Madbean Sharkfin (13700 filter/sample hold) help

[jez79]

Hi folks, looking for help with a sharkfin filter.

http://www.madbeanpedals.com/projects/Sharkfin/Sharkfin_2015.pdf

I've built the circuit on perfboard and with a little initial troubleshooting (forgot to connect a gnd hub) it mostly functions as it should. The envelope controls work fine as does the noise generator and lfo.
However, it seems there is something wrong with the filter sweep. With filter controls at minimum and very low volume input signal, the filter sweep is already rather high, so 'normal' signal sounds like going through an open wah. Engaging the controls to move the filter through it's sweep carries it from this midpoint ending way too high.
Seems like the voltage control filter is biased too high?

I've checked voltage on all power pins (even though there is no problem with the 741 sections) and the 13700 and associated bits (fet and bias power) are getting correct values.

I've audio probed the signal, and that is normal also.
The circuit is functioning fine except for the filter sweep starting way too high and never dropping below a partially open wah sound. I'm losing all the low end, and the filter range is limited.

I conclude that something needs to be done around the 3906 section feeding the 13700's pins 1 and 16. The only substitution i've made is 220n instead of 150n parrallel with 2.7M to ground.
I'm not sure i fully understand the minutia of the functioning of the 3906 beyond that it sends voltage control to the 13700 from signal from envelope or noise/lfo.
I've tried swapping the 13700 chip, but that was before discovering the missing ground hub. I should try swapping again.
Could a bad 13700 be responsible? If not, how do i lower the filter default so it's not partially open all the time.
Many thanks for any suggestions!

[jez79]

Swapped 13700 from a different batch with identical results.
I'll take voltage readings from the fet and pnp near the 13700 and report.
Any ideas appreciated!

[PRR]

> section feeding the 13700's pins 1 and 16.

Voltages there, and around the "mess" of pots and switches in Q3 Emitter, might be clues.

Basic check: C3 C4, and R8 R12, are the right values? No decimal slip? (I sure would mix 2.2K for 22K and be unable to see my goof.)

[jez79]

Here are some voltage values to help solve this puzzle
Component values were double checked and seem fine...

Voltages with no input signal (unconnected input, not grounded)

Q2
D 8.06
G -0.11
S 0.02

Q3
C -6.59
B 0
E 0

13700
1 -6.83
2 0
3 0
4 0
5 -0.11
6 -7.98
7 0
8 -1.13
9 0 (n.c.)
10 0 (n.c.)
11 8.06
12 0
13 0
14 0
15 0
16 -6.83

[PRR]

DC voltages look reasonable.

> Q3
> E 0

Does this change as you turn knobs or play?