactured the board then it's probably doable.

One thing is if you are doing ESD testing on this board it could affect the results because of the gap difference, any other testing should be mostly identical, the only other differences will be inductance differences from the solder and those will be minimal.

\$\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.