Internal battery resistance test tools

I have been actively working on a team releated prototype project (this applies to FRC approved batteries and other kinds of batteries) and find myself frequently needing to measure the internal resistance of batteries.

I was wondering if anyone has experience with the
All-Sun EM3610 meter and measuring battery internal resistance.

I need to be sure to account for the lead resistance on the battery so I find myself using 4 lead milliOhm meters measuring the resistance of the leads to the battery and subtracting that from a reading across the battery and the leads with the measurement from the charger. I’d like something more direct. Just hoping someone else can comment on the experience of using that product?

Well I just got that All-Sun meter.

I tested it against 1% tolerance: 200mOhm, 100mOhm, 10mOhm and 2mOhm resistors.

It read all values but the 2mOhm adequately and there it read 3mOhm on the 200mOhm scale.

I put it on a very lightly used FRC battery at near full charge and off charge for a few days and got 12.5mOhm to 13mOhm which sounds about right.

I’ll try some more batteries soon but this may not work well on my LiPo packs because they read 1mOhm per flat pack with 2 in series when charged.

1 Like

Awesome, thanks for testing this! That looks like a nifty tool.

Not nearly as accurate or nifty: We wrote some code in 2016 to calculate it at runtime:

The numbers that came back seemed reasonable (especially since they also included battery cables, connectors and main circuit breaker). But, never had much to cross-check it against besides the datasheet.

The CTRE Battery Beak has been used by quite a few FRC teams (mine included) for several seasons now. It measures internal resistance using DC loading, by a procedure that is similar to the one described in IEC 61951-1 (2005); that method is recommended for batteries that serve DC power loads.

The All-Sun tester that the OP referenced uses a different method, based on 1000 Hz measurements. That method is recommended for batteries that serve digital device loads.

For a little more depth on battery internal resistance test methods, look here.

For a lot more, maybe Mike or Omar from CTRE will weigh in.

Did the meter come with a manual? Does it indicate what the test current is?

What value did it give for each of the resistors you tested?

The concern is that the method used still includes the resistance of the two probes to the battery and at low internal resistance values, this can become a significant but inconsistent error.

This is a good way of measuring the source resistance (= battery internal resistance + cable resistances + connection contact resistances + breaker internal resistance). For FRC, this is probably at least as useful than just the battery internal resistance since one never uses the battery alone without all those other pieces and the resistances of many of those other pieces can change or degrade over time just like the battery internal resistance.

What is the current when the measurement is made? Is the current dependent on the components (loads) connected to the PDP? I am not a programmer and could not find what sets the current in your (very well documented) program but I could find where it is measured and used in calculations.

Is it possible for you to make this measurement many times, starting with tight bolted connections and a new SB connector then unplug and re-plug the SB connector before making the next measurement? The reason I am asking this is I have noticed some SB connectors where the silver(?) plating is noticeably/visibly worn off the contact areas and I was wondering what impact this has on the contact resistance. If it does have a significant effect, those aiming for the highest level of performance might want to crimp on new contacts periodically.

Sorry to Techhelpbb for taking this thread in a different direction.

FWIW, All-Sun’s product webpage for model EM3610 does not list lead/acid as one of the supported chemistries.

*


https://www.chiefdelphi.com/media/papers/3364 describes it in more detail.

At least for our purposes, there was nothing to “set” the current per say. We ran this while the robot was on the field doing it’s thing. We would track a short history of samples, and every loop analyze the sample set to look at the standard deviation.

If the robot was under the right conditions, we would use the voltage/current history buffers to do a best-fit line to calculate resistance. The criteria for “right conditions” was more than a 7A standard deviation of our measured current. (7 was kinda just a made up number, but was enough to ensure our V/I point cloud was spread out enough to make the information in the best-fit line meaningful). The slope of the best-fit line correlates to resistance.

Yup, I think i know what you’re saying, and it should be possible. This estimator was more designed to track degrading battery ability over the course of a match, but you could average out the estimates and try with a bunch of different input permutations (such as connector changes like you mentioned) to generate samples.

PM me for details, or feel free to comment on the paper thread if we want to move this conversation :slight_smile:

The meter manual is in not in English.
I’ll try to get one in English but there are a few things that could be the values you are asking about.

It’s got 2 wires to each probe hence the probe resistance is not much an issue. I tested by shorting the probes together and got zero Ohms. It’s likely it’s using the 4 wire method to measure the resistors but I can check that with another meter.

I don’t have the values from tests in front of me, but I should be back in my home state on Monday and I can post them. The resistor values are within 1% of the readings and I can provide the Mouser part numbers…for all except the 2mOhm where that value is really low on the 200mOhm range.

What’s interesting is they sell this on AliExpress as an automotive tester. I have to dig out my Battery Beak and see if the readings are different. However so far the readings for internal resistance I got with it from a 9.6V NiCd pack, 9.6V NiMH pack, 6.6V LiFePo4, 6.4V LiFePo4 and this FRC battery seem pretty close based on what the load circuits have been demonstrating.

Most important to me…these internal resistance readings are way closer than the internal resistance measurement function of my Turnigy Reaktor 250 charger which is only 2 wires and consistently off as much as 50-60% because of the leads to the battery. It only reads close on batteries with the short balance charging connector (makes sense cause it makes the battery leads really short).

My issue with the Battery Beak is the opposite. I don’t want to potentially do something bad to my Battery Beak testing battery chemistries that are not normally on FRC robots.

