Finding a sudden parasitic drain

[cigmas]

I can measure the voltage dropping 1 mV every few seconds on fresh batteries, resulting in the batteries reaching ~0V within a day. This is a relatively recent development on this electric toothbrush. With age, the rubber components on the external tube have deteriorated and the internals were drenched when I opened it up, so I suspect water damage could be an issue.

I wiped every surface I could with IPA and a cotton swab and an old brush. The problem persists.

Suggestions on how to find the culprit? Does anything look suspicious in the photos?

I have replaced the "non-replaceable" batteries on this electric toothbrush a few times, but was surprised to find the battery failing to hold a charge sooner than expected given the last battery replacement. Nevertheless I replaced the battery again, but this time it made little difference. The old batteries seemed ok too. I noticed some rust on the metal plates of the large rectangular object (transformer? see last pic).

[JBurdman7]

Check the caps for current leakage, especially before the switch

Check the diode(s) on that charging coil for reverse leakage, or de-solder the coil and see if the drain stops

[richnormand]

As JBurdman7 said for both cap and diodes.
I have seen  cases where an electrolytic cap would slowly bleed the battery (DMM) in particular.

However, I would first recharge the battery and then disconnect it from the circuit to see if it discharges on its own before blaming the rest.

If not put your DMM in series with the battery and the circuit to monitor the discharge current.
Water behind the switch?
Also worth considering is a leakage path on the pcb given the nature of the device (water/corrosion) although it does look clean and dry.
Usually, using a dry toothbrush you can brush away the corrosion and see the change in leakage current.

If the caps test OK I would then suspect the switching IC. A quick test with a freeze or heat usually trigger a change in leakage current.

Best of luck with it.

[wraper]

There could be corrosion left under any SMD component. Also IPA is a quite poor solvent to clean corrosion. Even plain water works better.

[cigmas]

Thanks for all the suggestions. How would I check for current leakage through capacitors or diodes?

I'll try scrubbing everything with water as well, just to be safe.

I suppose I could test self-discharge of the batteries on their own, but they are brand new LADDAs, so I doubt they are the problem, especially since the behavior is the same as with the previous batteries (made in Japan rechargeable Duracells).

What would measuring the leakage current in series with the batteries tell me? Is it just to monitor for changes while cleaning or trying temperature changes?

[Manul]

Electrolytic capacitor on the back of the board looks suspicious to me. You should desolder the wires, take out the board and look everything good.

[wraper]

Electrolytic capacitor on the back of the board looks suspicious to me. You should desolder the wires, take out the board and look everything good.

Yeah, looks like there is corrosion between the pads.

[richnormand]

Once you unsolder the electrolytics and check for corrosion underneath you can setup current monitoring with you DMM and see if the current dropped without the caps.

Another way to check for cap leakage on the bench is to put a voltage across the cap near its rated voltage and monitor the current. It is then easy to vary its temperature and see if it changes.
In an automotive application this is how I found a tantalum cap that would grossly leak in the knock circuit that would saturate an input op-amp in the sensor input filter. It was reading just fine at room temp (module is in the engine compartment).

[cigmas]

Sharp eyes, @Manul! I had not noticed the electrolytic capacitor behind the plastic divider. I think that "corrosion" is actually glue that was poured over that part of the board. Just to be safe I cleared the glue around the pads and scrubbed the area with water and IPA.



Thanks for the great tips. Here's what I've tried so far:

- Charged batteries and disconnected from the circuit to make sure they are not self-discharging. --> Ok, voltage (~2.7x V) was maintained overnight.
- Scrubbed the PCB with demineralized water and IPA with a brush, let dry overnight. Did not get rid of the parasitic drain.
- Measured the drain at ~25 mA with the DMM in series with the batteries.
- Heated with a hairdryer and noticed an increase in the current to ~32 mA within seconds. Heating the end of the board nearer to the charging coil (right, in pictures) makes no difference, but heating the opposite (left, in pictures) end of the board quickly sees a jump in current. No effect was noticed based on which side (top or bottom) of the board was heated, just which end.

Does this point towards a cause?

I don't have a bench power supply, so I could not test putting a voltage across the capacitor. I had previously measured a capacitance in-circuit with the flat batteries connected (I don't remember the value, maybe 8.8 uF or mF ?) but recently, and with the batteries disconnected, no capacitance is read (OL). Maybe that's something...

I have not yet removed the board. Aside from the 8 wires from coils etc, the board is secured by two plastic "rivets" which appear to have been melted or expanded after passing through holes in the board. In the first two pictures, they are whitish circles next to the Coil- and Coil+ solder points. Can the board be removed in a way that it can be reinstalled later? I could just squeeze the melted heads with pliers to remove the board, but I don't know how it could be secured later.

Anything else to try, especially without/before removing the board? If I can, I might de-solder the capacitor to measure it out of circuit. I don't have an ESR meter (yet).

[richnormand]

""but heating the opposite (left, in pictures) end of the board quickly sees a jump in current""

Is that the end with what looks like the on/off button?
Could that be leaking?  Would be easy to check by unsoldering the black wire to S+ without having to dismantle anything.

Or perhaps some component hidden by the motor?

[cigmas]

Is that the end with what looks like the on/off button?
Could that be leaking?  Would be easy to check by unsoldering the black wire to S+ without having to dismantle anything.

No, I meant the end of the PCB closer to the motor. I did not heat other parts like the switch or motor; I could try that too.

[richnormand]

Is that the end with what looks like the on/off button?
Could that be leaking?  Would be easy to check by unsoldering the black wire to S+ without having to dismantle anything.

No, I meant the end of the PCB closer to the motor. I did not heat other parts like the switch or motor; I could try that too.

Ah OK. So the end where U2 and the electrolytic are located.

[cigmas]

Ah OK. So the end where U2 and the electrolytic are located.

Yes, exactly.

I tried heating the switch and it doesn't seem to make a difference.

Given the size of the hairdryer relative to the board, it is difficult to tell precisely where the heating makes the most difference. I'm starting to think that pointing towards the U2 chip seems to raise the current the most quickly compared to anywhere else, but I don't know if that's just because the other side of the board is shielded by plastic parts. The electrolytic capacitor's soldered leads are also right next to U2.