reading voltage with multimeter changes result

[waspinator]

I'm trying to read a 24V signal using a voltage divider into a ucontroller, but the readings aren't consistent unless I measure the 24V signal's voltage with a multimeter.

I first had a 10k/1k divider, but it was bringing  down the voltage from 24V to 2.4V. I assume that's because the signal can't source the ~2mA needed, so I changed it to 1M/100k. That helped keep the voltage around 24V.

I have the signal going into a 74hc4067 mux and then into an adc pin on the micro.

A switch toggles the signal from 24V to ~0V, but the micro only reads the correct values when I'm also measuring the signal with a multimeter. Otherwise it sticks to one of the two values.

What could be some things I could do to troubleshoot the issue?

Thanks

[MosherIV]

Most dmm have input impedance of 10M

When you measure, you are effectively putting a 10M resistor in parallel with the resistors.

[Refrigerator]

Check your connections, make sure everything is making contact as it should. Especially if you're breadboarding.
Add a small decoupling capacitor across the 100k resistor, size depends on noise vs responsiveness, 1nF should be fine for a starter.
DMM probes are around 1 to 3nF, depends on how tangled yours are.
Schematic would be nice.
Could also be noise, i tend to get a lot of noise in my circuits when i have SMPS power supplies running.
Also it could be a grounding issue, especially if the curcuit is being powered from a switch mode PSU.

[waspinator]

Wouldn't adding a 10M resistor in parallel have pretty much no effect on the circuit? The effective resistance would drop from 1.1M to 1M, right?

This is a signal coming from a battery, and connections should be okay, soldered onto a PCB.

[neko efecktz]

Not solving your question but
10k /1k or 1M / 100k voltage divider is not 10:1 ratio
It is 11:1 ratio

To get an output of 2.4volts  with 24 volts in you need to use something like 18k / 2k.
or 180k / 20k or 1.8M / 200k

BILL

[rstofer]

All this voltage divider stuff is fine but when you run the tapped voltage through the mux and into the ADC, you need to think about the driving impedance required by the ADC.  Read your uC User Manual and it will tell you the maximum source impedance (usually less than 10k and sometimes as low as 2.4k).

So, pick your resistors to drop the voltage but make sure the parallel combination of the two resistors is less than the required source impedance.

You may not be able to find suitable resistors because, somewhere upstream in the 24V signal, the output impedance is very high.

In this case, you may need to add a buffer after the voltage divider and before the MUX.  Some kind of op amp will do the job if you wire it as a voltage follower.

http://hyperphysics.phy-astr.gsu.edu/hbase/Electronic/opampvar2.html

If you are sure your signal doesn't go negative, you can use a single-supply rail to rail op amp.  Pick one that will work from the uC Vcc.

If you want to find out what the source impedance for the 24V signal is, load it down with resistors until it drops to 12V.  At this point, the internal and external resistances are equal.

https://en.wikipedia.org/wiki/Maximum_power_transfer_theorem