BJFE - Emerald Green Distortion Machine [traced]

[Seiche]

Manfred, when did the resistor 470k on Q3s gate appear? It's not in the tagboardlayout or the other schematic.

[Manfred]

Manfred, when did the resistor 470k on Q3s gate appear? It's not in the tagboardlayout or the other schematic.

I already mentioned above, that I miss the resistor in the tagboardlayout or the other schematic.
IMO the resistor is essential for biasing the FET Q3.
You can find a 1 Megohm or a 470 Kiloohms resistor in other circuits from BJFE on that point for.

[Seiche]

IMO the resistor is essential for biasing the FET Q3.

that's surprising, because the tagboardlayout is verified, I understand.

[Manfred]

that's surprising, because the tagboardlayout is verified, I understand.

Now the question is: What is verified, the layout corresponds with the schematic or/and the circuit sounds as it should?
What is truth?
I ran a simulation using the 2N5952 FET-model with and without the grid resistor.
I fed 700 Millivolts/1000Hz into the circuit.
Without the grid resistor the output gaves a very low signal with heayily deformed curve shape.
I going to order some 2N5952s for making real trials.

EmeraldWithoutAndWithGridresistorDC_Values.jpg

EmeraldWithoutAndWithGridresistorSignal.jpg

[Seiche]

that's very interesting. Which program are you using?

[Manfred]

that's very interesting. Which program are you using?

An old Multisim version.

[Seiche]

Okay I built this in line IvIark's and Manfred's schematics (without 470k resistor) and it works and sounds great with an MP38A NPN Germanium.
It sounds similar to the Bearfoot one I played once.

However, after playing with it for a while, I think Volume and Treble should be 50kA and Drive and Voice should be 50kB.

I'll run some tests with the 470k resistor soon and report back.

[modman]

I found some notes on my old computer on this that are anonymous (I'm assuming from the same source as IvIark) and it has a 470k resistor that is not shown in the schematic and the following voltages:

V+ = 8.93
Vref = 7.99

Q1:
S= 5.82
D = 2.15
G = 0

Q2:
C= 3.03
E= 1.21
B= 1.34

Q3:
D= 4.91
S= 1.74
G= 0

Q3 is in line with Manfred's calculation. I will do some test with my build.

EDIT: I just measured my unit's voltages (without 470k) and they are similar to the ones posted above, and sounds pretty similar to what I expected (not strange and low volume, it is very loud!). Also Manfred is taking the output from the source, whereas the original circuit has the 1uF and output connected to the drain of Q3. Manfred, could you repeat your simulation for both? I don't think a 470k is needed.

[Seiche]

I think Volume and Treble should be 50kA and Drive and Voice should be 50kB.

So I tested this with knobs to make sure and this is not correct, Drive, Treble, Voice are better when they are 50kA, better range and spread.

Volume is fine as 50kA because it gives more control at the beginning of the dial, but then jumps at the end. 50kB works more evenly throughout, but gets loud quickly. So maybe it's up to taste.

What irritates me is that the pedal is incredibly loud. Drive does not only thicken but increases volume a lot.

What is also interesting is that 2n3904 and 2n2222a in place of the MP38A do not give much or any distortion at all. 2n5088 give a bit more but the low gain germaniums give the most drive. Maybe this is similar to how fuzz faces bias differently when Si transistors are used.

However, have a look at C14 in the old one: there is written "100k = 10pF", other schematic has this value as 100pF, don't know where that came from.
Maybe the solution the capacitor has 100k written on it, meaning 100,000 (pF) = 100nF? [...] if you have the unit in front of you, how could you have to guess?

I'm guessing it is one of those orange ones and in reality 100K = 10pF and the person tracing it has not seen one of those before.

[Seiche]

I don't think a 470k is needed.

So the 470k is definitely on the gutshot as Manfred mentioned (see below) and resistor count of 18 confirms this. The GND layer on this seems to be on the front and Vref on the back. The back gutshot confirms this resistor is going to Q3's gate.

I've included the 470k in my build and while there is not much difference in volume (it's still too loud, how would you propose to reduce gain?), there is more treble and clarity, but I cannot thoroughly test this right now, because my neighbours have already knocked twice

The change is subtle and should probably a/b tested with a switch. It's not as big a difference as Manfred's simulation suggested.

Voltages for Q3:
D= 5.31
S= 1.72
G= 0

[Manfred]

I'm guessing it is one of those orange ones and in reality 100K = 10pF and the person tracing it has not seen one of those before.

IMO K means the rated voltage class.
The 2N1308 transistor got a much smaller collector-base capacitance than e.g. the OC44 type has.
I guess that the 100pF capacitor in parallel to the collector-base path shall simulate this higher capacitance.

Have you heard differences using each capacitor value?

I going to order some 2N5952s for making real trials.

It is good that the topic continued, I lost the track of my order, the transistor are not yet delivered.

[Seiche]

I'm not using 2N1308, I'm using MP38A, I don't know if that makes a difference, the MP38A are very low leakage.

I'm pretty sure 100 = 10*10^0=10pF.

[Manfred]

I'm not using 2N1308, I'm using MP38A, I don't know if that makes a difference, the MP38A are very low leakage.

The 2N1308 has a low leakage and a small collector-base capacitance too.
I had Rangemaster in the brain, but I believe that it is somewhat out of place here.

I'm pretty sure 100 = 10*10^0=10pF.

You are right, for sure.

Also Manfred is taking the output from the source, whereas the original circuit has the 1uF and output connected to the drain of Q3. Manfred, could you repeat your simulation for both? I don't think a 470k is needed.

Sorry, I had that overlooked that, but I think it is solved.

[Manfred]

I've included the 470k in my build and while there is not much difference in volume (it's still too loud, how would you propose to reduce gain?), there is more treble and clarity

I will think about.

[Seiche]

Sorry, I had that overlooked that, but I think it is solved.

Maybe it would be of help to see the real output curves.

I suspect the output resistor in parallel with the 8.2k that is in line with the 22uF cap is in reality 20k/22k. This would give an equivalent resistance of 5.8k making the effect much more manageable in volume.

[Manfred]

Also Manfred is taking the output from the source, whereas the original circuit has the 1uF and output connected to the drain of Q3. Manfred, could you repeat your simulation for both? I don't think a 470k is needed.

I now understand what you mean.
Here you are.

[Manfred]

Also Manfred is taking the output from the source, whereas the original circuit has the 1uF and output connected to the drain of Q3. Manfred, could you repeat your simulation for both? I don't think a 470k is needed.

I now understand what you mean.
Here you are.

WithoutAndWithRg470kLoadOnDrain.jpg

(it's still too loud, how would you propose to reduce gain?)

Increase the 2k resistor to decrease the gain of the output stage.

OutputStageDetail.jpg

[Seiche]

hm sounds better without the 470k to me

I have also upped the 2k to 20k, but I think it should be increased even more (maybe also the 8k2)

[Belanger88]

hm sounds better without the 470k to me

I have also upped the 2k to 20k, but I think it should be increased even more (maybe also the 8k2)

That wouldn’t reduce the gain much more, you’d be better off increasing the 2k resistor on q2’s emitter as well

[marshmellow]

Yes, at that point you can just take out the cap and resistor completely. To further reduce gain you can put in a series resistor with the 22µ on Q2 emitter or take it out as well. Like Belanger said increase the emitter resistor or increase the 47k in series with the potentiometer, which limits the gain range.