2000 Gain Preamp circuit 20Hz/20kHz help?

[Titan]

Hi so I need to design a op amp circuit that meets the following criteria:

2V/mV (Gain 2000) max input of +/- 15mV
20Hz to 20kHz (19,980Hz Bandwidth)
50k ohm input impedance
Using cheap readily available lab components (student)*.

*Now in our lab all we have to hand are LM 741s so i've tried using these but am I right in thinking this is far too much gain/bandwidth for the LM741? I got as far as calculating my capacitors, 159nF and 159pF if I am correct (I know i will need to use caps of other values to get as close as possible in reality, this is purely simulation at this point but I do need to build it later).

My issue comes when working out my gain per stage. I planned on using 2 741s for reasons mentioned previous but my GBP is well over 1MHz so achieving a gain of 2000 sounds hard. What is the maximum gain I have per stage? 50? Any advice on how I do this using 741s? Or a better alternative op amp? (because no where is it stated that I should use 741, preferably cheap though because student).

Much appreciated

 

[ronsimpson]

Two sages of 44.7 gain, but you are right at the limit of the amp. or Three stages of 13.
Do you know how to make a single stage with gain of 44 or 13?
Do you have a SPICE program so you can check your work?

 

[Titan]

Hi Ron

Yes I have MultiSim 14, access given by uni.

I have drawn what I think it should look like for 2 741's (Note I wasn't sure of the gain so all caps/resistors are at default for now.) Does this look correct?



I had been working on the assumption that i'm too close to the 741's limit and had been looking for alternatives. Would the TL082CP with a GBP of 4MHz be more suitable?

 

[ronsimpson]

That a good start.
Inside the red circles are the components to set the low frequency point. The three RCs will almost add up so maybe start off with 10z and see what happens. So you know f=1/(2x3.14xRxC)
I removed C2. If you need to slow down the amp put a very small cap across R2.
Change R4 to 1K.
Now do a frequency sweep and see what happens.

 

[Nigel Goodwin]

I had been working on the assumption that i'm too close to the 741's limit and had been looking for alternatives. Would the TL082CP with a GBP of 4MHz be more suitable?

Pretty well ANYTHING is better than a 741, I don't want to turn in to AudioGuru but a 741 is a 50 year old low spec antique, to be fair the TL0xx range are pretty old as well, but at least have a fairly reasonable spec and excellent devices - I'm personally using TL072's at the present time.

 

[Titan]

That a good start.
Inside the red circles are the components to set the low frequency point. The three RCs will almost add up so maybe start off with 10z and see what happens. So you know f=1/(2x3.14xRxC)
I removed C2. If you need to slow down the amp put a very small cap across R2.
Change R4 to 1K.
Now do a frequency sweep and see what happens.
View attachment 121987

So I rearranged f=1/2piRC to get C= 1/2piRf. Resulting in C values of 159nF (1/2pi*50k*20) and 159pF (1/2pi*50k*20k) to calculate my capacitor values for filtering 20Hz-20kHz. How can I remove C2? I thought it was necessary for the high frequency point?

 

[Nigel Goodwin]

So I rearranged f=1/2piRC to get C= 1/2piRf. Resulting in C values of 159nF (1/2pi*50k*20) and 159pF (1/2pi*50k*20k) to calculate my capacitor values for filtering 20Hz-20kHz. How can I remove C2? I thought it was necessary for the high frequency point?

C2 is part of a low pass filter, and will stop any high frequencies going through with the default values you had. You don't need that, or it's resistor, add small capacitors across the feedback resistors.

Personally I think the specification is very poor - you're chasing 20Hz-20Khz - this is just the standard audio bandwidth, and not of any great consequence.

For a 'practical' design I wouldn't even consider calculating the coupling capacitors, just stick a value in that's 'more than large' enough, and perhaps a 'little' calculation on the feedback capacitors (which you don't seem to have) to roll off the HF somewhere above 20KHz.

If for any crazy reason you needed sharply defined filtering, then you'd usually do it in the filtering/tone control stages - but this was all in ancient history to remove rumble etc.

 

[audioguru]

I also agree that your antique 741 opamps should be kissed goodbye then buried. Does your school also teach that the world is flat?

Your opamps each have a voltage gain of only 2, not 44.7 that you want.
Your 20kHz lowpass filter has a very gradual 1x RC drop-off and the three 20Hz highpass filters combine and produce a very droopy drop-off.

 

[audioguru]

I think the preamp is for a microphone and the 20Hz highpass cuts the putt, putt, putt sounds from motorcycles and the 20kHz lowpass cuts ultra-sonic sounds from bats that dogs can hear.

 

[Titan]

C2 is part of a low pass filter, and will stop any high frequencies going through with the default values you had. You don't need that, or it's resistor, add small capacitors across the feedback resistors.

Personally I think the specification is very poor - you're chasing 20Hz-20Khz - this is just the standard audio bandwidth, and not of any great consequence.

For a 'practical' design I wouldn't even consider calculating the coupling capacitors, just stick a value in that's 'more than large' enough, and perhaps a 'little' calculation on the feedback capacitors (which you don't seem to have) to roll off the HF somewhere above 20KHz.

If for any crazy reason you needed sharply defined filtering, then you'd usually do it in the filtering/tone control stages - but this was all in ancient history to remove rumble etc.

Yes I agree the spec is poor but as is often with school work you wouldn't ever do it in real life for a frequency range that is the entirety of the human range of hearing.

I also agree that your antique 741 opamps should be kissed goodbye then buried. Does your school also teach that the world is flat?

