When this goes in, how much of it should come out again?

Announcement - Help us fight the BOTS! Please report all spam including stuff in your inbox!

Roland W

Oct 9, 2017

I am writing this post, because I am kind of wondering if what I see is right.

At the moment, I am implementing coulomb metric features into my LFP Powerwall controller, and when I see the initial
measurements, I see a certain amount of energy being charged into the pack, and a certain amount which is then again discharged.
All measurements start and end at the same Voltage point of a rested pack .
At the moment I seem to have a maybe 10% difference in result from in vs out, but that will improve, as I am still calibrating the sensor.
On my Chinese Coulomb meter which was installed before, the drift ended up to be around 2%.

You are all working with old cells. How much of a difference or loss of energy per cycle would you expect on a pack?
What is the typical roundtrip efficiency on your packs (cell base, DC in vs DC out)?

Last edited:


Active member
Dec 15, 2018
The best I can determine from my 40% DOD in middle voltage range of 18650 lithium-ion over 100(s) of cycles... my loss is <=2% based on a few years of operation data. For 2020 I had 18,030kwh in (Midnite Classic measurement). 15013kwh consumed out the AIMS inverters per cheap Amazon meters. That's a 3017kwh loss or 16.7% loss. This puts the AIMS Inverter at 83.3% efficient - which is perfectly reasonable based on other reports I've seen.

A 2% loss could exist due to fuzzy measurements or AIMS at 81.3% efficient (instead of 83.3) or half and half as I do have warm wires in the afternoons. But I'd be surprised if it was much higher than 2% just because I'm already at low 80% for AIMS on the gross measurements.

So I conclude that the battery charge/discharge loss at modest stress and 40% DOD is probably 1% or less on a daily cycle basis.

Happy to be wrong - just sharing my thinking. Wish I could actually measure this accurately.


Oct 8, 2016
Most lithium cells have a loss of less than 1% for new cells. I think im in the same area for my 2nd hand where its less then 1% for my daily use.

I hve tested this with colloumb meter and its darn close to exact if the meter is calibrated AND you dont stress the cells. When you start stressing the cells with high current the drift will be alot higher!


Feb 28, 2018
Just charge/discharge across the shunt, it should be pretty low like Daromer says.
If your system is doing balancing, there'd be some losses there.
Some types of current waveforms may not be accurately measured by the shunt system, eg peaky or high frequency (think inverter &/or charger mosfet switching noise).