0

I just came across this video: https://www.youtube.com/watch?v=fxC5w3f8FkY which says that I can duplicate a trace on the bottom layer and then at the end just via it back up to increase the current capability of that trace.

Say for example I want to carry approx 5A ish on a trace. A trace calculator gives me the following enter image description here

I would require 68mm trace width. Is it possible to just do 34mm on the top layer and 34mm on the bottom layer? I.e each trace would be carrying approx 2.5A each. How would that work if the target device pin is SMD? I would need to via it up to the single 34mm trace, wouldn't that be a bottleneck?

AlfroJang80
  • 3,103
  • 2
  • 32
  • 59
  • Are you saying that you get 5A from a single SMD device pin? What is the size of the pin? What is the diameter and plating thickness for your vias? – Elliot Alderson Dec 30 '18 at 17:39
  • @ElliotAlderson It's more of a theoretical question. I'm essentially designing a power distribution board that has a 12V rail input. This board is gonna drop that 12V down to 5V and power two microcontroller boards (with their own sensors) as well as a few other boards. Total current draw from all devices is going to around 3A tops I would say with around 1.8A for the microcontroller boards and the remaining 1.2A for LED driving boards (again max values). My plan is to have JST-ZH connector come in and split up quickly with one trace going to 5V regulators for the MCU boards. – AlfroJang80 Dec 30 '18 at 17:44
  • That 5A is meant to be a VERY conservative value for how much current I expect to get from the 12V rail at the JST-ZH connector. – AlfroJang80 Dec 30 '18 at 17:45
  • 2
    68mil is not 68mm! 1mil is 1 thousandth of an inch. Hence, 68mil is about 1.73mm. – user2233709 Dec 30 '18 at 17:56
  • @user2233709 Uh oh. That explains the doubts I was having on how this trace would actually connect to a single pin and not the entire row of connectors pins. – AlfroJang80 Dec 30 '18 at 18:00
  • The formula for resistance $ R = \rho\frac{l}{A}$ that means if you keep the cross sectional area the same you'll have the same resistance. Therefore, if you have two traces that are half the width they will be the same resistance. – Voltage Spike Jan 04 '19 at 21:21

2 Answers2

2

Two coplanar 68 mil tracks will rise 11'C with 2.5A on each track.

But the via will add resistance to the other track, and the via temperature rise will be higher so to prevent thermal gradient delamination stress, they should solder fill the vias.

But putting mirror tracks on opposite layers will raise the temperature of both tracks more than on the same layer.

The length does not affect the temperature rise since the power per length is constant and the total power loss is small. This is because the resistance per unit length is constant and Pd=I^2*R here 15 mohms/" @ 68 mil.

Yet trace length does affect voltage drop and inductance.

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 185
0

Such calculators are fine to calculate the required width of a “long” track. But you can still have a short section (something about 1mm-long) that is significantly thiner, for example between a via and a pin of a SMD device. The heat that is generated by this short section will be conducted to the whole track.

user2233709
  • 1,735
  • 9
  • 15