# Operational amplifier design problem

Discussion in 'The Projects Forum' started by rjnachshin, Apr 18, 2016.

1. ### rjnachshin Thread Starter New Member

Nov 10, 2015
12
0
Hi I'm trying to design a op amp circuit that meets these following requirements. The current design I have isn't meeting the requirements and I was wondering if you guys have any advice to give.

Zin: 47kΩ minimum Zout: 4.7kΩ maximum Voltage gain (with RL = 4.7kΩ at output): at least |AV| = 1000, but no higher than 2000. Gain shall remain constant within 3dB over the full frequency range of 20 Hz to 20 kHz. Note that AV is defined as being: = Output voltage level: capable of delivering 10 V peak-peak output to a 4.7kΩ load with low distortion over the frequency range of 20 Hz to 20 kHz. Power supply voltage available is +15 Volts DC. Note that only positive polarity is available, so amplifier must be designed for single-supply operation

First I have to get a gain of a 1000 using a 741 op amp but my design distorts the wave forms when I use 10m volts. The design produces a non distorted sine wave at lowers voltage but the gain is only around 900. Any advice would be very helpful.

I'm attaching my design.

2. ### dannyf Well-Known Member

Sep 13, 2015
2,196
417
Because the entire gain is in the 2nd stage.

Your questions can be easily answered by looking pretty much any opamp book or data sheet.

As to input impedance, 47k is typically the upper end for a real life circuit. Anything about that requires careful shielding.

Dec 4, 2009
254
33
Without going into details, 741 is not a single supply opamp! Consider gain bandwidth product.

4. ### crutschow Expert

Mar 14, 2008
13,509
3,385
With 47KΩ desired input impedance I presume this is a vinyl record magnetic cartridge amplifier(?).
For a 47kΩ input impedance you need to change R8 to 47k.

A 741 is a noisy old op amp with a low gain-bandwidth and slew-rate, and totally unsuited for this task.
For that you want to use a low-noise, low-distortion op amp designed for audio such as the TLOxx.

Divide up the gain between the two stages (giving each a gain of about 31.6.)

Increase the value of R4 and R6 by about 10 times each. They provide too low a load resistance for the 1st stage op amp output. (The minimum equivalent resistance should be ≥2kΩ.)

Decouple the op amp directly from its power pin to ground with a 0.1μF ceramic cap.

5. ### AnalogKid Distinguished Member

Aug 1, 2013
4,705
1,300
There are several things wrong with your schematic design that are independent of the specific opamp component used.

The required gain-bandwidth is 1000 x 20,000 Hz, or 20 million. That is an impossible number for a single opamp, and difficult for compound devices like an instrumentation amp. So you need two amplification stages in series as you have, but both have to contribute significant gain. For example, if each stage has a gain of 31.6, the resulting circuit gain would be 1000 but the gbw for each stage would be 633 kHz. This still is a large number, and too large for a 741, but not impossible for other parts.

Your input impedance at high frequencies is 1.6K, not 47K. At 20 Hz the input impedance is 2.4K, but that is because C4 is acting as an input attenuator at that frequency.

C1 and R5 form a 32 Hz highpass filter.

The 741 output voltage swing range barely meets your requirement, will take precise DC level adjustment to prevent asymmetrical clipping, and the adjustment will drift with temperature changes. Also, the DC level set by R4 and R6 is not in the center of the output voltage range.

Separate from al of that, the 741 has significant input stage harmonic distortion and even worse output stage crossover distortion.

ak

6. ### rjnachshin Thread Starter New Member

Nov 10, 2015
12
0
Thank you! I was warned I have to use another op amp to complete the design so i will start looking into TLO's

7. ### dannyf Well-Known Member

Sep 13, 2015
2,196
417
That would be the wrong warning or a wrongly received warning, as you certainly don't need another opamp to get the job done.

The more fundamental issue here is you haven't figured out the issues with your existing design.

8. ### DGElder Member

Apr 3, 2016
347
87
You realize your input signal is 10mVpk, not Vpp? A gain of 1000 (20Vpp) would exceed the amp's compliance with a 15V supply.

9. ### AnalogKid Distinguished Member

Aug 1, 2013
4,705
1,300
rjnachshin said: ed I have to use another op amp to c
Pretty sure he does. All of the (significant) voltage range issues aside, the 741 simply doesn't have the necessary gbw in a compound design, let alone in a single gain stage.

ak

10. ### crutschow Expert

Mar 14, 2008
13,509
3,385
You most certainly do if this is an audio preamp.
A 741 is very noisy and has insufficient frequency response and slew-rate for his application.

11. ### dannyf Well-Known Member

Sep 13, 2015
2,196
417
If you want a 1000x combined gain, you need each stage to contribute ~30x.

So take a look at the feedback arrangement for U2 vs. U1 and you will see that with the addition of 1 resistor + 1 capacitor you can give U1 the same gain so U2 doesn't have to take on all the gain responsibility.

You also can take out quite a few unnecessary parts and your circuit will then do what you want it to do.