Noise from preamp

[vsmGuy]

I just made a preamp for myself .. based on a LF351 I had lying around.

I did not AC couple the output from the amp however.

I get a constant noise at the background (like an Airconditioner heard from inside the room). We have central heating and I am sure the noise upstairs does not come down so much for the mic to pick up.

Is it because of me not ac coupling the output?

In that case .. would a 0.33uF polycarbonate be ok?

 

[audioguru]

The old LF351 opamp is not a low noise audio opamp. Of course it will produce hissss.
The output coupling capacitor has no effect at high hiss frequencies, its value affects the circuit's lowest bass frequency.

 

[vsmGuy]

I was thinking maybe the constant DC output from the preamp was contriibuting to the noise .. can it be the cause?

I seem to measure 2.5V with a cheap chinese coil meter even when I don't excite the electret. My Fluke measures ~1.7v but that keeps on changing. I am connecting this directly to the "mic in" of my laptop (I don't have a line level input)

If the output CD offset can never be a cause for the "hiss", should I replace the thing with a TL074? (The TL071 seems to be much more expensive - both are fronm TI - why?)

 

[Torben]

Schematic?

 

[bountyhunter]

The LF351/356 J-FET op amps are very good amps for LINE LEVEL audio applications where 0 dB signal levels are pretty high. For microphone or phono pre-amps, they are not good enough. The best amp used to be the Signetics 5532 type, don't know if they are still the best, but they are very good.

http://www.datasheetcatalog.org/datasheet/texasinstruments/ne5532.pdf

The other source of hum in low level amps is magnetic flux: is there a transformer within 12" of your amplifier circuitry?

 

[Hero999]

I don't see audioguru's problem with the LF351, it might not be as good as the TL071 but it's perfectly acceptable for most applications. If you're got plenty LF351s in your junk box then I don't think it's worth ordering new parts, especially if it's not part of a larger order.

 

[audioguru]

The complaint is that the LF351 produces noise in the preamp circuit. A better opamp made for low level audio will not produce noise.

 

[marcbarker]

Maybe the circuit is 'hooting' @ few 100 kHz? (due to bad layout maybe)

 

[Hero999]

I think it's more likely to be a ground loop problem or the lack of shielded audio cable than the op-amp.

 

[vsmGuy]

Aw, you guys made me draw a schematic. It was the usual stuff you know.

It's powered from a 9V PSU (7809).

Analog electronics is not my forte, so here is the schematic.

The input is twin wire unshielded, but the mic is 6'' away from the opamp input.
The output is a moulded stereo jack cable, the usual stuff.

Op amp is not decoupled (the input power supply is suitably bypassed) and the pins shown not connected in the schematic are not connected.

The schematic is truly the "true picture" of what I have... maybe I got some resistor values wrong from what it really is in my build, but I will have a closer look in the morning.

Should I upload a few audio samples?

 

[vsmGuy]

The LF351/356 J-FET op amps are very good amps for LINE LEVEL audio applications where 0 dB signal levels are pretty high. For microphone or phono pre-amps, they are not good enough. The best amp used to be the Signetics 5532 type, don't know if they are still the best, but they are very good.

http://www.datasheetcatalog.org/datasheet/texasinstruments/ne5532.pdf

The other source of hum in low level amps is magnetic flux: is there a transformer within 12" of your amplifier circuitry?

I had just bought a lotta 351's. Got a few TL071 with me but they are reserved for some "better" experiments

There are no Transformers within 12 inches, but I connect my mic to a laptop , although it has a SMPS but no transformer to be sure.

How do I find if the noise is injected from within the opamp, picked up be leads or is ambient noise just amplified by the preamp?

 

[marcbarker]

Disconnect one end of C1 and see if hiss stops.

 

[Torben]

I'd try removing the 9V power supply connection from between the input and the input cap. Seems to me that any ripple on the supply will also be amplified and, if it's full-wave rectified, that would make it 120Hz--which might sound a lot like an air conditioner.

