I apologize if this is posted in the wrong board but please forgive me I am new. I am attempting to precisely measure a zirconia oxygen probe sensor input (DC mV signal). My issue is resolution. Everything I have been able to find thus far is scaled for 0-5VDC or 0-10VDC at which point the resolution is not granular enough to see small changes. The effective voltage range I am trying to measure is only from about 1000mV-1200mV with required precision of a least 1mV but preferably greater. Any insight would be greatly appreciated.
Which controller do you use? An Uno can measure 0-1.1V with the internal 1.1V reference. With 10 bit ADC that's about 1mV resolution.
For more bits use an external ADC with typically 24 bit like HX711. There exist more solutions with more electronic circuitry (opamps...).
Theoretically, the best you can do with it 10-bit ADC (0-1023) would be a little better than 1.2mV (with a 1200mV reference).
You can use an external reference.
Ebay has modules which have good ADC, and here you can find libraries (that is, software) for them. Usually they are not expensive.
I would suggest you use an external A/D converter like this one.
Not just for the extra bits but for the delta sigma conversion. A ΔΣ converter will average the incoming signal throughout the conversion. This type of measurement is inherent less susceptible to noise on the input signal.
Why You Need an Analog Front End and How to Set It Up | Nuts & Volts Magazine -- offset and amplify your signal's 0.2V range into the full range of your ADC.
Thanks for the suggestion, I've since found that resolution is not my largest challenge. I also need input impedance ≥10MΩ. I've found several ready made options satisfying one or the other, but none for both. As I am not college educated in circuitry, I honestly have no idea the relationship between the two, if there is a relationship. Any light one could shed on the subject would be greatly appreciated.
I was browsing the data sheet for the A/D converter you recommended, It appears to be exactly what I have been searching for. Thank you for your time!
Happy to help. Keep in mind for the best noise rejection you want to use the slowest sample rate.
Then you'll have to fight dust, humidity and other environmental pollution. I think that you can not compete with professional (chemistry?) equipment.
If the A/D converter input impedance is not high enough for you, you will need to add a fet input opamp buffer before the A/D input.
You will have to consider layout, guard rings etc to deal with possible leakage over the hardware surface.
I purchased the mcp3421ev board and it works very well at 3.75 sps and a gain of 1, but with the current set-up, the data sheet says it should have 2.5MOhm impedence. As previously mentioned i know only slightly more than enough to be dangerous. Ive been advised that the life of the zirconia oxygen probe will be significantly reduced if the impedence is below 10M. I googled fet input opamp buffer and it looks like i have some learning to do before i can try to tackle the data sheet to figure out what im looking for. Any suggestions for PN's and/or decent learning resources? Thanks again for your time.
Bryan
Look at this design guide.
I find it odd that the above guide put a 10Meg resistor at the input when you information suggests the higher the better.
The zirconia sensor i am referring to is a very particular type of zirconia oxygen sensor specific to the heat treating industry. It is installed inside the heating chamber and used to calculate the carbon potential of the atmosphere. The manufacturer informed me of the impedence desired and claims lower impedence will lead to significant loss of probe life. Their sensor is found here: http://www.supersystems.com/files/WebManuals/GoldProbe/
See my #10. Even if you find an instrumentation amplifier with less leakage current you'll have problems in making use of that feature. Do you have a reliable meter for resistance above 10M?
I don't share @DrDiettrich's concern regarding the 10Meg Ohm input. The link I posted had an example ckt. However your sensor outputs only millivolts so there needs to be some gain added to the buffer amplifier.
We designed aircraft instrumentation that measured thermocouple signals. They had a much lower output impedance but were in the same mv range. The key is to heavily filter the incoming signal. You might ask the mfg if loading the sensor with a capacitor will shorten its life.
You will need a film capacitor for their low leakage capability.
You are trying to measure a signal that varies between 1000 & 1200 mV to 1mV res.
The MCP3421
can measure a differential voltage so you could offset the readings with a fixed voltage (but you will need a stable reference)
however single ended with a gain (PGA) of 2 you will read 2000 - 2400 mV to a sufficiently high resolution (as its 18 bit). Or even 16 bit at 15SPS.
The input IMPEDANCE is then 2.25 / 2
however
"This input impedance is due to 3.2 pF internal input sampling capacitor"
so the input RESISTANCE may be much higher.
I'd suggest if you try to add external buffering you will introduce a lot of drift.
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.