Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 19, 2024, 03:31:11 AM

Login with username, password and session length

Recent Posts

 

Fet version of the JCM800

Started by KMG, January 16, 2011, 01:42:05 PM

Previous topic - Next topic

KMG

Schematics (pdf)

Stage by stage frequency response in comparison with the original tube version (stage2 with different gain positions)

Ready to transfer PCB (pdf)

Upper side assembly (pdf)

Bottom side assembly (pdf)

Assembled board

Controls and input jacks (pdf)

Project page (with some samples)
http://milas.spb.ru/~kmg/jcm800fet_en.html

DJPhil

Excellent work!
That's a very well thought out and documented project. I bet you'd be able to sell boards or kits for it. I don't think I've seen such a well made DIY high voltage FET tube sim circuit before.
I'll bet that you'll have folks who are trying to make their own asking about the power supply board soon. It might get tricky for those of us working from 115VAC. I bet I could sort something out from older tube circuitry, after all, it makes sense. :)

You've got a really good article on your site about FET simulation as well! (Google translate link to english)

Thanks for all this, I've got some reading to do.  :tu:

KMG


KMG


J M Fahey

IM_PRESS_SIVE !!!!!!!!!!!!!!!!!!!!!!!! :tu: :tu: :tu: :tu:
You have nailed it *exactly*.
Congratulations !!
Thanks for sharing your awesome work.
Can you explain how did you record this?
Did it go through Cab+mike or is it all emulation too?
Thanks.

KMG

Photo of tube prototype:

Samples recorded by simply swapping connectors from one board to another (connectors pinouts of boards are the same) via poweramp & cabinet.
Recording Path:
E-MU Tracker Pre->preamp->5W Fet PA->4x12 cabinet->Sennheizer E906->E-MU Tracker Pre

PS Any thoughts who is who on samples?

Tubis

Some questions to an old post:

1. The 2SK215 circuit has the gate diod feed by the drain. The LMD150 circuit by the source. Why the difference?

2. Why do the circuits need a negative supply on the source and why do that supply need to be buffered?

3. How crucial is the type of diods? You use different types for the two circuits?

4. Does the circuit need to be high voltage or can you get good results with 10times lower voltage?

KMG

Quote1. The 2SK215 circuit has the gate diod feed by the drain. The LMD150 circuit by the source. Why the difference?
Different VA curves.
I used 2SK216, 2SK215 has max 180V drain voltage.
Quote2. Why do the circuits need a negative supply on the source and why do that supply need to be buffered?
Because with autobias with source network tied to GND I can`t get stage gain, frequency response & top/botton clipping points same as for tube prototype.
Quote3. How crucial is the type of diods? You use different types for the two circuits?
In source network you can use germanium/shottky diodes, they has similar knee curve.
Quote4. Does the circuit need to be high voltage or can you get good results with 10times lower voltage?
I started with low voltages but understood that this can`t give "true tube sound".

phatt

Quote from KMG;
"I started with low voltages but understood that this can`t give "true tube sound"

Yes I'm quite convinced that high voltage plays a big part even though I've never found a good explanation of that point it makes perfect sense that even a banana would benefit if you run it from big voltage. :cheesy:
Phil.

Tubis

Whats your take on this circuit?

http://gtlab.net/gtlab4/wp-content/uploads/2009/03/soldano1.jpg

How close to tube can that sound?

And why the neagtive supply on that? Dosent that get eaten upp by the drainfeed anyway?

KMG

I started low voltage fet experiments with Victor Kampf idea.
http://www.sugardas.lt/~igoramps/article68/article.htm
But sound was not close enough to tube one & I moved to high voltage experiments.
First was Bogner Ecstazy preamp. Even without grid current emulation it gives results close to tube sound.
Fet version:
http://milas.spb.ru/~kmg/files/projects/xtcfet/pedal/samples/denn/kmg_xtc_mix.mp3
Tube version:
http://milas.spb.ru/~kmg/files/projects/xtcfet/pedal/samples/denn/denn_xtc_crunch_bluesmix.mp3
I tried to use Kampf`s idea with 2SK216, but in this case stage fall to FET saturation or has very fast beginning of grid limiting depending on diode biasing.
After discussing with author of this article http://gtlab.net/gtlab4/archives/216 (where this schematic http://gtlab.net/gtlab4/wp-content/uploads/2009/03/soldano1.jpg was published), I used his idea with 2SK216 & it works fine.
But I don`t tried his schematics in original.
When I start experiments with LND150 I tried to use Kampf`s idea again & it works fine with this type of FETs.
If you want to try JFET soldano you must use JFETs with cutoff voltage 0.6-0.8 Volts.
Stage tuning procedure (translated from article).
1. Set voltage at drain 6-7 Volts by adjusting source resistors (1k*).
2. Connect left pin (by schematic) of gate resistor to VSupply. Check drain voltage, it must be 2.5-3 Volts. If it high decrease resistance of R(22k*) from -bias supply, othervise increase it.

Tubis

#11
So many questions :-[

In the source network...there is a diod called ES1J.
Google says its a silicon fast recovery. Is it crucial or would any old silicon diod do?

What ratings does all the diods in the source demand when it comes to voltage and current?
Im now talking about the lnd150 circuit.

And finaly,  wouldn't it be nice to model the power amp with lnd150 high voltage as well? I can se on your home page that you done a 12V poweramp model. But wouldn it be nice to be able to run that with the preamp at the same voltages?

I like your work by the way....got me all exited. And i happen to have a 100 lnd150 collecting dust at home.

KMG

QuoteIn the source network...there is a diode called ES1J.
Google says its a silicon fast recovery. Is it crucial or would any old silicon diod do?
Germanium/shottky diodes added to source network to round sharp edges then stage enters/exits from cutoff state by summing VA curves of diodes & fets.
Then previous stage outputs full clipped voltage span, peaks on gate may be up to -160 volts.
Because LND150 has gate protection zener, fet source follows gate with voltage less than gate voltage by amount of protection zener breakdown voltage.
Diodes in source network become reverse biased. But maximum reverse voltage for BAT54 is 30 Volts. So I added high voltage fast diode to source network.

Ilya-v

KMG,

According to this final gain stage schematic:


R2 & D1 control how (softly/round/when) the grid is clipping when the input is reaching the bias point (-1.5)?

C1 is miller capacitane emulation.
What exactly D2, D3 and D4 are doing in terms of 12ax7 biasing equivalence?
R? and C2 are the same as Cathode resistor and Capacitor as in the 12ax7?

KMG

QuoteR2 & D1 control how (softly/round/when) the grid is clipping when the input is reaching the bias point (-1.5)?
R2 defines "softness" of grid clipping. Voltage drop on D1 & R3 defines point of clipping.
QuoteC1 is miller capacitane emulation.
Yes.
QuoteWhat exactly D2, D3 and D4 are doing in terms of 12ax7 biasing equivalence?
As I wrote above:
Quoteshottky diodes added to source network to round sharp edges then stage enters/exits from cutoff state by summing VA curves of diodes & fets.
D4 added to protect shottky diodes from destructive breardown
QuoteR? and C2 are the same as Cathode resistor and Capacitor as in the 12ax7?
No, left resistor defines stage gain below cutoff frequency, right R defines gain above cutoff, RC defines cutoff frequency.