30 MOhm voltage divider - is it OK?

Hi, my project will use lithium coin battery as backup and I want to measure it's voltage. Problem is the operating voltage when powered from main source will be less than 3V. I plan to use a voltage divider to measure the coin cell and measure it up to once a day to check if it is still OK. Three 10M resistors should create 2:1 voĺtage divider and drain neligible 100nA from the battery. I plan to "buffer" the divider with 100nF ceramic cap. Do you think it is OK? I have read wet air and/or dirt on PCB may cause considerable "leakage" on such high impedance signals. Should I do something to minimize chance something like that happens? For example solder the resistors "laying" instead of "standing" to increase distance between its legs on the PCB?

attach a schematic of the circuit for clarity. thx

Three 10M resistors

Is rather silly, when a fingerprint on a resistor can easily measure 10M.

I would use a FET to switch the power into a more realistic potential divider. Say 1M total.

Why use a potential divider ?

100nA is fine I think, just keep things clean.

If fingerprints on 1M resistors were really an issue, every valve amp out there would be misbehaving!

I'd be more worried about the leakge of the ceramic cap of any MOSFET used.

I think you'll be OK... Go ahead and give it a try! If there are problems you'll know rather quickly.

You might want to calculate & compare the self-discharge/shelf life to the drain at 30M just to see what practical effect it will have.

Then of course, there's the question of how long you'll need to run on the back-up battery when you loose power.

I have read wet air ...

Humidity isn't a problem until it gets to 100% (relative) and you get condensation and "wet water".

...and/or dirt on PCB may cause considerable "leakage" on such high impedance signals.

Certain solder fluxes can be conductive.

MarkT: If fingerprints on 1M resistors were really an issue, every valve amp out there would be misbehaving!

Not what I said. I said a 10M not a 1M.

Did you read this. Leo..

Thank you for your inputs. To answer some questions: I want to use 30M voltage divider because it drains ~100nA ~ 1mAh/year. 200mAh coin should last 200 years which seems good compared to its shelf life — and my life expectancy too. Saving complexity of a transistor powered voltage divider looks like a good deal. I think I will try it and see if it will give readings of reasonable accuracy.

Grumpy_Mike: Is rather silly, when a fingerprint on a resistor can easily measure 10M.

Can you be more specific about this? How you make such fingerprint and how you measure the resistance to get 10M? At high voltage/frequency maybe? I used my cheap DMM to measure resistance of my 10M resistor (fingerprints already included) and got readings between 10.05 and 10.07 MOhm (the DMM is cheap, it takes long to settle). I tried touch and rub the resistor by bare fingers but the reading didn't change. I even used a hand creme to make my fingers "more printing" but still nothing. Another of my 10M resistors read 10.15 +/- 0.01 MOhm so I concluded my fingerprints has resistance many GOhm...

If you wet your finger with brine you might be able to have a substantial effect until it dries, but Gohms is much more like what you'd expect for a random greasy splodge than Mohms.

Once in the nA range and 100M+ resistances then leakage from pcb surfaces starts to become an issue.

Hi, What model Arduino controller will you be using to measure the cell.

Thanks.. Tom.. :)