Battery beak 2.0 with June 2015 firmware
Versus
All-Sun Em3610

On AndyMark crimped 1foot 6AWG leads.
Measured at the connector.

Very lightly used battery Beak reads 17mOhm 4 times and 18mOhm once.
Rates battery at charged to 77% of capacity.

Em3610 tried 5 times as well 15.2-15.6mOhm with probe tips held firm against blades of connector by hand.
I got my lower reading before because with these probes I can get against the battery blades themselves removing the cable and connector as other sources of low resistance in the circuit.

I am charging that FRC battery on my Turnigy Reaktor 250 charger with a RadioShack 15A 13.8V power supply set at 0.5A while I am gone. The charger display says the battery is taking 160mA right now.

Probes touched together:


3 tests per setting
Range 20Ohm: 0.0 Ohm, 0.0 Ohm, 0.0 Ohm
Range 2000mOhm (2Ohm): 0.0 Ohm, 0.0 Ohm, 0.0 Ohm
Range 200mOhm (0.2Ohm): 0.0009 Ohm, 0.0012 Ohm, 0.001 Ohm = Average error 0.00103 Ohm

Hence on the lowest value resistance I have below of 0.002Ohm this is going to read incorrectly.

Passive resistance measurements:


Mouser Part #: 588-650FPR002E
  Maker:        Ohmite
  Resistance:   0.002Ohms (2 mOhms)
  Tolerance:    +/-1% = +/-0.00002Ohms
  Power rating: 5Watts
  Package:      Through-hole
  Datasheet:    https://www.mouser.com/ds/2/303/res_60_4term-1265426.pdf
  Qty to test:  1

  Readings:
    Resistor A:
     Range 200 mOhm
     Attached leads by hand pressure
     3 readings: 2.5-2.6 mOhm, 2.6-2.7 mOhm, 2.5 mOhm

Mouser Part #: 588-12FR010E
  Maker:        Ohmite
  Resistance:   0.01Ohm (10 mOhms)
  Tolerance:    +/-1% = +/-0.0001Ohm
  Power rating: 2Watts
  Package:      Axial
  Datasheet:    https://www.mouser.com/ds/2/303/res_10-1265495.pdf
  Qty to test:  2

  Readings:
   Resistor A:
     Range 200 mOhm
     Attached leads by hand pressure
     3 readings: 10.2 - 10.3 mOhm, 10 mOhm, 10.2 mOhm
   Resistor B:
     Range 200 mOhm
     Attached leads by hand pressure
     3 readings: 10.1 mOhm, 10.2 mOhm, 10.1 mOhm

Mouser Part #: 588-12FR100E
  Maker:        Ohmite
  Resistance:   0.1Ohm (100 mOhms)
  Tolerance:    +/-1% = +/-0.001Ohm
  Power rating: 2Watts
  Package:      Axial
  Datasheet:    https://www.mouser.com/ds/2/303/res_10-1265495.pdf
  Qty to test:  2

  Readings:
   Resistor A:
     Range 200 mOhm
     Attached leads by hand pressure
     3 readings: 101.8 mOhms, 101.7 mOhms, 101.5mOhms
   Resistor B:
     Range 200 mOhm
     Attached leads by hand pressure
     3 readings: 101.6 mOhms, 101.2 mOhms, 101.2-101.3 mOhms

Mouser Part #: 588-13FR200E
  Maker:        Ohmite
  Resistance:   0.2Ohms (200 mOhms)
  Tolerance:    +/-1% = +/-0.002Ohm
  Power rating: 3Watts
  Package:      Axial
  Datasheet:    https://www.mouser.com/ds/2/303/res_10-1265495.pdf
  Qty to test:  2

  Readings:
   Resistor A:
     Range 2000 mOhms (exceeds 200 mOhm range)
     Attached leads by hand pressure
     3 readings: 198 mOhms, 198 mOhms, 198 mOhms
   Resistor B:
     Range 2000 mOhms (exceeds 200 mOhm range)
     Attached leads by hand pressure
     3 readings: 199 mOhms, 199 mOhms, 199 mOhms

I am definitely not making these readings more stable by simply pushing the leads into a non-conductive surface with my hands, but in fairness, the meter comes without clips and this is what will happen if you push the probes onto the connections.

± 1.3% variation in the readings from the Em3610 is pretty impressive.

The difference between the Beak and the Em3610 is only just over 10% which is also pretty good correlation considering they are probably measuring using different current levels.

Not just different current levels. They are using different technologies.

Beak uses DC ΔV/ΔI

EM3610 uses 1000Hz AC

As others have noted it’s also likely because of the different methods of testing internal resistance between the Battery Beak and the EM3610.

I don’t have any more time to play with this tonight. I’ll try to connect my oscilloscope across a passive resistor under test with the EM3610 and see if I can determine what it’s doing.

I don’t know if merely placing a fixed resistance across the battery is a good test when one considers that the ESC are chopping the DC to power the motors who’s windings are clearly not passive resistive loads. However the measurements produced don’t deviate so much that I am as confused as I was when I was trying to reconcile the internal resistance measurements from the Turnigy Reaktor 250 and it was doing stuff like this:


Battery            Voltage     Reaktor 250     EM3610      Approx Error % (EM3610/Reaktor250)
FRC lead acid      13.8        92mOhm          12.7mOhm    63% (accounted for readings of wires for 45 mOhms)
LiFePo4 custom     6.4         139mOhm         75.7mOhm    53%
LiFePo4 Zippy RC   6.6         56mOhm          29.5mOhm    53%
DeWorld NiMH       9.6         405mOhm         251mOhm     63%
NiCd pack          9.6         410mOhm         220mOhm     53%

Reaktor 250 charger readings are 1.8x - 1.9x higher or 1.6x - 1.7x higher than the EM3610 readings.
The increased error tracks circuits with very different wire gauge sizes in the path to the battery assembly.

