There have been a few threads on how to get a full high gain no-holds-barred design from the Jet City JCA20H amplifier.
I find this topology does a good job of going clean-ish when the volume from the guitar is backed off nearly all the way but has more than enough overdrive to use with my strat. I have a bleed cap fitted on the guitars volume BTW.
Compared with the original first three gain stages setup of x69 x23.6 x59 the individual gains are now x55.7 x47.6 and x54.6 so no valve is pushed too hard to spec. What has happened is a attenuation between stages has been removed and the treble voicing moved to after the tone stack. If we had infinite supply voltage and valves that didn't arc then the original cascaded gain to the start of the tone stack was around a factor of 12.5 thousand, with the modifications it would be around 118 thousand.
The mods are presented as achievable without cutting tracks or removing the PCB from the chassis. Simple to try and simple to put back. From the official schematic here are the changes made.
C27, 1uF. Fit an additional 1uF in parallel to make it 2uF.
R28, 220k. Fit an additional 220K in parallel to make 110K.
C26, 20nF. Fit a 22nF 400V cap in parallel to make 42nF in total.
R25, 2K2 fit a 1uF 25V cap in parallel.
R53 1Meg. Sort circuit this part
Gain Pot 1Meg Log. Fit a 62K or 68K resistor across the outer track connections.
R53 to Wiper of Gain pot. 1.8Meg and 120pF in series (was100k and 1nF)
C23, 470pF Short circuit this part.
R14 470K Wire 220pF and 220K in series to restore some of the treble lost shorting out C23.
R12 47k (NEG Feedback Resistor, Power Amp section) Fit 330pF in parallel to quash harshness.
And two mods from The Valve Wizard Pages
R20 100k. Fit an additional 100K in parallel to make 50K
I also fitted a neon between pins 2 and 3 as protection.
The amp also has a depth or resonance modification fitted which warms the amp in the correct place.It also has a triode switch which mellows the tone and an unbalanced phase splitter to allow 2nd Harmonic distortion in the output stage.I'll detail the output stage modifications in another post.
I found that the treble bleed at the gain pot helped establish a usable clean sound lifting the mid to treble but not the bass which often pushes the stage into grit. For now I have fitted R-3 as 100k and C-3 as 1nF. However this may change as it is still work in progress. (UPDATE: Fitted R-3 = 1.8Meg and C-3 = 120pF as final value)
Below is the response at the cathode follower with the gain pot set to about 2. The lowest plot is without a bleed modification. The green upper plot is 100k and 1nF which gives a 12dB lift beginning at the lower mid frequency. The intermediate red trace is 100k and 220pf.
Dumping area for various material. May or may not be useful or correct.
Thursday, 19 January 2012
Wednesday, 4 January 2012
Jet City JC20H output harshness
The Jet City JC20H is an EL84 based guitar amplifier which has had some criticism for sounding a bit harsh or brittle. Mine was no exception. The stock EL84s proved to be substantially mismatched and were ditched for some matched JJ Tesla. However the brittleness proved unmoveable despite many changes to the pre-amp section.
The solution was with the power amp section. A 330pf capacitor across the 47K negative feedback (NFB) resistor marked R12 in the service schematic. This removes the unnecessarily high bandwidth of the output stage, and will also increase high frequency stability. Whatever the cause, the brittleness is gone and the amp sounds a whole lot better. Roll off is 1/ 2Pi * R *C , so 330pF and 47K will equal about 10kHz which is ample for a guitar amplifier.
The simulation below shows the frequency response with and without the 330pF capacitor. The peaking at 14kHz is reduced by about 9dB. Scale is 3dB per division vertically and 2Hz to 20kHz.
Tuesday, 3 January 2012
Pots that drive me up the wall.
I get annoyed at pots in guitar amps (potentiometers), usually the volume pots. They have two attributes I hate. Noisy over time and the inability to get from zero to low volume in an orderly fashion.
Firstly let’s look at issue of a noisy pot, you move the volume and get great crunches out of the speaker. I got to thinking about this the other day and came up with the following plan. I’ll replace the 1Meg volume pot (log scale for volume) with a 2.2Meg stereo log pot. I’ll then wire the two pots together in parallel, so wiper to wiper, top of pot to top, ground to ground. This will give me a 1.1Meg Log pot. This way I have two wipers to go wrong, but I should have less chance of a disconnection occurring as one of the wipers should have a good connection at any one time. Something to give a go.
This then set off a train of thought. What about wiring just the wipers and ground connection together. This should give a modified Log response with hopefully a slightly more gentle lift from zero and a bit more granularity in this area. So a pot and a parallel sub-pot.
If I plot out the response from a theoretical Log pot and then as a “sub-pot” I get the following change plotted as the green curve. This assumes zero source impedance and an infinite load. I do get a slight easing in the very low volume setting area, followed by a slower curve into the 60% rotation which then quickens up to meet at 100% rotation.
Now anyone familiar with a Marshall JCM800 amplifier will tell you that 0-4 is zero to full power, after this you are soon overdriving the power section. This method will give slightly more control outside of the power amp overdrive section, which is where I want it.
You can also play with the same idea with a linear pot, say the Treble tone control in a passive TMB (treble, mid, bass) tone stack. Here the parallel pot is marked as X1.
We will get a degree of log law faking but we also modify the end to end resistance which will play with the impedance sensitivity of the other controls. To be fair it isn't too far from a fixed resistance at the same point, but different enough.
A common treble value for the pot is a 250K linear. If we substitute in a 250K lin stereo pot wired as a sub pot then the following changes take place. Note that the mid scoop is reduced, and at maximum treble boost the scoop centre frequency changes from ~250Hz to ~350Hz. There is also a more linear change with pot rotation. You can also get a scoop reduction by just replacing the parallel pot with a fixed 50K resistor but that can’t be as much fun.
Firstly the notmal TMB circuit moving the treble control from 0 to 100% in 10% steps. 6dB / division.
And now with a 250k stereo pot wired as a sub-pot.
Feel free to play and adapt.
Roosville.
Friday, 9 December 2011
Day One.
For a while I have realised that I draw a lot from the web without really putting back. I decided a blog would possibly serve to address this whilst keeping the stuff I want to keep fresh for myself in one place and not distributed on message boards or bits of paper. If anyone finds anything here useful then it’s served a purpose and I’ll be happy with that. Stuff will appear here that hasn't been fully thought through or proven, so please bear that in mind. I am sure I'll see some "seen & done that before" in which case, well done.
Roosville
Roosville
Subscribe to:
Posts (Atom)








