IET
Decrease font size
Increase font size
Topic Title: JFET Guitar Amplifier Advice
Topic Summary: Should I use a source bypass capacitor?
Created On: 21 August 2012 12:29 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 21 August 2012 12:29 PM
User is offline View Users Profile Print this message



amillar

Posts: 1918
Joined: 28 May 2002

I'm building a simple JFET guitar amplifier for home - just an impedance buffer really so that I can plug guitars straight into the mic input of my mixer. Most circuits I can see on the web are common source amplifiers but with no bypass capacitor on the source resistor. Although I'm not an expert on discrete circuits (to put it mildly!) wouldn't a source bypass capacitor be a good idea? Or doesn't it make any difference in practice? Anyone here got any experience?

See, for example, the circuits at www.till.com

Thanks,

Andy

-------------------------
Andy Millar CEng MIET CMgr MCMI

http://www.linkedin.com/in/millarandy

"The aim of argument, or of discussion, should not be victory, but progress." Joseph Joubert
 21 August 2012 01:20 PM
User is offline View Users Profile Print this message



ectophile

Posts: 526
Joined: 17 September 2001

I'm no expert on analogue electronics either, but my immediate thought is that if you added a capacitor, it could muck up the frequency response by presenting a frequency-dependent impedance to the source.

In particular, if you stick a capacitor in parallel with R1, then the impedance will drop at high frequencies.

-------------------------
S P Barker BSc PhD MIET

Edited: 21 August 2012 at 11:00 PM by ectophile
 22 August 2012 11:57 AM
User is offline View Users Profile Print this message



amillar

Posts: 1918
Joined: 28 May 2002

Ok, with a bit more browsing and thinking I think I understand what the idea is now - perhaps someone could confirm I've got this right?

With the source resistor NOT bypassed a common source JFET amplifier works like this: The source voltage follows the gate voltage (i.e. this bit is a source follower). This defines the current flowing through the source resistor, Vin / Rs, this current will then be flowing through the drain resistor Rd*. Voltage gain will therefore be Rd / Rs.

With the source resistor bypassed the gain is just transconductance (gm) x Rd*, hence gain is dependent upon the gm for that particular device and the linearity depends upon the linearity of gm against Vin.

So for a single stage audio amp with no other feedback you DON'T want to bypass the source resistor, you want to use it to set your gain.

Am I right? I will award myself a Malteser** if so!

* Actually should be Rd // Rload, but I'm keeping it simple for the moment
** Other chocolate based confectionery is available


-------------------------
Andy Millar CEng MIET CMgr MCMI

http://www.linkedin.com/in/millarandy

"The aim of argument, or of discussion, should not be victory, but progress." Joseph Joubert
 23 August 2012 04:27 PM
User is offline View Users Profile Print this message



ANFierman

Posts: 136
Joined: 25 July 2008

Consider yourself awarded a virtual chocolate based confection.

You might like to have a play with the various circuits you've been looking at by entering them into the free web-based simulation tool:

www.circuitlab.com.


Beware though that the overall frequency response will be heavily influenced by the guitar pickup and the circuitry built around it internal to the instrument.

There's a Circuitlab example of what some people call a tone stack here:

https://www.circuitlab.com/circuit/xr4zdg/guitar-tone-stack-01/

Searching their site for "tone stack" will turn up some more bits and bobs.


Alternatively you could try the excellent, free, unlimited mixed-mode spice simulator, LTspiceIV from:

http://www.linear.com/designtools/software/#LTspice

Both can be run on Linux, Mac and Windows. Circuitlab can be run on Android and iPad too.

-------------------------
Andy Fierman

---------------------------
http://signality.co.uk
---------------------------
 23 August 2012 05:06 PM
User is offline View Users Profile Print this message



amillar

Posts: 1918
Joined: 28 May 2002

Originally posted by: ANFierman
Consider yourself awarded a virtual chocolate based confection.


Many thanks!

I've moved on a bit now, having had a cunning thought in the early hours of the morning that I could make a phase splitter (by taking the output from the source and drain sides) and feed it as a balanced signal into the mixer input - may give me a bit of extra voltage swing and will be interesting to try anyway. Have come up with a VERY cunning plan (if I say so myself, and assuming it works ) to keep the impedances equal given the phantom power circuit.

The other free simulator (the one I use at home normally but not at the moment as I've just rebuilt my PC) is PSPICE at http://www.electronics-lab.com/downloads/schematic/013/

But as I'm on holiday this week I'll probably just knock up the circuit - will probably be quicker than installing the simulator and remembering how to drive it!!

It's a bit sad that I only get to do proper engineering in my holidays now Ho hum, the joys of management...