Basically I’d take these readings over the 2 wire readings from the Turnigy Reaktor 250 charger across the board.
The Reaktor only behaved itself with the balance charging connectors in use.

The seller has sent me the English manual for the EM3610.
It’s attached here.

I found these 3 things interesting:

Page: 7
“For measurements in the 200m range, set the range switch in the 200m range position and short the two test probes, the display will show a reading. This reading must be subtracted from all the measurements in the 200m range.”

Notably this means the resistance reading of the 2 mOhm passive resistor is now too low, instead of too high.
The rest of the measurements from the passive resistors now seem just a little more accurate on that 200m range.
Except for the 200mOhm resistor which I measured on the 2000m range.

This also then applies to the FRC battery measurements which were done on the 200mhm range.
This means at the battery terminals I was reading:
12.5 mOhm - 1 mOhm on average = 11.5 mOhm
to
13 mOhm - 1 mOhm on average = 12 mOhm

Which for a battery where we commonly believe the internal resistance averages 11 mOhm that’s very close.
On the end of the AndyMark lead kit the measurements then become:
15.2 mOhm - 1 mOhm = 14.2mOhm
15.6 mOhm - 1 mOhm = 14.6mOhm

Manual says the error for the EM3610 could be 10% so let’s take the average of 14.4mOhm.
14.4 mOhm + (14.4 mOhm × 10%) = about 15.8mOhm if the EM3610 is low

The Battery Beak manual says it has an error of about +/- 1 mOhm no mention of rounding.
So 17 mOhm might be 16 mOhm if the Battery Beak is high.

So pretty close between the EM3610 and Battery Beak.

Page: 8
“Do not apply a voltage higher than 50V between the test probes; otherwise the meter will be damaged.”

Page: 9
“To avoid damage to the meter, do not apply any AC signal between the test probes.”

Moving on…

I was curious about the 11 mOhm average I have seen used for FRC batteries (look around ChiefDelphi it’s come up in several posts over the years). The InterState batteries on AndyMark’s website right now say 12 mOhm-15 mOhm internal resistance I don’t see if they mention how they measured that. The SLA1116 datasheet link on Andy’s website is broken right now I got that from the PowerPatrol datasheet below that.

The FRC battery I bought from AndyMark that I have been testing is an ES17-12.
The datasheet is attached to this post.
It clearly says: “Internal Resistance (at 1kHz): Approx. 12 mOhm”

I looked at the datasheets for all these other FRC compatible batteries listed here:
https://www.chiefdelphi.com/forums/showpost.php?p=1396978&postcount=1
I found the 11 mOhm measurement only on the Yuasa NP18-12 datasheet as attached.
None of the other datasheets mention more detail other than tested at temperature 25C (my shop is 68F-71.5F as measured by AcuRite 06044 sensors and graphed on their App on my Android phone and I can set them up to read in Celsius but internally they read in Fahrenheit so just do the conversion: 20C-22C).

Generally if you test the internal resistance of the battery when it’s just off the charger and still settling you’ll read as high as 19 mOhm on the Battery Beak with this same battery I have been testing in these posts. So when it completely settles and hits room temperature (for me) is when the Battery Beak reads 17 mOhm. So if my room temperature is a little lower than 25C that might account for reading a little less than 12 mOhm they probably measured at 25C.

Basically I’d love to know what methods and tools all these other battery manufacturers were using to measure these batteries and post those results on their datasheets.

Not going to have time for any oscilloscope readings tonight either.

EM3610+ENGLISH.pdf (149 KB)
ES17-12.pdf (343 KB)
NP_18_12_DataSheet.pdf (261 KB)

It was tough to measure the output voltages from the EM3610 meter on the 100 mOhm resistor A from the last post.

Attached is the best I could do with my Rigol DS2072A. Notice the cursors read 1kHz between the pulses but they are very faint with a ton of background noise at such a high sensitivity. I confirmed that this signal disappears when the meter is removed though the noise remains. Also notice this was done with a Rigol 10X probe and on Channel 1 and it’s been adjusted to 10X so no need to divide or multiply. The voltage scale is true. I confirmed it with a battery just to be sure. I was not able to see the impact of the meter at all with the FRC battery attached. Maybe with a AAA battery but needless to say I don’t think this signal is much an issue for the battery to deal with voltage wise. In the picture the reading for that resistor can be seen on the EM3610 display.

I’ve also attached a picture of all the batteries I’ve mentioned in previous post, including a LiPo battery I haven’t tested with the EM3610 because it’s internal resistance is extremely low. It reads: 7.0 mOhms on the EM3610 meter at the power connector end. That same battery reads as 2 mOhms per pack x 2 from the balance connector on the Turnigy Reactor 250 internal resistance measurements. That extra 3 mOhms seems fairly reasonable.

