Author Topic: Battery Impedance Measurement - ESR  (Read 6084 times)

0 Members and 1 Guest are viewing this topic.

Offline Jay_Diddy_BTopic starter

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Battery Impedance Measurement - ESR
« on: October 25, 2020, 01:46:07 pm »
Hi group,

In another thread, forum member ogden asked about modifying my '100 kHz ESR Meter Adapter' to measure the ESR of batteries or cells. This is possible because my ESR meter is able to measure charged capacitors, and therefore batteries, without discharging them.

Link: https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg3294324/#msg3294324

Is there any interest in a 1 kHz version of the ESR adapter to measure cells or batteries?

Regards,
Jay_Diddy_B
 
The following users thanked this post: BravoV, ogden, gnavigator1007, Jacon, sandalcandal

Online BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: Battery Impedance Measurement - ESR
« Reply #1 on: October 25, 2020, 01:48:20 pm »
YESSSS !!!!!

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #2 on: October 25, 2020, 05:33:20 pm »
I would love to have main function of Fluke BT510 for a cheap. Anybody else interested? :)
 

Offline bob91343

  • Super Contributor
  • ***
  • Posts: 2675
  • Country: us
Re: Battery Impedance Measurement - ESR
« Reply #3 on: October 25, 2020, 08:16:02 pm »
Why not just use a coupling capacitor of known ESR and subtract?
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: Battery Impedance Measurement - ESR
« Reply #4 on: October 25, 2020, 10:54:50 pm »
Would be interesting if you can determine SoC and SoH of a battery by measuring its ESR over a discharge cycle, and over its life.
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #5 on: October 26, 2020, 02:30:56 am »
Using real component of impedance for battery assessment has been around for decades, but its practical usefulness requires some care to distil out what is being sought.  A few companies jumped on the bandwagon to make a product, but few testers are able to extract a valid measurement that has any benefit.

One hassle is if the tool has to connect to a live battery cell in a power system.  The power system can be quite noisy, dependant on power supply and load equipment and level of charging and discharging.  The cell or monobloc under test is also paralleled by a complex impedance configuration that may include other strings, as well as power supply and loads.

The DUT temperature, and its chemistry and capacity, all affect an impedance measurement as well as SOC.

The main industrial application was for telecommunications lead-acid batteries, and related to trending over time to indicate when end-of-life replacement was appropriate, given that other measurement schemes like performing a capacity measurement cycle are invasive and expensive.  But trending is a long term measurement campaign, and doesn't give any valid initial or early measurement value except for pretty degraded cells and only when comparing to a known benchmark value (somewhat similar to comparing ESR from different eCaps).

The increasing dominance of Li-Ion and the ability of its bms to adequately monitor health and soc from just cell voltage has meant that impedance measurement is not applicable for that growing field of application.
« Last Edit: October 26, 2020, 02:35:47 am by trobbins »
 
The following users thanked this post: Someone

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #6 on: October 26, 2020, 04:05:43 pm »
The increasing dominance of Li-Ion and the ability of its bms to adequately monitor health and soc from just cell voltage has meant that impedance measurement is not applicable for that growing field of application.
I wonder how BMS from just cell voltage can measure cell impedance.
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #7 on: October 26, 2020, 04:55:12 pm »
It doesn't.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #8 on: October 26, 2020, 05:15:31 pm »
It doesn't.
Bingo. That's my point. Generic public usually are fine using trial & error approach, some needs multimeter, others - spectrum analyzer, VNA or ESR meter.
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #9 on: October 26, 2020, 06:00:59 pm »
Ogden, are you commenting on Li-Ion and inferring that a BMS with battery voltage and temperature monitoring can't adequately represent the SOC of the cells?
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #10 on: October 26, 2020, 06:04:16 pm »
Ogden, are you commenting on Li-Ion and inferring that a BMS with battery voltage and temperature monitoring can't adequately represent the SOC of the cells?
No. I am implying that you do not know or care about difference between high current capability and normal or deteoriated cells. They all can have SOC that looks normal for BMS but behave differently under high load.
 

