Unable to acquire flat 0V signal from op amp

[Plasmateur]

Hello,

I am trying to wire an op amp and send two identical signals to the inverting and non-inverting inputs for a sanity check.

The signals are
100kHz
1Vpp

If I have wired everything correctly, then I should simply get a flat voltage - (+IN)-(-IN) = 0V

Attached is the diagram.

The op amp is here: http://cds.linear.com/docs/en/datasheet/lt1226.pdf

The Null pins need to be grounded in order for it to work.

I am unsure as to why I receive a signal between 260-600mV, which looks like shark fins on my scope.


Also, when switching out resistors Rf & Rg for 530R (shown in picture not drawn by hand)
if my inverting input is off, I receive a jagged signal around 1Vpp.
However if my non-inverting input is off, I receive a signal of 4Vpp.
I'm unsure why.

 

[Richard Crowley]

What precisely do you mean by "two identical signals"?
Where did these "two identical signals" come from?
How do you know that they are "identical"?

What happens if you REALLY input exactly the ONE SAME signal into both the inverting and non-inverting input?
That is the only way you will genuinely have "two identical signals".

Furthermore, if you seek complete perfect cancellation of the inverting and non-inverting signals, you will need to precisely match the value of the resistors.  Have you done that?  What is the actual measured value of all the resistors?

When I look at the "Simplified Schematic" diagram on page 8, it appears that if you ground the "null pins" (assumed to be pin 1 and pin 8?), it will render the input stage of your op-amp inoperable, so it is a wonder that you are getting anything form the output.

[Andy Watson]

What gain have you set? The first line of the datasheet says "Gain of 25 Stable".

[Plasmateur]

What precisely do you mean by "two identical signals"?
Where did these "two identical signals" come from?
How do you know that they are "identical"?

What happens if you REALLY input exactly the ONE SAME signal into both the inverting and non-inverting input?
That is the only way you will genuinely have "two identical signals".

Furthermore, if you seek complete perfect cancellation of the inverting and non-inverting signals, you will need to precisely match the value of the resistors.  Have you done that?  What is the actual measured value of all the resistors?

When I look at the "Simplified Schematic" diagram on page 8, it appears that if you ground the "null pins" (assumed to be pin 1 and pin 8?), it will render the input stage of your op-amp inoperable, so it is a wonder that you are getting anything form the output.

Thank you for taking the time to reply.

These signals come from a SDG2042X dual output AWG. Before the signals go to the breadboard I check to make sure they are at the same phase and Vpp on the oscilliscope, but that is done via BNC. I just checked on the breadboard with some probes just to make sure.

I will attempt to use the same signal on both inputs of the opamp....and I still get a signal.

[amspire]

You do have the negative supply pin of the opamp going to a negative supply rail and not 0V?

What are the supply voltages?

[Plasmateur]

What precisely do you mean by "two identical signals"?
Where did these "two identical signals" come from?
How do you know that they are "identical"?

What happens if you REALLY input exactly the ONE SAME signal into both the inverting and non-inverting input?
That is the only way you will genuinely have "two identical signals".

Furthermore, if you seek complete perfect cancellation of the inverting and non-inverting signals, you will need to precisely match the value of the resistors.  Have you done that?  What is the actual measured value of all the resistors?

When I look at the "Simplified Schematic" diagram on page 8, it appears that if you ground the "null pins" (assumed to be pin 1 and pin 8?), it will render the input stage of your op-amp inoperable, so it is a wonder that you are getting anything form the output.

Thank you for taking the time to reply.

These signals come from a SDG2042X dual output AWG. Before the signals go to the breadboard I check to make sure they are at the same phase and Vpp on the oscilliscope, but that is done via BNC. I just checked on the breadboard with some probes just to make sure. It appears the voltage to the non-inverting input is slightly less. Using the same signal on both inputs yields the same result.

Without a proper power supply, I think it might be hard to find the resistance of each resistor, but I can attempt this. Perhaps I can find a battery laying around.

Perhaps I had the setup wrong when I grounded both nulls. I just set up everything on the breadboard again.

If what I am seeing on +IN and -IN is true, then the difference being displayed on my oscilloscope it wrong. So I'm a bit at a loss.

[Plasmateur]

You do have the negative supply pin of the opamp going to a negative supply rail and not 0V?

What are the supply voltages?

I have two 5 volt power supplies: https://jet.com/product/uxcellr-AC-110V-220V-to-DC-5V-3A-Switching-Power-Supply-Driver-for-LED-Strip/9c548ca7bd2a490baa8c70135ac3cdee

V- tied to ground on one and V+ tied to ground on the other. Checked on the scope. So I have +/- 5V going to the op amp on pins 4 and 7 which are V- and V+

[Plasmateur]

What gain have you set? The first line of the datasheet says "Gain of 25 Stable".

Ahh, perhaps that is the problem. I have a gain of 1 set. If I have a 2.5k Ohm resistor as Rf and RG that should set the gain to 25 if the other two resistors are 100 Ohm.

[Plasmateur]

I kicked it down to 1 kHz and 50mV and I believe I set the gain to about 5.

I believe I need the correct Caps, .1uF. But used what I had - 20uF.

The signal is a bit noisy.

I then set one channel to be 180 degrees out of phase with another and got a nice looking sine wave with .220Vpp

Then I turned one channel off and unplugged it and got rubbish. I'm a little confused about that.

[amspire]

Good luck getting a 1GHz opamp to work in a breadboard. You will be much better off using Dead Bug breadboarding.

Have a look at this, especially the photos at the end:

http://cds.linear.com/docs/en/application-note/an47fa.pdf

If you can start with a low frequency opamp like a TL071, you will get much better results.

By the way, watch your scope tip - you can break it that way.

[Plasmateur]

Good luck getting a 1GHz opamp to work in a breadboard. You will be much better off using Dead Bug breadboarding.

Have a look at this, especially the photos at the end:

http://cds.linear.com/docs/en/application-note/an47fa.pdf

If you can start with a low frequency opamp like a TL071, you will get much better results.

By the way, watch your scope tip - you can break it that way.

Thank you thank you and thank you!

[Audioguru]

A TL071 (and many other ordinary opamps) will also oscillate on a breadboard because it has a response up to 3MHz and the stray capacitance between all the wires and rows of contacts on the breadboard cause positive feedback. A TL071 will also oscillate if it feeds a shielded audio cable (the cable capacitance causes phase shift) unless the opamp circuit feeds a 100 ohms resistor in series with the cable.

I have never seen an opamp with such a high frequency response (microwaves) and low gain as the LT1226 opamp. It oscillates without any stray capacitance when its gain is less than 25.