When read on the same power connector I used for the EM3610 measurement I got 3 repeated 37 mOhm measurements on that Turnigy Reactor 250 charger internal resistance measurement. Which is 5.3 times too high following the same pattern I have been seeing with an even greater severity.

Hopefully tomorrow I will have access to at least one more unused FRC robot battery and maybe another Battery Beak to grab some confirming measurements.




I got to test some Team 11 FRC robot batteries tonight with the EM3610, my Battery Beak, their Battery Beak 2.0 and their Battery Beak 1.0.

So firstly I discovered that their Battery Beak 1.0 always reads 0 mOhms internal resistance regardless of the battery under test. Not sure if that’s a quirk of the older version or it somehow got damaged. It still rates the percentage of the battery but I’d take that with a grain of salt. I did notice that with the both additional FRC batteries and all 3 Battery Beaks I got consistent readings for each unit but they were off of each other. For example on the brand new, never used, Genesis battery my Battery Beak rated it 95%, Team 11’s Battery Beak 2.0 rated it 98% and their Battery Beak 1.0 rated it 94%. During the tests the battery was not moved any great differences or experienced any sudden temperature changes.

So this is really going to be comparisons of Team 11’s Battery Beak 2.0 internal resistance measurements with my Battery Beak 2.0 and the EM3610 with some additional batteries.

The one battery provided to me by Team 11 was an ES17-12 which is a year newer than my ES17-12 and was used as part of their competition battery set last year. It had an AndyMark crimped Anderson connector tail on it that was about 10" long.
The other battery is a brand new Genesis NP18-12B and it had an AndyMark crimped Anderson connector tail on it that was about 12.5".

Measurements with EM3610 on the battery blades directly, using the 200m scale.

Meter turned off then on with 200m range selected.
Probes shorted together gave a read of 1.6 mOhm when should be subtracted from the subsequent readings.
On the Genesis battery we got: 15 mOhm, 15 mOhm, 15 mOhm, 14.9 mOhm, 14.9 mOhm and 15.3 mOhm.
Averaging: 15.0 mOhm from which we substract the 1.6 mOhm and get: 13.4 mOhm.
EM3610 turned off.

Meter turned off then on with 200m range selected.
Probes shorted together gave a read of 1.4 mOhm when should be subtracted from the subsequent readings.
On the Team 11 ES17-12 battery we got: 12.7 mOhm, 13 mOhm and 12.8 mOhm.
Averaging: 12.8 mOhm from which we subtract the 1.4 mOhm and get: 11.4 mOhm.
EM3610 turned off.

Previously in this topic my ES17-12 battery was measured at 11.5 mOhm.
So basically a comparison between my lightly used ES17-12 and Team 11’s more heavily used battery is very promising.
Both ES17-12 batteries had slightly lower internal resistance than the Genesis which had 13.4 mOhm which is basically 2 mOhm more.

There is no way to perform this test on the battery blades themselves with the Battery Beaks because they don’t have probes.

Measurements on the Anderson connector tails.

For the Team 11 ES17-12 battery on the Anderson connector:

  1. My Battery Beak read 17 mOhm which was identical to my ES17-12 battery with a 2" longer tail.
  2. Team 11’s Battery Beak 2.0 read 18 mOhm (remember that the Battery Beak says it has a +/-1 mOhm accuracy).
  3. With the EM3610 turned on, set to 200m and the probes shorted I get one reading of 1.3 mOhm and another of 1.4 mOhm that needs to be subtracted the average of the readings: 14.1 mOhm, 14.6 mOhm, 14.3 mOhm. The average is 14.3 mOhm minus 1.35 mOhm is: 12.95 or 13 mOhm for this battery’s internal resistance and some cable/connector resistance. It seems reasonable that the cable/connectors are at least 1.5 mOhm.

For the Team 11 Genesis battery on the Anderson connection:*

  1. My Battery Beak read 21 mOhm which is 4 mOhm more than with either ES17-12 battery.
  2. Team 11’s Battery Beak 2.0 read 22 mOhms (again 1 mOhm higher than my Battery Beak).
  3. With the EM3610 turned on, set to 200m and the probes shorted I get the following readings: 1.6 mOhm, 1.6 mOhm, 1.7 mOhm with an average of: 1.63 mOhm. The readings of the EM3610 were: 16.7 mOhm, 16.6 mOhm and 16.4 mOhm with an average of: 16.57 mOhm. Subtracting 1.63 mOhm from 16.57 mOhm gives us a measurement of: 14.94 mOhm or 14.9 mOhm. When we take 14.9 mOhm and subtract the blade measured value of 13.4 mOhm we get again about 1.5 mOhm accounted for by the cable/connectors.

Observation:

The 2 Battery Beak 2.0 measured the ES17-12 batteries and cable/connectors as: 17 mOhm - 18 mOhm.
The EM3610 measured the ES17-13 batteries as: 13 mOhm - 14.4 mOhm.

The 2 Battery Beak 2.0 measured the Genesis battery and cable/connectors as: 21 mOhm - 22 mOhm.
The EM3610 measured the Genesis battery as: 14.9 mOhm.

The Battery Beak 2.0 tested generally read a higher internal resistance and cable/connector resistance than the EM3610.
The difference seems to be about: 5-7 mOhm.
The values the EM3610 reads at the battery blades is within the specifications for these batteries.

