Noise testing in transistors

[Dr.EM]

I'd like to choose some low noise transistors out of the ones I have for a particular application. I have some BC559C models and 2N3906, 2N4403 and BC177. I'm thinking the 559C and 3906s are the best candidates (both quote max 4dB). So, i'm looking for a way to test the respective noise levels between various transistors, same model and different, to select the lowest noise ones. I have a basic oscilloscope, op-amps, regulated supply and multimeter but no specialist/sensistive equiptment. Any suggestions appreciated.

 

[audioguru]

Hi Dr. EM,
A transistor has its lowest noise with a certain amount of collector current and source resistance.
Old datasheets had much more detail than ones shown today and had graphs showing the interaction of current and source resistance on a transistor's noise.

Make an opamp with a gain of about 100 from a low noise opamp. Use it as a preamp that feeds an LM386 small power amp with a gain of 20 to listen to noise from your transistor under test that is a common emitter anplifier. Vary the source resistance and collector current.

 

[Nigel Goodwin]

Or, for that matter, build your intended application and then try the different transistor types in it - see which are the quietest - that way you're selecting under the correct conditions.

 

[Dr.EM]

Thanks for the suggestions, I can try audios method, but Nigels suggestion does make sense, and that is how I've thought of doing it. What has made it impossible is that the application is a microphone, and its very difficult to keep everything quiet around it whilst testing. This is the circuit:

"Better electret microphone phantom powering circuit" https://www.epanorama.net/circuits/microphone_powering.html



Is there a way to test this for noise without the microphone capsule in place? The capsule uses 0.5mA and has around 10v across it, would placing a 20k resistor where the capsule is shown work to test it? The reason for this is i'm sure i'm getting more noise than I should be (I used BC177s in my current mic but am now realising they arn't that quiet compared to others) and with 2 capsules still left (these are the nice panasonic WM-61As, I finally found them in the UK) I want the next ones to be as low noise as possible.

 

[Nigel Goodwin]

Thanks for the suggestions, I can try audios method, but Nigels suggestion does make sense, and that is how I've thought of doing it. What has made it impossible is that the application is a microphone, and its very difficult to keep everything quiet around it whilst testing. This is the circuit:

"Better electret microphone phantom powering circuit" https://www.epanorama.net/circuits/microphone_powering.html



Is there a way to test this for noise without the microphone capsule in place? The capsule uses 0.5mA and has around 10v across it, would placing a 20k resistor where the capsule is shown work to test it? The reason for this is i'm sure i'm getting more noise than I should be (I used BC177s in my current mic but am now realising they arn't that quiet compared to others) and with 2 capsules still left (these are the nice panasonic WM-61As, I finally found them in the UK) I want the next ones to be as low noise as possible.

A BC177 isn't a low noise transistor, it's just a standard PNP one, try something like a BC159, which is a low-noise PNP designed for audio preamps.

 

[Dr.EM]

Yeah, I would like to use something like that, but unfortunately can't find them at my usual supplier:

**broken link removed**

Is BC159 the European number for the NTE159, I saw that used in a similar circuit.

Also looks like my BC559s arn't suitable due to thier low 30v rating. Is my idea for using a 20k resistor as a load to check transistors in the circuit ok? I have quite a few 3906s and might be lucky with some.

 

[Nigel Goodwin]

Yeah, I would like to use something like that, but unfortunately can't find them at my usual supplier:

**broken link removed**

Is BC159 the European number for the NTE159, I saw that used in a similar circuit.

No idea on NTE numbers, "Towers International Transistor Selector" book doesn't even mention them?.

Actually, BC159 is only rated at 20V, all the low noise ones in my 1973 Mullard databook are only 20V (such as BC109).

Also looks like my BC559s arn't suitable due to thier low 30v rating. Is my idea for using a 20k resistor as a load to check transistors in the circuit ok? I have quite a few 3906s and might be lucky with some.

You can certainly try it and see?.

 

[Dr.EM]

This is where i've seen the NTE159 and google seems to find some datasheets:

http://www.mp3forkidz.com/mic/phantxy.html

Thats the circuit i'll most likely use, is there any reason for such large 220uF caps, most other designs use 10uF?

I tried some 3906s out; some are definately quieter than others, and i've identified a pair where they contribute very little extra noise over that of the mixer's mic amp.

 

[Optikon]

Yeah, I would like to use something like that, but unfortunately can't find them at my usual supplier:

**broken link removed**

Is BC159 the European number for the NTE159, I saw that used in a similar circuit.

Also looks like my BC559s arn't suitable due to thier low 30v rating. Is my idea for using a 20k resistor as a load to check transistors in the circuit ok? I have quite a few 3906s and might be lucky with some.

Can you get BC850C / BC860C ??

The "C" designator indicates noise binning. I think these are < 1.0 dB

 

[RadioRon]

I think NTE is a reseller of parts to the repair market. I bought some from them once and the parts came in little bags obvioulsy meant for display in a store. The ICs are indeed marked NTE so I'm not sure how they buy them, but I suspect they are not the original mfr, or else they have a supply deal with a foundry. Seems to be mostly older stuff and popular replacement parts they carry, from what I've seen.