Capacitor mic pre amp hiss

[timtimtim]

My AT4022 worked well with my home built preamp using one 5534 with balanced inputs, but the 5534 hiss was too high, so I bought an AD8099 which at .95nV/root Hz is one of the quietest op amps available, much quieter than 5534 at 3.5nV, but when I tested it the AD8099 isn't much quieter than the 5534, the hiss is still well above the mic hiss (which is inaudible over the AD8099 hiss). This is most disappointing and confusing, the AD8099 should be 11.3dB quieter than the 5534, why isn't it? So several questions:

Is there something about my circuit which makes it give more hiss than normal? I can't show you my circuit cos I havn't got a scanner, but one mic O/P goes into the op amp -ve I/P via a C and a 5k6 with a 56k from -ve I/P to OUT, and the other mic O/P goes to +ve I/P via a C and a 5k6 and a 56k to ground to make it equal level to the -ve I/P. Higher values of resistor gave slightly more hiss.

Looking on the net, many circuits use a separate op amp for each mic O/P, combined later. Would this give lower hiss? If so why?

Another question: would it be possible to use a 4:1 or higher ratio transformer at the I/P giving a 12dB improvement in signal to pre amp hiss? Any advice on the best circuit to use for minimum hiss would be much appreciated.

This low-noise business is a whole new world, I've seen some very wierd circuits on the net.

Another thought: why can't you use say 8 identical op amps in parallel, all driven off the same input, their outputs summed, thereby giving a 9dB reduction in hiss relative to the signal?

 

[mshilarious]

Thermal noise of 5k6 resistor: -117dBV. That is your noise floor. Use an instrumentation amp configuration instead, that has no series resistance on the inputs, and gain before the series resistance feeding the differential amp.

 

[timtimtim]

Thanks for that mshilarious. I've built an instrumentation circuit but I havn't been able to test it yet cos I'm getting hf instability which I havn't managed to fix yet (any tips?)
One question in the meantime: you say any series resistor generates hiss. Why then don't the feedback resistors, especialy the one which goes between the two -ve inputs, generate hiss? Is it because smaller value resistors give less hiss, and that resistor can be made low in this circuit?

 

[mshilarious]

They do generate more noise, but that noise is shunted by the smaller series resistance. Remove both series resistors such that the amp is only connected to ground and feedback and see how much worse the noise gets.

You can lower noise by decreasing series resistance, but that resistance sets input impedance so there is a limit to how low you can go. This is why an inamp works well, it separates input impedance from the differential amp and doesn't sum until after there is gain.

HF stability with an inamp is achieved by bypass caps in the feedback loop as well as filter caps on the input. The differential filter cap should be 10x the shunt caps to ground. 100pF and 1nF work well.

 

[timtimtim]

They do generate more noise, but that noise is shunted by the smaller series resistance. Remove both series resistors such that the amp is only connected to ground and feedback and see how much worse the noise gets.

You can lower noise by decreasing series resistance, but that resistance sets input impedance so there is a limit to how low you can go. This is why an inamp works well, it separates input impedance from the differential amp and doesn't sum until after there is gain.

HF stability with an inamp is achieved by bypass caps in the feedback loop as well as filter caps on the input. The differential filter cap should be 10x the shunt caps to ground. 100pF and 1nF work well.

I gave up with the AD8099 because it requires a very low resistor from its +ve input to gnd to work at all, and it takes too much current and doesn't give much output swing, but the inamp circuit works very well with 5534s and the hiss is far lower than before, probably low enough?

The mic hiss is now definitely above the amp hiss, but possibly the amp hiss still adds a little bit to total hiss, especialy at hf, so are there any op amps with lower noise than 5534 which will work at 7.5v supply, and which will behave like a normal op amp? I don't need op amps with an ft of 3GHz like the AD8099.

Is it possible to use a transformer to step up the input voltage a little? If so can you recommend one?

Thanks very much for your help

 

[mshilarious]

I gave up with the AD8099 because it requires a very low resistor from its +ve input to gnd to work at all, and it takes too much current and doesn't give much output swing, but the inamp circuit works very well with 5534s and the hiss is far lower than before, probably low enough?