To that end I’ve attached the Genesis battery data sheet. On Page 9, in the section Impedance there’s a graph and it says this: “The internal resistance (impedance) of a battery is lowest when the battery is in a fully charged state. The internal resistance increases gradually during discharge, Figure 7 shows the internal resistance of an Genesis NP battery measured through a 1,000 Hz AC bridge” which is inline with the 1kHz testing of the EM3610. Also again in the glossary it says: “Internal Impedance - The resistive value of the battery to an AC current, expressed in ohms. Normally measured at 1 khz at full charge.”

Conclusion:

I think the EM3610 is not bad for a $40 something internal resistance meter. It’s economical enough and it gets close to reproducing expectations from the data sheets for the FRC batteries. From this experience I am now much more confident that the measurements I have for all these other batteries (where the manufacturer provided none of these details) is roughly accurate enough for comparison of the batteries and to the impact of the battery internal resistance on my various projects. I don’t think I need to do any more testing of this product.

The largest weakness of the testing I did in this topic is that I have only one EM3610, 1 of each of the non-FRC batteries, and only one Rigol DS2072 oscilloscope. I can only work with what I have and there’s insufficient motivation on my part to keep testing.

One other thing I noticed in doing this testing was that for days my EM3610 was inside my work area hovering around 70F. When I tested Team 11’s batteries I put it in my warmed up car. Drove it a few miles and carried it into the school at about 18-22F temperature each time the meter was outdoors. When this meter was in my work area shorting the probes together on the 200m range consistently produced measurements that averaged 1 mOhm over dozens of measurements. Once I got to the school and measured the batteries in the cooler electrical area about 5 minutes after walking in the door. I noticed that this measurement after a dozen or so attempts was now around 1.6 mOhm. I think this is caused by the temperature shift. So be sure to short those probes together once on the 200m range each time you turn on the meter. I expect that this initial measurement each time will negate this sensitivity and while the English version of the manual does not warn about this, the instructions indicate that you should do that initial measurement and it is implied each time you put the meter into that 200m range. (I have the EM3610 back in my work area since 11PM EST last night and the shorted probe initial readings are back to the 1 mOhm range).

NP18-12B-product-data-sheet.pdf (1.11 MB)

… perhaps because there could be a slightly different contact resistance in the switch each time you change it.

I thought of that as well.

Dozens (over 70) of measurements with the probes shorted at the same temperature greatly reduces the risk it’s the greatest driving factor.

I all but thought it was a meter specific value till it changed when I took it to the school. I can easily do that test again right now.

At room temperature currently: 70.3F, 25% humidity
mOhm readings with probes shorted on 200m setting, after each reading I break contact between the probes:

(1.1 + 1.0 + 0.9 + 1.0 + 1.2 + 1.2) ÷ 6 = 1.067
Turned off then on again (range selector to off then back to 200m):
(1.1 + 1.1 + 1.1 + 1.0 + 1.1 + 1.1) ÷ 6 = 1.083
Turned off then on again:
(1.1 + 1.2 + 1.1 + 1.1 + 1.1 + 1.0) ÷ 6 = 1.1
Turned off then on again:
(1.1 + 1.2 + 1.1 + 1.1 + 1.1 + 1.0) ÷ 6 = 1.1

Lowest reading: 0.9, Highest reading: 1.2, Average: 1.05, Difference: 0.3 mOhms

That makes > 100 shorted probe measurements.

Then it was put outside in the dark at currently: 31F, 68% humidity, 22F dew point, 29.57inHg pressure for 30 minutes.

mOhm readings with probes shorted on 200m setting, after each reading I break contact between the probes:

(1.3 + 1.4 + 1.3 + 1.4 + 1.4 + 1.4) ÷ 6 = 1.37
Turn off then on again:
(1.5 + 1.5 + 1.4 + 1.4 + 1.4 + 1.2) ÷ 6 = 1.4
Turn off then on again:
(1.3 + 1.4 + 1.2 + 1.4 + 1.5 + 1.5) ÷ 6 = 1.38
Turn off then on again:
(1.5 + 1.5 + 1.5 + 1.5 + 1.5 + 1.5) ÷ 6 = 1.5

Lowest reading: 1.2, Highest reading: 1.5, Average: 1.95, Difference: 0.3 mOhms

EM3610 meter not turned off nor range changed.
Measuring 100 mOhm resistor A:
(102.2 + 102.1 + 102.3 + 102.1 + 102.5 + 102.1) ÷ 6 = 102.22 - 1.5 = 100.72 mOhm on 1% resistor.

My freezer is set to 10F and measures as 9F on the top shelf by independent digital RadioShack thermometer. It’s in autodefrost so humidity reads as 72%. In 20 minutes next to my HungryMan dinners. Wiped off some condensing humidity on the probes first.

(1.3 + 1.4 + 1.4 + 1.5 + 1.5 + 1.5) ÷ 6 = 1.43
Turned off then on again:
(1.5 + 1.5 + 1.5 + 1.6 + 1.6 + 1.5) ÷ 6 = 1.53
Turned off then on again:
(1.5 + 1.5 + 1.5 + 1.5 + 1.5 + 1.5) ÷ 6 = 1.5
Turned off then on again:
(1.4 + 1.5 + 1.6 + 1.5 + 1.5 + 1.5) ÷ 6 = 1.5

EM3610 meter not turned off nor range changed.
Measuring 100 mOhm resistor A:
(102.7 + 102.7 + 102.5 + 102.5 + 102.6 + 102.5) ÷ 6 = 102.58 - 1.5 = 101.08 mOhm on 1% resistor.