Offline Jay_Diddy_BTopic starter

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: Battery Impedance Measurement - ESR
« Reply #11 on: October 26, 2020, 06:55:38 pm »
Hi,
I have posted the modification required to my ESR Meter Adapter to 1 KHz for testing cells in this message:

https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg3296986/#msg3296986

I am not a battery chemist, so I don't know how to relate ESR to SOH and SOC. I think that these have to be compared against 'learned' measurements.

Regards,
Jay_Diddy_B
 
The following users thanked this post: BravoV, ogden

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #12 on: October 26, 2020, 07:48:20 pm »
I am not a battery chemist, so I don't know how to relate ESR to SOH and SOC.

Voltmeter is better tool for SOC measurement - if coloumb-counting BMS no avail. When you know ESR of new battery, you may estimate SOH by level of ESR degradation. Main use of ESR meter - sorting cells for battery packs. Healthy battery pack needs cells of similar capacity *and* ESR, thus current handling capacity.
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #13 on: October 26, 2020, 11:27:05 pm »
Ogden, for clarification, can you confirm your comments are just about lead-acid?

Certainly for lead-acid, the methods by how SOC and health/degradation can be estimated are a lot more complex, and very dependant on application.  Adding ESR monitoring to a suite of monitoring tools that include voltage, temperature, load cycling, and maintenance cycling and OCV testing, can be a benefit - but extracting predictions whether that be existing SOC or service life or capacity degradation can be complex, especially if the load profile during life is ad-hoc and varied - and not all batteries are premium quality, and so many users often just see the result of, or have to manage, poor outliers.

ESR monitoring is a complex task.  It can assume the user knows the ESR degradation profile of a particular battery product over its service life, and that the chosen test frequency aligns with the extracted ESR value following a known ESR variation characteristic with SOC - which can be quite dependant on cell capacity/size and construction technique.
« Last Edit: October 26, 2020, 11:34:11 pm by trobbins »
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #14 on: October 27, 2020, 02:21:53 am »
Ogden, for clarification, can you confirm your comments are just about lead-acid?

Not only. Using ESR meter you can sort your Alkalines, NiMH's and Li-Ion cells as well.

Quote
Certainly for lead-acid, the methods by how SOC and health/degradation can be estimated are a lot more complex, and very dependant on application.  Adding ESR monitoring to a suite of monitoring tools that include voltage, temperature, load cycling, and maintenance cycling and OCV testing, can be a benefit - but extracting predictions whether that be existing SOC or service life or capacity degradation can be complex, especially if the load profile during life is ad-hoc and varied - and not all batteries are premium quality, and so many users often just see the result of, or have to manage, poor outliers.

 :palm: I said already: voltmeter is better tool for SOC estimation. You are mistaken what ESR meter adapter is about. It is just slightly better tool than multimeter in DC current mode measuring short circuit current of the cell. ESR meter may come to rescue for those cells/batteries which are too powerful and may destroy DMM (fuse). I said already - it is for ESR estimation and comparing, sorting cells for battery packs. Hopefully I do not have to repeat more
« Last Edit: October 27, 2020, 02:27:42 am by ogden »
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #15 on: October 27, 2020, 03:59:15 am »
Ogden, you may not appreciate that there has been at least 2 decades of research and practical experience with using 'impedance measurement' of lead-acid batteries in the international telecommunications community.  Quite a few published papers, and quite a few commercial devices.  The measurement they use is effectively an ESR measurement at a particular frequency (circa 400Hz to 1kHz range).  That experience and the devices are pretty much unknown of for DIY and many other commercial users, so I can appreciate you are not aware of its application.

I would be very cautious about stating that one type of tool is better than another for some form of parameter measurement.  I wouldn't say a voltmeter is a better tool for SOC measurement - as a blanket statement - as every 'result' (like what is the SOC now) needs to be related to the battery tech and battery history and application (temperature; discharge rate; allowed end-of-discharge voltage).
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: Battery Impedance Measurement - ESR
« Reply #16 on: October 27, 2020, 10:18:42 am »
Note that a BMS could measure ESR and impedance over some frequency range by monitoring ripple current caused by the loads (BMS almost always measures pack current to begin with) and simultaneously, the cell voltage. Such measurements are quite limited when the BMS cannot itself act as a dummy load (which is obviously a mostly unwanted and expensive feature).

