A one JFET two NPN buffer design

[Gus]

here is a screenshot of a sim

18VDC just for fun
JFET input with a "helper" NPN follower stage for high input resistance and good current drive the NPN supplies most of the drive
Constant current circuit in the emitter source leg set at .525VDC/220 ohms

Also look at Geofex for the onboard preamp a different constant current setup



1ma 1k 1V
1ma 10k 10V

Ask questions if you don't understand what is going on. 
This is posted to get people thinking about the follower in the sim (I first saw a circuit like that in a Sony microphone schematic)and the use of a constant current circuit

things can be adjusted
input resistance
CC section
power supply voltage

[booj]

Bucu I through 2n5089's

[Gus]

Found the Sony schematic
http://recordinghacks.com/images/mic_extras/sony/C38B.jpg
That follow should also work well as a guitar buffer

[blackieNYC]

Would the sim tell you the THD with 100-150 mv AC input?
With 9v?

[PRR]

> the THD

Super-duper clean to 2V output.

[thelonious]

Nice filename 

Any advantage/disadvantage to increasing C5? Seems like I usually see >10u in that spot. It's just for smoothing Vref, right? Is 1u at C5 enough when you already have a big filter cap like C4?

[PRR]

> increasing C5? Seems like I usually see >10u in that spot.

And often with 10K resistors. Since this only has to supply the low-low-current gate, Gus used 100K resistors. 10u with 10K is like 1u with 100K.

> you already have a big filter cap like C4?

C4 is directly on V1. In this tutorial plan there isn't even a resistor between. V1 could be a clean battery, or could be a crappy trashy voltage converter. If V1 is bad enough then C4 could be huge and you'd still have trash.

In real-life it might be wise to throw 100 ohms between the power source and C4.

(In SPICE, C4 is useless because unless simulated otherwise, V1 is "perfect".)

In either case.... the path from power through J1 drain and Q1 collector has high supply rejection. The drain+collector impedance is over 10K, the emitter impedance is 10 ohms. Crap is cut-down 1000:1.

BUT crap that gets to the JFET gate goes right through. So that one node needs to be clean. And since it is a low-current node, it makes sense to do a filter (R5 C5) just for that node.

[thelonious]

10u with 10K is like 1u with 100K.

Ah, got it. Thanks!

[Gus]

Part of the reason I posted this is because I get a laugh when I read about the Klon buffer and how people like it and now people clone it(clone a textbook circuit?).  Boring
It is a standard textbook circuit than anyone that has a clue should know about.

So I posted this circuit as something a little different.   Pick your input resistance adjust the constant current for output drive wanted/needed. "Helper" NPN for the JFET you can use a lower IDSS device and supply more current than the IDSS limit with the NPN  part.  I even posted a link to a Sony microphone circuit using this type follower.  

Class A, constant current, high input resistance all discrete transistor follower with a JFET

Then there is a nice low power circuit R.G. has at GEOFEX.  I am surprised being an R.G. post people have not wrote about building it.

Then there are NS app book circuits JFET and PNP gain of 1 and greater than X1 gain

There are other followers I posted here as well a transistor pair, bootstrapped transistor and opamp.

The TCE preamp thread is interesting

[Gus]

Would this get more reads if I named the thread SUPERBUFFER?

With all the buffer threads I thought something different might be noticed.

R.G.s coin cell powered buffer is also something I am surprised about with the lack of posts about it.  I was trying to get people thinking about buffers and the different ways of building them

In the circuit in this thread I did not care about current draw and can turn up the (constant current)CC if needed.  There are different ways of building a CC.  I simmed something easy to build that people might have parts for.

The interesting thing with this circuit (I saw this type circuit in the Sony schematic I linked) is how the JFET and BJT work together.  
The Vbe drop holds J1 source current fairly constant
The NPN supplies most of the current this allows more current than a JFET can supply due to IDSS.  Say you want 2ma but the JFET has an IDSS of .6ma.

This is on my list of circuits to build. It should work.

Note the Sony circuit is at 9VDC you should be able to drop a J210(EDIT I meant J201) and use a 2.2meg etc for the gate to divider node to set the input resistance
The gain stage to transformer Sony gain circuit might work nice as a DI with some adjustments

[duck_arse]

j210 you say? finally! I'm in.

[aquataur]

Constant current biasing is (amongst all other weird and wonderful buffer designs) well described in Ray Marstonīs "FET Principles and Circuits, Part 1 , Nuts and Volts Magazine, which you will find on the web.

You need elevated supplies, as with many constant current designs. IMHO, in queue with all the distortion devices for a guitar, this does not pay.
But it may be useful for something else. Good stuff Gus.

have fun,

-helmut