Instrumentation Amplifier Input

[foetusmachine]

I have an instrumentation amplifier (LT1789-10) that is being supplied off by 5V-0V. I have the reference pin being tweaked to ~2.5v:

Under normal conditions, the LT1789-10 is being supplied with a DC input of up to ~200mV (2v output).

However, there are some situations where the non-inverting input will be subject to the full rail supply voltage (5V DC). I am not concerned with the validity of the output when this happens.

The absolute max input voltage of the LT1789-10 is 36V and the high impedance of the input (1.6G ohm) means minimal current will flow into the inputs.

Does this mean that the INAMP will be okay, except that the output will be out of spec when the input runs to 5V.

[ajb]

If you're within the datasheet margins, especially by as much as you seem to be here, then there's not much cause for concern.  One thing to check is if there is a V_in-to-V_supply limiting rating.  The datasheet says 36V for both V_supply and V_in separately, so that's probably not the case, but always good to check the rest of the datasheet for traps like that. 

Some devices may need time to recover from a saturated/invalid condition before they operate normally again, so that's something else to consider, depending on your requirements.

[dobsonr741]

The data sheet suggests the common mode range did not reach the positive rail. Bias goes dramatically up around 4.5V.
I assume you are operating it from a single 5V supply, is that right?

[foetusmachine]

Correct. Supplying it from a single 5v supply.

Refer here: https://www.eevblog.com/forum/projects/milliohm-meter/msg5147226/#msg5147226

This is the basic circuit. When there is no DUT in place, is when the input voltage to the INAMP rises to 5V.

[dobsonr741]

5V on input would not damage it, and the normal range does not seem to go that high to impact accuracy, so IMO it's OK.

[SiliconWizard]

The datasheet says that the input voltage should not exceed the supply voltage minus 1.5V in the OP's case, at first glance. They do not say if it's just to ensure proper operation, or if going higher could damage the amplifier.
They do not give a functional diagram either, so we can only guess its internals. It's probably safe to assume that you can apply any voltage up to the supply voltage + a diode threshold. They give an application schematic in which the input voltage could go up to the supply voltage. So, that should be ok. But the datasheet could have been clearer.

[foetusmachine]

Would it be worth throwing a forward biased Zener on the input of the current source to shave a couple of hundred millivolts off the 5V supply that leads into the LT1789's non-inverting input?

This means the LT1789 wouldn't need to do any work internally, assuming it might have had to do some work?

[magic]

The datasheet should specify absolute maximum rating on the input pins, but it doesn't.


There is only a simplified diagram, figure 1 on page 18.

It looks like you can't exceed V+ because the input is clamped by a diode. As long as you stay below V+, this diode is not a problem. There is a maximum input current rating of 20mA, which may refer to the maximum current this diode can take. There is also note 3: "this parameter is not tested"

Then there is 5.7kΩ resistor to an internal PNP. The PNP is either of a type that can withstand many volts of reverse base-emitter voltage, or of a type that can't. If it can, no problem. If it can't, it could (and should) be protected by a diode between the base and emitter. Either way, 36V maximum differential input voltage is specified, so this issue ought to be sorted out one way or another.

Nothing else is shown on the diagram which could possibly cause problems.


My bet is that you are safe. But encountering this chip myself would tempt me to see how much hassle it is to contact Linear Devices support and ask for clarification...

[foetusmachine]

I've decided to provide input protection that limits the voltage at the InAmp while the DUT is not in place:



This drops the voltage at the input of the InAmp to ~3.85V and draws ~320uA across the diode and resistor when the DUT is not present.

When the DUT is present, the protection diode and resistor draw between 6pA - 75nA, depending on the DUT. I can compensate for this to a degree when i set the current source, but the error there will be negligible overall.