I have three Sony/Murata VTC6A cells I intend to use in a Convoy 3x21b. I know these are higher discharge cells than necessary for that light but they’re the only set of three 21700s I have which are the same manufacturer, model, capacity, etc.
From reading here and elsewhere, I get the impression that using closely matched cells matters more when they’re used in series than it does when used in parallel, but that I should still try to ensure the capacities and internal resistance are close.
I checked internal resistance using my XTAR VC4SL. I’ve noticed I can get different readings just by retesting the same cell repeatedly. So to account for variation from one test to another, I tested each cell twelve times – three times in each slot of the charger.
Two of the cells seem to be closely matched – both averaging 16 mOhms. The third one averages 22 mOhms. That’s nearly 40% higher internal resistance than the other two.
Thanks. I know the measurement isn’t that precise on chargers, which is part of the reason I tested them repeatedly and in different slots. It may not be giving me completely accurate readings, but the average of multiple tests will at least give me some comparative values.
I just finished running another 12 tests per cell, and the results are similar. Two of the cells average 15-16 mOhms, and the third averaged 23 mOhms this time. So I think it’s safe to say one of the three has ~40% higher internal resistance, even if the test equipment isn’t very accurate.
If that’s not enough of a difference to be concerned about, great. They’re going into the 3x21b.
Parallel self balance. They become ‘one’, charging or discharging. Only problem is if you hook them up and the voltages are substantially different.
Series do not. So they can get out of balance charging or discharging. Matching becomes more important.
What everyone else said. Those are perfect. That 6mOhm higher internal resistance on that one cell, even if we pretend the charger is perfectly accurate, could be a fingerprint on a terminal or the cell being 2°C colder than the other ones
Just to clarify – I’m not pretending that the resistance testing on my charger is ‘perfectly accurate.’ The fact that it returns different values with the same cell when repeatedly tested only seconds apart proves that.
But averaging the values from multiple test passes is a legitimate way to make comparisons between items even when the test equipment is less than perfect. The more tests that are done, the higher the confidence level is. This is a known technique that I learned about decades ago in school. (I don’t mean testing internal resistance specifically; I mean the general methodology of working around limitations in test equipment through repeat testing.)
These may not be giving me the exact internal resistance of each cell, but concluding that two of them have similar internal resistance to each other and that one of them has higher internal resistance relative to the others is valid. I’ve now checked internal resistance 48 times for each cell. The tests have been spread out over 16 hours. All batteries were stored together prior to each round of testing and were at the same temperatures. Based on Jeffgoldblum’s comment about fingerprints, I cleaned the positive and negative terminals of each battery prior to the most recent round of testing and the results mirror the prior rounds.
One of these cells is averaging ~6 mOhms higher measured resistance than the others. That’s about 40%, which is statistically significant. I don’t know what their actual internal resistances are. My flawed test equipment can’t be relied on for that. But what it can tell me is that one of the cells is not quite like the others with respect to internal resistance.
I have been repeatedly reassured on this thread that this is fine for my intended use. I believe you. And again, thank you.
If one cell is at 2 mOhms
and the second cell is 6 mOhms
does that mean cell # 2 is 300% worse ??!
Hell no it means they are both perfect… best ever !!!
This is just bad math and the wrong way of looking at things .
The better way of doing percentages in this case would be to ask yourself at what point is a cell no longer any good …If you say a cell is unusable at anything over 300 mOhms, ( pick a number) then these cell’s percentage wise are 99% good or 1% bad .
Percentage of what??? is the question ?
You’re testing perfect cells 48 times ? The result is they are 100 and 50 percent perfectly fine.
The real question is , how much resistance is too much resistance .
Q: What is the normal internal resistance of an 18650 battery?
A: It will depend on your batteries’ different conditions and usages, also on how you measure the IR. You may test some of your new 18650 cells measuring around 50-80 milliohms. And there is reference list online about 18650 cells’ IR and ranking:
The general rule when uaing cells together is to :
Use identical model number batteries. And
From the same batch you purchased.
For example if you have a new battery and another that is 3 years old, even if they are the same model number, its not ideal to use them together.
These are general guidelines. If you use three different batteries, different age, different capacity, different amp rating, your light will still operate. Its just not an ideal or recommended situation.
"Parallel " you can put any 3 name , size, etc together ( my opinion I always use the same batch cells) – The one thing you need to remember is that they are all at the same state of charge when connecting together - If you have one at 3.3v and the other two are at 4.2v — when they self balance they’ll exceed the max charge rate of the lower cell
We’re getting well away from my original question, which has been answered thoroughly now. But if you want to talk about general testing methodology, that’s fine.
Assuming the test equipment is accurate, it means that the second cell has 200% (not 300%) higher internal resistance. That’s not “better” or “worse” – it’s just the difference in that measurement. It only becomes better or worse when you apply those measurements to a given use or purpose.
For use in parallel in a modern flashlight, yes – for that use, they’re both ‘perfect.’ That doesn’t mean that one doesn’t have 200% higher internal resistance than the other. It’s just that it doesn’t matter for this purpose. There might be other uses for these two cells in which that difference matters.
It’s not “bad math.” 200% higher in the example you proposed is correct math. I was asking if a roughly 40% higher internal resistance in one cell makes it inappropriate for use with two other cells in a particular flashlight. Apparently, it doesn’t.
I understand what you’re saying. That differences in internal resistance don’t matter much as long as all the cells have internal resistance below some particular threshold. That’s fine.
I appreciate the reference material you found and linked here. It reinforces the answers I’ve already gotten from you and others on the thread. Again, my question has been answered. Thank you.
Your thinking is generally sound (not that you need me to tell you that;-)
BTW: If you’d like to do better with internal resistance measurements, you can, and save a lot of effort/work specifically as well. I did, and I would never ‘go back’:
@Desertcat Thanks for pointing me to that thread. Looks like the SkyRC MC3000 gives much more consistent internal resistance measurements than the XTAR I’m using. I haven’t read through the entire thread, but I did see some concerns expressed about resistance measurements with the latest firmware. Has that been resolved?
Not to the best of my knowledge. The only way in which it is resolved in my case is to remain on FW Rev 1.15, which is what my unit arrived with installed, and where the ‘issue’ does not exist. I’m not aware of any reason for me to update to any other version, so it’s easy for me: ‘do nothing’. If it matters, if / when it is ‘resolved / understood’, you’ll likely hear about it first in the thread I linked to.
There was a big change in the measurements between 1.15 and 1.17. Which remain in 1.18.
Indicated IR dropped a bunch (IIRC like from numbers around 40 to around 10 mOhms).
Some correspondence with SkyRC indicated that the change in 1.17 that caused this was to address a problem with IR measurements. So I am not sure that they think that there is a problem. Firmware updates over the last few years have been few and far between. So I am not sure I would count on this being addressed in the short term.
But, as @Desertcat indicated, when running V1.15 the issue does not exist. 1.15 is available for download and the device allows flashing earlier versions of the firmware versus what you are on. So that is an easy resolution.
One of the members who had good data from using the MC3000 was going to do a direct comparison between the versions. But @dmenezes hasn’t been around for a few months. Maybe when it get to be winter in the southern hemisphere he will have more time for us…
