Offsetting mv input.

[Stanfi]

I have a situation. I have a chip that to which I am going to input 0-30mv and it will give me 0-10V on the output.

My load cell with give me 0 to 36mV However the bottom 5mv is "dead load". The signal will never go beneath that. So in order to increase the resolution on my output, I want to offset that voltage.

They way I am currently doing it is by using a NMR106 12 in 5 out Isolated DC/DC converter. In series with the isolated 5V output I have a 49K resistor and 500ohm potentiometer. across one leg and the wiper of the pot, I will get 0-50mv.

I connect this adjustable 0 to 50mv so that is negative in series with my 0 -36 mv load cell singnal, and it will allow me to offset my dead load.

This does work, but my question is, if there is a better way to do it?

Thanks

 

[IRstuff]

What's wrong with an analog 5V reference? They draw very little current and are very quiet. For that matter, can't you pull a 5V reference from the ADC?

TTFN

 

[Skogsgurra]

The normal way of doing this is to add some positive offset to the negative input of the opamp that does the signal conditioning. The offset can be derived from your supply voltage, as IR says. Can you not do that with your "chip"?

If you cannot use an ordinary op amp configuration, your solution seems to be as good as any. I can see that you need to add an isolated bucking 5 mV signal in series with the load cell signal before you input it to your chip.

I can also see that you cannot refer the bridge output to a negative voltage if you do not have one. Or do you?

It would help a lot to know something about your "chip". Type for instance.

On the other hand - why bother? The 5 mV dead meat is only about 15 percent of full range. You will not have a much better resolution if you shave that off. But you will have more problems with added noise and shifting zero. I would leave it as is and scale/offset on the 10 V side.

 

[Stanfi]

The chip I am using is a Burr Brown INA125

 

[Stanfi]

The reason that we are wanting to offset the mv singal is that the 0-10v signal is going into an A to D converter with on 11 bit resoultion. We are trying to increase the performance on the A to D by getting rid of the dead load voltage.

I like the idea of offsetting the output side of the 0-10. I I can get my signal to increase above 10 that is what I will try to do.

 

[Skogsgurra]

OK, that's an instrument amplifier. Which is a set of opamps with resistors built in.

You can do the offset adjustment by connecting the IAref pin (5 on the ordinary DIP package) to an external voltage. Preferably a stabilised negative voltage derived from your negative spply.

If you run with a single supply, you cannot do that. You could offset the bridge by connecting two resistors parallel to opposite bridge sides (not recommended if you are not very familiar with these things) and compensate for the reduced gain. Again, you do not lose much if you leave things the way they are.

 

[felixc]

If you do more than one unit, please note that the "zero load" output of the load cell will probably be different from cell to cell. And unless the gain has been carefully adjusted at factory (which means a high-end load cell) the gain of your load cell will change from cell to cell.

 

[Stanfi]

Ok, what I decided to do was use the INA125. the mv input from the load cell goes directly into the chip. I set my gain for 0 to 36mv to 0-13volts.

I am not using the chip reference for exitation, but stayed with my LM723 voltage regulator as I needed a minimum of 12V excitation.

I am using also using the LM723 to power at NMR107 12V in 12V out isloated dc dc coverter. Then using a 2K resistor in series with a 1K pot. I am able to get 0 to amlost 5V for an offset on the 0 to 13 side.

Seems to work well, and a bench showed good stabilty.


Thanks to everyone for your input!

 

[Stanfi]

With my design I applied the offset voltage to pin 5. Which on the board with a meter worked great. However, I am connecting this voltage to an Analog input on a ZWORLD OP7200. It appears that the internal grounding of the Anaolog Inputs cause a problem with my offset circuit.

I did some tests and I can apply a negative offset voltage to pin 10 and apply the offset on the postive side. This seems to work well, but is it a good way to do it?

Thanks