INA317: DC Offset-Nulling method

Hi Jagbir,

a useful method is shown in figure 1 of datasheet of XTR106:

Kai

Kai,

Can use these techniques for my application while using INA317 ? Generally speaking, out of XTR106 and INA317, which is better suited for a strain gauge sensor signal conditioning.

Hi Jagbir,

the method with the trimmer is normally used, if there's no other way to adjust the offset error.

Which method is better also depends on the resources of your circuit. Can you show a schematic of your circuit?

Kai

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/14/8171.Schematics.jpg

Generally speaking, out of XTR106 and INA317, which is better suited for a strain gauge sensor signal conditioning ?

Hi Jagbir,

here is your schematic:

Take care, the output voltage of INA317 cannot fully go down to 0V. See section 8.2.2.9 of datasheet. Is your input voltage unipolar or bipolar?

There are still some other issues in your circuit:

- I would decrease C7 to 10nF and insert a suited isolation resistor between the output of INA317 and C7. Have you already run a stability analysis in TINA-TI?

- TVS and Schottky diodes at the inputs will considerably increase the input bias currents and cause additional input offset voltage effects which might be hard if not impossible to be removed by calibration because of uneven temperature drifts.

Kai

Kai,

I will decrease the C7 value. The INA317  has a +/- 4KV ESD protection, I will need higher levels of ESD protection,up to 16 KV.

Jagbir

Hi Jagbir,

what leakage currents through the Wheatstone bridge can you allow? And at what ambient temperature will the circuit run?

Kai

The ambient temperature is 21* C. I don't have a number on the leakage current, however, I am planning to correct/compensate the dc offset  using 2 methods: Zero the o/p in Software and keeping the INA317 gain relatively low.

Ok, Jagbir.

I would like to repeat the other issue:

"Take care, the output voltage of INA317 cannot fully go down to 0V. See section 8.2.2.9 of datasheet. Is your input voltage unipolar or bipolar?"

Kai

My bridge excitation voltage as well as the input voltage is uniplolar, I have 3.3V and GND (0V).

Hi Jagbir,

have you read section 8.2.2.9 of datasheet as I have suggested you on April 30?

Kai

Kai,

yes I need to place an extra resistor on the low side of the bridge since I have I don't have a differential supply. What exactly is meant by ".. bridge output voltage is within the common-mode range of the amplifier inputs.."

Thanks,

Jagbir

Hi Jagbir,

do you know this appnote?

tidub00.pdf

Can you do the offset adjustment in software?

How much offset adjustment do you need to carry out? Is the offset voltage bigger than the maximum signal span?

Kai

Kai,

I can perform the offset in software.

Currently, I have connected pin5 (REF) to the ground. Since my power supply is uni-polar, do I need to provide a mid supply voltage on pin 5 or is it is ok to connect pin 5 to ground with a uni polar supply?

Thanks,

Jagbir

Hi Jagbir,

it depends a bit on the offset voltage adjustment range. Yes, I think it would be a good idea to bias pin 5 with a pseudo ground potential.

Do you have a free ADC channel to sample this pseudo ground? Would ease the offset adjustment in software...

Kai

Kai,

I do have an extra ADC channel. So you are proposing to sample the pin5 (Ref) and then perform an offset adjustment in software.

How is this approach superior to measuring the output of INAMP and the perform an offset adjustment in software.

Thanks,

Jagbir

Hi Jagbir,

now exactly happens what I wanted to avoid: We are discussing your circuit in two different threads...

I would recommend to create a pseudo ground voltage, maybe by the help of a simple voltage divider. Then, buffer this pseudo ground with a unity gain follower and feed pin 5 of INA317 with this voltage.

What voltage should be used for the pseudo ground?

The voltage should be as low as possible and as high as necessary. It all depends on your offset voltage originating from the bridge. So, if the offset voltage is only marginal, you can use 1V, or even lower. But if your offset voltage is huge and varying from bridge to bridge, then pseudo ground should eventually be set to midsupply. Or you adjust the offset voltage directly at the bridge, as I discussed earlier, then you can choose a much lower pseudo ground voltage again.

If you can sample the buffered pseudo ground voltage with a free ADC channel and do the offset adjust in software, then you need not to generate a very stable pseudo ground voltage. A simple voltage divider with the voltage directly coming from the supply voltage can do. That's the advantage from this method!

Kai

Thank you Kai and Jagbir. I will close this thread, reply to reopen.