Simple idea to test and prove or disprove one way or the other, at any rate.


Torben

 

[audioguru]

The opamp is inverting with an input impedance that is so low that it shorts the signal from the mic. The opamp should be non-inverting with a high input impedance.

The 1k resistor that powers the mic injects noise from the power supply that is amplified by the opamp. This resistor should be 10k and be decoupled with a 1k resistor in series from the supply and a 47uf capacitor to ground as a filter.

The mic is low level so should be connected to the preamp with shielded audio cable.

We are guessing that the noise is hiss or is hum. What is it?

 

[Torben]

The opamp is inverting with an input impedance that is so low that it shorts the signal from the mic. The opamp should be non-inverting with a high input impedance.

The 1k resistor that powers the mic injects noise from the power supply that is amplified by the opamp. This resistor should be 10k and be decoupled with a 1k resistor in series from the supply and a 47uf capacitor to ground as a filter.

The mic is low level so should be connected to the preamp with shielded audio cable.

We are guessing that the noise is hiss or is hum. What is it?

Oops--I had missed where the OP mentioned that it was a preamp for an electret mic. My bad. Audioguru's suggestion then makes much more sense than mine.


Torben

 

[Hero999]

What voltage is the transformer, the value of the filter capacitor and is the bridge rectifier half or full wave?

The LM7809 might be behaving erratically, if the input voltage isn't high enough.

 

[vsmGuy]

The problem is that the current preamp is soddered pretty tight on a veroboard. I will prototype a copy on a breadboard and upload some voice samples as well.

Audioguru, thanks for the insights; I for one believe its a hum - almost not hiss; but audio samples would prove my point better.

Should I upload wav's or MP3's of samples from the mic?

 

[audioguru]

Hum comes from the power supply or through the air to unshielded low-level wires.

 

[unclejed613]

gain of 100, and going into what is already a mic preamp in the computer........ going with a much higher value input resistor, or going with a noninverting amp would be a good start. knock the gain down to more like 10. also if you stay with the inverting amp, bypass those 47k resistors with a couple of 1uf or higher caps. you could be amplifying thermal noise from the 47k resistors, which will definitely add to the noise. you are probably also amplifying noise from the 7809.... is the 7809 properly bypassed with 10uf caps? i've seen 78xx regulators get noisy when the bypass caps dry out, because the regulator is primarily an "amplified zener", and the zener used as a voltage reference inside them creates a fair amount of noise, which is one of the reasons for adding the bypass caps at the input and output of the regulator. so you actually have several sources of noise here. the regulator and the voltage divider for the noninverting input you can do something about with a few electrolytic caps, but you can't do much about the input transistors of the op amp. an LF351 is (compared to a 741) a low noise device, but there are other devices out there that put the LF351 to shame. cutting the gain down to 10 rather than 100 will reduce the noise quite a bit. from what i see you're using an electret with a built-in jfet preamp (which is the only reason you would feed it with DC), so you really don't need a gain of 100 anyway. your PC input also has a mic preamp, so a gain of 100 stage before a mic preamp is going to be very noisy. go with a gain of 10 (replace the 1k resistor with a 10k), and if you want it to be variable gain (not neccesary, just adds a bit of flexibility to the circuit), replace the 100k resistor with a 100k pot. that way if you find a gain of 10 too much, you can turn it down from there.

 

[audioguru]

gain of 100

No.
The gain would be 100 if the microphone has a very low impedance. But the impedance of the mic is about 4.7k ohms which is in parallel with the 1k resistor that powers it. So 4.7k//1k= 767 ohms which is in series with the 1k input resistor of the opamp. So the gain is 100k/1767= 56.6.

bypass those 47k resistors with a couple of 1uf or higher caps.

No.
Only the bottom 47k resistor should be bypassed.

you're using an electret with a built-in jfet preamp (which is the only reason you would feed it with DC), so you really don't need a gain of 100 anyway.

No.
The FET is an impedance converter and has no gain. The output of an electret mic is very low and needs a preamp with a gain of 100 or more.