This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
I have a circuit that I originally put together a couple of years ago and it seemed to work quite well. It is an amplifier for a low impedance sensing element (around 25 ohm) that puts out a very small dc voltage. The current batch seems to exhibit a fairly large start-up offset that slowly changes and eventually stabilizes. This takes several minutes.
For testing purposes I disconnected the 217 from the second stage and removed the sensing element, shorting the inputs together. The amp is powered by +/- 12V, has a 220 ohm gain resistor for a gain of about 46, and has a 10k resistor from each input to ground. The reference pin is grounded.
On power up the output will show about a 8-10 mV reading that decreases to around 4 mV after 10 minutes. Is there something I can change to reduce this? Was I just lucky that the first batches didn't do this?
Confused,
Jerry
Hello Jerry,
Can you please post a schematic of the circuit?
Also, can you please take the measurement without the 10k resistors to ground? In other words, please try shorting the inputs and tie them directly to ground.
Hello Pete,
My old software won't let me cut and paste so I did this in paint:
For test purposes I disconnected the output from the second stage amp and replaced the sensor with a piece of wire. V+ is 12V and V- is -12V. There are also 1 uF and .01 uF caps on each supply. With the 10k resistors in place, I measure -24.2 mV on each input and the previously mentioned -8 mV at Vo that slowly drifts to about -4 mV. I can deal with an offset as long as it is constant, but the drift is a problem. I saw in the spec sheet that low input impedance can cause oscillation, but I don't see any on a scope.
Shorting the inputs to ground eliminates the voltage on the inputs, but doesn't change the voltage on the output. It is still drifting from about -8 mV to -4.
Further review of the specs in the parametric search engine raises more questions:
What is the distinction between single supply and dual supply instrumentation amplifiers?
The 217 has one of the largest Output offset specs (5000/G uV) and one of only 2 amps with an Output offset drift spec (20/G uV/deg C). These specs don't seem to be on the data sheet? I could probably live with an amp with lower bandwidth. (around 400 khz) Is there a better choice?
Thanks
Jerry
Hello Jerry,
The distinction
is generally with respect to the input voltage range.
Single-supply instrumentation amplifiers can typically accept inputs
voltages close to, if not including or below, the V- rail.
Perhaps the confusion surrounding offset is because the web-based selection guide explicitly differentiates between input stage and output stage offset. Please note that both terms will still be ‘referred to the input’ or RTI.
So, let’s start by looking at an example datasheet. The INA110, for example, shows the following:
In the product data sheet (PDS) the initial input offset voltage (referred to the input) is described as an equation, if you will. In the example above it is 250uV+(3000uV/G) for the ‘B’ version of the device. The first term (250uV) depicts the initial input offset voltage due to the input stage of the instrumentation amplifier. The second term (3000uV/G) depicts the initial input offset voltage due to the output stage of the device. A similar equation is given for the initial offset temperature drift (2uV+(50uV/G)).
Now, looking at the online selection tool we see the following for the INA110:
I have modified the columns to simplify the illustration. What is perhaps causing the confusion here are the terms “Output Offset” and “Output Offset Drift”. These are really referred to the input, but represent the offset due to the output stage of the device.
The INA217, for example, in a gain of 46 the maximum initial offset voltage referred to the input is Vios=250uV+(5000uV/46)=359uV. Multiplying the input-referred offset by the gain (46V/V) will now yield the initial offset that is observed at the output. For this example that would be 16.5mV. This is only the initial offset and does not include offsets due to shift, drift, etc.
For selecting a different part I will assume that your supply voltage (+/-12V) and gain (46V/V) are required. Perhaps you should look at the INA128/9. Looking at the high grade part, the initial offset voltage RTI is (50uV+500uV/G) and the drift is (0.5uV+20uV/G). For a gain of 46V/V this corresponds to an initial offset voltage referred to the output (RTO) of 2.8mV and shift (RTO) of 43uV/C. Unfortunately the BW at a gain of 46V/V (33dB) is ~350kHz, which is lower than your 400kHz requirement.
Concerning the drift of
~4mV over 10 minutes, I still do not have an explanation for it. Our initial thought was self-heating, but the
calculations do not yield such a large drift (I calculated just under 2mV). Also, we would expected self heating to occur in much less time. I will have to think about it more and will let you know if I have any ideas.
Hi Pete,
Thanks for the additional info. I will take a look at the INA128. For now I purchased a few 217's from a different distributor and they seem to work more like the ones I have used in the past. On five that I have swapped out, the initial offset is less than 4 mV and the drift is less than 0.5 mV . While this isn't perfect, it is acceptable and I will be able to finish this batch of sensors. Perhaps there is this much lot to lot variation?
Thanks again
Jerry
Hello Jerry,
Glad to hear you're obtaining better results and will be able to finish the batch of sensors. I am interested in obtaining one of the devices that is exhibiting the odd behavior. Please send me an e-mail and I will write back with my mailing address. I do not suspect lot-to-lot variation, but it is a possibility.