Found Repeatable, Precise (Accurate?) DC IR measurement method :)

Whew. At long last I believe I've found a way to measure Prius HV battery module DC Internal Resistance that gets results similar to what we see from the Prius itself. I'm not ALL the way there... but want to document what I have while it is completely fresh.

(This is ~continued from an earlier discussion at HV Monitoring / Reconditioning Questions and Insights | PriusChat )

THE CONCEPT

• Obtain an accurate measure of the resistance to be used for the test. This needs to be accurate to a fraction of an ohm. For every 0.1 ohm this is wrong, the IR will change by about 1 mOhm.
• Measure the "open" (at rest, no drain) module voltage. Vopen.
• Set up so you're ready to
  • Record a rapid series of voltage measurements. Ten per second is enough. In half a second, the answer will be off by ~2 mOhm or more.
  • Connect the resistor in the circuit for about one second (while recording)
• IR is change in voltage divided by change in current. In this case:
  • The instantaneous voltage change when the 1 watt resistor is attached
  • Divided by the calculated current flowing when attached
• VERY quickly after that, the numbers change due to battery chemistry and discharge. But there's a brief blip of time where what is measured is due only to the battery "ohmic" resistance.

TOOLS NEEDED

• A one ohm high power resistor of pretty good quality. High power: 100 Watts is good. Quality: the main thing is good temperature stability. I got amazon product id B008IDAD76 (sorry, link won't post even with JS off )
• You may want to keep it cool, although that's not strictly necessary for an IR test. I intend to use this instead of the 'traditional' headlight bulb test however, and cooling can be a great help. Solution: Odorless Mineral Spirits. Non-conductive, high boiling point, very safe. I picked up a gallon at Walmart for under $20. The IR measurement only involves having power connected for 1-2 seconds, so it won't heat uip anyway.
• The hard part: an accurate LOW resistance meter, and accurate volt meter. An old bench meter can do the trick. I'll explain below why this is important. (I was gifted a used Fluke 8808A. There are MANY meters that can do the trick.) Reading ohms to hundredths (0.01) is pretty important. 0.001 would be better even if not so accurate at that point. 0.1 ohm is really pushing it.
• Two clip-on leads plus a heavy duty wire (10A or more) with alligator clips at both ends. Should be low resistance. Probably not from a 12 V car battery charger. I use amazon product ID B07L29V4Z2
• It may well be time for you to discover some nice free software that's new to me: Test Controller. This amazing tool knows how to talk to a LOT of different electrical devices: power supplies, meters, oscilloscopes, load generators, etc etc. See Test Controller introduction (at the bottom of that page is a link to a discussion/Q&A thread at EEVblog as well.)

PHYSICAL SETUP

• Measure total resistance of the meter cables, meter cables plus alligator clip cable, plus 1W resistor, and all three. The math should work out.
• Record the combined resistance of the alligator clip-cable and resistor. My setup: the resistor plus clip-cable is 1.041 ohm. My meter leads add another 0.136 for 1.177 total.) Rtest = 1.041
• Now set the meter to measure DC volts, ideally in a range that includes 9V or so. You want to avoid "auto range" as that slows it down a LOT. My meter can easily to ten readings per second at "Medium" speed.
• Fire up TestController, and practice logging voltage readings. Be sure that works! (Or perhaps you have a meter that logs internally. Test it!

TESTING ONE MODULE

• Connect the meter to the module
• Clip one end of the alligator-clip cable to battery -
• Clamp one end of the resistor to battery + (my meter clip easily holds both the module bolt and the resistor lead.)
• Now you should see the battery voltage on the meter, and have the loose end of the alligator-clip cable ready to attach to the loose end of the 1W resistor. Vopen = 7.869

Now we're ready to run the test...

• Turn on voltage logging. (This will go by WAY too fast to just look with your eyes.)
• Connect the alligator clip to the resistor for a count of one. Disconnect.
• Stop the log.

... and do the analysis:

• Find the measurement where the voltage first drops below Vopen. Vload = 7.802
• The difference (Vopen - Vload) is the voltage drop due to the resistance. dVbat = 0.067
• Current draw ("I" is current) is Vload / Rtest -- I = 7.495 Amp
• Internal Resistance is dVbat / I -- IR = 8.9 mOhm

I've done this to a dozen modules and the results "make sense" to me. I've NOT yet reworked my battery pack, so it remains to be seen if my answers are at all similar to the Prius.

CAUTION: AFAIK IR varies with state of charge. It is lowest near 50%. The official UL test for IR suggests testing at 85% and 30%. I hope to test all modules at least at a similar SOC.

 

Oops. 1 ohm, not 1 watt