TL072C upgrade for audio interface

Jump to Latest
#21
Had a bit of a poke around, many pickups have about 7k resistance and tons of inductance. Both resistance and inductance will turn current noise into voltage noise, the high inductance is the main reason to avoid bipolar opamps, but with a FET input the current noise becomes negligible, leaving the 7k Johnson noise from the pickup and the voltage noise of the opamp to worry about.

7k gives about 11nV/√Hz Johnson noise, so any FET opamp with voltage noise substantially less than that is a good choice, 5nV/√Hz would be a good figure to aim for. That then leaves refrigeration as the only way to reduce noise further!

OPA1677/1678 is a plausible choice (4.5nV) and cheap, and 2mA consumption per amp. TLE2071/2072 are a lot more expensive, noisier, and have phase-reversal...

[ overall preamp noise is comprized of:
1: source resistance Johnson noise,
2: amp input voltage noise,
3: amp input current noise * input impedance
these sum as √(sum of squares) ]
 
  • Like
Reactions: JMFahey
#22
Skezza said:
Are there any other components I could replace or am I likely just grasping at straws?

If it's a pointless operation then I'll more than happily post afterwards to say so, save someone else doing the same work.
The OPA1642 was the chip I had in mind though!
Click to expand...
The point is, me thinks, that too much soldering/de-soldering increase the risk of butchering up the PCB for no benefit or even worse performance.
IMHO, to check properly, you'd need a second identical device in its stock state, anyway.

Replacing of the TL072 with OPA1642 may or may not give any real noise improvement in the application but is maybe worth a try. However, while unlikely, it could even be the circuit clearly degrades in signal headroom, see @abraxalito's comment in post #14. To find out, I'd certainly measure and note DC voltages at all opamp pins to see how it is biased.
 
#23
Mark Tillotson said:
Had a bit of a poke around, many pickups have about 7k resistance and tons of inductance. Both resistance and inductance will turn current noise into voltage noise, the high inductance is the main reason to avoid bipolar opamps, but with a FET input the current noise becomes negligible, leaving the 7k Johnson noise from the pickup and the voltage noise of the opamp to worry about.

7k gives about 11nV/√Hz Johnson noise, so any FET opamp with voltage noise substantially less than that is a good choice, 5nV/√Hz would be a good figure to aim for. That then leaves refrigeration as the only way to reduce noise further!

OPA1677/1678 is a plausible choice (4.5nV) and cheap, and 2mA consumption per amp. TLE2071/2072 are a lot more expensive, noisier, and have phase-reversal...

[ overall preamp noise is comprized of:
1: source resistance Johnson noise,
2: amp input voltage noise,
3: amp input current noise * input impedance
these sum as √(sum of squares) ]
Click to expand...

I've been told there is usually a 250 kohm or 500 kohm volume potmeter inside the guitar. Even at maximum volume, it will add noise by injecting thermal noise current into the guitar pick-up. I have no idea how much extra noise pick-up iron losses will give, as I never saw any information about those.

Still, 5 nV/√Hz will be more than good enough, assuming the noise with the volume pot turned all the way anticlockwise is of little practical importance.
 
#27
JMFahey said:
In a nutshell: improving "components" will do nothing.
If you feel up to it, improve design.
Click to expand...
Ok so how would I start? Obviously if we're talking whole new PCB that may be above my skills but general PCB mods including wiring in extra components are food and drink so if you've got any suggestions, I'll more than happily get stuck in and post the results here
🙂
 
#29
KSTR said:
The point is, me thinks, that too much soldering/de-soldering increase the risk of butchering up the PCB for no benefit or even worse performance.
IMHO, to check properly, you'd need a second identical device in its stock state, anyway.

Replacing of the TL072 with OPA1642 may or may not give any real noise improvement in the application but is maybe worth a try. However, while unlikely, it could even be the circuit clearly degrades in signal headroom, see @abraxalito's comment in post #14. To find out, I'd certainly measure and note DC voltages at all opamp pins to see how it is biased.
Click to expand...
I have three 🙂

Like I said, these cost pennies. I use them as USB Soundcards as well.
 
#30
Mark Tillotson said:
Had a bit of a poke around, many pickups have about 7k resistance and tons of inductance. Both resistance and inductance will turn current noise into voltage noise, the high inductance is the main reason to avoid bipolar opamps, but with a FET input the current noise becomes negligible, leaving the 7k Johnson noise from the pickup and the voltage noise of the opamp to worry about.

7k gives about 11nV/√Hz Johnson noise, so any FET opamp with voltage noise substantially less than that is a good choice, 5nV/√Hz would be a good figure to aim for. That then leaves refrigeration as the only way to reduce noise further!

OPA1677/1678 is a plausible choice (4.5nV) and cheap, and 2mA consumption per amp. TLE2071/2072 are a lot more expensive, noisier, and have phase-reversal...

[ overall preamp noise is comprized of:
1: source resistance Johnson noise,
2: amp input voltage noise,
3: amp input current noise * input impedance
these sum as √(sum of squares) ]
Click to expand...
What would make the OPA1677 or 78 a better choice over OPA1642?
 
#31
KSTR said:
The point is, me thinks, that too much soldering/de-soldering increase the risk of butchering up the PCB for no benefit or even worse performance.
IMHO, to check properly, you'd need a second identical device in its stock state, anyway.

Replacing of the TL072 with OPA1642 may or may not give any real noise improvement in the application but is maybe worth a try. However, while unlikely, it could even be the circuit clearly degrades in signal headroom, see @abraxalito's comment in post #14. To find out, I'd certainly measure and note DC voltages at all opamp pins to see how it is biased.
Click to expand...
Ill do that tomorrow 🙂
 
#33
Wel
Mark Tillotson said:
At full volume the pot resistance is parallel to the ~7k of the coils, so it won't contribute significant noise. I've no information about iron-loss noise alas.
Click to expand...
Well, the 7k DCR inside pickups is never "alone", always in series with a big inductor which becomes the main impedance component, by far.
So 250k (Fender) 500k (Gibson) volume pot is always in parallel with that.
Plus another 250k-500k pot in series with a typically .022uF cap to ground, which is the crude tone control.
 
#34
Skezza said:
Is there a particular iteration of the OPA1641 worth buying?
Click to expand...
No, buy any, it's all the same. I picked up a few pieces at Mouser (1641&1642). It is a cheap OPA, but it performed very well in DACc with PCM56 and AD1865 in I/V circuit. The sound is on the brighter side, much more detail is heard than with the TL071/072 (also bought from Mouser, and I also have Chinese copies). The AD1655/1656 turned out to have a slightly more neutral sound, and they are also not expensive. Each is sonically much better than the TL071/72. That is my extremely subjective observation. :rofl:
 
#35
I think ill order a few and then do some tests and upload the results. More than happy to demo my workings for the world 🙂 Just working out how I can do a faithful A/B test.

The amp chips don't seem too expensive. Around £2-£5 each. Is that about right?
 
#37
I've looked at:
OPA1642
OPA1677
OPA1678

Which appear to range from 2-5. I am looking at the SOP packages not DIP.

I am wondering whether I might be better installing a SOP to DIP adapter.... Would allow me to test many and way easier 🙂
 
You must log in or register to reply here.
Top Bottom