Selecting transistors for lowest noise

[Adam Smith]

I'm building a simple single transistor guitar petal to buffer long cable runs.

schemeo here...
http://generalguitargadgets.com/pdf/ggg_buffer_bjt_sc.pdf?phpMyAdmin=78482479fd7e7fc3768044a841b3e85a

No big deal to knock up in a couple of hours. I was reading something somewhere about people selecting transistors for the lowest noise. I realized I have no idea how this is done. I've matched transistors for mic pre front ends and the like, but I don't build many guitar circuits. So, I'm assuming you just build the circuit, then swap out transistors whilst looking at the scope and simply find the quietest one in circuit, like you would do with a tube mic pre. Is there a better way to select transistors from a bunch for lowest noise?

 

[joaquins]

Go for a BC550 or 549 that are already low noise and you shouldn't have trouble, no much difference with 5089 thogh. It's a follower stage, I can't imagine what can go wrong and what you need to do to have problematic noise... What I see is that input Z es around 100K that it's a little low for a guitar, so I wouldn't expect to be so transparent, it will load your guitar if connected direct to it, if it's used at the output of any active stage it would be much transparent and you won't have problems. Mostly, to connect direct to the guitar jfet input is used because higher impedance, you can go over a couple of MΩ without suffering as you would do with BJT.

In the other hand I'd go (for really long cables) for an opamp with an output resistor inside the feedback loop, to compensate cable capacitance... known schem I guess, but bit more complex than a single transistor. depending on your needs and preference best option would be a TL071 with BD139/140 output driving the cable, but it has sense if you are driving 40/50m cable. An 5534 could drive more than a TL without output help but is BJT input...

In any way, if you just want to know wich transistor to this circuit, put any 5089 and you shouldn't have noticiable noise from it, if you have noise I'd look for something else before start changing transistors as a trained monkey.

JS

 

[Gus]

IMO that circuit needs some help the noise is not the issue
the lower than 110K input resistance will take away the high end sparkle caused by the interaction of a the passive guitar and cable and input resistance interaction.

look at the last post in this thread for a NPN buffer with stability resistors and a better emitter voltage.
http://www.diystompboxes.com/smfforum/index.php?topic=99308.0
I have posted a few different buffers over the years at that forum
If you drive long cables you need to think about stability.
The circuits are buffers so a 2n5088 or mpsa18 etc

The bump is the interaction I am posting about that goes away with low input resistance and when you add series resistance by turning down the volume control.  The stuff before the input is a simple single coil guitar sim

 

[JohnRoberts]

I'm building a simple single transistor guitar petal to buffer long cable runs.

schemeo here...
http://generalguitargadgets.com/pdf/ggg_buffer_bjt_sc.pdf?phpMyAdmin=78482479fd7e7fc3768044a841b3e85a

No big deal to knock up in a couple of hours. I was reading something somewhere about people selecting transistors for the lowest noise. I realized I have no idea how this is done. I've matched transistors for mic pre front ends and the like, but I don't build many guitar circuits. So, I'm assuming you just build the circuit, then swap out transistors whilst looking at the scope and simply find the quietest one in circuit, like you would do with a tube mic pre. Is there a better way to select transistors from a bunch for lowest noise?

Yup, that's a simple buffer...

Generally you would start by specifying a part that will be low noise "as used". A significant characteristic of guitar pickups is that they are relatively high impedance compared to microphones or low impedance sources, so a transistor that might be good for a mic preamp would likely not be optimal for a guitar pickup.

Input noise associated with bipolar transistors is generally noise current and noise voltage. The noise current will get multiplied by source impedance and add as another noise voltage term so both need to be considered. If you study low noise resistor data sheets you will see graphs of constant noise figure showing a region of impedance where noise voltage and  noise current add to be X dB noisier than the transistor alone.  Generally finding a device that has the lowest NF when terminated with the source impedance of a guitar pickup should be quietest.

I'll save you the time and trouble of looking up transistors, you are probably better off looking at a JFET which will generally have very low noise current so work better with high impedance guitar pickups.

That said the voltage signal coming from a hot guitar pickup can be a couple volts, so the need for extremely low noise devices in your buffer are not that apparent.

I will second the other comments that you do not want to load down the guitar pick up.. even 100k ohm is marginal... Perhaps consider a JFET or darlington connected bipolar to get the input impedance up.

Stability in a unity gain follower should not be very hard, but good layout, perhaps a small resistor in series with the output to decouple from cable capacitance might be useful.

This seems like a perfect project to start with.

Fist step might be to experiment with sonic effects of loading your guitar pickup with a simple resistor to see how much loading is audible and undesirable to you. Don't take our word for it, try a 50k, 100k, 500k, etc and listen.

JR

 

[Gus]

Wow

Just find my posts at diystompboxes about buffers.  I posted Jfet, jfet bjt, opamp, bootstrapped opamp, boostrapped NPN bootastrapped NPN with a bottom bump, szikla pair.  I also posted sims of what loading does to the response and what series resistance does to the response.  Don't forget cable capacitance.

IMO a min load to still get the high end sparkle with a tele strat etc is around 470k (ever look at the input resistors on the npn buffers in jfet switched effects).  I tend to like 1meg for a buffer however with a gain stage I like to make the input resistance change with gain.

A boostrapped NPN buffer can be low noise and be design to let a transistor drop in and work correctly a jfet circuit can trip a beginner.