Cheers,

Andy

-------------------------
Andy Millar CEng MIET CMgr MCMI

http://www.linkedin.com/in/millarandy

"The aim of argument, or of discussion, should not be victory, but progress." Joseph Joubert
 24 August 2012 12:59 PM
User is offline View Users Profile Print this message



ANFierman

Posts: 136
Joined: 25 July 2008

Hmmmm,

You probably don't use simulation enough to be too bothered about it this but I gave up on PSpice years ago. The link is to version 9.1 which is very old now but even the up to date version still has a lot of convergence problems and dubious tricks to try to avoid them. Plus of course the demo version is either heavily node or - nowadays - time limited.

I have found SIMetrix to be about the fastest and most well behaved spice simulator with a very generous free Intro version. The GUI is nice too.

www.simetrix.co.uk

LTspiceIV is a bit slower but I've found it about on a par with SIMetrix for accuracy and convergence. I can also make it do things that are tricky to achieve in SIMetrix. Overall both SIMetrix and LTspiceIV are two of the best spice simulators available at any price.

Whatever, it sounds like you're going to have fun on your holiday anyway!



-------------------------
Andy Fierman

---------------------------
http://signality.co.uk
---------------------------
 24 August 2012 07:49 PM
User is offline View Users Profile Print this message



amillar

Posts: 1918
Joined: 28 May 2002

Since I don't do any design work in the day job any more I haven't personally used simulation professionally for a few years (except for some really trivial cases). My team does still use PSPICE (since we have it and are used to it) but I agree that convergence can be a problem. They also now use the Proteus simulator built into ISIS (since we use ISIS for our schematic capture anyway) which so far seems slightly better perhaps than PSPICE, but to be honest our serious simulation is moving much more over to MATLAB.

For home use I find the 9.1 version fine - but then I've been simulating audio circuits in various forms of SPICE for nearly 25 years so 9.1 is modern to me!! I remember the days when we would leave our machines running overnight to plot the time response of a two pole filter...

But many thanks for the tips, I will look into SIMetrix for my own use at least.

-------------------------
Andy Millar CEng MIET CMgr MCMI

http://www.linkedin.com/in/millarandy

"The aim of argument, or of discussion, should not be victory, but progress." Joseph Joubert
 24 August 2012 10:10 PM
User is offline View Users Profile Print this message



ANFierman

Posts: 136
Joined: 25 July 2008

I'd recommend having a play with LTspiceIV too. The GUI seems a bit clunky but is quick to pick up and makes it easy to do some very powerful things.

There's a good user group and a couple of very helpful websites. The author, Mike Engelhardt, is very helpful too: bug reports usually get fixed within a couple of days. Amazing for a free tool!

Don't be put off by it being free or supplied by a vendor (so maybe it's tweaked to make their parts look good). That's not the case. LT use it to design their own devices and have made some major changes and fixed lots of the original spice bugs. I won't go on about it except to say it is an industrial strength simulator.

MATLAB and similar FOSS tools such as Scilab, Scicoslab, Octave and Sagemath are great for generic simulation but for specific electronics you have to try hard to get better results than spice or Gnucap. Accepting of course that any simulation can only be as good as the models! That said, you can simulate most things in spice as long as you can get your head round how to represent it as an analogue or mixed signal system. There have been a few articles in EDN etc., about using LTspice to model an analogue computer because it's removes most of the inaccuracies of a real analogue computer.

I use LTspice to simulate linearised models - based on the work of Dr Vatche Vorperian and extended by Christophe Basso - of switch mode PSUs, class D amplifiers and the like which gives me results that are every bit as accurate as switching simulations but well over 100 times faster.

-------------------------
Andy Fierman

---------------------------
http://signality.co.uk
---------------------------
 11 May 2013 04:24 PM
User is offline View Users Profile Print this message



kengreen

Posts: 400
Joined: 15 April 2013

It's a bit late in the day but I think I know the answer to the original question. The bypass capacitor across the Source resistor is there to produce a standing bias - as you will find in many circuits starting with valves. The purpose of the bypass C is to prevent the development of a signal voltage at the Source . Omit the capacitor and you have a negative-feedback circuit which will reduce the stage gain and modify both the input and output impedances of the stage while at the same time reducing distortion produced by the non-linearities of the transistor.

As a side issue by tailoring the value of the bypass capacitor you can make a fequency weighted stage (i.e. a Bass-lift stage) but beware the changing impedances and phase-shifts that go along with it.

Yes, you can use the circuit w/o the capacitor as a phase-splitter but there will be a loss of balance between the two signals and thatg will depend on the characteristics of the two following amplifiers.
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.