Note that the English version of the EM3610 manual says the recommended operating environment is from 0C to 40C or 32F to 104F. So it makes perfect sense the resistance reading is a little off (few tenths of a millOhm) after a short trip to 9F.

The EM3610 manual also says it can be operated in environments of <80% humidity. It was simply stored in the freezer off to cool, the test and resistor were done in my work area at that temperature and humidity.

Now last night I took my RadioShack digital thermometer out of the freezer and put it 1 foot from the EM3610 so they both had more than 10 hours to stabilize at my work area temperature. I know this RadioShack thermometer reads humidity with errors as high as +8%/-4% from experience (I have 3 of these units total). It currently reads 28% humidity versus the AcuRite indoor unit 11 feet from it reading 24%. There is a cool mist wicking humidifier in the room but also a forced air infrared heater directly under than humidifier. My AcuRite indoor units as far as I have seen to date are off on humidity by only 2% maximum (I have 5 indoor units in 5 rooms). The RadioShack thermometer reads 68F and the AcuRite reads 70F and that I accept because the AcuRite is more in the path of the heater exhaust and air from other rooms than the RadioShack unit. My outdoor AcuRite reads as: 36.5F, 27F dew point, 69% humidity, 29.52inHg pressure.

mOhm readings with probes shorted on 200m setting, after each reading I break contact between the probes:

(1.0 + 1.0 + 0.9 + 1.2 + 1.2 + 1.1) ÷ 6 = 1.07
Turned off then on again:
(1.0 + 1.0 + 1.1 + 1.2 + 0.9 + 0.9) ÷ 6 = 1.02
Turned off then on again:
(1.2 + 1.1 + 1.3 + 1.2 + 0.9 + 1.2) ÷ 6 = 1.15
Turned off then on again:
(1.0 + 1.1 + 1.3 + 1.1 + 0.9 + 0.9) ÷ 6 = 1.05
Turned off then on again (forgot on 4th try to test resistor):
(1.1 + 0.9 + 1.0 + 1.0 + 1.3 + 0.9) ÷ 6 = 1.03

EM3610 meter not turned off nor range changed.
Measuring 100 mOhm resistor A:
(102.0 + 101.7 + 102.0 + 102.2 + 101.7 + 102.0) ÷ 6 = 101.93 - 1.03 = 100.93 mOhm on 1% resistor.

I placed the EM3610 in front of an Optimus 800W Quartz infrared fixture. It had both bulbs turned on for the full 800W and I placed the meter vertically against a white box 1 foot from the unit. I put the RadioShack digital thermometer next to it and raised the temperature to 100F and because heat rises it was likely right at the 104F specification towards the top of the meter where the transreflective LCD was slightly darkened. I let it cool for 2 minutes to clear the LCD. Then I repeated the tests as quickly as I could.

mOhm readings with probes shorted on 200m setting, after each reading I break contact between the probes:

(0.9 + 0.7 + 0.6 + 0.8 + 1.0 + 1.0) ÷ 6 = 0.83
Turned off then on again:
(0.9 + 0.8 + 1.1 + 1.1 + 1.4 + 1.0) ÷ 6 = 1.05
Turned off then on again:
(1.2 + 1.2 + 1.1 + 1.1 + 1.1 + 1.0) ÷ 6 = 1.12
Turned off then on again:
(0.9 + 0.8 + 0.9 + 1.0 + 0.8 + 0.8) ÷ 6 = 0.87

Just to be sure about this reading I put the meter back into the infrared heater path this time with the RadioShack digital thermometer on top of it. I raised the temperature till it read 102F, 15% humidity.

mOhm readings with probes shorted on 200m setting, after each reading I break contact between the probes:

(1.0 + 0.9 + 0.9 +1.0 +1.0 +1.0) ÷ 6 = 0.97

EM3610 meter not turned off nor range changed.
Measuring 100 mOhm resistor A:
(100.0 + 100.0 + 99.8 + 100.0 + 100.1 + 100.1) ÷ 6 = 100 - 0.97 = 99.03 mOhm on 1% resistor.

So that’s what happens when you heat the EM3610 up to some reasonable temperature I could see in an FRC competition being in a poorly ventilated gym on a hot summer day and the limits of the specified operating temperature for the meter.

Now the EM3610 meter and RadioShack thermometer were left next to each other where they started their day in my work area for at least 6 hours. When I got the meter next the RadioShack digital thermometer read: 70F and 27% humidity. The meter readings were taken kickly as shown below:

(1.1 + 1.2 + 1.3 + 1.2 + 0.9 + 1.2) ÷ 6 = 1.15
Turned off then on again:
(1.0 + 1.1 + 1.2 + 1.3 + 1.2 + 1.3) ÷ 6 = 1.18
Turned off then on again:
(1.3 + 1.3 + 1.0 + 1.2 + 1.3 + 1.3) ÷ 6 = 1.23
Turned off then on again:
(1.4 + 1.5 + 1.4 + 1.3 + 1.5 + 1.3) ÷ 6 = 1.4

EM3610 meter not turned off nor range changed.
Measuring 100 mOhm resistor A:
(101.7 + 101.7 + 101.7 + 101.7 + 101.7 + 101.8) ÷ 6 = 101.72 - 1.4 = 100.32 mOhm on 1% resistor.

