Can anyone say what the noise level of a TL072 is like when used as the input stage of a microphone amplifier that's meant to be "good" but not hi-fi or studio quality? I'm thinking of a general-purpose system that will normally be used in informal gatherings for speech and live music.
The old TL07x opamps are classed as low noise but the term is relative and the 18nV/√Hz figure is 20 times that of the best ultra low noise types. Calculating the noise in uV and dB is a simple process but what does it actually sound like? Terrible? Acceptable? .....
I'd test it myself by swapping opamps in an existing setup but I don't have anything handy at the moment. So I'll appreciate hearing what others have to say from experience. Reasoned estimates are also welcome but no wild guesses please.
The old TL07x opamps are classed as low noise but the term is relative and the 18nV/√Hz figure is 20 times that of the best ultra low noise types. Calculating the noise in uV and dB is a simple process but what does it actually sound like? Terrible? Acceptable? .....
I'd test it myself by swapping opamps in an existing setup but I don't have anything handy at the moment. So I'll appreciate hearing what others have to say from experience. Reasoned estimates are also welcome but no wild guesses please.
The TL072 is only low noise for high source resistances. 18nV/rtHz vs 5nV/rtHz makes the NE5532 much quieter for most applications
At the moment, the matter is really of academic interest, in case the information becomes useful later. There's no practical reason to use a TL072 instead of other lower-noise opamps like the NE5532, RC4558, etc. which are also dirt cheap. I've designed the PCB to accept just about any dual opamp will standard pin-outs.
Anyway, here's the basic schematic:
Anyway, here's the basic schematic:
Attachments
Depends on your mic and application.
Any powered electrete capsules delivers a noise level of its own that far exceeds TL072 noise. In that case even LM358 will do!🙂
A dynamic capsule is another cup of tea. With an impedance of 200~600Ohm these passive transducers require very low noise op-amps for best noise performance.
Any powered electrete capsules delivers a noise level of its own that far exceeds TL072 noise. In that case even LM358 will do!🙂
A dynamic capsule is another cup of tea. With an impedance of 200~600Ohm these passive transducers require very low noise op-amps for best noise performance.
CRATE sold a low-price band mixer with TL072 inputs. Used with typical dynamic mikes, set for PA gain, I thought the hiss in an empty room was irritating. The owner agreed but said it was not a problem what he was doing. Indeed when we filled the room with high school students, the hiss went un-noticed, even with the students "quiet". And when the band started nobody heard any hiss.
So it depends. But "gatherings" quieter than teenagers may find the hiss irritating.
So it depends. But "gatherings" quieter than teenagers may find the hiss irritating.
I've got 4558 and 4580 mixers. On $3 dynamic mikes 2 m away from source (hammond organ) I find the hiss irritating. Lots of gain at that distance. I was doing solo recording with a PC line input, audacity software. Don't know how 4558 are on a standard SM58 dynamic mike, I don't have one. I changed the 4558 mixer to 33078 and hiss was much better: after I eliminated the resultant oscillation with local Power Supply disk caps and 33 pf around the feedback resistor. 4558 is slow slew rate, didn't need the caps.
The 4580 mixer (with phantompower) is hissless on a phantom powered condensor mike, KSM27. That mike has ~1 v outputs.
You can hear the recording on inbojat.tumblr.com the yellow bird track. The hissy channel is the $3 dynamic mike, the silent one is my first KSM27 condensor mike. The $3 dynamic mike was slightly better than the dynamic mikes that came with the Sony TC250 tape recorder, but not much. That is a 70 W hammond H182 I'm playing, at low volume, maybe 3-5 watts. In my music room.
The 4580 mixer (with phantompower) is hissless on a phantom powered condensor mike, KSM27. That mike has ~1 v outputs.
You can hear the recording on inbojat.tumblr.com the yellow bird track. The hissy channel is the $3 dynamic mike, the silent one is my first KSM27 condensor mike. The $3 dynamic mike was slightly better than the dynamic mikes that came with the Sony TC250 tape recorder, but not much. That is a 70 W hammond H182 I'm playing, at low volume, maybe 3-5 watts. In my music room.
Last edited:
Thanks a lot, PRR and indianjo. Those are exactly the kind of information I hoped for. And what you said about the hiss level is about what I expected.
indianajo's experience with 4558s can be extrapolated to the TL072, With the application I have in mind, the distance between the mike and the sound source will certainly be much less than 2m.
It reminds me of the time I repaired a mixer for an amateur recording studio some 25 years or so ago when conditions here were even more primitive than they are now. The 4558 input opamp on one channel had gone bad but I didn't have any with me. I replaced it with a TL072 and warned the owner that they'd probably notice more hiss than before.
