one more "CGS 35 help please" thread, sorry...

[LucienR]

Hi!
I just built a version of this filter, with my own pcb design. I'm new to this, and as it is right now i just can't tame the resonance. It basically acts as a VCO, the self oscillation is incredibly loud, no matter how i tweak the trimpots and the resonance pot. The input can hardly be heard, and acts more like a CV input than an actual input. I tested the input mixer stage, works as it should. I built it with a few component changes, due to the fact that i'm completely broke and can't place an order right now, but i guess i should still be able to solve this with some generous help from you guys...
I used the eurorack version as a model
https://static.miraheze.org/sdiywikiwiki/5/5d/Cgs_schem_cgs35euro_steiner.gif
and the schematic I draw on Kicad is attached (a main pcb, one for the input mixer stage and one for the CV in and output. i added a picture just to give you an idea of the whole design).


So here are the changes i made due to unavailable components :

- The three 2n2 caps are replaced are 3n9. I guess it should just lower the frequency center, without altering anything else, right?

- The 2k2 trimpot is 1k

- The CV reject trimpot is 50k

- The transistors are BCXXX instead of 2n. Yes, they are correctly oriented (I had a hard time soldering them, as on the Kicad default footprint the pins are incredibly close, i but i somehow found a way around as you'll see if you zoom on the attached picture. It's ugly, but connections are good.)

- I'm using a 12v power supply, so 8 diodes instead of 12.

All the resistor values are good.

One thing i noticed : on my first attempt, i forgot the 1k resistor between the two 2n2 caps (3n9 for me) and the resonance pot. Surprisingly, it was far from ideal but still worked better than it does now, it couldn't self oscillate but still sounded like a filter (for the BP in at least).

Any ideas?
Thanks!

---

Capture d’écran 2022-10-01 à 14.56.25.png
Description:
Filesize:	305.01 KB
Viewed:	109 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

Capture d’écran 2022-10-01 à 14.57.10.png
Description:
Filesize:	151.52 KB
Viewed:	120 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

Capture d’écran 2022-10-01 à 14.57.44.png
Description:
Filesize:	43.57 KB
Viewed:	107 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

IMG20221001150619.jpg
Description:
Filesize:	4.07 MB
Viewed:	108 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[blue hell]

[LucienR]

Ok thanks a lot for the advice. Seems that it was the way to go. I first tried 2k2, it was still a bit uncontrollable, I went up to 10k, which is much higher than suggested, so maybe there's still something wrong with my circuit? but it works much better now. I'm not familiar at all with the process of calibration, i'll do it by ear.
Thank you!

[blue hell]