Observations:

  1. It should be noted that having a good connection whether between the meter probes or the meter probes and the circuits or batteries under test is very important when you are talking about millOhms. The probes on this meter are cylindrical and have 4 points on the end like a human molar tooth. I’ve been touching the sides of the probes together, rotating them slightly between readings and when touching I am applying at least as much pressure as I would expect a good pair of alligator clips would apply. This detail is kind of obvious but it’s worth noting for anyone that may try to test a different meter. Sometimes I touched the probe tips to each other’s sides perpendicular, other times I made an angle.

  2. The EM3610 tends to start each reading a little low then eventually climb and stabilize on the reading value I have provided and while this is apparent it’s pretty fast (< 2 seconds). Sometimes when I release the pressure as I let go, the last reading climbs but I ignore this because as mentioned in #1 above, this is a symptom of seeing very small resistances and having a weak connection. It’s really no different than not adequately tightening down your FRC robot lugs and crimps.

  3. I also rotated the 100 mOhm resistor I was measuring each time I measured and I tried to measure that resistor as close to the body of the part as possible because otherwise I’d be measuring the lead resistances extending to the body of the part as well.

  4. The good part about the inconsistency of my testing location is that it simulates repetition in a real environment with various people over time. The Battery Beaks tested have the slight improvement of having a connector that they can depend on to remove a bit of that uncertainty but those Anderson connectors can still get dirty, damaged or be improperly crimped. The opposite side of the issue is that the Battery Beaks work great with the connectors but now if you need to probe around in the circuit you have to make some sort of leads for it.

Ran out of the 1440 minute (24 hour) limit to finish setting up my last post…so:

Initial work area temperature readings:
Lowest reading: 0.9, Highest reading: 1.2, Average: 1.05, Difference: 0.3 mOhms

Stored outside for 30 minutes around the lowest suggested operating temperature of 32F:
Lowest reading: 1.2, Highest reading: 1.5, Average: 1.35, Difference: 0.3 mOhms
(I noticed I made a simple typo on data entry for the average I can’t fix in the post above that I have fixed here.)

(These didn’t make it into the last post…)

Stored in freezer for 20 minutes measurements:
Lowest reading: 1.3, Highest reading: 1.6, Average: 1.45, Difference: 0.3 mOhms

Return to work area temperature from freezer:
Lowest reading: 0.9, Highest reading: 1.3, Average: 1.1, Difference: 0.4 mOhms

In front of heater:
Lowest reading: 0.6, Highest reading: 1.4, Average: 1, Difference: 0.8 mOhms

Return to work area temperature from heater:
Lowest reading: 0.9, Highest reading: 1.5, Average: 1.2, Difference: 0.4 mOhms

Observations:

  1. We can see that for any given time we turn on the EM3610 with the act of turning the range dial from the off position to the 200m position it does not actually cause any wild swings. The simple act of making and breaking the connection between the probes, as a required step before making measurements on the 200m range setting, causes a deviation at each measurement that can be as little as 0.3 mOhms and as much as 0.8 mOhms. Heating the meter certainly seemed to break the trend of this lowest and highest measurement difference being 0.3 - 0.4 mOhms.

  2. We can see that, for each block of measurements, if we take the difference between the lowest and highest average measurements between them we get a deviation of 0.45 mOhm on over 150 measurements in just these last 2 ChiefDelphi topic posts alone.

  3. In doing this test we’ve also demonstrated what the EM3610 does at either end of it’s recommended temperature operating range. We can see that when getting too cold it reads a few tenths of a milliOhm too high despite subtracting the initial shorted probe measurement (though that helps). We can see that when getting too hot it reads a few tenths of a milliOhm too low despite subtracting the initial short probe measurement (though that helps).

  4. In addition to all this I actually did put the EM3610 on the other 2 resistance ranges and touch the probes together at the temperatures I was testing and in all cases I got 0.0 mOhms. So either those ranges are simply to inaccurate to see the shift of a few tenths of a milliOhm or there’s some other compensation involved in those ranges. However, it’s not important to me and probably much less import to FRC because the FRC robot batteries and good wiring should in most cases neatly fit in the EM3610’s 200m range.

  5. Notice I used a the passive resistor for these latest measurements. I did so because I didn’t want to deal with the slow discharge of a battery that will happen at even my stable work area temperature over the course of more than 24 hours it has been. As the FRC battery discharges it’s internal resistance may change even from just sitting there. I used the same resistor in each group of measurements and that 100 mOhm value is about mid-scale for the EM3610’s 200m range setting. We already knew my EM3610 meter had some issues near the bottom of the 200m range with the 2 mOhm resistor from previous measurements. We can now additionally see from all these measurements that with connection losses that can be a few tenths of a milliOhm at each attempt so such a small resistor as 2 mOhm is approaching the limits of my EM3610’s capacity to reliably measure (accuracy and repeatable accuracy).

  6. It’s worth noting that when you first connect the probes to themselves, or to something you are measuring, the EM3610 on the 200m range tends to read lower than the reading it will stabilize at. This may account for the very few really low measurements when the EM3610 was being tested over 100F. How the meter settles doesn’t bother me as much as the fact that it means one should wait for the reading to stabilize as I did before the record it.

Conclusions:

Nothing found here changed this from before in this topic:

"I think the EM3610 is not bad for a $40 something internal resistance meter. It’s economical enough and it gets close to reproducing expectations from the data sheets for the FRC batteries. From this experience I am now much more confident that the measurements I have for all these other batteries (where the manufacturer provided none of these details) is roughly accurate enough for comparison of the batteries and to the impact of the battery internal resistance on my various projects. I don’t think I need to do any more testing of this product.

