Hi everyone,
I designed this version of the instrumentation amplifier as a line input, but these values should give a gain of 60 (around 35dB) cause I initially wanted to use it as a microphone preamp input section. If I want to use it as a regular line input, should I give it unity gain? 2? 4? Also I added two bypass caps (C5 and C6) in between each rail and the 0V (power ground). Do I need another set of these for each IC or is this enough? Do I need a 47uF electrolytic decoupling cap from rail to rail like you might do in a single supply circuit? Will I need to hook up the power (0V) and signal (GND) grounds together? (The PSU will be on another board)
I designed this version of the instrumentation amplifier as a line input, but these values should give a gain of 60 (around 35dB) cause I initially wanted to use it as a microphone preamp input section. If I want to use it as a regular line input, should I give it unity gain? 2? 4? Also I added two bypass caps (C5 and C6) in between each rail and the 0V (power ground). Do I need another set of these for each IC or is this enough? Do I need a 47uF electrolytic decoupling cap from rail to rail like you might do in a single supply circuit? Will I need to hook up the power (0V) and signal (GND) grounds together? (The PSU will be on another board)
Last edited:
I'm sure you do not want IC1AB driving 37 ohms per side.
As a mike preamp, R12 R13 100r really degrade your noise figure. Although maybe not with '5532 which is OSI more like 6K?
'5532's input bias current in 100k resistors will give a good tenth-volt of DC offset. The symmetry tends to cancel but not 100% even for peas in the same pod. I would expect to want some DC blocking before going to a pot or next box in the chain.
I figure gain is 45 (33dB). I'm not sure if that is hot mike or lame line.
As for "line level".... it depends. Sony miniDisc? Fairchild 660? Into what? Consumer cassette? Broadcast transmitter? You usually want a trim.
This problem has been approached thousands of times before and hundreds of those attempts can be found on the internet.Plagiarize "study" all you can.
As a mike preamp, R12 R13 100r really degrade your noise figure. Although maybe not with '5532 which is OSI more like 6K?
'5532's input bias current in 100k resistors will give a good tenth-volt of DC offset. The symmetry tends to cancel but not 100% even for peas in the same pod. I would expect to want some DC blocking before going to a pot or next box in the chain.
I figure gain is 45 (33dB). I'm not sure if that is hot mike or lame line.
As for "line level".... it depends. Sony miniDisc? Fairchild 660? Into what? Consumer cassette? Broadcast transmitter? You usually want a trim.
This problem has been approached thousands of times before and hundreds of those attempts can be found on the internet.
Few will need gain over x4 these days, except perhaps for a lower gain phono stage source.
Definitely scale up all of the feedback resistors. Where is your volume control?
The power supply ground and the signal ground have to connect together at some point.
On this board, add 10uF to 100uF at each rail to ground near the power input connector.
Use around 0.1uF at each IC package power pin to ground.
Definitely scale up all of the feedback resistors. Where is your volume control?
The power supply ground and the signal ground have to connect together at some point.
On this board, add 10uF to 100uF at each rail to ground near the power input connector.
Use around 0.1uF at each IC package power pin to ground.
Last edited:
What type of microphones are to be supported? Some require phantom power...
Also D5,D6 may introduce some distortion depending on signal levels at that point.
Also D5,D6 may introduce some distortion depending on signal levels at that point.
The are across the opamp's input, which should be "zero", until the opamp is in gross overload. I have no idea why they are there but for once they do no harm.
Agreed they do little at DC. Maybe as loop gain falls off with frequency they do more to approximate a soft fuzz pedal effect. I measured the frequency spectrum of a cymbal picked up with a condenser mic. Short term harmonics extended up well above 35kHz, don't remember exactly how far but still have the 24/192 wav file if we want to check (my mic capsule was not equipped with the wide bandwidth preamp body or the harmonics likely would have gone even higher). IMD could perhaps find its way down into the humanly audible audio band. I know most people wouldn't worry about such things, but if the diodes are not needed maybe they shouldn't be there.
Last edited:
The external diodes do not conduct until the '5532 (any functional opamp) is totally freaked-out.
The '5532 already has diodes (diode-connected transistors) in that position.
The '5532 already has diodes (diode-connected transistors) in that position.
Yes, in principle they are there to protect against ¨gross transients, such as plugging a preamp there with a floating capacitor having 35V DC at the other end or a brain dead (drunk/stoned) Guitar Player plugging his 100Watt Marshall tube amp speaker out into a Mic Preamp input, etc.
Or plain old lightning falling nearby.
As a minor side note: my eyes hurt trying to read those fuzzy low contrast schematics, is it Eagle or KiCad?
Ugh! .... what´s wrong with black on white drawings?
Such as:
To boot, Forum Software fuzzies it even more trying to make it fit into reply window.
Or plain old lightning falling nearby.
As a minor side note: my eyes hurt trying to read those fuzzy low contrast schematics, is it Eagle or KiCad?
Ugh! .... what´s wrong with black on white drawings?
Such as:
To boot, Forum Software fuzzies it even more trying to make it fit into reply window.
Last edited:
Ok, thanks for the feedback, yes, I did an internship at a guitar pedal and amp repair company so there may be some unnecessary things such as the D5 and D6 diodes. So to recap, this design was initially meant to be a mic preamp input stage, but I want to use it as a pro line in, such as the input of a compressor, an EQ, or any other line level input that you'd typically use in a studio. I think I'll keep the 100k input impedance in case the devices I build ever need to interface with valve equipment (though I doubt it). This is basically a more complete instrumentation amplifier such as the one used by Douglas Self in Small Signal Audio Design (p.516). so for balance R6/R9 must be equal to R7/R8, so if R5=R4 then Gain = (1+(2*R5)/R3)*(R9/R6). I guess I could use a dual gang pot instead of R8 and R9 to keep the symmetry? And replace the R3, R4, R5 resistors with 1K each?
The 0.1uF (100nF) caps near going from each voltage pin of the IC to ground are the bypass caps, right? But these 10uF to 100uF caps you're saying to put from the each power rail to ground, are they electrolytic caps to decouple the power rail to ground or am I missing something?
Also I just realized maybe I could use IC2B as a buffer in the signal path to the main circuit board, or would a buffer make more sense at the beginning of the main PCB?