Yusynth EMS filter Issues

[matinnin]

i placed an order several months ago for 2N3819 - showed stock - looks like its discontinued now - Central is making a BF244 that looks to be its replacement - but i can get the NTE312 at Frys it appears -

Using the NTE312 will be fine Right?

Thanks,

[LFLab]

You can use several FETs, but VC response of Q works very well with 2n3819 (IIRC one also needs to change a resistor for better VC Q response)

Have you tried ebay? (not from china, but there are some people from europe and US offering them, and those usually really are NOS and real).

[matinnin]

I have many projects and several Yusynth modules under my belt, but this one is giving me fits. not sure why, seems like everyone else didnt have this much trouble.

first: I did not find the 2N3819 (i have now , waiting for them) so built and used that NTE312 - and of course i started with a module in complete oscillation - discovered after the fact the pins are not the same, and corrected this.
then i had something working.
then Res came on way too early - have settled on 560K on R41 for the moment as it gives room, and doesn't self oscillate until almost the top of the pot.

CV res input needs to be attenuated way more

second :GAIN !!! i am still fighting with the output/input i have attempting to reconcile as the thing is extremely over driven, on the scope the inputs start to clip the input wave about a third of the way up on the pot - i have changed some resistors - started comparing with the ARP filter as it turned out perfect and i l LOVE it. it also uses the TL074 so i thought it comparable -
R8 100k (changed several times and still doesnt seem right)
R9 10k (can't seem to find a correct value for this)
R46 56k (this emulating the ARP filter) think i will raise this
R47 56k (this emulating the ARP filter) think i will raise this

Frequency pot also seems to do strange things as you open it up when it gets the end of the pot it seems to "pull back for a hitch" and then return back to opening up only where it fell out ? not sure how to explain - easier to see on scope

summary:

Res is still a question, as i have the 2N3819 on the way and will see when it gets here - i may come down on my 560K - to 500k and they may give me a little more res, just dont want self oscillation too early on the pot

the audio is not right- with the resistors i've changed so far, when you turn either input up and get to about 2/3 you can see the waveform clip out and loose any shape - eventually everything ends up a clipped square wave
cant see it being only an out issue - must be issue mixing at the TL074
maybe some bigger POTS ? for the Inputs and the cv inputs?

thanks

[matinnin]

sequencer running a modulating Res and Freq trying to find that sweet tearing squelchy sound - but alas... it just isnt doing it for me. when i get the 2n3819 in ill test, i like the idea of placing a socket in there and being able to change it out- i have ruined my board desoldering resistors so i'll have to burn a new board , Grrr

---

FUNK U Edit 1_Track 2_Take_1.wav
Description:	Yusynth Diode EMS filter	Download
Filename:	FUNK U Edit 1_Track 2_Take_1.wav
Filesize:	18.47 MB
Downloaded:	682 Time(s)

---

[wackelpeter]

increasing the pot values for resonance or the Input doesn't have any big effect as the resonance pot just works as a voltage divider, so here playing with the value of R41 is the more important factor…

When you say you have a clipping at your Output, does this occur wihtout resonance too?

have you looked at pin 14 of the TL074... with the Signal input pot turned all the way around at max Input you should have the same Level as the Input here…

also worth to look again if R9 and R8 are the correct values… or try to increase R8 or lower R9 a bit…

Also check the resistor values of R46 and R47 and look with a scope if the Output is already clipping before R46... also check all values in the amplifier circuit Q1-Q5 and the resistors around them…

[LFLab]

I did add input attenuators to my build, and usually have them turned up only half way.

[wackelpeter]

i did that too…

Maybe worth a note my final gain stage has a 10k as Input resistor and 20k in the Feedback path…

Just had this VCF on my bench again as with the new Transistor and Diode mathcer from Elektrouwe (Muchas gracias for this brilliant idea and design) i decided to replace all the diodes with those ultra-close matched ones to improve Performance...

still has some flaws, Maybe i chi´me here in, like the resonance Performance being not so great… firstly the typical resonating Sound could be a bit more, it could be more accentuated and it sounds really good only in a few sweet spots and then the gap between resonant sounds and self oscillation is a bit too narrow for my tastes, especially when you consider that the resonance does not much kick in before…

Also at the best possible sounding settings the CV bleedthrough at the Output is higher as expected… when i turn the bleedthrough at the Minimum the filter sounds not as good as before… guess i'll Change the trannies in the differential amplifier section too and look for Closer matched resistors for those which are required...

[matinnin]

Well, i got my 2N3819 - put it in, ....
constant oscillation. Put the 680K back in R41

NO CHANGE

Switched out the 2N3819 - No Change

Swapped out the CA3046 - No change

wow this is starting to become my white Whale.

i think i'm going to have to move over and build the MOOG clone and set this aside for while. probably just build another one completely
This is approaching the most difficult build for some reason, can't really see why.

So, I had a functional filter with the NTE312 and adjusted R41 - ??
these are Fairchild 2N3819

right now i am seeing 0 - -1.24 v on Gate of JFET when turning Res Pot

Grrr.

[matinnin]

well, i put it down, and looked at again tonight and , just dont get it . oscillating constant, i see conflicting data sheets on pinout on these - so i flipped it and i get the same results
i got mad and put a 2M in R41, didnt do anything, jumped out R41, and filter does nothing.

so disappointing ..