Resistor tolerance


I build my 3 IotaWatt with standard 5% tolerance through hole resistors. (i think 5% had them lying around in a big bag :wink: - could als be 1%)

I was wondering if changing them to 0.1% would improve it (not that i’m complaining :wink: when i made them i measured between 100 and 104W for a 102W lightbulb )

Since i am using SCT013-030 the resistors over the CT aren’t there.
So there are 4 locations left

  • the led: quite sure the tolerance doesn’t matter here
  • the voltage devider on the VT: i guess that it isn’t very important to have low tolerance here
  • the vref shunt: unsure, maybe this is important?
  • the LM358: unsure, maybe this is important?

The reason i’m asking is to know if the time and money (ok, 0.6euro per resistor isn’t that expensive) is worth it?


Correct, lower it if you want a brighter LED and visa-versa.

Right. As long as you calibrate the VT, it doesn’t matter. In the production units however, we do use 0.5% resistors and the result is that the standard issue VTs are usually within 1% without calibration.

Nope, but the tolerance of the shunt itself is critical. The production units use a 0.2% shunt. This is the whole enchilada. Everything is calibrated to that standard.

I assume you are talking about the 4.7K resistors in the voltage divider that feeds the LM358. That’s not critical. The bias is compensated dynamically, but I’d recommend at least 1% for that.

I used to do thru-hole, but found that SMT is actually a whole lot easier and better once you get a few basic tools. But this is particularly better with SMT. Resistors are pretty much 1% by default and .5% don’t really cost much more.

You’re right that the burden resistors need to be .5% or better to get consistent results.


Thanks for the clarification! My shunts are 1V and 0.35%, so a bit worse than yours, but still ok i guess :wink:
offtopic: i was thinking of trying some smd stuff on the next holyday :wink: