Post Go back to editing

Amplifier gain offset configuration help

Hello,

I have somewhat of a general op-amp configuration question and need to get some feedback from someone more experienced in analog design than me.  I have the below schematics:

This is a single stage in an analog input channel for a product.  The input will have a constant offset of 1.65V.  This is because the sensor for this input channel is a wheatstone bridge.  If I want full DC performance my first question is, can inject an offset voltage of 1.65V to the feedback divider so I'm not gaining my steady state DC input content.  This looks like it works and I've found some app notes on the net talking about this, but I'm not the greatest analog engineer so I need some feedback.

This would be an idea configuration for me if it does in fact allow me to gain up all the content on the input that deviates from the 1.65V stead state input.  The second part to this (assuming this is in fact a legitimate design tactic) is how do I figure out the gains?  I've calculated a transfer function, but I can't factor it in a way to get Vo/Vi like you normally would.

So two questions:

1.  Does this make sense / is this a legit technique to accomplish applying gain to an input signal that has an offset?  I am trying to get away from AC coupling so this input can work from absolute DC to my cut Fc of the LPF.

2.  How do you figure out what your gains are now that there is an offset in the gain feedback path?

Rob

Rob,

You can write the nodal equations and the gain will be (1 + R1/R2).  Think of it

this way;  the 1.65V is a artifical ground.

The output will be  Vout = (Vin-1.65)*(1+R1/R2) + 1.65

Harry

• Hi Harry,

That is a lot easier than I expected.  Great answer.  Is there a recommended way of providing this reference voltage to the divider so that I don't mess with the impedances?  Should I use a high voltage reference or a high precision resistor divider buffered with an amp to drive this 1.65V ref. voltage to the gain divider?  I would think the output impedances of either cold affect the circuit.

• I actually tried to use the 8237 as my instrumentation amp and had a ton of problems with it.  I switched to the typical 3 amp configuration type of in amp and that works way better.  The 8237 had a clock sneaking out among some other problems so I had to give up on it.  As you mentioned though I am driving the ref. pin of the new instr. amp with a low output impedance buffer and that works great.

It sounds like I'll probably just grab a 3.3V reference and divide it in half by buffering a voltage divider.  This shouldn't be an issue for my design and is probably the simplest and most efficient way.

The InAmp guide looks great.  I'll spend some time reading through that.  Thanks for the help!  This has gotten me back on track I think.

• Rob,

Usually, people use a resistive divider from a reference voltage and then buffer w/ an op amp.

Here's another solution, but it depends on the highest signal frequency you have.

An instrumentation amp is Vout = (In1 - In2)*gain   +REF.  (But usually REF has to

be driven by a very low impedance to maintatain common mode rejection.

Some InAmp topologies get around this requirement.  See the AD8237 data sheet.

See fig 71.

Also see the InAmp guide:   http://www.analog.com/en/education/education-library/dh-designers-guide-to-instrumentation-amps.html

Harry

• Rob,

A lot of people have successfully used the 8237, so next time, post a question and let's figure out

why it's not working for you.  For a good general purpose op amp that's RR I/O, I would use