Just a quick (fingers crossed ) whiz about diff op amps (again).
Was running into a situation with my single supply op amp on 5v (inv and non inv gain both 3.9) where it was outputting 0v when it should have been saturated.
Working great under conditions where input voltage diff is about 0-1.5 v but in one operating situation where voltage difference is say up to 8v and i would have thought output would have been right up against the supply (maybe im delduded) but instead its outputting 0v.
The reason is that i want it to have that kind of resolution when input difference is say ~1v. It doesnt need to deal with 8v difference....but it does need to show that it's 'off the scale' so to speak.
Its a MAX4389 just btw if that would make any difference.
Basically just wondering if this is supposed to happen with amps or whether it might be some other issue with the rest of the circuitry.

Thanks for reading (if you have been)

D

 

Without looking up that particular chip, I can say that a lot of op-amps invert their output when their input is outside the allowable range. Try following the rules and you'll get better results. Do not tell an amplifier with a 5 volt supply to deal with an 8 volt input.

 

The output certainly wasn't inverted and 'following the rules' means that the circuit is pretty much useless in the lower voltage difference. Surely you can saturate an amp without breaking the laws of the universe
Cheers,

D

 

We're obviously having difficulty communicating.

Applying 1.5 volts to an op-amp with a 5 volt supply is not breaking any rules.
If the output was supposed to be "saturated" and it was wrongly delivering 0 volts, what does saturated mean in this case, and what would you call the opposite of what you expected besides inverted?

 

Any fix that you can suggest given that I need around that gain (3.9) for voltage differences of less than 1.5 but still do the business with 8v or so?

D

 

You can clamp the inputs using diodes.
But, can you post your circuit?

 

I think I can fix this problem elsewhere in my circuit now.

Attachment is exact copy of this part of my circuit.
Having another issue now. This circuit works grand when Vreg input comes from a separate source (ie sense is isolated from load) but ideally, it needs to run from same source as load.
I have tried powering high and low of shunt but to no avail.
When i try to power load and sense circuitry from same source, the amp gives no change in output (again, working great using separate sources)

Help much appreciated as always

D

 

How are the grounds configured?
What is the Load?
You may be getting sag or noise on the power supply for the load.
Do you have a 22-100uf cap plus a 0.1 uf bypass cap on the input to the regulator.

And what did you mean by trying power on high and low side of the shunt resistor?

 

For sepatate supplies, Load ground is -ve of one battery. Other grounds are -ve of the second battery.
For single supply, all grounds are -ve of power source.
Load ranges from about 3-10 Ohmsish.
I tried drawing input of VR from above shunt, essentially +ve rail or below shunt (11.8 volts or whatever it is). Thought powering sense circuitry in parallel with shunt (and load) was causing a problem.

Would interference differ largely between sources? sense circuitry runs fine on 9v battery regulated without caps. But powering with regulated 12v causes probs.
I ran outta time at end for adding caps but I think I ended up just putting in 1 220ut (input-ground) just for heck of it

Thanks for response,

D

 

Bear in mind that many opamps can't drive their output more than a volt or so below the +V rail, and/or pull the output very close to 0V.

 

It is an MAX4389 rail to rail amp and gets to as-good-as-makes-no-difference 0v so that is not proving to be an issue.

Rather than being interference related, is there some sort of loop in the circuit that allows the power supply to mess with the amp in some way? In mean time, I will swap power supplies to see what happens

D

 

I am virtually certain that the problem is nothing to do with interference from the source.
powering the op amp without regulating (powered in parallel with load), output behaves normally. When I add a voltage regulator (either parallel to load or in parallel to both load and high side shunt) the op amp output stops working.
This is when the amp is not powered by the vreg and happens whether or not there is load on the vreg (ie adc and rest of circuit).

Can anyone help with this? Is there a 'shortcut' caused by the VR which causes the amp to lose control of the output because the voltage difference into the amp remains constant.

Many thanks,

D

 

What is your Vin on the regulator? Also, have you probed the inputs (at the op amp V+ and Vin) while this happens?

 

Vin on VR is between 11 and 12 volts, depending on the load.
V+ and V- still kick in and around the ball park. Tested a minute ago: they changed slightly, presumably due to load as i had the adc powered up from the VR but they only changed by about 0.1v.

Did some other tests where i powered the Op amp in parallel with load (about 12v) but i also powered the PIC chip in parallel with the load but i regulated this, even without caps. This setup works fine.
The regulated voltage itself doesn't cause a problem but it doesn't like powering the op amp with it.

thanks,

D

 

Where are you getting your 12V from?

Also, did you read the datasheet? It is rail to rail, but not when you use it as a differential op amp.

 

I am taking the 12v from between the high side shunt and load ie in parallel with the load.

Also, it is working just hunky dory when I:
-Power the op amp and PIC from a separate power supply, isolated from the load supply
-Power the Amp with the 12v (as described: parallel with load) and power the PIC from the same place but I regulate it to 5v
I haven't tested the upper limit but the lower limit of the op amp output is down at 0.01 and 0.02 volts. And I did scan through the datasheet and it says that it can only handle inputs that are less than Vcc+0.3v...But when Vcc if 5v (from separate supply, the op amp works great, even though the inputs are still 11->12v

Thanks again,

D