Testing CMOS batteries right! (PWJ235/shorty)

CMOS and other coin batteries run low after some time and need replacing. But how to check them correctly? It turned out that many batteries show good voltages but are still bad.
It is necessary to test batteries (not only coin cells) with an electric load connected to show their potential to deliver power. A high internal resistance may block current to flow but the battery still shows a good voltage.
My "tester" consists of a small 3V motor and a white LED in parallel. The white LED will only light up whan the voltage is more than ca. 2.5V. A good battery lets the motor spin fast and the LED glow bright.

[@stevenbirch]

A good tip! This phenomenon caught me out before. Without such a tester, you can use a multimeter on the "amps" range to very briefly check the short-circuit current of the cell.

[@IvanStepaniuk]

Probably not the most orthodox, but I always short them with the probe tip for half a second right before measuring. Good ones spring back to ~3v immediately.

[@PlaywithJunk]

Not working for these.... they jump back to 3.0V immediately when load is off. It's like they actually built up a real resistor inside. You need to connect a load while measuring.

[@raulcrudele1]

You re absolutely right, when unloaded these batteries show 3V, I would suggest measuring with a resistor in parallel, for example 1k, then watch the voltage drop!

[@PlaywithJunk]

That will surely work. But I like my Motor-LED-Tester :-) It makes BRRRRR when the battery is good.

[@graemedavidson499]

I use the “Lo Z” (low impedance) mode on my Fluke for measuring batteries for the very same reason!

[@fresheFresse]

MMs in Lo Z mode have around 10k of impedance which is still very high for a lithium battery

[@joruss]

That's not fair driving motor with them :D They're still plenty ample to drive a blinky led 😁

[@SeanBZA]

Yes got some Taidiran primary lithium cells that are fine, when testing with one multimeter with gigohm impedance on the 3V range, but flip to the 30V range, and the voltage drops to under 1V. 30V range is 10M input resistance, so internal resistance of the cell, now around 40 years old, is around 20M. It was used for battery backup, so was replaced with a CR2032 cell in a holder, as those are cheap to get, and also will do 10 years as battery back up on a CMOS RAM chip and RTC. did around 20 years powering a clock though, with a 4700uF 35V capacitor as bypass for the once per hour chime. Replaced when the chime went very wonky, using a 3V micropower regulator, and a 3Ah pouch cell from a power bank. Just charge it every year now.

[@dbmaster46446]

mhmm, i just hold them on my tongue

[@PlaywithJunk]

That doesn't work with these batteries. I tried it... The 3V bad ones taste like the fresh ones. I can't tell a difference for sure.

[@dbmaster46446]

@@PlaywithJunk funny calling it "taste" but i get what you mean

[@CEzikMaj]

Just touch both wires while measuring voltage, if it drops - it's drained, your skin resistance is enough for that.

[@mumiemonstret]

Just note that you are not supposed to touch the insulator with your fingers, since they leave a residue that is somewhat conductive. But you could still pinch the battery without touching the insulator.

[@beamer.electronics]

These cells are only meant for Ma loads, most cells (as far as I know) fail due to an ever-increasing internal resistance. Nice demo though, thank you.

[@PlaywithJunk]

a good coin cell of this type has a few ohms internal resustance. The bad ones have k-Ohms!

[@ixushardware]

@@PlaywithJunkdo you measure it with multimeter then?

[@PistonAvatarGuy]

Weird, never had one fail to work in a computer unless it was below something like 2.8 volts.

[@PlaywithJunk]

HP uses KTS and JHT lithium batteries and those are all crap. Dead after 3 years while Panasonic or Maxell last 10. And they all get the "High-Z" resistance problem.

[@Chriva]

An RTC takes microamps if not nano so this is way overkill. The cells are rarely even rated at more than 10 mA peak draw so an LED and motor is way, way too heavy. Typical Energizer cell is rated at 6.8 mA~ over 400 Ohms as an example

[@Sven_Dongle]

What about stacking them for higher output?

[@benbaselet2026]

@@Sven_Dongle Stacking implies a series connection, resulting in the same current.

[@Sven_Dongle]

@@benbaselet2026 I understand basic electronics, pal, I meant that there is a use beyond single cells for RTC's.

[@mumiemonstret]

The test is not for determining if it will work right now in e.g. an RTC application. It is for determining if it will work for one day or several years.

[@PlaywithJunk]

I once stacked 100 of them and got 300V. But a few of them must have been bad and besides a light tingle, there was not much to feel. Currently I have more than 100 larger coin cells that should be much better. I assume it should be good enough to power a 10-15W 230V input power supply. A simple one with no PFC correction should work with DC too.

[@haralamc]

Thats why weighing scales are such a head ache, they turn on fine 0 fine then go to shit when you put something on them

[@M0GLU]

Yes correct and it shows the Voltage is not equal to the power.

[@PlaywithJunk]

Voltage is nothing without current...

[@wadamt16]

Will using LoZ on DMM still show good voltage on those small bad batteries?

[@PlaywithJunk]

I don't have a low-z multimeter but I think it would be good to identify the bad cells. Voltage will certainly drop.

[@gglovato]

usually low-z is only for the AC range

[@Ltech-ludditetechnologies]

Hahaha. Whirly confirm good.

[@PlaywithJunk]

Whan it goes BRRRR the battery is good. It's the machine that goes BRRRR :-)

[@kelvinstokes996]

I just put the leads of an LED directly across. If it glows brightly, the battery is good (enough). If the battery is weak, the LED will be dim / off.

[@PlaywithJunk]

I have a LED too. But a motor that goes BRRRR makes more fun! :-)

[@TedBackus]

take an old USB cable, cut the non USB end off, and expose the + & - wire ends. tape or affix + to the + side of the battery, affix the - to the - side of the battery & plug the USB end in to a wall charger for 5 seconds or so. it will dump a fast charge in the coin battery. just dont hold it too long or it might pop.

[@PlaywithJunk]

And that would do what? These batteries are not discharged, they built up a too high internal resistance and that won't be fixed by charging. Also these batteries don't "pop". They explode violently as I know from my own experience. That was not a smart advice...

[@TheEmbeddedHobbyist]

the same way I never fault find on my car with a voltmeter, fell fail of this years ago. the voltmeter said that there was power to the rad fan, but it would not turn. could not measure the voltage when the fan was connected due to the water proof connector used. then was told to use a 12V bulb, so I now use a stop and tail light bulb. you get a 5W and 21W load in one place with a built in indicator. loaded the fan connector with the 21W section and the 12V disappeared. a high resistance joint will appear fine with a digital meter. This also makes me laugh at the battery testers that claim to be able to show capacity as a percentage, all this from a little bit of current and a voltage measurement. 🙂

[@PlaywithJunk]

Also always test car fuses with the ohmmeter. I had one that looked perfectly good, I thought the problem was elsewhere and the garage found the fuse at fault. And I paid the bill...

[@tveasy5172]

You just need measure current on mA mode of the same tester)))

[@sascha5668]

Einfach die Zunge nehmen und fertig. Dafür brauche ich keine laborausrüstung

[@PlaywithJunk]

Funktioniert nicht! Hab's gerade ausprobiert, der Unterschied zwischen einer schlechten Batterie mit 3.1V Leerlaufspannung und einer Guten ist kaum spürbar. Die Zunge ist hier zu empfindlich, es kitzeln beide gleich fest... Ausserdem ist eine LED und ein kleiner Motor wohl keine "Laborausrüstung" ;-)