The owner was effusively grateful and reported later that the hiss had actually gone down. This was totally unexpected and I'm not basing my current judgment on that one incident.
indianajo's experience with 4558s can be extrapolated to the TL072, With the application I have in mind, the distance between the mike and the sound source will certainly be much less than 2m.
It reminds me of the time I repaired a mixer for an amateur recording studio some 25 years or so ago when conditions here were even more primitive than they are now. The 4558 input opamp on one channel had gone bad but I didn't have any with me. I replaced it with a TL072 and warned the owner that they'd probably notice more hiss than before.
The owner was effusively grateful and reported later that the hiss had actually gone down. This was totally unexpected and I'm not basing my current judgment on that one incident.
Speaking of the 4580, various datasheets all give the equivalent input noise voltage as 0.8µV under the same conditions. This is with RIAA equalization and will be worse with a flat response.
However, when I compare the flat noise density between datasheets from different manufacturers, there's a considerable difference. The density curve in the datasheet by JRC shows it as 3.3nV/√Hz at 1kHz whereas the TI datasheet gives about 6nV/√Hz. Both are measured at the same power supply voltage, temperature, gain and source resistance. This is a very significant difference.
Thoughts?
Here's a composite image from two datasheets. (The supply voltage on the TI datasheet is given as ±1.5V but this must be a typo. All other data are given at ±15V)
However, when I compare the flat noise density between datasheets from different manufacturers, there's a considerable difference. The density curve in the datasheet by JRC shows it as 3.3nV/√Hz at 1kHz whereas the TI datasheet gives about 6nV/√Hz. Both are measured at the same power supply voltage, temperature, gain and source resistance. This is a very significant difference.
Thoughts?
Here's a composite image from two datasheets. (The supply voltage on the TI datasheet is given as ±1.5V but this must be a typo. All other data are given at ±15V)
Attachments
Peavey used JRC 4580 in the unity12 mixer I used in the recording in post 9. It strikes me as lower in hiss than the 4558 at high gain. Plus it has high drive current, which is useful for a mixer output where the cable might be longer than 2 m.
I swapped a 33078 for a JRC 4580 on the input card peavey cs800s, I thought it was equivalent in hiss so I put the 4580 back.
I didn't try 4580 for the RA88a disco mixer at 50x gain with the really hissy 4558's , I didn't have it then. I didn't use 5532 because newark(farnell usa) didn't have them at the time I was experimenting, and 2068 was twice the cost of the 33078 at $.38 each. The RA88a disco mixer is now quiet enough, equivalent in hiss to my "legendary" dynaco PAS2 preamp. That is on the 50x RIAA input. At 100 less watts burned.
I swapped a 33078 for a JRC 4580 on the input card peavey cs800s, I thought it was equivalent in hiss so I put the 4580 back.
I didn't try 4580 for the RA88a disco mixer at 50x gain with the really hissy 4558's , I didn't have it then. I didn't use 5532 because newark(farnell usa) didn't have them at the time I was experimenting, and 2068 was twice the cost of the 33078 at $.38 each. The RA88a disco mixer is now quiet enough, equivalent in hiss to my "legendary" dynaco PAS2 preamp. That is on the 50x RIAA input. At 100 less watts burned.
Last edited:
TI makes billions of logic chips, some analog on the side.
JRC, for years, was mostly about AUDIO chips.
JRC may have a cleaner process. Or TI may get the low number 4 days a week but has bad days and wrote the sheet so they would not be rejects.
If you have a big project, you will get large numbers of chips from multiple production runs and makers, and test test test. For a one-off, get a few extra and see if there is a "Wow" difference.
However proper topology and design will generally work OK with non-dirty chips.
For a microphone pre-amplifier, methinks the OPA1612 cannot be beat---the combination of low noise, low THD, high common-mode/power supply rejection, fast slew rate, and rail-to-rail operation is simply the BEST. YES, they are a bit expensive (~$5-6 each) but compared to the cost of the rest of a system, it's well worth it.
Yes, of course. The OPA1612 is high on my list of opamps for low-level audio. But the thread is about technical curiosity, call it an intellectual exercise if you will, regarding the TL072.
On a side note, when noise is less important than a balanced combination of other factors like low cost, low bias current, low power consumption, etc. , this old device is still a good candidate.
On a side note, when noise is less important than a balanced combination of other factors like low cost, low bias current, low power consumption, etc. , this old device is still a good candidate.
Well, I still think the TL072 is a poor choice for a microphone preamplifier. As previously mentioned, an NJM2068 (at 60¢ each) would perform better in that application.
The op-amp's noise translates into about 35 dB(A) of equivalent acoustic noise with a 1.8 mV/Pa dynamic microphone.
- Status
- Not open for further replies.