I could think this would be quite practical for low-frequency (near DC, say, below a Hz) ESR under heavy accelerations. R = dV/dI, and the resulting number is highly useful calculating power loss and heating in pack, to limit available power in case some of the cells is going worse than the others, preventing excessive heating. I'm sure some EV battery systems do this.

For li-ion, AC ESR and complex AC impedance at 10kHz, or over frequency sweeps, are discussed in academy, and used as tools by the cell manufacturers to control quality, but every time someone claims they are super useful for battery system designers (those who buy the cells and construct packs) or end-users, I have to ask, how, and have not seen any kind of answer to that question so far. When I test cells, I go for DC resistance (measured with pulses below 0.2Hz) because that gives some direct insight about power loss and heating which are both important parameters, whereas AC Z only correlates with the information I want.
« Last Edit: October 27, 2020, 10:22:02 am by Siwastaja »
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #17 on: October 27, 2020, 05:07:52 pm »
Ogden, you may not appreciate that there has been at least 2 decades of research and practical experience with using 'impedance measurement' of lead-acid batteries in the international telecommunications community.  Quite a few published papers, and quite a few commercial devices.  The measurement they use is effectively an ESR measurement at a particular frequency (circa 400Hz to 1kHz range).

This is what Jay Diddy is offering here - ESR meter for batteries which works at (surprise surprise) 1KHz frequency.

Quote
I wouldn't say a voltmeter is a better tool for SOC measurement - as a blanket statement - as every 'result' (like what is the SOC now) needs to be related to the battery tech and battery history and application (temperature; discharge rate; allowed end-of-discharge voltage).

Voltmeter is better tool for SOC estimation compared to ESR meter. Live with that. Obviously one measuring voltage shall know specs of the battery - it's charged/discharged voltages as minimum. Yet this exact knowledge will not help you measure SOC using ESR meter. Try to convince me how ESR meter is good at SOC measurement when I give you one abused 2.5AH NiMh cell which can take max 2AH charge and another fresh 2AH NiMh cell, both fully charged up-to 2AH. Abused cell will have way worse ESR compared to fresh one while SOC for both is fully charged!
« Last Edit: October 27, 2020, 05:13:42 pm by ogden »
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #18 on: October 28, 2020, 12:54:40 am »
For li-ion, AC ESR and complex AC impedance at 10kHz, or over frequency sweeps, are discussed in academy, and used as tools by the cell manufacturers to control quality, but every time someone claims they are super useful for battery system designers (those who buy the cells and construct packs) or end-users, I have to ask, how, and have not seen any kind of answer to that question so far.
One application for large Li-Ion storage systems has used Li-Ion modules connected in series in a rack format, and then multiple parallel strings/racks, for relatively high DCV in to large bi-directional inverters for grid connect.  The Li-Ion modules are all racked and very convenient for such an application that extends in to the MW and MWh range.  The BMS in that situation just uses cell voltage and temperature, and string current, monitoring/logging and parameter prediction (ie. SOC), and is based on each module containing the cell monitoring and equalising circuitry, and a common total system PC manager.

I have no doubt that research would have been done on the impedance spectrum and how that changes with SOC and age - that is fundamental to how impedance measurement can be distilled down to a single frequency measurement scheme, and what to expect from that measurement scheme.  That was the situation for lead-acid VRLA by telecom companies who were after monitoring and maintenance improvements on their extensive use of battery storage, and which voltage monitoring alone was not providing sufficient assurance.  In the long run, most telco's deemed the old ways of monitoring and cycle testing was too costly compared to a simpler periodic replacement of a total dumb battery population, and only buying from premium audited manufacturers, such that there fault rate was considered low enough.
« Last Edit: October 28, 2020, 01:00:04 am by trobbins »
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Battery Impedance Measurement - ESR
« Reply #19 on: October 28, 2020, 06:00:21 am »
For li-ion, AC ESR and complex AC impedance at 10kHz, or over frequency sweeps, are discussed in academy, and used as tools by the cell manufacturers to control quality, but every time someone claims they are super useful for battery system designers (those who buy the cells and construct packs) or end-users, I have to ask, how, and have not seen any kind of answer to that question so far.

