instrumentation amplifier built with LM358N opAmp... why doesn't it work?

[snaggy]

It's a flaw in my design of course, since it doesn't work in SPICE either!
Sorry if this is a stupid problem, but it's my first little project..

The problem of course is that U2A doesn't feedback correctly, voltages on + and - inputs are different and output doesn't make any sense to me..

thanks!

 

[ericgibbs]

It's a flaw in my design of course, since it doesn't work in SPICE either!
Sorry if this is a stupid problem, but it's my first little project..

The problem of course is that U2A doesn't feedback correctly, voltages on + and - inputs are different and output doesn't make any sense to me..

thanks!

hi,
Looks OK, but there is no supply to U1B.???
Post your LTS asc file and I will try it

Hi,
This is the LTS sim. the problem is with the input voltages as you are using a single OPA supply.
The input voltages you show try to drive the output below ground.
Change them over.

EDIT:
The sim shows an input voltage sweep of +2V thru +4V

 

[snaggy]

there's no supply because it's a dual opAmp package, so you don't need to expllicit the supply to both opAmp.
Here's the asc file (hope I exported it right, I'm using MultiSim 11).
Please remember though that I realized the circuit on breadboard and it doesn't work either, it's not simply a simulation issue (the program simulates it correctly, actually)
thanks

 

[schmitt trigger]

U2a's inverting input (pin 2)is at 1.67 volt
U2a's non-inverting input (pin 3) is at 1.25 volt

The opamp is doing its thing, amplifying this differential voltage times its huge DC gain, and swinging negative as far as it can go...within 4.53 mV of grund.

 

[snaggy]

I knew it was something simple.. thanks!

anyway, I have a further question: since my amplifier is used with a temperature sensor (lm335) which gives me about 3 volts at room temperature, but an lm358 has a max output high voltage ov 3.5 volts, the amplification is not much (about 1 volts max, when the first opamp gives 2.5 and the second 3.5)

How can I better amplify a signal around 3 volts, with reference to a base voltage of 2.9, using lm358? (I don't want to use voltage dividers because I think I'd lose resolution, right? it's only 10mV for °C and I'm measuring differences of a few °C...)

thanks!

 

[schmitt trigger]

There are ways to do exactly what you want to do with a single opamp. Thomas Frederiksen's "Understandig Opamps" gives such a technique...let me see if I can find it.

 

[ericgibbs]

I knew it was something simple.. thanks!

anyway, I have a further question: since my amplifier is used with a temperature sensor (lm335) which gives me about 3 volts at room temperature, but an lm358 has a max output high voltage ov 3.5 volts, the amplification is not much (about 1 volts max, when the first opamp gives 2.5 and the second 3.5)

How can I better amplify a signal around 3 volts, with reference to a base voltage of 2.9, using lm358? (I don't want to use voltage dividers because I think I'd lose resolution, right? it's only 10mV for °C and I'm measuring differences of a few °C...)

thanks!

hi,
This is one way to amplify the 10mV/Cdeg signal referenced to +2.9v
For a 2.9Vt to a +3.1Vt input signal you get 0V thru +3V output.
ie: 200mV input change equivalent to 20Cdeg change.

I Did explain in post#2 why your circuit did not give the required output.

 

[MRCecil]

I knew it was something simple.. thanks!

anyway, I have a further question: since my amplifier is used with a temperature sensor (lm335) which gives me about 3 volts at room temperature, but an lm358 has a max output high voltage ov 3.5 volts, the amplification is not much (about 1 volts max, when the first opamp gives 2.5 and the second 3.5)

How can I better amplify a signal around 3 volts, with reference to a base voltage of 2.9, using lm358? (I don't want to use voltage dividers because I think I'd lose resolution, right? it's only 10mV for °C and I'm measuring differences of a few °C...)

thanks!

The good and honorable Mr. Gibbs has shown you the basic problems with your design. However, he didn't have the LM358 .sub file to allow a full evaluation. Such is the limitation of some simulators and taking the time to search for the Spice model, which can be time consuming in the least and fruitless most often, not that I am an expert user of LTSpice by any means.

I had the luck to stumble upon a working Spice model of the LM358. The results I got were less than favorable for a functioning and accurate instrument, which you are questing, as I thought might be the case.

I will include the results in attachments below. These will be of a simplified design with a more accurate, less noise filled 2.9V reference and fewer total parts. One will be with the single-ended supply and the other with a dual supply. The differences are pronounced.

Please note that the in the first case the limitations of attempting operation of the LM358, with a single-ended supply from 0V to 200mv, including the initial offset and the linearity; the datasheet says operation near ground for a single-ended supply not to ground. In the second case, one can see a true linear response from as little as 1.5V applied to pin 4 (gnd or V-).

The schematics are annotated for clairity. Bottom line is it does not appear a single ended supply on an LM358 for this use will yield accurate results. I did try other amps, and the TLC277IP had the least non-linearity at the lower end single ended of those I tested.

EDIT: The dual supply schematic (on right) should read -1.5V on neg. rail input of U3 vice -5V.

 

[MRCecil]

hi C,
Thanks for the kind thought, but I do have the LM358 in my LTS lib.
[member of the Yahoo LTS users group]

The circuit I posted was based on a CA3140 opa model, I was just trying to make the OP aware in the way he had connected his voltage input sources, he was attempting to drive the output below 0V, which he could not do because of the single supply.

I am always interested in other members *.asc files and their techniques, so I would ask that you post the LTspice *.asc file.

My apologies for the improper assumption...I should have thought beyond a cursory glance at your schematic given your experience displayed on this forum.

Yesterday I was looking for an amp that would swing below the lower rail, but was unable to locate one. We're 500 miles away from home and that distance from all my reference material, and will not be returning until October. All I have is my calculator, laptop and the slow WIFI in this Montana campground.

Yes, the CA3140 was a good choice for demonstration purposes as displayed in your plot above. The LM358 needed ~300-400mv more room to get into the linear range from my tests.

I'll be happy to send the .asc file. It should be attached.

Cheers,
Merv

 

[ericgibbs]

hi C,
Have downloaded your asc file, will run it later.
Ref the below 0V for an OPA used with a single supply, I have never seen such a device.?
I find that most often the OP is trying to get a voltage that swings from close to 'just above' 0V and upto say +3.3V or +5V to suit the analog input of a PIC/MCU.
Depending upon the sensor that is outputting the source signal, the range over which the OP is using the sensor output, is usually well away from the 0V level and the OPA saturation voltage.
So the OPA output voltage is linear over the range of interest.

I like the ageing CA3140/CA3240, its cheap and has a very input impedance and will swing very close to 0V, with a saturation of approx Vs-1.5V.
Regards

 

[muktinathgogoi]

have a problem frnz.... i am using instrumentation amplifier for my ECG signal display in cro...
i m using lm358 opamp to design it... but when i check it in dc supply it will work properly , but in ac supply it does not works.. my ecg signal are not comming in the output..
is there any problem regarding the opamp input power?? or some other reason?? plz reply if you know...