[FR] Documentation to increase V in

[slepiavka]

My Labrador arrived yesterday - congratulations on bringing this amazing project to life. It's insane how a tiny, cheap PCB can have better support and documentation than many full-blown commercial solutions.

Anyways. I saw an Amazon question regarding higher input voltages, where you suggested to [...] Just solder on an additional resistor in parallel with R3 and R4 and then rebuild the software [...]. I can mess around with thorugh-hole, but this circuit is TINY! How would you suggest for a user to go about this? Not sure where a resistor would even physically fit (assuming I could even SMD solder).

Maybe I'm missing something: is there a drawback to just using a good old voltage divider in the kOhms range? I don't speak signal processing, but I understand that dividers can introduce some capacitance/inductance? I'm not dealing with high frequencies or anything fancy.

[EspoTek]

Hi mate. Don't be imtimidated by SMD soldering! It takes a bit of skill to do it with an iron, but it's dead simple with paste.

Just get yourself a syringe of cheap solder paste (keep it refrigerated), a knockoff 858D from AliExpress and some bent-nose pliers. You'll be able to swap out little 0603 components like the resistors with no effort. Just make sure you don't set the fan speed too high, or you'll start blowing the other components away and make a great big mess.

Or, as you mentioned, you can use a voltage divider in the kOhms range. The inductance/capacitance wouldn't make a difference at the low frequencies Labrador can measure, but the additional resistive load could change things! Also, make sure you use resistors with enough wattage to handle the higher voltages you're passing through them. And please, don't kill yourself by being reckless with 120V/240V AC...

[slepiavka]

Thank yo so much. To be clear... When you said "additional resistor in parallel with R3 and R4", do you mean literally squishing another resstor right next to R3 sharing the leads and another one on R4? .

(Thanks for the heads up; this is for an old 30v circuit.)

[EspoTek]

I have no idea what I meant when I wrote that comment 3 years ago, but either method should be OK.

[photonicsfox]

@slepiavka indeed, scaling v down is easy with resistive dividers to scale Vin down if Vin is > V design of the input stage.

You are correct that inductance and capacitance can cause issues for higher frequencies. For frequencies < 10KHz this shouldn't be an issue as long as you keep the total impedance of the divider low enough and do not use resistors with high inductance/capacitance. Also, for anything that is not expected to drift with temperature, or remain accurate over time, I would avoid carbon composite resistors and only use precision resistor networks that are monolithic.

I am just starting to familiarize myself with the Labrador circuit.

The upper limit on the resistive impedance for this application will be set by both the desired frequency response, the maximum voltage you want to measure and the characteristics of the input stage of the front end of the ADC. Especially the input impedance/input current requirements of the input stage.

Now that I glanced at the schematics, it looks like at least for the two scope channels, EspoTek is buffering the input signal through an LM324 opamp (so that sets the requirements).

The LM324 is a decent part, max bias current at room temp is < 251nA, and < 501 nA at extreme temperatures. Given this, personally I would make sure that the current available from the divider is at least 1 microAmp at the lowest voltage you want to accurately measure. e.g. for 0.1V, 1 microAmp would require the total divider impedance to be 100 KOhm.

Note, the disadvantage of using a single resistive divider is that dynamic range could be an issue. Hence on nice DMMs from hp/keysight, the DMM will have multiple dividers to cover a large dynamic range.

If you wanted large dynamic range I would suggest using a few different resistive dividers. e.g. generally I prefer to not load down whatever I am testing by more than 10 microAmps. For 48V this would require an impedance of 4.8 MOhms!