Li-Ion cell impedance is specified at 1KHz. Check some datasheets of popular 18650's, the same about specs of rechargeable NiMh. I am sorry that you are so uninformed about how useful and widespread battery ESR measurement is in certain industries. Search web/YT for "fluke battery analyzer", #1 tool for hi capacity UPS technicians. No wonder there are literally no consumer and DIY battery ESR devices. - Measuring < 1Ohm impedance is not straightforward nor low cost, one needs to understand and interpret results too. Generic public do not need to balance cell packs anyway, many just discard whole pack or whole tool  instead of just one cell.

Yes, one *can* do many things - measure battery impedance using pulsed load, he can check small cells using short circuit current measurement too. Idea here was to add battery function to already existing tool, make it even more useful.

Why don't you guys use your heads thinking about how to make it work, not about how it may fail.
« Last Edit: October 28, 2020, 06:02:12 am by ogden »
 

Online trobbins

  • Frequent Contributor
  • **
  • Posts: 754
  • Country: au
Re: Battery Impedance Measurement - ESR
« Reply #20 on: October 28, 2020, 06:45:39 am »
Ogden, the schematic and comments by Jay_Diddy_B for 1kHz measurement are available.  Are you going to make it, along with the required Kelvin style probes, and road test it on batteries and report back?

I was able to set up the ESR meter for reasonable accuracy below 10 milliohm using Kelvin probes suitable for discrete part testing.  For in-situ battery applications, the Kelvin probes are typically robust spring loaded spikes (as per bed of nails pcb test jig) so that lead (Pb) posts and tinned copper intercell connectors can be probed successfully through oxide layers.  Older lead-acid battery terminals and links can sometimes exhibit a hard oxide layer (due to electrolyte migration) that can be difficult to probe and often requires initial scrapping.  I have measured many many banks of lead acid battery with an automated impedance and voltage measurement tool, just as much as with a standard voltmeter.
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: Battery Impedance Measurement - ESR
« Reply #21 on: October 28, 2020, 08:08:09 am »
For li-ion, AC ESR and complex AC impedance at 10kHz, or over frequency sweeps, are discussed in academy, and used as tools by the cell manufacturers to control quality, but every time someone claims they are super useful for battery system designers (those who buy the cells and construct packs) or end-users, I have to ask, how, and have not seen any kind of answer to that question so far.

Li-Ion cell impedance is specified at 1KHz. Check some datasheets of popular 18650's, the same about specs of rechargeable NiMh. I am sorry that you are so uninformed about how useful and widespread battery ESR measurement is in certain industries. Search web/YT for "fluke battery analyzer", #1 tool for hi capacity UPS technicians.

Yes, I know all this, I know how to measure this. The question was, unless you are a cell manufacturer, what do you do with the number once you have obtained it? "I was tasked to measure it" doesn't count.

My experience is that quite crappy EOL cells can show acceptable 1kHz impedance, and you need to measure it over the whole SoC, and/or DC ESR, capacity, and possibly self-discharge current to decide if the cell has gone bad.

Evaluating SOH reliably would require some more complexity than just measuring 1kHz AC impedance and using some good/bad thresholds. As trobbins has explained above, lead acid industry has come with some quite complex algorithms estimating the SOH. But li-ion works differently and needs different algorithms. SOH estimation has been widely discussed in academy, I have skimmed through a lot of papers.

The reason why 1kHz AC impedance doesn't correlate with real-world requirements too well is simple: the cells have capacitor-like behavior due to their similar construction to an ultracapacitor. Ion transfer takes over at lower frequencies (well below 1kHz), but that is what is actually needed - if the load was happy taking just 1ms pulses, they would have used capacitors, not batteries. Real loads need DC power from the battery, and the ion transfer is supplying it. Even further; most real loads have enough bypass capacitance on their inputs that they pull insignificantly low currents around 1kHz and beyond.

But when we measure 1kHz AC Z, we are measuring the capacitance of the cell.

This is what makes capacitors and batteries so different. A capacitor is just a capacitor, but a battery is more equivalent to a voltage source with a series resistor, PLUS a capacitor with another (possibly smaller) series resistor in parallel. Add inductance to the mix, and you'll get much more complexity to the ESR over frequency curve sweeps than sweeping capacitors. And it becomes hard to see which part of the curve is caused by just the capacitance of the cell, and which is relevant for evaluating the cell's capacity to supply DC power, which is all what finally matters.

