8
\$\begingroup\$

If I have a PCB with standard 1.4 mil copper thickness. I have used a trace width calculator based on IPC-2221 to determine that a 2.72mm trace is enough to handle 5 amps for an "external layer in air".

I have seen many times on power supply and relay board that people add solder to unmasked traces to increase the safe current flow.

I am using Sn63Pb37 solder. My finish is ENIG. I know that solder is much less conductive than copper.

My question is how much does this help? If I lay a fat convex blob of solder over the entire length of the track what ballpark level of increase in safe current flow can I expect?

I am sure an exact answer relies on various circumstances but a conservative value would be helpful.

I have seen in various claims on the internet that it can increase the amount of safe current by 40%-100%, but the sources seem to contradict each other. I would like to get an educated answer.

Is there any reason why doing this is a bad idea?

[EDIT]

I am going to attempt to answer this myself based on the information at PCB Tinning to increase current about how to figure out conductivity based on cross section and KalleMP's information about solder conductivity. Please correct me if I get any of this wrong.

Assuming a 1.4 mil thick copper trace with a width of 1 mm(to make the math simple) and a convex layer of solder going .5 mm tall at the highest point.

The cross section of the trace is a rectangle 1 mm wide by 1.4 mils(0.03556 mm) tall which gives an area of 0.03556 mm2.

Assuming a circular shape a 1mm wide and .5mm tall blob's cross section is a half circle so I use pi * .5^2/2 to get an area of about 0.3927 mm2.

If I go my what KalleMP said about solder being about 9-14% the conductivity of copper I can conclude that the effectiveness of the solder should be the equivalent of 0.035343-0.054978 mm2 of copper which is a 99-154% improvement, less depending on how thick the solder blob is.

This answer seems consistent with some of what I have heard from others, but I have also read of people doing actual tests and finding it closer to 40%. Does anyone see any fault in my reasoning? Are there other factors that would reduce the improvement such as the connection between the solder and copper?

[Note] at first I posted a different solution with an error, I think I have it right now.

\$\endgroup\$
8
  • \$\begingroup\$ I vaguely remember this sort of question before, perhaps not at eeSE. I think the consensus was that solder is a poor conductor compared to copper so the benefits are there but often less than one might hope for. They might be recommended for thermal mass in pulsed applications at a guess. \$\endgroup\$
    – KalleMP
    Commented Oct 24, 2015 at 16:59
  • 2
    \$\begingroup\$ When I've needed to increase the current-carrying capacity of a track I've soldered a thick piece of copper wire along it. It's easy to find the maximum current from the wire diameter. \$\endgroup\$ Commented Oct 24, 2015 at 17:00
  • 1
    \$\begingroup\$ Solder will melt long before a copper trace will fuse, if liquid solder in your device is very bad (it can shake free) then I would not use that option. \$\endgroup\$
    – KalleMP
    Commented Oct 24, 2015 at 17:00
  • \$\begingroup\$ Regarding the melting of the solder, I am using a fuse to limit the current to below whatever my end design can handle. \$\endgroup\$
    – HighInBC
    Commented Oct 24, 2015 at 17:01
  • 2
    \$\begingroup\$ The conductivity of typical solders is about 9 to 14% of copper depending on alloy. Special solders may be better. farnell.com/datasheets/315929.pdf \$\endgroup\$
    – KalleMP
    Commented Oct 24, 2015 at 17:08

3 Answers 3

3
\$\begingroup\$

I can't give you a definitive answer but adding a thick layer of solder over top of a copper trace on a PCB does increase the current-carrying capacity significantly, especially for pulse loads.

That thick layer of solder increases the thermal mass of the copper trace. The increased thermal mass reduces or prevents localized hot spots in the thin layer of copper trace.

Testing is required for your specific application. Testing that I have done on some of my products showed that I was able to safely increase the current rating on a single-sided PCB with 1 oz copper by about double. This is on boards wave-soldered with 63/37 solder.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ I was thinking only of electrical conductivity. It did not occur to me that it would even out hot spots. Given that this is for an LED display there will certainly be pulses of high load with much lower load most of the time. Good information. \$\endgroup\$
    – HighInBC
    Commented Oct 24, 2015 at 19:58
4
\$\begingroup\$

It does work, but it is hard to model reliably and you should use empirical testing to see how well it works for your boards, process, environment, solder.

There can be a lot of variance from little things that are hard to predict and measure. For example, are you soldering by hand or running though a wave machine? The resulting solder profile cross sectional area can be off by 50% or more depending on which process you use- and neither of them resemble a half circle :) .

EEVBlog feauted a video where he tests exactly this question...

EEVblog #317 - PCB Tinning Myth Busting

And a similar video...

Does putting solder on high current PCB tracks help?

\$\endgroup\$
1
  • \$\begingroup\$ I agree a half circle was over optimistic. I may just get some callipers out and measure. \$\endgroup\$
    – HighInBC
    Commented Oct 24, 2015 at 23:51
2
\$\begingroup\$

One reason people may be giving numbers like 40% rather than 100% could be as simple as different starting points. Saying that soldered 1oz traces have roughly double the current carrying capability is functionally the same as saying that the solder has the same current carrying ability as a 1oz copper trace. Thus, if you add solder to a 2oz copper trace, you will only get a roughly 50% benefit, which falls close enough in line with the 40% number given.

\$\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.