The largest weakness of the testing I did in this topic is that I have only one EM3610, 1 of each of the non-FRC batteries, and only one Rigol DS2072 oscilloscope. I can only work with what I have and there’s insufficient motivation on my part to keep testing."

I can now add that I would also now conclude that when operating the EM3610 I have at the extremes of it’s recommended operating temperature I would expect it to read as much as 1 mOhm too high on the cold side near 0C (32F) and 1 mOhm to low on the hot side near 40C (104F) with it reproducing fairly consistently the FRC battery datasheet expectations of internal resistance in the 20C-25C range (68F-77F). This includes when someone walks in with the meter from say a cold winter night or out of hot vehicle during the summer. It is important based on the 150+ readings above to short the probes together when the 200m range is selected at some point before you leave that range setting. It’s a little less important if the temperature in the room is stable but changes in temperature could make that value you need to subtract from the reading harder to guess at.

If I sit back after doing all this and think about it. I think that the Battery Beak 2.0 (which measures differently from the EM3610 using a purely resistive load) is probably a little less accurate than +/- 1 mOhm as it claims, probably more like +/- 2 mOhm maximum. Just comparing Team 11’s with my Battery Beak 2.0 shows a consistent 1 mOhm difference at the same temperatures and same batteries. We have no idea how, or if, the Battery Beak 2.0 rounds the milliOhm measurements it makes. I would guess they likely took that value from a 1% resistor they use as a load and it doesn’t account for the additional deviations from that value that their measurements of the resistor in operation likely have, nor variations in the contact resistance of the Anderson connectors over usage.

If I look at all these EM3610 measurements (knowing it measures resistance and internal resistance with a 1kHz AC test) the accuracy and repeatable accuracy I believe is within and sometimes better (at least for my EM3610) than what the data sheet leads one to believe it can do. I don’t have another EM3610 to test and frankly I won’t be buying one as this wasn’t actually ever intended to suggest FRC teams buy an EM3610 as much as to give me confidence in the EM3610 I have and shed some light on some things I’ve heard in FRC for all the years we’ve used this style battery (we used to use drill batteries back when I started in 1996). If I took my EM3610 out into the field. I can now predict what negative impact temperature can have on the 200m range readings (+1 mOhm cold, -1 mOhm hot), and that when using the 200m range if I short the probes say 6 times before taking a measurement I would be fairly safe assuming the offset I need to subtract from the reading I am trying to get could be as much as 0.5 mOhms higher than my lowest initial probe shorted measurement. I can’t predict what this will do over time so: let’s say it’s close to the accuracy of +/- 2mOhm from the reading you get under workshop conditions worst case and better (worst case +/- 1 mOhm right now) if you confine the meter to the 20C-25C (68F-77F) range (and I’m being rough on my EM3610 as the measurements would conclude it’s slightly more accurate than that).

It also occurs to me that inconsistency of touching the EM3610 probes together could be overcome by other design choices in: the probe coatings, geometry and connections, in how to do the calibration or in how the measuring system behind the probes behaves. There are milliOhm meters that cost in the ranges of $100, $500, $1200 that can manage to accurately measure in the tenths and hundredths of milliOhm. However they aren’t necessarily internal resistance (impedance) tools and they may be easily damaged by hooking them across a battery because the Kelvin method they use would put their internal power source across another power source and it may damage the meter’s internal power source. Given the $40 and the operating temperature range and a little more robust durability I feel like this meter is a great value point. Sure maybe I can’t be sure I can measure a 2 mOhm resistor (though I can guess if it’s close to 2 mOhm with the measurements I do get) but, honestly, I can’t safely design my circuit anyway where a change of 2 mOhm in the battery internal resistance is a big deal because the battery’s internal resistance can easily increase 2 mOhm from internal heating during charge or discharge. I mean I’ve got a 2 mOhm difference between the charged Genesis battery at Team 11 and the charged ES17-12 I have; and they are both barely or not even used batteries. So at least for FRC batteries +/- 2 mOhm or less ought to be fine. For my other project with Team 11 it’s also not a big deal because those motors are also brush motors with similar power needs.

In the end when it comes to making these measurements: there are really 3 common ways. To use a resistor and hopefully the Kelvin measurement method to eliminate the tricky business of getting a clean reading for it. To use the 1000 kHz AC method as clearly the ES17-12 and Genesis NP18-12B folks did based on their data sheets. These might produce slightly different readings but neither is necessary any more right depending on the load you wish to power. However there are even more accurate ways that involve lab work. I was looking for a good workshop tool that could read internal resistance and was just fine with the 1000 kHz AC method. I don’t want to send batteries to the lab to get information close enough to work on my circuits. I don’t want to send batteries to the lab every time I notice something a bit strange. I don’t want to measure values with tools that are likely to make me falsely think something may be wrong and honestly the Turnigy Reaktor 250 charger basically did that when I made 2 wire internal resistance measurements with it. There was just too much difference between the balance connector readings and that measurement across the whole pack with that charger it didn’t make sense.

Unless someone else has something to request I will consider this post a closed matter.
I may test a CEM DT-5302 Kelvin 4 Wire MilliOhm meter sometime, but I don’t have one right now.