Problems with the ADS1115 A to D converter

Hi I have a problem with the ADS1115 converter I need to cover a voltage range of up to 100v. I am using a single voltage divider that gives approx 5v with 100v. I calibrated the conversion equation with a 40v battery and a fluke multimeter. I then checked the voltage of other batteries I had of different voltages, all gave different results between the Arduino ana the multimeter where am I going wrong?

I could reduce the range of the voltage divider as it is to test the voltage of a battery through its working voltage so for example a 20s LiPo battery goes from 60v to 84v, if I alter the divider to work with those voltages how much over 5v will the ADS1115 take before it will no longer worked for example set the divider for a 10s battery max 41v and connected a 20s max 84v. This would happen if the wron divider was used for the battery that was being tested

Is the output from the ADS1115 a straight line, a curved onerous curved at the ends but straight in the middle. I have searched the specs pdf and cannot find any obvious answers there. Kendrick

Kendrick: Hi I have a problem with the ADS1115 converter I need to cover a voltage range of up to 100v. I am using a single voltage divider that gives approx 5v with 100v.

You should have used a divider that drops to 4volt (4.096volt). That, with a PGA gain of 1x, should give a ~20% higher resolution.

The ADS1115 is linear, and so is a voltage divider. Assuming the resistors are stable at that voltage.

Post a diagram, pictures of the setup, and your code if you want help. Leo..

What are the values of your resistors?

These have to be carefully chosen: low enough so the circuit's output impedance towards the ADS is not too high; high enough so they don't burn due to the current. 100k sounds like a good value for the high side resistor, that way you have 1 mA to keep the ADC happy and a reasonable 100 mW dissipation at 100V.

Also the input impedance of the ADS pin is 15MΩ so you can't go too high with the resistor or it starts to affect you.

For a voltage divider I use a 10k 20 turn screw trimpot. I set the pot to 20 to1 approx 500ohm to 9k5. I tried fixed through hole resistors and it took 30seconds or more for the number to settle, I don’t have the equipment to try SMD resistors. The voltage value settles immediately with the pot do you recon I have too low a pot and should have chosen a 100k pot? Kendrick

ADCs and DACs are linear with very few exceptions (and the datasheet will tell you if its doing something like log encoding of 8 bit audio for instance). One of the key specifications for ADCs and DACs is the deviation from linearity, measured in LSBs (or even in fractions of an LSB).

100volt on a 10k pot.
Most small pots don’t like to dissipate 1watt.

As wvmarle suggested, use a 100k resistor between voltage source and analogue input.
And use a ~4k1 resistor (or your trimpot as a variable resistor) from analogue-in to ground.
Leo…

When I used fixed value resistors for the voltage divider I used 1% tolerance 1/8W. Might this hav been the cause of my problems with the reading settling down on startup, would a 1w or greater resistors give me a more stable result. Kendrick

What actual values did you use?

Power rating doesn't affect stability as long as the resistor doesn't get warm in the first place (and if they get warm, you're using a too small value resistor).

Tolerance of the resistor doesn’t matter (will be fixed with software).
Stability of the resistor/pot does. Best to use two fixed metalfilm resistors (no pot).
If you use 100k as top resistor, then the resistor could dissipate <=0.1watt.
1/8watt is borderline, both for power and maybe voltage (1/8watt is tiny).
Use common 0.4/0.5watt metalfilm.
Leo…

I made a fixed resistor voltage divider from available resistors 5k1 1/4w 11100k 1/4w. It was fine at 40v but not quite right once up to 80v. I've just ordered 1/2w versions when i get them I'll put them in the place of the 1/4 w resistors. kendrick

Kendrick:
I made a fixed resistor voltage divider from available resistors 5k1 1/4w 11100k 1/4w.

11100k = 11M1. That can’t be right.

For 80V you’d need <1:15 so >80k for the second resistor. That’s about 1 mA, good for 80 mW of heat dissipation between the two resistors. 1/4W should do just fine, probably not even getting warm to the touch (try with the power disconnected!)

Why don’t you explain what actually happens?

Kendrick: For a voltage divider I use a 10k 20 turn screw trimpot. I set the pot to 20 to1 approx 500ohm to 9k5. I tried fixed through hole resistors and it took 30seconds or more for the number to settle, I don’t have the equipment to try SMD resistors. The voltage value settles immediately with the pot do you recon I have too low a pot and should have chosen a 100k pot? Kendrick

Hmm, a pot is not the way to do this - too much scope for mechanical creep changing the ratio, and 10k is too small a value, you're getting self-heating and maybe electrolytic effects on the wiper contact point.

Use precision fixed resistors, 1% or better, don't use carbon(*), use metal film. Use resistor values that will not self-heat appreachably - so 220k/10k perhaps. That's only 45mW or so (only a few degrees temperature rise, not 10's of degrees)

(*) The tempco of carbon resistors is shocking, not suitable for measurement applications.

Sorry I put in an extra 1 it should have read 5k1 0.25 w 100k 0.25 w, With that I was getting voltage creep if that’s the correct turn the voltage started reading one value at switch on then slowly increased till it settled after about 30 seconds.
The 0.5w resistors arrived, I have made a voltage divider with them, it works there is no creep, the 5k1 is carbon 5% though I have metal film 1% on order.
I’ll work with what I’ve got for now.
Kendrick

The 5% rating is tolerance - the actual value may be up to 5% more or less than 5k1.

The temperature rating is a different thing - looking at Digikey I saw 400-1500 ppm/°C (can be both directions: increasing or decreasing resistance with increasing temperature) or even higher for carbon against 25-100 ppm/°C for metal film.

For now I think I’ll be ok with the setup I have. I may change to 1w metal foil resistors when they come the reading difference between a low and high voltage battery is 0.1v but various connections not solder points may be the cause of that I have enough ADS115 boards free and resistors I’ll build up a prototype with all solder connections. Kendrick