Your opamps each have a voltage gain of only 2, not 44.7 that you want.
Your 20kHz lowpass filter has a very gradual 1x RC drop-off and the three 20Hz highpass filters combine and produce a very droopy drop-off.

No they don't teach the Earth is flat but professors being old and recycling their content means 741's stay in use. It doesn't specify that I use 741's here so I was considering trying TL082CP instead.

I have however started again with 741's in 3 stages 4-4-125. I have attached this circuit but I remember being told as much gain as possible should be in the first stage anyway.

EDIT:

Input is set to 10mV. I get an output of 14V because i only gave 15V supply, my mistake. Ignore that clipped wave.

Looking at the bode plotter is not what i expected though. It’s my first filter circuit but I expected a sharp drop off outside the frequency range permitted, instead it’s very gradual is that correct?

 

[Ylli]

Input is set to 10mV. I get an output of 14V when I expected 20V (to achieve 2V/mV) so something is obviously still not correct.

Do you really expect to get an output greater than your supply voltage? (Or does this program assume peak-peak values)

 

[Titan]

Do you really expect to get an output greater than your supply voltage? (Or does this program assume peak-peak values)

Wow i was being especially dense there. That’s exactly why its clipping... I’ll lower input to 1mV, was just the default for some reason.

Thanks.

EDIT: It’s the bode plot that is more “useful” anyway. Not what i expected though. It’s my first filter circuit but I expected a sharp drop off outside the frequency range permitted, instead it’s very gradual is that correct?

 

[Nigel Goodwin]

Wow i was being especially dense there. That’s exactly why its clipping... I’ll lower input to 1mV, was just the default for some reason.

Thanks.

EDIT: It’s the bode plot that is more “useful” anyway. Not what i expected though. It’s my first filter circuit but I expected a sharp drop off outside the frequency range permitted, instead it’s very gradual is that correct?

Yes, it's not a 'filter' really, or at least not a very good one.

Filters are rated in dB's per octave, and these are pretty well as low as they come.

You can easily improve it somewhat by adding filtering to every stage, instead of only one for each.

 

[audioguru]

The bottom part of the output is clipping because the 50k value of R5 does not match R6 parallel with R7 (almost 1k ohms).
The max input bias current of a lousy old 741 opamp is a whopping 0.5uA which produces 25mVin R5 which the opamp amplifies 126 times to produce an output offset voltage of 3.15V.
The output offset voltages of the first two opamps cancel because the second opamp is inverting and is the same circuit as the first opamp.

 

[Titan]

The bottom part of the output is clipping because the 50k value of R5 does not match R6 parallel with R7 (almost 1k ohms).
The max input bias current of a lousy old 741 opamp is a whopping 0.5uA which produces 25mVin R5 which the opamp amplifies 126 times to produce an output offset voltage of 3.15V.
The output offset voltages of the first two opamps cancel because the second opamp is inverting and is the same circuit as the first opamp.

I've scrapped the LM741's and got a couple of TL082CP's on order (I know its 2 op amps in 1 chip but these things are easy to lose!). I will explain why i moved away from 741s in my report so it's all worth it. I have simulated the below, it looks ok to me. Thoughts? Again this is my first experience with filters so i'm not sure how it should behave.

 

[ronsimpson]

so i'm not sure how it should behave.

The first amp has high gain while the second is only 14.
The first amp must have a GBW of 3mhz.
I would reduce the gain in the first amp.

 

[audioguru]

Your filters are very simple so they will not do much. At 20Hz and 20kHz the output level will be down a little. At 10Hz and 40kHz the output level will be down only a little more. Then at 5Hz and 80kHz the output level will be down a little more. Then at 2.5Hz and 160kHz the output level will be …..

 

[Titan]

The first amp has high gain while the second is only 14.
The first amp must have a GBW of 3mhz.
I would reduce the gain in the first amp.

The TL082CP has a GBP of 4MHz so I thought being under 3 would be fine? Also my lecturer said to put as much gain as possible in the first stage so thats why I did that. I will reduce the first stage gain later and see if it makes any difference.

Your filters are very simple so they will not do much. At 20Hz and 20kHz the output level will be down a little. At 10Hz and 40kHz the output level will be down only a little more. Then at 5Hz and 80kHz the output level will be down a little more. Then at 2.5Hz and 160kHz the output level will be …..

I know its a large bandwidth but how would I make the filters more effective? The examples I find in textbooks for example are very simple, usually only 1 op amp with a gain of say 10 and small bandwidth.

 

[Nigel Goodwin]

The TL082CP has a GBP of 4MHz so I thought being under 3 would be fine? Also my lecturer said to put as much gain as possible in the first stage so thats why I did that. I will reduce the first stage gain later and see if it makes any difference.



I know its a large bandwidth but how would I make the filters more effective? The examples I find in textbooks for example are very simple, usually only 1 op amp with a gain of say 10 and small bandwidth.

Refer to post #13, adding simple filtering (low and high pass) to every stage will improve things somewhat with little effort.

But there's no real need for anything else, unless the specifications require it.

It would also add the bonus of removing the DC coupling between stages.

 

[Titan]

Refer to post #13, adding simple filtering (low and high pass) to every stage will improve things somewhat with little effort.

But there's no real need for anything else, unless the specifications require it.

It would also add the bonus of removing the DC coupling between stages.

Thanks. the final thing I need to do is have it work single rail. My textbook says I can accomplish this by connecting pin4 to ground and having my single rail supply go through a resistor equal to R1 (as well as being connected to pin 8 still of course). I have done this and it no longer works, despite looking like the figure.

Is the TL082 not able to operate in this way?

Not at my computer with spice software at the moment, but it says to do it like this:

https://imgur.com/E4J7l