Internal resistance (IR) tells you more about a battery's condition than capacity does. High IR means voltage sag under load, heat, and wasted energy. Here's how to measure it with whatever gear you have.
What internal resistance is
Every battery has some resistance to current flow. That resistance is measured in milliohms (mΩ). When current flows, the voltage drop across that resistance is wasted as heat. A fresh cell might read 20 to 100 mΩ. A degraded cell can read 500 mΩ or more.
IR matters most in high-current applications. A cell with 200 mΩ IR drops 2V at 10A discharge. That's 20 watts of heat instead of power to your load.
What counts as good IR
These are approximate ranges for common cell types at room temperature:
- Quality 18650 NMC (new): 50 to 150 mΩ
- Quality 21700 NMC (new): 20 to 60 mΩ
- LFP prismatic (100Ah+): 0.2 to 0.5 mΩ
- LFP cylindrical (38120, etc.): 5 to 20 mΩ
- Used 18650 (Grade A): under 200 mΩ
- Used 18650 (marginal): 200 to 400 mΩ
- Used 18650 (retire): above 500 mΩ
Manufacturer specs use AC IR measured at 1kHz. DIY methods give DC IR, which runs higher. Don't compare your DC reading directly to the datasheet — compare cells against each other within your batch.
Method 1: Dedicated IR meter (best)
A YR1035+ or similar 4-wire milliohm meter is the most accurate and fastest method for spot-checking cells. It clips onto the cell terminals, applies a small AC signal, and reads resistance directly. Total time per cell: under 5 seconds.
Test range up to 200 ohms with a minimum resolution of 0.01 milliohms and less than 5 milliohms.
Check price on Amazon
Some capacity testers (OPUS BT-C3100, SkyRC MC3000) include IR measurement built into the charge/discharge cycle. This is slower but means one tool does both jobs.
Four bay charger and tester. Supports all chemistries. 3A charge rate, up to 50W charger power, discharge at 2A.
Check price on AmazonMethod 2: Multimeter with load resistor
If you don't have a dedicated IR meter, you can calculate DC internal resistance with a multimeter and a known load resistor.
What you need
- Multimeter (preferably with 4-digit resolution)
- A load resistor — 1Ω at 5W for 18650s is a good starting point
- Alligator clips or a cell holder
Pretty versatile multimeter. Accurately measures AC & DC current, voltage, capacitance, and more.
Check price on Amazon
Load ResistorsCheck price on Amazon
Procedure
- Charge the cell to a known state (50% SoC is conventional; voltage varies less at midpoint).
- Measure open-circuit voltage (Voc) with no load. Record it.
- Connect the load resistor across the cell terminals.
- Immediately measure the loaded voltage (Vload) — do this within 1 second. Voltage continues to drop as the cell heats.
- Calculate current: I = Vload / R (use your measured resistor value, not the label)
- Calculate IR: IR = (Voc − Vload) / I
Example
Voc = 3.72V, R = 1.00Ω, Vload = 3.58V
I = 3.58 / 1.00 = 3.58A
IR = (3.72 − 3.58) / 3.58 = 0.14 / 3.58 = 39 mΩ
That's a healthy cell. The same math works for any resistor value.
Method 3: Two-point voltage comparison (quick screen)
If you just need to quickly sort a batch, measure open-circuit voltage, then measure again under a small load (even just a flashlight bulb). Cells that show a larger voltage drop under load have higher IR. This won't give you a number, but it quickly separates bad cells from good ones in a batch.
Temperature effects
IR is strongly temperature-dependent. A cell at 0°C can show 3 to 5x higher IR than the same cell at 25°C. Always test at room temperature and let cold cells warm up for at least 30 minutes before testing.
IR vs state of charge
IR is lowest at mid-charge (30 to 70% SoC) and rises at both extremes. Testing at a consistent SoC level — 50% is conventional — gives the most reproducible results when comparing cells.
Using IR to grade a pack build
When building a series-parallel pack, match cells within each parallel group to within 20 mΩ of each other. High-IR cells in parallel pull less current than low-IR cells, creating imbalance. Over time, that imbalance causes uneven degradation. For a series string, IR mismatch causes voltage sag differences between cells, which can trigger BMS cutoffs on the weakest group first.
Use the Cell Database to look up typical IR specs for name-brand cells before buying salvaged batches — it helps you know what you're aiming for.
Rule of thumb: estimating safe current from internal resistance
So we can pretty much see a general trend. While it's true that all of these cell manufacturers consider a different amount of waste heat per cubic centimeter to be acceptable in a given cell, they all kind of hover around the 90% efficient mark. So personally, I think that's a good rule of thumb. If you want to rate the cells similar to how the manufacturers rate them, but you don't know the official rating, you can just measure the resistance and then consider the size and its losses. And the fact that these cells are roughly the same size. And they're using basically the same chemistry. So the differences in how the cell manufacturers rate them are really just for marketing. It's to find the best balance of marketing and actual performance.
If you want to use something more tied to the size of a cell, then 0.35W waste watts of heat per cubic cm of cell seems to be the average acceptable continuous amount of heat according to cell manufacturers.
Will using these estimates provide a perfect, scientifically accurate rating for an unknown cell? Absolutely not, but its a whole lot better than simply guessing, or assuming the cells you have are brand new and can perform at the rated current.
Worked examples: current from internal resistance
30 mOhm 18650
This cell represents something on the level of an LG MH1
An 18650 cell is about 18 millimeters in diameter and 65 millimeters long, which is a radius of 0.9 centimeters and a length of 6.5 centimeters. Treating it as a cylinder, its volume is pi times radius squared times length, so pi times 0.9 squared times 6.5 which is about 16.54 cubic centimeters. With a limit of 0.35 waste watts per cubic centimeter, the maximum total waste power is 0.35 times 16.54 which is about 5.79 watts. If the cell has 30 milliohms internal resistance, that is 0.03 ohms, and resistive waste power is current squared times resistance, so current is the square root of waste power divided by resistance, giving the square root of 5.79 divided by 0.03 which is about 13.9 amps.
15 mOhm 21700
This cell represents something on the level of a Molicel P42A
A 21700 cell is about 21 millimeters in diameter and 70 millimeters long, which corresponds to a radius of 1.05 centimeters and a length of 7.0 centimeters. Approximating the cell as a cylinder, its volume is pi times the radius squared times the length, which is 3.1416 times 1.05 squared times 7.0, giving a volume of about 24.25 cubic centimeters. With a waste heat limit of 0.35 watts per cubic centimeter, the maximum allowable waste power is 0.35 times 24.25, or about 8.49 watts. If the internal resistance of the cell is 15 milliohms, which is 0.015 ohms, the resistive heating is equal to current squared times resistance, so the allowable current is the square root of the waste power divided by the resistance. Taking the square root of 8.49 divided by 0.015 gives a current of about 23.8 amps.
4 mOhm 21700
This cell represents something on the level of an EVE 50PL. A 21700 cell is about 21 millimeters in diameter and 70 millimeters long, which corresponds to a radius of 1.05 centimeters and a length of 7.0 centimeters. Approximating the cell as a cylinder, its volume is pi times the radius squared times the length, so pi times 1.05 squared times 7.0, which is about 24.25 cubic centimeters. With a waste heat limit of 0.35 watts per cubic centimeter, the maximum allowable waste power is 0.35 times 24.25, which is about 8.49 watts. If the internal resistance of the cell is 4 milliohms, that is 0.004 ohms, and resistive waste power is current squared times resistance, then the allowable current is the square root of the waste power divided by the resistance. Taking the square root of 8.49 divided by 0.004 gives a current of about 46.1 amps.
FAQ
- Can I measure IR with a standard multimeter's resistance mode?
- No. Standard multimeter resistance mode applies a tiny current that won't give you a useful reading on a battery. You need to use the voltage-drop method above, or a dedicated IR meter.
- My reading seems high — is the cell bad?
- Check temperature first. Cold cells read high. Also check your connections — poor contact adds resistance. If it's still high at room temperature with good connections, the cell is degraded.
- Do LFP cells have lower IR than NMC?
- Large-format LFP prismatic cells (100Ah+) have extremely low IR — fractions of a milliohm — due to their large electrode area. Cylindrical LFP cells are in a similar range to cylindrical NMC. Compare within the same form factor.
