I have the following scenario:



There are two halves to my system -- one at 12V with analog circuitry (shown in drawing), and the other at 5V which includes an arduino micro (off dwg to right separated by 50 ft of wire). They share a common ground and are battery powered. At the 12V side, I sits an element that needs tested for continuity (I must detect open, short, or pass). Passing means the resistance is loosely within target range (I don't have to measure it's value with any great accuracy).
The drawing is based on a 4-wire (Kelvin) resistance measurement (hoping the low current flow through the long cable would minimize losses in measurement). Since the voltage is so small, I would think putting an amp between the DUT and micro controller might work, but I can't send more than 5V to the arduino and the range of possible voltages for the DUT doesn't let me amplify them linearly, either. Ideally, I'd want an system in the middle that would do the measuring and output as follows:

• Pass: 0.005-2V -> 2.5V-ish output
• Open circuit: Anything higher than passing V -> 5V output.
• Short circuit: Anything less than passing V -> 0V output

Any suggestions? Maybe some kind of modified window comparator that outputs one of three different states depending on input?

Rules:

1. No more than 50mA through the DUT (preferably no more than 25mA)
2. Only one available line (besides ground) to the aruduino (no two signal solutions, hence using the A/D for three states).
3. I can add a 5V regulator at the 12V side if needed but the fewer parts, the better.

 

I have the following scenario:

View attachment 89613

There are two halves to my system -- one at 12V with analog circuitry (shown in drawing), and the other at 5V which includes an arduino micro (off dwg to right separated by 50 ft of wire). They share a common ground and are battery powered. At the 12V side, I sits an element that needs tested for continuity (I must detect open, short, or pass). Passing means the resistance is loosely within target range (I don't have to measure it's value with any great accuracy).
The drawing is based on a 4-wire (Kelvin) resistance measurement (hoping the low current flow through the long cable would minimize losses in measurement). Since the voltage is so small, I would think putting an amp between the DUT and micro controller might work, but I can't send more than 5V to the arduino and the range of possible voltages for the DUT doesn't let me amplify them linearly, either. Ideally, I'd want an system in the middle that would do the measuring and output as follows:

• Pass: 0.005-2V -> 2.5V-ish output
• Open circuit: Anything higher than passing V -> 5V output.
• Short circuit: Anything less than passing V -> 0V output

Any suggestions? Maybe some kind of modified window comparator that outputs one of three different states depending on input?

Rules:

1. No more than 50mA through the DUT (preferably no more than 25mA)
2. Only one available line (besides ground) to the aruduino (no two signal solutions, hence using the A/D for three states).
3. I can add a 5V regulator at the 12V side if needed but the fewer parts, the better.

I think you could just gain it up (X 60) and send that signal to the A to D. That would give you about .7 to 4.5 volts as a pass.
Use a good rail to rail op amp with low offset and put like 51 ohms in the output so it doesn't oscillate driving the cable.
If the supply voltage variation is to much for your tolerances you might need to regulate it.

 

If you want to measure the unknown resistance I would use a constant current source of about 20 mA. You mention a DUT resistance of between 0.5 Ohm and 2.8 Ohm so with a constant current of 20 mA source the DUT will output between 10 mV and 56 mV. I would think about setting up an IA with a gain of about 50 resulting in an output of about 0.5 Volt to 2.8 Volts. Run that into your uC and use your code for the pass fail criteria. Regardless of IA gain I would use a constant current source and generate a voltage drop across the DUT and measure the voltage drop. Then write in your pass/fail criteria.

Actually depending on what you want to do you could take the output of the IA and run it to a Window Comparator. Below the window is a short, in the window is a pass and above the window is a fail. That said I would just run into the uC and do all in the code.

Just My Take....
Ron

 

Thank you both for such quick responses. So it looks like I'll need to regulate my driving voltage(I'm using a lead-acid 12V, and allowed for 11.8 - 12.8V). Regulated, I should have more consistency to have a cleaner input range to amplify. A gain of 50-60x should get me far enough from the 0V low of short condition.

But (and forgive me if this sounds too noob -- I don't know amplifier circuits yet) wouldn't that gain give me an insanely high output for the short condition (5v regulator x gain of 60 = too big for the poor arduino)? Or is there some way to limit/clip the amplifier output to 5V?

@Reloadron: By IA, do you mean Instrumentation Amplifier? If so, would something like the MAX4208 or AD8226 do? If so, I've no where to start after looking at the data sheets (way over my pay grade til I get some more basics down). Any recommended primers or example circuits for these babies? I'm a quick learner but this isn't my comfort zone yet.

Thanks again for all your help.

 

By IA, do you mean Instrumentation Amplifier? If so, would something like the MAX4208 or AD8226 do? If so, I've no where to start after looking at the data sheets (way over my pay grade til I get some more basics down). Any recommended primers or example circuits for these babies? I'm a quick learner but this isn't my comfort zone yet.

Yes, IA would be an instrumentation amplifier, any one of several would do and the AD8226 would likely work out fine as well as others. Good IAs for instrumentation type work have become inexpensive and many require very few external components with a single resistor or pot used to set the gain. In my example I just use a 20 mA constant current source which can be easily built. In my approach I use a constant current source. Give me today and I will try and draw something up which should give you some ideas. Normally if we want to measure an unknown resistance (Runk) we pass a known current through Runk and measure the voltage drop across Runk.

Ron