Unstable Op amp

[bee]

This is a circuit i have been workin on,,, Its a strain guage (SG 120 ohms)in a wheatstone bridge,op amp as a differential amp then the voltage is fed into micro and transmitted as a value ,I have been workin on this for some time now, It works but the voltage is unstable ,So when the strain guage is tensioned to give a reading of 3 volts it rises and falls by .4 of a volt,, this then give false reading in the micro
Iam stuck and have tried capacitors and different resistors
Any help would be much appreciated
cheers
bee

 

[JimB]

So when the strain guage is tensioned to give a reading of 3 volts it rises and falls by .4 of a volt

I assume the 3v is at the output of the second op amp.
What is the voltage when the strain gauge is in its resting condition (ie no load)? Could the load be varying?
How are you measuring this voltage, with a DVM I assume.?

this then give false reading in the micro

Have you tried measuring the voltage when the opamp output is disconnected from the PIC?

How is the opamp powered?
My first thought was that your input voltages may be a bit close to the 0v rail, but the opamp datasheet inferrs that rail to rail is OK.

How quick is the 0.4v rise and fall?
Could this just be amplified noise?
Can you look at it with an oscilloscope?

How are the connections made to the strain gauge?
How far is the gauge from the bridge completion resistors?


More questions than answers I am afraid.

JimB

 

[bee]

JimB

1 Yes the 3 volts is from the second amp
2 The voltage when guage unloaded is .6v measured with good dvm bench type
3 Yes i have tryied with pic disconected and no change
4 The amp is powered from the 5 v and 0v rail
5 The rise and fall is around .5 second
6 The connections are soldered to sg
7 The sg is around 3 inchs from bridge circuit
8 I looked with scope thats how i come to place caps were i did

I can take more readings from circuit if needed

Thanks for taking the time to look jim

bee

 

[ericgibbs]

This is a circuit i have been workin on,,, Its a strain guage (SG 120 ohms)in a wheatstone bridge,op amp as a differential amp then the voltage is fed into micro and transmitted as a value ,I have been workin on this for some time now, It works but the voltage is unstable ,So when the strain guage is tensioned to give a reading of 3 volts it rises and falls by .4 of a volt,, this then give false reading in the micro
Iam stuck and have tried capacitors and different resistors
Any help would be much appreciated
cheers
bee

hi bee,
This is the output response of your circuit when using LTspice, the plot shows the problem that you explain.

 

[JimB]

JimBs comments in RED

1 Yes the 3 volts is from the second amp
2 The voltage when guage unloaded is .6v measured with good dvm bench type
And is the voltage stable, or are there jumps when the sg is unloaded?
3 Yes i have tryied with pic disconected and no change
4 The amp is powered from the 5 v and 0v rail
5 The rise and fall is around .5 second
Is this a precise cyclic thing or apparently random?
6 The connections are soldered to sg
7 The sg is around 3 inchs from bridge circuit
OK so there are no long leads in the low signal part of the circuit which may be susceptible to Electro Magnetic interference?
The circuit is all soldered together and not built on a breadboard with possibly poor contacts?
8 I looked with scope thats how i come to place caps were i did

Eric
What op amps are you using in your simulation, just some generic opamp or an MCP6296 which is according to its datasheet is rail to rail on input and output?

To what do you atribute the "step" in your simulation?

JimB

 

[ericgibbs]

JimBs comments in RED
Eric
What op amps are you using in your simulation, just some generic opamp or an MCP6296 which is according to its datasheet is rail to rail on input and output?

To what do you atribute the "step" in your simulation?

JimB

hi Jim.
The generic opa is set for rail to rail operation.[ and below 0v input]
I will look at the sim more closely to identify the step problem.
Do you use LTspice.? I could post the 'asc' files.

This modified circuit works OK.

 

[JimB]

Eric

Yes the modified circuit seems to behave much better, but why did the original sim do what it did? and is this an indication of our OPs problem, I am not convinced.


No I am not a simulator user.

If I was into electronic design professionally it is most likely that I would be, these days electronics is really just a hobby.

I believe that simulators should be reserved until someone knows how to use electronic hardware and test equipment at a practical level.
So often on the board here I notice students playing with simulations with no idea of the real world implications, not good in my opinion.

JimB

 

[ericgibbs]

hi Jim,
To highlight the problem I have changed the 0V for the OPA's to -5V, so that the 'true' biassing state shows on the outputs.

I have tried a range of rail to rail OPA's with the same result.

If he trims the SG to 120R, the outputs are below zero.?
[The 4k7 and120R in the bridge I have set to 0.01%]

 

[bee]

Update on this circuit,
Yes i am guilty of building it on a breadboard,
But that was not the problem,
At the end of the board i have the transmitter doing its thing with the pic
Now fast running out of ideas i started disconecting ,pic,transmitter, and the problem is still around but now only 80mv to 100mv
this i can live with, I have moved the tx module down the other end of board and that is the result
i should have realised this as the tx is set to transmit every .4 of a second
i will move some caps around today and see if i can improve it

Thanks heaps Jim and eric

cheers
bee

 

[JimB]

When I commented about breadboard, I was thinking in terms of contact resistance in the bridge circuit.
Now you mention what I assume is a radio transmitter (UHF) all bets are off!
Breadboards are impossible at RF.
I dont know how you intend to build this in its working version, but something with a solid groundplane (eg unetched circuit board) combined with 1000pF decoupling capacitors on the signal lines will go a long way to eliminating the problem.

JimB

PS where abouts in Australia? I was in Perth on business last week.

 

[bee]

I have built a version of this on a vero board with just one of the amps in the mcp6292,And it works not to bad but i wanted more gain so thats why iam going with the both amps ,I will build this on a circuit board now
I have it working in front of me with the transmitter working and the output voltage is fluctuating around 50mv ,And that is great I have been messin around with this thing for a couple of years on and off
Yes the transmitter is 433mhz this works quite well for the job just upsets the amp hey

I am in melbourne on the mornington peninsula place called Frankston,, We had a big storm last night and trees are down everywere

cheers
bee

 

[bee]

I have found a instrument amp ad625 single resistor for gain,And it worked nicely not quite up to the rail voltage only to 3.4v this i can live with and sort it out in the pic
However when i connect to the transmitter module it pulls the voltage down to around 2v,Not sure what i can do this never happened with the mcp6292 amp but that was very unstable,The ad625 is stable just not sure what to do if anything
Any help or thoughts appreciated
cheers
bee

 

[crutschow]

I believe that simulators should be reserved until someone knows how to use electronic hardware and test equipment at a practical level.
So often on the board here I notice students playing with simulations with no idea of the real world implications, not good in my opinion.

I disagree about whether students should use simulators. I believe they are a very good way to get familiar with the operation of circuits before you build a breadboard and zap a device because the design was incorrect. Many of the people on this board have only a minimal idea of how circuits work and a simulator is a great way to learn. Certainly simulators can be misused, but they are an easy way to get familiar with the circuit basics such as ohms law, parallel and series circuits, AC circuit behavior, etc. without have to build a circuit.

 

[bee]

I have this working now just as i wanted it
Thanks for all your help

bee