The old LM301 used external compensation (the cap) while the TL071 and many many are internally compensated and unity gain stable. So take the cap out.
I always implement option 3, which I think is more in line with the extensive answers of Marcel, thanks, option 3: R6 is 100K, R4 is 22K C4 is 150nF and C2 is 33uF.
Correction/addition to my earlier post:
With a FET op-amp and when you don't want to keep it original, you can increase the impedance of the DC feedback. It reduces the noise gain/increases the feedback around the op-amp, thereby reducing its noise contribution and its distortion, reduces the noise current R4 injects into the virtual ground, and it allows you to use a film capacitor for C2 instead of a tantalum electrolytic. Putting a small capacitor across R6 helps to keep things stable and to suppress ultrasonics, assuming a unity-gain-stable FET op-amp such as the TL071.
R6: 100 kohm in parallel with 10 pF
C4: 150 nF
R4: 47 kohm
R5: 39 kohm
C2: 2.2 uF
C1 and R3 as is (680 nF and 22 kohm)
I think this is the lowest noise option so far.
With a FET op-amp and when you don't want to keep it original, you can increase the impedance of the DC feedback. It reduces the noise gain/increases the feedback around the op-amp, thereby reducing its noise contribution and its distortion, reduces the noise current R4 injects into the virtual ground, and it allows you to use a film capacitor for C2 instead of a tantalum electrolytic. Putting a small capacitor across R6 helps to keep things stable and to suppress ultrasonics, assuming a unity-gain-stable FET op-amp such as the TL071.
R6: 100 kohm in parallel with 10 pF
C4: 150 nF
R4: 47 kohm
R5: 39 kohm
C2: 2.2 uF
C1 and R3 as is (680 nF and 22 kohm)
I think this is the lowest noise option so far.
Marcel, is the use of a Fet input opamp in a shunt feedback configuration not always the better option? The increased current noise of a BJT input Opamp will create extra noise, especially in this case, with a relatively high input resistance. Joost Plugge
That depends on the impedance levels and the noise specifications of the op-amps. If you want to make an inverting amplifier with a reasonably high input impedance, you will end up with a fairly high impedance level at the negative op-amp input, and a FET op-amp will then usually give less noise than a bipolar op-amp.
Hello Marcel, another quistion pops up. In your post with the drawings you stated: The noise current inoise just flows completely into the virtual ground, no matter what. But is this true? I thought the virtual earth is formed by a very high resistance, the input impedance of the op-amp, but with a very very low voltage across it, hence it looks like a virtual earth. If the current noise always flows in the virtual earth I can't explain the differance in noise performance between BJT and Fet input op-amps, or am I missing something? Or is only the produced voltage noise of importance in a shunt feedback op-amp circuit. Joost Plugge.
I'm not sure if I understand the question. In an inverting amplifier with an op-amp, when the current flows into the virtual ground, physically that means that the feedback loop keeps the voltage at the virtual ground close to zero while the current flows into the shunt feedback resistor. Hence, it causes the output voltage to change.
I used an ideal current-controlled voltage source as an abstract model for everything to the right of R3 to keep my calculations simple. It's not a physical model, but just an abstraction: something with a low input impedance, low output impedance and a transfer from input current to output voltage.
I used an ideal current-controlled voltage source as an abstract model for everything to the right of R3 to keep my calculations simple. It's not a physical model, but just an abstraction: something with a low input impedance, low output impedance and a transfer from input current to output voltage.
So in real life most of the the input noise current is flowing through the input resistor in a shunt feedback configuration, so in such a configuration, like in the 405, it is wise to keep an eye on the produced noise current. My guess: because of the high input resistor of 22K, choose a Fet input op-amp with a low voltage noise.
No, physically, the op-amp's noise current flows through the feedback resistor. The input resistor only comes into play when you calculate how much noise voltage you would have to apply at the input to get the same amount of noise at the output as the op-amp's noise current causes.
I see, the other end of the input resistor is also at a very low voltage, so the current will flow trough Rf, thanks! A little off topic! But I think we answered the original question 🙂.
Getting back on topic, namely regarding changing the LM301A to a TL071:
When you compare the QUAD 405-2 schematics to the QUAD 405 schematics, you see several changes made around the op-amp:
A. The QUAD 405-2 has a TL071 rather than an LM301A
B. The 3.3 pF compensation capacitor is removed, as Mooly already mentioned
C. It has a 4.7 nF capacitor across the Zener diode that stabilizes the -15 V (that is, from pin 4 of the TL071 to ground)
D. R9 is changed into a short circuit
E. R10 is reduced from 1 kohm to 560 ohm
F. The optional clipper that limits the output power is implemented differently, namely with Zener diodes
As far as I know, change C is related to stability. I guess it's stability of the op-amp during power-up and power-down.
