Help me with my JFET / BJT design

[RichardJones]

OK... I tried to attach an image.. but its in emf? format.

Simply put... the pre-amp was simulated with 2N3819's and BC547's.

It has a cascode 2N3819 with a 560 ohm source resistor and a 10K drain resistor.

Running from a +36v supply... that places the upper 2N3819 drain at around 18v.

This is taken from this point into 3 paralleled BC547 emitter followers with a common 270 ohm resistor.

Output is taken from this point via a 1uf cap and a 10K load resistor.

Simulating the circuit... I get 0.0001% THD with 10v peak output at 1Khz.... but at 10Khz get 3.53% THD...

Why the huge difference?

Should I be running more current in the cascode?

btw... I need a voltage gain of 10.

Thanks!!!!

 

[JensRasmussen]

Simulations

Hi Richard

First of all it would help with a schmatic

The THD simulation is difficult, since it's based on samples, and a FFT (Fast fourier transform) The problem is problably the simulator and not your amp.

Also the THD needs a center freq to do the calculations. What program do you use?

My experience is that simulations are less than perfect regarding THD, but a nice tool to make sure all your DC levels are right.

\Jens

 

[mlloyd1]

yeah, we need a schematic.

and I think most of us have come to the conclusion that distortion numbers on a simulator are about as close to worthless as you can get. there was at least one thread on this a while back.

sorry

mlloyd1

 

[RichardJones]

schematic?

I am using a SIMetrix demo software.

Ahh... I think I figured it out:

 

[jam]

What happens if you replace R4 with a current source?

Regards,
Jam

 

[RichardJones]

Advantage?

The output darlington's don't really add a lot to THD I think... so what would be the advantage?

Would it be better to put one on the Source of the cascode?

Thanks!

 

[jam]

Possibly,

Why don't you try both and report your results.

Regards,
Jam

 

[RichardJones]

split supply

oh... well since the source of the FETS are about 1v above ground.

And btw I changed the drain resistor to 12K and the source to 680. Better symmetrical clipping...

I would have to go with a split supply to build a CCS for the source of the input FETS.

I could try a CCS on the output NPN....

 

[RichardJones]

CCS on output

Remember that I am simulating all of this... and don't have anything built..

the CCS didn't change anything wrt the 560 ohm resistor.

the simulated distortion is dominated by the input FETS.

 

[jam]

I suppose you could cap couple the input and use a voltage divider to pull up the gate of the fet to avoid building a split supply.

Regards,
Jam

 

[RichardJones]

confused?

What would increasing the voltage on the gate of the FET do?

Please explain!

p.s. This is kinda my first intro to JFETS and BJT... I have been designing with tubes and MOSFETS b4 now.

 

[jam]

The j fet is a depletion mode device so it self-bias but is limited as to how much voltage swing is allowed at the input. If you bias the input at a higher voltage, this will allow a wider voltage swing at the input before clipping.

Regards,
Jam

 

[Rob M]

Just curious -- how do you get THD figures using SIMetrix?

 

[RichardJones]

THD figures

Run a transient analysis... display the curve and select it. Its under the 'Measure' menu on that window.... select transient and select distortion full.

something like that anyway...

you can also do a FFT on it and see how the THD spectra looks like.

Simulated this circuit has a nice distortion progression.... dominated by 2nd until close to clipping then 3rd creeps in.

 

[Rob M]

THD figures

Thanks! I've been eyeballing it from the spectrum, but it'd be nice to also have a single number to keep track of changes.