# Making a high current PCB (power PCB)

I want to make a power PCB with 1 oz copper thickness. The problem is that I want this PCB to stand up to 50 A and I don't know what to do to keep the traces a reasonable width.

Using the KiCad PCB calculator, I calculated that for 1 oz copper I need 68 mm trace width (something impossible) for 50 A current.

Is there any technique that I can do to keep the traces up to 5 mm width or something close to that?

The only solution I thought of is to solder stripped wire (for enhancement) to the traces that I want to pass high current.

Will that work?

The PCB will be double-sided.

• How many layers of copper on your PCB and how long is the 50 amps present for continuously? How warm can you allow your copper tracks to become? Feb 28, 2022 at 13:54
• In Kicad you can play with the "temperature increase". Maybe 20/30°C aren't that bad. Else you need to increase the copper thickness or as saw on high current designs increase the trace "thickness" with solder
– Mat
Feb 28, 2022 at 14:04
• Is it possible to make big pads and solder big wires instead? Feb 28, 2022 at 14:09
• Sounds like you need thick copper or to use busbar (wire soldered on will do too). Feb 28, 2022 at 14:13
• Is it possible to use different copper thickness on both sides? Feb 28, 2022 at 21:57

## 50A traces are not practical with 1 oz Cu.

You will also have problem interfacing connectors or wires to PCB that requires some care and attention to design specs.

It requires 400 mil traces for a 20'C rise with 10 A thru 1 oz Cu.

I suggest you use busbars along the shortest path possible rated for 100A.

https://electronics.stackexchange.com/search?q=busbar+pcb Searching will show an answer for a tiny 100A busbar.

https://saturnpcb.com/saturn-pcb-toolkit/ Thermal calc avail.

https://www.e-fab.com/products/stiffeners-busbars/ e.g. supplier

• Good tool, especially because it shows the standards it's referring to. Heating is surely a physical phenomenon, but if you were calculating creeping distances etc. Feb 28, 2022 at 21:48
1. Run traces in outer layers.

2. Add for example 1.5 mm wide, 5 mm long paste mask (and solder mask) openings to allow solder to effectively thicken the tracks and the openings allow better cooling too, as solder mask would be an insulator. Unless making opening cause a risk of short circuiting.

3. If you don't add paste on the tracks, simply removing the solder mask yields a bit of benefit, still risking making shorts though.

4. If temperature rise is acceptable, there's still voltage drop, which you might be able to overcome having a voltage feedback "sense" function in the power supply.

5. The obvious: Use 2 oz copper and run tracks in both/all layers.

6. Out of scope: Once you run out of options (currents or voltages are too high to practically run on a PCB) you need to be adding wires, cables and copper bus bars. The components designed for the scale allow cable or screw connections. Often times you surely can get away with using a PCB for those connections, but failures can be catastrophic and dangerous. It's easier to react to failures when you have enough room for fault currents etc, so I do recommend taking this step rather than pushing a PCB to the limit.

• " as solder would be an insulator." Did you mean "solder mask"? Or is solder a really terrible conductor of heat? Feb 28, 2022 at 15:36
• 5. Replacing the high-current traces with sufficiently thick (single core) copper wires is a technique I've seen sometimes. Just like the OP suggested. Feb 28, 2022 at 20:42
• Thanks, I meant solder mask. And yes, wires or copper bus bars are ultimately the solution, but I considered it out of scope, maybe worth adding after all. Feb 28, 2022 at 21:40
• Especially because going too far with a PCB might put people and property to risk. Feb 28, 2022 at 21:46
• Solder (pure lead-free tin) is rather poor (compared to copper) conductor of both heat and electricity. Leaded solder (otherwise a good thing for one-off manual PCB population) is even worse and its melting point is lower (also bad at high current density setups). Feb 28, 2022 at 21:54