5534s are certainly quiet enough for a capacitor mic, and usually good enough for dynamics. Whisper tracks with a ribbon and you'd need quieter.

The mic hiss is now definitely above the amp hiss, but possibly the amp hiss still adds a little bit to total hiss, especialy at hf

Yes it will. Mic noise is mostly 1/f (pink) whereas amp noise is white. So at some point they can crossover (depending on the mic's overall noise level) and the amp will become the dominant noise source.

so are there any op amps with lower noise than 5534 which will work at 7.5v supply, and which will behave like a normal op amp? I don't need op amps with an ft of 3GHz like the AD8099.

Probably you should just go straight to dedicated inamps, like INA217, THAT1512. Those are really quiet (at high gain) and will save you a lot of parts.

Is it possible to use a transformer to step up the input voltage a little? If so can you recommend one?

Yes, but you have to raise the input impedance of the amp otherwise your mics will see an excessive load. Once you go to an amp like INA217 it's pretty hard to get much quieter. The trafos are good at getting low noise performance out of noisier amps (~8nV/Hz^2).

It's really down to how much you want to spend on transformers. Neutrik NTE4 is cheap and small but you can't hit it too hard or it will saturate easily. On the other end is Jensen JT-13K6-C, in between are Cinemag and Lundahl offerings. A 1:4 mic input transformer is a rather common item . . .

The 1:4 will have an impedance ratio of 1:16 (the Jensen is 1:5 and 1:25), so the input impedance of the amp after the transformer has to be 16 times the desired load on the mic. See the schemo in the Jensen datasheet. The transformer also gives you minimum gain of 12dB (14dB for the Jensen).

 

[timtimtim]

Thanks a lot for all that.

Unfortunately I can't use either of those inamps because they don't go down to 7.5v supply. Also I don't want more than gain 10 so I think they won't be so quiet at that gain?

So I want a transformer. My question is what is the lowest impedance you can safely put on the mic outputs? Some of the transformers I have looked at say 200 ohms which I guess is far too low?

 

[mshilarious]

Thanks a lot for all that.

Unfortunately I can't use either of those inamps because they don't go down to 7.5v supply. Also I don't want more than gain 10 so I think they won't be so quiet at that gain?

So I want a transformer. My question is what is the lowest impedance you can safely put on the mic outputs? Some of the transformers I have looked at say 200 ohms which I guess is far too low?

A transformer doesn't have an impedance per se, although they are wound with an impedance in mind. Transformers reflect impedance across them according to the square of their turns ratios, such that the input impedance on the primary of a step-up transformer will be the load on its secondary divided by the square of the turns ratio.

So for example let's say you find a transformer that is 200:5K, that is a turns ratio of 1:5. If you have an amp with 25K input impedance and you put that transformer in front of it, the input impedance will be 1K.

Read this for all you'd want to know on transformers with a much more comprehensive explanation on impedance:

http://www.jensentransformers.com/an/Audio Transformers Chapter.pdf

 

[timtimtim]

That transformer information is fascinating, but the more I read the less I wanted to use a transformer. The R/C damping on the secondary for instance to control the hf response - I wouldn't know where to begin on that. Anyway I think my amp noise is probably inaudible below mic noise, so a transfo isn't necessary. A more elegant solution would be a very low noise inamp, but unless one exists which works at 7.5v, that isn't going to happen.

 

[Minion]

It sounds like it is either your curcuit or your layout/grounding scheme that is causeing the hum or possibly the PSU.....

Ive built several Mic preamps with NE5532 "s and they are dead quiet , no noise or hum that is detectable with your ears , even at 56db gain they are dead quiet .....

Do you have a schematic you can post ??

 

[timtimtim]

No hum, its the hiss I'm concerned about. I've done careful tests now, putting the amp output into a steep-cut high-pass-filter with fc 9kHz, to isolate the highest frequencies where the amp noise is most noticeable, and judging auraly I reckon the combined amp and mic hiss at hf is at least 6dB, possibly 10dB, louder than the amp hiss alone. This means the amp is adding between 0.5dB and 1.2dB to the amp hiss, which, if my guesstimation is accurate, is negligible.