D and E must be related to the optional clipper, but changing R9 into a short could also be needed for op-amp stability.
Assuming you want to change the op-amp to TL071 and don't want the clipper, I would remove the 3.3 pF compensation capacitor, add the 4.7 nF capacitor across the Zener for the negative supply, change R9 into a short circuit and change R10 into either 560 ohm or (preferably) a short circuit.
When you compare the QUAD 405-2 schematics to the QUAD 405 schematics, you see several changes made around the op-amp:
A. The QUAD 405-2 has a TL071 rather than an LM301A
B. The 3.3 pF compensation capacitor is removed, as Mooly already mentioned
C. It has a 4.7 nF capacitor across the Zener diode that stabilizes the -15 V (that is, from pin 4 of the TL071 to ground)
D. R9 is changed into a short circuit
E. R10 is reduced from 1 kohm to 560 ohm
F. The optional clipper that limits the output power is implemented differently, namely with Zener diodes
As far as I know, change C is related to stability. I guess it's stability of the op-amp during power-up and power-down.
D and E must be related to the optional clipper, but changing R9 into a short could also be needed for op-amp stability.
Assuming you want to change the op-amp to TL071 and don't want the clipper, I would remove the 3.3 pF compensation capacitor, add the 4.7 nF capacitor across the Zener for the negative supply, change R9 into a short circuit and change R10 into either 560 ohm or (preferably) a short circuit.
The original question on page 1was about changing the input sensitivity, we end up with noise analyses and me learning a lot!
My best advice would be the following, I did this many ttimes during repairs/overhauling/modifying. R6 is 100K, R4 is 22K C4 is 150nF and C2 is 33uF. This gives a reduction of 3 times, so a sensitivity of 1.5V. To keep close to Quad designs, replace the old fashioned LM301 with a still period TL071 and remove the compensation cap C3. Although the voltage across C2 is very low, connect the plus of a polarised Cap to ground (mass), or use a non polarised capacitor.
If you want a 1V sensitivity: R6 is 150K, R4 is 22K, C4 is 100nF and C2 is 47uF.
If you want a 1V sensitivity: R6 is 150K, R4 is 22K, C4 is 100nF and C2 is 47uF.
It's a good option, but why do you prefer it over a lower-noise option with 15 times smaller C2?
I only tried/tested your suggestion in a simulator, not in a real amplfier, so that is why I adviced to keep close to the Quad design parameters.
Fair enough. The only combination I ever tried in real life was
C1 = 680 nF
R3 = 22.1 kohm
R6 = 221 kohm (probably shunted with a very small capacitor)
C4 = 68 nF
R4 = 68 kohm
R5 = 56 kohm
C2 = 2.2 uF
using a TL071 and modifying the decoupling, compensation and R9 and R10 like in a QUAD 405-2. This specific set of values is unusable for mmmalmberg, because it only changes the sensitivity to 0.75 V.
I've also used the trick with the extra R and C in the connecting cable. The driving circuit was a semiprofessional mixing desk that used a tree-pin XLR connector for a stereo unbalanced output - completely contrary to all standards, but at least it gave me plenty of space for the capacitors.
C1 = 680 nF
R3 = 22.1 kohm
R6 = 221 kohm (probably shunted with a very small capacitor)
C4 = 68 nF
R4 = 68 kohm
R5 = 56 kohm
C2 = 2.2 uF
using a TL071 and modifying the decoupling, compensation and R9 and R10 like in a QUAD 405-2. This specific set of values is unusable for mmmalmberg, because it only changes the sensitivity to 0.75 V.
I've also used the trick with the extra R and C in the connecting cable. The driving circuit was a semiprofessional mixing desk that used a tree-pin XLR connector for a stereo unbalanced output - completely contrary to all standards, but at least it gave me plenty of space for the capacitors.
OK So we're all in agreement here? yuk yuk... But seems close. So this "connect the plus of a polarised Cap to ground (mass), or use a non polarised capacitor" is in reference to C2, yes? I truly appreciate all the contributions here. Most of it above my pay grade but all toward the same end - thanks🙂
Yes C2, sorry about the deviations in noise behaviour, but sometimes it works that way.