Why Opamp with potentiometer?

[sagh]

It seems a good idea but I can't use this kind of idea for my whole circuit..

 

[sagh]

It seems a good idea but I can't use this kind of idea for my whole circuit.

 

[alec_t]

I can't use this kind of idea for my whole circuit.

I don't see why not.
I may have misinterpreted your .asc file.
Is the whole schematic showing just one amplifier, or (as I assumed) do Q2,6,7,8 form one amp and Q9,12,13,14 form a second amp?
What is the purpose of R2,R10 which link the non-inverting inputs of U2,U4 ?
Why R3,R6,R7,R11 ?
Are U1,U3 really necessary ?
V3,V4,V5 seem redundant. Unnecessary items slow down the simulation.

 

[sagh]

The whole schematic is just one amp. Actually this amp is a differential one. V3 and V4 may seem redundant but V5 not, because this is a differential amp. R3, R4, R7 and R11 aren't necessary I just put them to check the current.

 

[alec_t]

Here's the CMRR plot for your amp (albeit using LT1001 instead of LM358). Note the irregularity at low frequencies, due to tiny variations in the (very low) common mode gain. This irregularity can be eliminated by removing R2,R10,R15,R18.

 

[sagh]

Thanks, You put all of the elements in a better place and of course make the schematic much better. Why do R2,R10,R15,R18 make irregularity in low frequency?

 

[alec_t]

Why do R2,R10,R15,R18 make irregularity in low frequency?

I really don't know; but that's what Spice says .

 

[sagh]

I think you are right about LM358. when I use it instead of LT1001 for plotting CMRR vs frequency I got a wrong result . but the most interesting part is when I do transient analysis with LM358 to get the diffrential gain the result is correct.

 

[MrAl]

Hello,

Are we sure the LT1001 model properly models (or at least somewhat) the CMRR?
Check that first for one amplifier and compare to the data sheet. I thought i tried that device and had problems, but maybe it was another part number, i cant remember now.

 

[alec_t]

Here's the CMRR for the LT1001; as per datasheet and as per sim. Pretty good agreement IMO.

 

[sagh]

Hello,

Are we sure the LT1001 model properly models (or at least somewhat) the CMRR?
Check that first for one amplifier and compare to the data sheet. I thought i tried that device and had problems, but maybe it was another part number, i cant remember now.

With LM358 I have the negative CMRR. That's why I think something is wrong with the LM358 model. while in transient analysis when different SINE voltages are applied I have a correct gain. what things in datasheet should I check for LT1001?

 

[MrAl]

Hi,

Did you test the LM358 in the same circuit Alec used to test the LT1001?

 

[sagh]

Hi,

Did you test the LM358 in the same circuit Alec used to test the LT1001?

Here is the CMRR for LM358. It's like the plot in datasheet.

I get negative CMRR when I use LM358 instead of LT1001 in my circuit. why?

 

[MrAl]

Hi,

Well going by the plot itself, it tells me that the CMRR gets extremely BAD after about 1MHz.
Going by the data sheet, the LM358 bandwidth is limited to 1MHz with a gain of 1 so testing higher than that will result in very poor performance anyway, even if the CMRR was very good.

 

[alec_t]

Your negative CMRR result may be due to the following problem with the LM358 models

Helmut Sennewald 8/3/2004
None of the models has correct quiescent current.
The NATIONAL and ONSEMI part fail totally for Vdd < 1V.
The NATIONAL part has negative supply current for Vdd < 10V.
The ONSEMI and the ST part don't draw the load current from the supply.
The supply current of the ST part is far off for Vdd < 2.5V .

 

[sagh]

Hi,

Well going by the plot itself, it tells me that the CMRR gets extremely BAD after about 1MHz.
Going by the data sheet, the LM358 bandwidth is limited to 1MHz with a gain of 1 so testing higher than that will result in very poor performance anyway, even if the CMRR was very good.

The thing that I got from your post is LM358 would be a bad choice for my circuit due to the limitation in bandwidth and having poor performance in high frequency. but I really don't understand why I have negative CMRR while I get a correct differential gain. I hope the wrong result is because of LM358 models' problems that alec mentioned.

 

[sagh]

Your negative CMRR result may be due to the following problem with the LM358 models

I hope so. What are NATIONAL part, ONSEMI and ST part? I have no clue about these things. :-(

 

[audioguru]

National Semiconductor (NATIONAL) invented the LM358 dual and the LM324 quad low power opamps a long time ago. National was purchased by Texas Instruments a couple of years ago.
ON Semiconductor and ST Micro Semiconductor companies produce copies of the LM358. Their spec's are all the same.

 

[MrAl]

The thing that I got from your post is LM358 would be a bad choice for my circuit due to the limitation in bandwidth and having poor performance in high frequency. but I really don't understand why I have negative CMRR while I get a correct differential gain. I hope the wrong result is because of LM358 models' problems that alec mentioned.

Hello again,

I fail to see what you are talking about. Dont you read all the posts? I told you that you should not test the LM358 beyond 1 Mhz, but even if you still want to you can still see the "CMRR" but it will be very BAD. A negative CMRR would indicate an extremely bad CMRR, so bad that the common mode gain becomes greater than the differential gain rather than the other way around. But that's not REALLY the case until you exceed the recommended operating spec's for the op amp anyway as i tried to point out already.

You are not seeing a negative CMRR, at least not up to the useful range of the op amp, so i can understand why you would say that you are. If you see positive for MOST of the test range and negative near the end, why would you call the entire result negative? That doesnt make any sense. And for an extreme example, would you like to test the LM358 at 1GHz (1 gigahertz), 10GHz, and 100GHz too? Im sure it will look pretty bad at 100GHz. But then why stop there, lets test at 100THz soo ha ha

Did i miss something or misunderstand you? Just let me know then and i'll see if i can agree then, but from what i am seeing so far the CMRR is positive.

See the attachment showing the part you are referring to as 'negative' in red and below 0db. Note that the op amp frequency limit is around 1Mhz.

I might also add that any of these CMRR tests are going to be purely academic anyway because no model can accurately show the CMRR unless we know the range of possible mismatch of the internal transistors. Since we dont really know that, we can never model the CMRR.
What we can do however, is vary certain parameters of the models or the circuit to see what it's effect on the op amp CMRR are.

 

[sagh]

You are not seeing a negative CMRR, at least not up to the useful range of the op amp, so i can understand why you would say that you are. If you see positive for MOST of the test range and negative near the end, why would you call the entire result negative? That doesnt make any sense. And for an extreme example, would you like to test the LM358 at 1GHz (1 gigahertz), 10GHz, and 100GHz too? Im sure it will look pretty bad at 100GHz. But then why stop there, lets test at 100THz soo ha ha

If I use LM358 in the circuit that I attached to this post CMRR is negative . As You can see from 1Hz untill 100khz CMRR is -10db then it decades and becomes worse. while I did test LM358 in a simple amplifier in post number 233 and saw it worked very well at least untill 1Mhz.