Teensy resolution - measuring low voltage

Status
Not open for further replies.

GOB

Active member
Hi All,

I am in a lab testing a transducer being developed.

The transducer takes an excitation voltage of 5 to 24V.

The transducer output is 0-100mV.

I have been told that for this sensor, at 5hte 5V excitation, I need a resolution of about 0.1 mV. Is this possible with the teensy 3.2 or teensy 3.6?

Speed is not much of an issue, it can be relatively slow (<100Hz).

Thanks!
 
0.1 mV for a 100 mV signal requires 10 bit resolution (log2(100 mV/0.1 mV)) -- the ADC can do that (it has 16 bits. but realistically only ~ 13 are 'good'). With a 1.2 V reference, the LSB is 0.14 mV. If you use the PGA (programmable gain amplifier) you can scale your input signal by 8x (so it is up to 800 mV), and et better resolution.

Note that 0.1 mV is a very small signal, you likely you will have noise and offset issues. Filtering can reduce the noise (and equally the update rate); depending on your sensor, you may need to do something additional if you need to eliminate offsets.
 
Thanks, Jp.

I did the calculation as well so I know 16 bits should be good, but I see that even National Instruments devices at 16 and 24 bits specify a resolution on the order of 200-300 uV when at low ranges. I assume this is due to noise.

Is the PGA built into the teensy? Any examples somewhere of using it?

The output I have seen so far from another source shows minimal DC offset with this sensor.

Thanks again for your help, appreciate it.
 
Using a Teensy 3.6 ADC in 16-bit mode with a lot of averaging (net output rate of 30 Hz) I saw a stability somewhere between 50 and 100 uV. At these levels all the analog circuit issues matter, such as thermocouple junctions between each and every connection and capacitive and inductive noise pickup from all kinds of sources.
 
Status
Not open for further replies.
Back
Top