So I opt for measuring capacity and DC ESR. Which requires doing a full charge-discharge cycle to a cell taken apart from the pack. Which is very tedious, but gives you the right answers you need.

Analysing SOH in-circuit is interesting, but far more complex than measuring 10kHz Z "because that's what industry does".
« Last Edit: October 28, 2020, 08:35:40 am by Siwastaja »
 

Offline Jay_Diddy_BTopic starter

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: Battery Impedance Measurement - ESR
« Reply #22 on: October 28, 2020, 08:51:03 am »
Hi group,

About a year ago I posted some measurements made using different techniques. You can find the measurements here:

https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg2798498/#msg2798498


I compared the HP4328A (very small signal) with a load step at 1kHz and 10Hz.

Today I also added some measurements of 18650 cells made with the 1 kHz version of the ESR Meter adapter.

Link: https://www.eevblog.com/forum/projects/esr-meter-adapter-design-and-construction/msg3299596/#msg3299596

Regards,
Jay_Diddy_B
 
The following users thanked this post: Siwastaja, ogden

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: Battery Impedance Measurement - ESR
« Reply #23 on: October 28, 2020, 01:23:08 pm »
Thanks.

To sum up to my best knowledge,

* EIS is used by cell manufacturers for quality control. This is a scan over a broad frequency range measuring complex impedance.

* DC ESR (as a curve set as a function of SoC, temperature, and ageing, not single datapoint) may be used by system designers for thermal & efficiency calculations because I^2R losses follow DC ESR. Getting reliable values require full discharge cycles, total PITA for trying to assess the condition of single cells built in packs.

* 1kHz AC ESR (real part of Z measured at 1kHz) - sometimes 10kHz - is reported in some datasheets for tradition. It can be extracted as a single datapoint from the EIS scan. Some small sweatshop style cell manufacturers may use only this for quality control instead of EIS because of the lower cost of measurement equipment. It seems to be widely relevant on fake science sites like Battery University, and obviously discussion forums, as well. Cheap (compared to EIS tools) meters exist to measure this. Can be measured in seconds using small test currents compared to DC ESR; even in-system. Ease of measurement is indisputable, but rectal temperature of the pack designer is easy to measure as well, it doesn't mean it's a very useful parameter.

Maybe 1kHz AC ESR is somewhat useful, maybe not; show me. Despite common Internet advice, it isn't some magical "this is how it's done professionally" word of God, but a single datapoint which carries limited correlation to the parameters we actually are interested in (namely those defining if the cell can be used safely in the application, at what energy storage capability, and at what power loss).

I have designed cycling systems for li-ion analysis because that obviously gives most insight to the parameters of interest, directly, with the cost of being slow and requiring bare cells removed from the packs. The latest has hardware support for EIS scans up to some 5kHz but firmware hasn't been written for that and there has been no interest for this feature by those paying for the work.

A low-cost EIS scanner might be very interesting, IMHO.
« Last Edit: October 28, 2020, 01:35:14 pm by Siwastaja »
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: Battery Impedance Measurement - ESR
« Reply #24 on: October 28, 2020, 01:45:45 pm »
For example, this EIS scan:
https://www.researchgate.net/figure/EIS-curves-of-lithium-ion-batteries-with-different-SOC-and-maximum-discharge-capacities_fig1_313830344

shows how cells 1, 2 and 3 give almost same (within 5%) Zre between 5kHz to about 1Hz but then the curves differ significantly, some 50-70% difference in Re below about 0.1Hz.

If you make the decision at Z_1kHz, you completely miss the fact they behave so differently at the frequency range actual loads use.


In this appnote,
https://www.analog.com/en/design-center/reference-designs/circuits-from-the-lab/cn0510.html

figure 8 explains the regions of typical EIS scan.

In an ultracap, parts marked "conductance and skin effect", SEI, and "electrochemical double layer" exist. In a li-ion cell, additionally, the part marked "mass transport" exists.

All these parts affect the performance in-application, and all except "conductance and skin effect" change as cells age.

This makes battery cells more complex to measure than capacitors, and warrants tools such as SEI scanners, which are professionally used, despite "Battery Universities" and internet forums giving the impression that 1kHz ESR meter is the sliced bread or even better.
« Last Edit: October 28, 2020, 01:55:22 pm by Siwastaja »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf