Measuring impedance

I received my Battery Impedance Meter SM8124A and did some tests. It isn’t consistent with HKJ’s result that is 60 mOhm for 35E. Is my meter faulty? Any others that got this impedance meter, does it measure correct? What results do you get for Samsung 35E? In its specification paper, it is 35 mOhm.

Samsung 18650 35E 3500 mAh Measured: 20.0 mOhm at 4.13v (4.117v with another multimeter)
Samsung 18650 35E 3500 mAh Measured: 19.9 mOhm at 3.76v HKJ -> 60 mOhm
Samsung 18650 35E 3500 mAh Measured: 19.3 mOhm at 3.65v HKJ -> 60 mOhm

Keeppower 18650 3500 mAh Measured: 39.4 mOhm at 3.78v HKJ -> 90 mOhm
Ikea Ladda 2450 mAh NiMH Measured: 18.0 mOhm at 1.36v HKJ -> 40 mOhm

Your reading of 20mOhm seems consistent with the 35E’s spec of <35mOhm. This is measured at 1kHz AC.

HKJ may be measuring DC impedance, since it says “Calculated internal resistance.”

DC and AC impedance are different metrics, and should not be compared to one another.

More info: BU-902: How to Measure Internal Resistance - Battery University

also keep in mind it is not REALLY a resistor in the battery

that is an approximation

at high currents or high temperatures it will change on a dynamic basis

as well as after high loads or temps have been applied.

Your measurement is correct for ACIR (AC internal resistance). I measured Samsung 35E 19,7 miliohm at 3,84 V with SM8124A.

Samsung 35E DCIR measured according IEC 61960-3, par 7.7.3 is about 29 miliohm at 3,84 V and 25°C.

Lygte has sometimes some incorrect data, his „calculated resistance“ is wrong. His pages are however still invaluable source of information, no one has such extensive interactive graphs, ….

The data is not wrong, but the resistance is measured another way and I do explain how I measure it.
Recently I have been adding a resistance curve to many battery reviews, that value is supposed to more or less match what chargers measure, except charger are very imprecise.

I explain a bit about resistance/impedance here: Internal impedance of a battery

From BU-902: How to Measure Internal Resistance

Conclusion: HKJs computed resistance seems to be more relevant for BLFers at least in high, medium and low mode. In Turbo it may be DC impedance that counts most. I.e. the more heat the mode produces the more is DC relevant. So for sustained lumens output AC should be the most realistic meter.

Well gentlemen, let´s dive little bit deeper into the ocean of resistance measurements of the Li-ion cells.

I agree that different methods can give different results. So it is important to have the same measurement methods. I also agree that DCIR is more relevant, for various reasons.

ACIR 1 kHz (also called impedance) measurement is usually no problem as there is lot of dedicated equipment like SM8124A, YR1030, or YR1035 which are not expensive and are easily accessible and accurate enough.

Situation with DCIR measurement is different, people usually don´t use unified method, some even don´t distinguish between ACIR 1 kHz and DCIR.

I measure DCIR according INTERNATIONAL STANDARD IEC 61960-3:2017, paragraph 7.7.3 "Measurement of the internal DC resistance".

Full name of this standard : Secondary cells and batteries containing alkaline or other non-acid electrolytes – Secondary lithium cells and batteries for portable applications – Part 3 : Prismatic and cylindrical lithium secondary cells, and batteries made from them

Frankly, I haven´t found more suitable standard for Li-ion cells measurements than IEC 61960-3.

Description of the DCIR measurement :

7.7.3.1 Measurement

The battery shall be discharged at a constant current of I1 = 0,2 It A. At the end of a discharge period of 10 s ± 0,1 s, the discharge voltage U1 under load shall be measured and recorded. The discharge current shall then be immediately increased to a value of I2 = 1,0 It A and the corresponding discharge voltage U2 measured under load and recorded again at the end of a discharge period of 1 s ± 0,1 s.

All voltage measurements shall be made at the terminals of the battery independently of the contacts used to carry current.

The internal DC resistance, RDC, of the battery shall be calculated using the following formula :

RDC = U1 – U2 / I2 – I1 (Ohm)

Where I1 , I2 are the constant discharge currents U1 , U2 are the appropriate voltages measured during discharge.

Actual measurement was done in 3 consecutive pulses with 30 s pause for cell cooling down :

Voltage drop measured at the cell :

Samsung 35E DCIR is 30,1 miliohm at 3,82 V and ambient temperature 25 ± 1°C. ACIR at the same conditions is 20,3 miliohm (measured with YR 1030)

Equipment : DC electronic load Maynuo M9712, DMM GW Instek GDM-8351, 4-wires testing fixture BF-2A, thermometer Omega HH520 with K thermocouple.

DCIR of the cell is changing through SOC course, especially at low SOC, therefore it´s important to mention the voltage in the protocol. DCIR is temperature sensitive, so ambient temperature reading is also important to display.

True 4-wire measurement is must for both ACIR and DCIR measurement.

Bellow you can compare HKJ´s values with results measured according IEC 61960-3 standard :

Docware thanks for posting data with excellent description of used DCIR measure method. :+1:

The question is if the precise absolute value of DC (AC)IR is crucial for you. I think that in many cases there is more useful direct comparison of the cells measured under the exact same conditions. So HKJ results seems to me to be consistent, maybe they are lacking a better resolution to show for example that LG M36 have slightly lower DCIR than MJ1. And of course you never avoid some accidental errors and so it is always good to compare your measured values with others. For example I have continuously measured for “brand new” Sanyo GAs initial DCIR of about 45 mOhm. It is interesting that it drops after few cycles down to ca 38 mOhm and then start to rise agin.

From my point of view the most useful information is about the trend of internal resistance rise during use/cycling of the cell. But of course this information is unavailable in most cases and so you must perform own measurement.

Pajda, thank you for your kind words. And thank you for the DCIR 10s description in other thread.

Unfortunately ZKE testers are not convenient for DCIR measurement (measurement lags and unsufficient resolution)

HKJ data - you are right, there is some weird consistency here. :-)

Anybody who understand DCIR measurement knows that such method, resolution and results is not OK. Especially for HKJ, as his renowned pages are in my opinion probably the most respected pages regarding the Li-ion cells testing.

For example such measurement is not possible to use for tracing of the cell resistance trend during use or cycling/aging. Temperature measurement is also little bit issue, but not so much as internal calculated resistance.

Frankly, I am surprised that nobody is using IEC 61960-3 standard for DCIR.

Yes, it´s always good to have possibility to compare measurement, therefore would be nice to have all using one standard method, which obviously is not the case today.

I had similar issues with Sanyo GA and 20700B DCIR values like you.

Bellow is a new table with added SONY VTC6 and Sanyo 20700B. As all cells were on the table, I have made 10s DCIR measurement to compare both methods, IEC 61960-3 and 10s DCIR and added ACIR to have all data gathered.

Docware, it is not only kind words but you are really doing great job. I mean both your explanation of technical problem and using the proper techniques to measure should be a guide for us.

I thing that the main reason why almost nobody is using IEC method is that it requires “better” equipment than average of common DIYers have. Most of common battery chargers and ZKE battery testers cannot do this measurement. The problem is also that almost all 18650/21700 cell manufacturers does not show the IEC value in their datasheets :frowning: . Also there is a question which I mentioned earlier if it is “a must have” information for you to know precisely measured value under one of the artificial tests. I think that for some people it is enough to know only the absolute difference for comparison. Like “the M36 have slightly better DCIR than MJ1” and so in this case even HKJ measured values do the job. But again it always will be better to work with precisely measured values under standardized tests.

By the way, your 10s DCIR values are almost identical to mine (just for the info, resting voltage for my “50% DoD” is ca 3.7 V in average). Great thanks for this check :+1:

I was conscious that cells are not exactly on 50 % SOC, but wanted to try 10s method knowing that in this SOC zone is DCIR nearly not changing. One thing is speculation about something, but real testing and measurement is usually more interesting and valuable.

