Hello,

I am following this video,

https://training.ti.com/ti-precision-labs-adcs-selecting-and-verifying-d...

and I would like to plot the open loop gain as Texas did. Texas uses the Tina TI simulator but I would like to use the LTSpice. Unfortunately the simulation takes a long time. I never passed the step "Find the DC operating point for AC analysis.". Is there anything I could do to speed up the simulation?


Best regards,

 

hi waulu,
This is an example of the way it can be done in LTS.
E
Check out this video

 

Try changing C1 to 1 and L1 to 1meg.
The values of 1T for both are very high and may be causing convergence problems.

 

Hello ericgibbs,

Thank you for your reply.

Isn't that the closed loop gain?

From the datasheet, Figure 1,

https://www.ti.com/lit/ds/symlink/o...54153&ref_url=https%3A%2F%2Fwww.google.com%2F

I have the Open Loop Gain and it's very different if I use that method.



Best regards,

 

Try changing C1 to 1 and L1 to 1meg.
The values of 1T for both are very high and may be causing convergence problems.

Thank you crutschow. You are right, 100 Meg is already enough to simulate it.

 

hi waulu.
Please post the plots for the TI version of the sim, I would be interested in comparing.
E

 

hi waulu.
Please post the plots for the TI version of the sim, I would be interested in comparing.
E

Hello ericgibbs,

I don't know if I understood well your request. I post the result of LTSpice and the graph from the datasheet. If you mean the result with the Tina TI I don't have it and I never used it before. I can try it later this week anyway if it's relevant.

Thank you.

 

hi waulu,
This is an example of the way it can be done in LTS.
E
Check out this video

I checked the video, there was one last step that I needed to perform. The result is more close to the manufacturer open loop gain, but there is still a small difference,

 

hi waulu,
I will get the OPA340 model and run it in LTS, later.
E

 

there was one last step that I needed to perform.

What was that?
I don't see it from comparing your two simulations.

 

hi Carl,
I think in his first sim he just plotted one Node, the 2nd run the Vfb/Vinm.
E

 

I post the result of LTSpice and the graph from the datasheet.

If you start your plot at 0.1Hz it will look more like the datasheet.

 

hi waulu,
This is what I get with the TI Spice OPA340 model , running in LTSpice.
E

 

Hello ericgibbs and crutschow,

hi Carl,
I think in his first sim he just plotted one Node, the 2nd run the Vfb/Vinm.
E

Yes, that's correct.

hi waulu,
This is what I get with the TI Spice OPA340 model , running in LTSpice.
E

Yes, with that method we have the same result. But that's not what the manufacturer presents in the datasheet. On the other hand, the method that Texas TI presents in the video with the coil and the capacitor, already produces a result similar to the datasheet.

This is a wild guess, maybe it's due to how manufactures create the spice model? Analog creates the spice model in a way that the results are more reliable if simulated that way and Texas TI creates the spice model in a way that the results are more reliable the other way.

Best regards,

EDIT: To easily compare the results,

From datasheet,


Using Texas Instruments method,



Using Analog Devices method,

 

hi waulu,
This guy is very good with LTS, this vidio covers checking manufacturers models.
Check it out.
E
BTW: run your sim from 1Hz thru 100MHz

 

Alternative method:

 

hi waulu,
This guy is very good with LTS, this vidio covers checking manufacturers models.
Check it out.
E
BTW: run your sim from 1Hz thru 100MHz

Hi ericgibbs,

Yes, very interesting video. It's important to check the model always.

From 1 Hz to 100 MHz,

Alternative method:
View attachment 222055

Hi Bordodynov,

The phase looks better than the Texas Instruments method, only the gain is a little bit lower. Thank you very much, it's another option.

Best regards,

 

I have been reading about the stability criteria for AMP OP configured as unity gain buffer. I don’t seem to understand how to conclude if the system is stable or not through the Bode Plot.

I have attached two documents, the first one (op_amps_everyone.pdf) it seems to state that the criteria is the phase margin (higher the better) in the open loop gain, page 87.

The second document (Voltage-Feedback-Op-Amp-Gain-and-Bandwidth.pdf) it seems to state that the criteria is when the open loop gain cross the closed loop gain, it must have a slope of -20dB/dec or less, page 6.

Best regards,

 

The stability criterion as mentioned in both referenced documents is identical - however they are described in a different way.
The keyword is "loop gain".
* At the frequency where the open-loop gain Aol crosses the closed-loop gain Acl (horizontal line, continued to the right), the loop gain is zero. (Note, the dB-difference in the plot between Aol and Acl is the loop gain in dB).
* At this frequency the phase shift must not be larger than -360 deg (identical to zero deg). This is the stability limit, which for a 2nd-order system is identical to a loop gain slope of -40dB/dec (-12dB/Oct.) at the zero-crossing.
* In the BODE-diagram the phase inversion for negative feedback (-180deg) is not included - therefore, the mentined stablity criterion allows a maximum of only 360-180=180deg .
* For stability reasons, we often require a phase margin (distance to the stability limit) of app. 60 deg.
* Be careful when calculating/simulating the loop gain directly (including the neg. sign at the inverting opamp terminal) - now the critical phase is at -360 deg (0 deg).
* Comment: The above described case applies for the non-inv. opam stage only. This is, because only in this case, the closed-loop gain Acl is identical to 1/beta (1/feedback factor).
For inverting circuits it is recommended to use the loop gain criterion in its original form (measurement/simulation of the gain around the loop when it is opened at a suitable point)

 

The stability criterion as mentioned in both referenced documents is identical - however they are described in a different way.
The keyword is "loop gain".
* At the frequency where the open-loop gain Aol crosses the closed-loop gain Acl (horizontal line, continued to the right), the loop gain is zero. (Note, the dB-difference in the plot between Aol and Acl is the loop gain in dB).
* At this frequency the phase shift must not be larger than -360 deg (identical to zero deg). This is the stability limit, which for a 2nd-order system is identical to a loop gain slope of -40dB/dec (-12dB/Oct.) at the zero-crossing.
* In the BODE-diagram the phase inversion for negative feedback (-180deg) is not included - therefore, the mentined stablity criterion allows a maximum of only 360-180=180deg .
* For stability reasons, we often require a phase margin (distance to the stability limit) of app. 60 deg.
* Be careful when calculating/simulating the loop gain directly (including the neg. sign at the inverting opamp terminal) - now the critical phase is at -360 deg (0 deg).
* Comment: The above described case applies for the non-inv. opam stage only. This is, because only in this case, the closed-loop gain Acl is identical to 1/beta (1/feedback factor).
For inverting circuits it is recommended to use the loop gain criterion in its original form (measurement/simulation of the gain around the loop when it is opened at a suitable point)

Thank you LvW,

For the Bode Plot in the post #17, I have a phase margin of 220º correct? So it should be very stable.

Just to test the theory I would like to add a capacitor on the output that would produce an oscillation on the amplifier, but apparently it's a little bit hard. I can produce overshoot but nothing very dramatic. When I check the phase margin, it's always at least 180º. Should I conclude that I cannot make it oscillate? According to the datasheet, with a capacitive load higher than 1n I should have problems.





Best regards,