JFET Mini Amp build

[jammybstard]

Not  a stompbox as such; but it's grown out of a preamp I made a couple of weeks ago which stared me off with JFET design (Mine field).
I've started building a mini practice amp so far I've built a clean power stage, based on a TDA7375 that will drive a 4Ohm 6" Jensen Speaker. (Its a Car stereo chip amp that can drive 4Ohms)
And I've done my sums and breadboarded the preamp this morning:



At the front is an 2sk170 Common Drain unity gain Buffer and then two 2SK170 common source gain stages; the SOT 1Meg Pot in between alows me to set up how much and when the second JFET starts to clip.
It all seems to be working ok! but it's a bit Boomy compared to my SE-5A Valve-amp! I'll add a tone stack once this is working right but at the moment I'm looking for some feedback on the design so far; I've never really designed anything like this before and I'd be interested to know if you think I've done anything stupid and any ideas on how to make it less boomy?

[familyortiz]

Looks very cool... I'm glad it's working for you!
One thing that stick out are the values for you coupling caps (100nF). Generally I find that these are values greater than 1.0uF and the actual corner frequencies will be defined by: Fc = 1/ (2*pi*R*C).
When the C is a bit low, then you are attenuating some highs, which might be the cause of the boominess. If possible, try using higher values and see if this makes a difference.
It would be cool to run it thru a frequency sweep to check it's freq response!  Good luck.

[jammybstard]

One thing that stick out are the values for you coupling caps (100nF). Generally I find that these are values greater than 1.0uF and the actual corner frequencies will be defined by: Fc = 1/ (2*pi*R*C).
When the C is a bit low, then you are attenuating some highs[/quote]

You sure? I thought capacitors in series blocked low frequencies; 1/(2*Pi*1M*.1u) = 1.58Hz (Fc)
Bottom E is 82HZ and human haering goes down to about 20/30; perhaps I could reduce it and block the subF's.

I would like to simulate it but I cant find a 2SK170 in PSPICE (I havent looked very hard though!)

[familyortiz]

jammy,
you're exactly right, caps block low freqs but the impedance (ac resistance essentially) is inversely proportional to the value, i.e. a higher cap value will have less impedance for the same frequency, than a lower value. The cap and resistor will form a corner frequency and your design is ok where 1Mohm is part of the pole, exactly as you've calculated. Where I would question the .1uF is at the output coupling. In either case, I would try to change this value only to see what effect you get.

[jammybstard]

Yes your right Yortiz I have got that wrong eitherway; the input impedance of the power amp is a lowly 7K5 which would bring the Fc right up towards 200Hz at max volume.
I'll need to go to at least 1u!
Well spotted.

I'm wondering if the value of CS on the third  JFET could effect the boom!

[familyortiz]

The value of Cs could have an effect, but I realized that you are gaining things up by 22 x 22! Could this be the culprit?

[PRR]

> 4Ohm 6" Jensen Speaker

Car-speaker or voiced for guitar?

> it's a bit Boomy

Without knowing some key values in your amp, I'm guessing it is "flat". I see 100nFd driving 1Meg: this is -3dB at 1.6Hz and -1dB at 3Hz. Good hi-fi practice.

Guitar amps often slope-down the bottom octave or so. Maybe -3dB at 150Hz. Or several -3dB at 100Hz.

Try changing some 100nFd to 5nFd or 2nFd.

Indeed the coupling out to the power stage needs re-checking.

> I'll add a tone stack once this is working right

Ah! No, since it IS working and not-wrong, add your tonestack now. Overall gain will drop a lot. Also the distortion profile will change. Depending on tonestack, "Boomy" may become better or worse.

> I would like to simulate it but I cant find a 2SK170

The JFET stage is perhaps best breadboarded.

You _should_ learn to rough-out the frequency response on a napkin. When you have many stages and odd values, SPICE may be convenient. You "know" the naked JFET stage has response from zero to MegaHertz. And you say it has gain of 22. In pSPICE, there is a pseudo-part called "GAIN". Drop one in a sim, change the number on it (the gain) to "22". Add 1K8 in series with the output to emulate the JFET stage output impedance. (Assuming 22 is correct) this will compute the gain and the frequency response.

[jammybstard]

> 4Ohm 6" Jensen Speaker

Car-speaker or voiced for guitar?

Guitar: http://jensentone.com/mod6-15.php

I dont have it yet; At the moment I'm testing through a 10inch 8ohm HH (Laney, Not a great speaker!) I also tried through a 16ohm G12 Greenback which I thought sounded a bit better; but that could have been for severzal reasons; (And most things sound better through a greenback) I'll have to get the Jensen in soon as it's a key part of the sound.

> it's a bit Boomy

Without knowing some key values in your amp, I'm guessing it is "flat". I see 100nFd driving 1Meg: this is -3dB at 1.6Hz and -1dB at 3Hz. Good hi-fi practice.

Guitar amps often slope-down the bottom octave or so. Maybe -3dB at 150Hz. Or several -3dB at 100Hz.

Try changing some 100nFd to 5nFd or 2nFd.

That's interesting (Most of my earlier fettling has been with HI-FI and flat was the aim; It's a nice easy part to swap out and audition)
Would you change the cap between the buffer and the gain? or between the two gain stages? or does it matter?

> I'll add a tone stack once this is working right

Ah! No, since it IS working and not-wrong, add your tonestack now. Overall gain will drop a lot. Also the distortion profile will change. Depending on tonestack, "Boomy" may become better or worse.

Point taken!
I was planning a fender-style tone-stack with Bass and Trebble controls and a fixed SOT Mid resistor; I'm waiting on parts for that.
It should mostly be together by the end of next week.