IEC - frankly, I would expect that at least professionals like you or HKJ use professional IEC methods. Both of you have equipment needed to do the job.

Question of „must have“ precisely measured values : it is natural to strive to have real precise values. Especially for your aging testing is knowing correct precise DCIR values very interesting to evaluate the trend, isn´t it ?

HKJ´s calculated internal resistance values are no acceptable, nor the method. Maybe 10 years ago, but not now, in 2019.

In the meantime I analysed some measurements and have made conclusion that even IEC 61960-3 standard method is not correct for true DCIR measurement.

When we analyze zoomed DCIR 10s measurement, we learn that biggest voltage drop happens in first 0,2 s, where we measure 35,53 miliohm. Than is voltage slowly falling thanks to capacity discharging and DCIR value is accordingly slowly raising. At 1s we measure DCIR 37,94 miliohm.

But, isn´t DCIR meant as pure resistive constituent of cell internal resistance ?

At zoomed IEC measurement we can see similar course of the voltage drop - DCIR at 0,2s is 34,67 miliohm, at 1s 36,91 miliohm :

So, in my opinion, TRUE value of DCIR is here measured let say at 0,3 or 0,4 s after voltage drop start . In my measurement is voltage logged at speed 10 measurements per second. Would be nice to have DMM with higher measurement speed to confirm that at about 0,3 - 0,4 s is voltage falling stabilized and other dropping is only result of the capacity discharging.

My method was defined in 2011 when I started doing these tests and I have not change it since to get comparable result for all the years. The method was designed to compare battery at a fairly high load and still do that
For me it was also important to make the test automatic, i.e. moving batteries between different types of test equipment was not an option if I wanted to run this for a long time. Lets see in 7 years if you are still measuring IR for many different batteries and posting them and if your values are still comparable to the ones you post today.

Well HKJ, thank you for your candidness. You are right, I measure IR now completely different way than last year. :-)

I understand that there is constant development when you are measuring, still something new, different. And understand that is difficult to cover up these changes in such quantity of your measurement. But quantity should not outweigh quality.

Your work is higly appreciated in the Li-ion cells testing world, which is obligate. From your answer I understand that you know that your IR measurement is not perfect as should be.

Though when 4-wire measurement was not comply with the requirements, you was able to do upgrade :

https://lygte-info.dk/info/batteryTestImprovement%20UK.html

Lately I have included another IR measurement in my reviews, it do not follow the ISO, but gives result close to what chargers are supposed to measure (But don’t due to contact resistance), except I do it for a full discharge cycle.

The change to four terminal was a minor change. It made high current measurement independent of connection resistance, but did not change low current measurement. Anyway many are still using two terminal measurements.

Generally is true 4-wire measurement considered as a basic presumption for such measurement. You are right, some people still don´t understand that presumption and use 2 wire or pseudo 4 wire measurement.

Another IR measurement - if such another measurement is for example test of Samsung INR18650-36G 3350mAh (Purple) than we can choose from two contradictory results :

1) Calculated internal resistance : 0,08 ohm

2) Internal resistance in the range cca 31 - 36 miliohm

Which one do you recommend ?

By the way, chargers are known for wrong IR measurements.

The results are not contradictory, but simply shows the result with two different methods, that are not supposed to give the same result.

The curve will be close to the ISO value, but slightly higher, because I measure over longer time and starts from zero current.

Yes, but lots of people use the measurements anyway and once in a while a charger measures mostly correct, especially if you know how to use the measurement.

OK, just want to finish my DCIR exercise here :-)

All measurements were done on the same LG MJ1 cell at nearly identical conditions (voltage, temperature)

DCIR at 1A discharge :

DCIR at 2 A discharge :

DCIR at 3 A discharge :

Summary table

My conclusion is that at time 0,3 s we measure true DCIR value of the cell, further is result gradually affected by capacity discharge, however results of all 5 measurements from 0,3 s are solidly consistent.

Basic assumption for DCIR measurement : 4-wire measuring, precise discharge equipment with accuracy 0,1% or better, voltage measurement resolution 0,1 mV or better, logging interval 0,1 s or better.