GCI Apostle : 2V variation in transistors biasing depending on the power supply.

[eh la bas ma]

Hello !

I recently built the GCI Apostle. I found the gain texture very muddy so I tried to experiment with the 2N5457 biasing.

This circuit use a charge pump going up to 36V or so, and transistors should be biased around 18V.

https://kurtballou.com/docs/N.E.W.%20Apostle%20Build%20guide.pdf

I noticed a 2V difference if i use an EHX power supply (9.6V, 200mA) or an ENGL power tap :

https://www.thomann.de/gb/engl_power_tap.htm?shp=eyJjb3VudHJ5IjoiZ2IiLCJjdXJyZW5jeSI6NCwibGFuZ3VhZ2UiOjJ9&reload=1

If I connect the EHX power supply and probe the 9V DC, I read 9.86V. At R27, the end of the charge pump circuit i read 37.15V.
I bias all transistors to 18V, as instructed, with the EHX.

Then I use the Engl power supply with the 500mA  or 100 mA cable, I read 9.41V on DC jack, 35.32 at R27, and  all transistors have droped down around 16.40 V ...

Is it to be expected ?

I mean there is 0.50V difference between the two power supplies, can it become almost 2V at the end of the charge pump ?

What is the best option here, set it to 18V with the Engl and it will increase up to 20V with a slightly more powerful power supply, or the opposite : let it decrease to 16.20 V with the Engl ?

I see on datasheet that max voltage for these transistor is 25V :

https://www.onsemi.com/pdf/datasheet/2n5457-d.pdf

When i first biased the transistors, I remember reading 36V on some of them, before I turned down the trimmer.
On datasheet, I read : "Stresses exceeding Maximum Ratings may damage the device."

36V even briefly, could really damage it ? How would I notice some damage ? Would they sound a bit more muddy ?

I noticed two of them (Q3, Q4)are showing some oscillations when the effect and the gain are engaged. Q4 oscillates between 8 and 18V, Q3 oscillates between 20 and 23V. Both are back to 18V when I switch off or turn down the gain.

Q1 isn't oscillating. I tried to swap it with Q3 and Q4, and I get the same oscillating results.

Should I worry about this ?

Thanks for your help, all suggestions and observations are welcome !

[antonis]

I mean there is 0.50V difference between the two power supplies, can it become almost 2V at the end of the charge pump ?

By what number is 9 multiplied to give 36..??
Find it and multiply 500mV by that number..

[eh la bas ma]

Thanks for your reply.

Of course, if you put it like that... I would have thought the charge pump was regulated somehow, to avoid such voltage intervals using various power supplies.

So if I understand charge pump circuits correctly : the more powerful it is, the greater the offset wil appear, using power supply with slight differences in voltage. That's quite a downside...

I just tried some smd MMBF5457, and i get the same sound results and same voltage oscillations, so I guess I didn't damage any transistor.

I opted for biasing Q3 during gain and signal activation, to get an oscillation narrowed between 15 and 18V. It feels a bit more precise.

I wonder if some clipping diodes could help tighten the gain texture ?

I am not very familiar with the process of adding clipping diodes in a circuit that doesn't have any...

What would be the best location ? Across the volume pot pin 1 and 3 ? Or maybe a couple of 1n4148, anti parallel between Q3 and Q4, one side connected to the signal path, the other side to ground ?

[FiveseveN]

tighten the gain texture

What does that mean?

If you want less interstage bass, i.e. less like fuzz and more like modern metal, start by making the source resistor bypass caps smaller. Look at an SLO/Recto circuit for clues.

[antonis]

So if I understand charge pump circuits correctly : the more powerful it is, the greater the offset wil appear, using power supply with slight differences in voltage. That's quite a downside...

I presume you mean "voltage multiplication" instead of "power"..
I can't see any downside here 'cause it isn't some kind of "regulator" - just a fair voltage multipler..
(after all, it's just a switched capacitor(s) DC-DC converter..)

P.S.
IMHO, those 100pF Drain decoupling caps "neutralize" Drain-Gate inter-electrode capacitance multiplication (Miller effect) which effect, for such a high gain circuits, it might be needed..

@ Justin: Try to desolder those caps (just one leg) from respective oscillating transistors to see what happens..