0

BCbox's VCA "Click be Gone" Mod
Moderators: altitude, adafruit_support_bill, adafruit, hamburgers, phono

Please be positive and constructive with your questions and comments.

Re: BCbox's VCA "Click be Gone" Mod

by crustypaul on Tue Aug 17, 2010 8:36 am

Altitude wrote:Here is the other thing that came to mind (and I may very well be talking out my ass here): the 4066. After building several other Roland clones from the same era (9090 and MB808) I have noticed that all 4000 series CMOS parts are not created equal. )


I just tried 4066's made by NXP, ST and TI, and there was no difference that i could hear.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

Re: BCbox's VCA "Click be Gone" Mod

by guest on Tue Aug 17, 2010 2:26 pm

another possibility
is q1 through q4
this is a current mirror
to replicate the ba662 input stage

a mismatch in betas
would make the current profile much different
guest
 
Posts: 3155
Joined: Fri Feb 17, 2006 5:35 am

Re: BCbox's VCA "Click be Gone" Mod

by altitude on Tue Aug 17, 2010 2:41 pm

guest wrote:another possibility
is q1 through q4
this is a current mirror
to replicate the ba662 input stage

a mismatch in betas
would make the current profile much different


So match them as close as possible?

altitude
 
Posts: 995
Joined: Wed May 11, 2005 5:17 pm

Re: BCbox's VCA "Click be Gone" Mod

by guest on Tue Aug 17, 2010 2:44 pm

all 4 dont need to be matched
but q1 and q2 should be close to each other
and q3 and q4 should be close to each other

ultimately a better current mirror would be nice
guest
 
Posts: 3155
Joined: Fri Feb 17, 2006 5:35 am

Re: BCbox's VCA "Click be Gone" Mod

by crustypaul on Tue Aug 17, 2010 3:53 pm

well, i've just swapped Q1 & Q2 and Q3 &Q4 for exactly matching pairs, and if anything the clicking has got a bit worse :(
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

Re: BCbox's VCA "Click be Gone" Mod

by guest on Tue Aug 17, 2010 4:06 pm

does the click change any
if you turn the accent all the way down

you can try increasing r132 to 220ohms
maybe 470ohms if that doesnt work
but i dont expect that to change much
guest
 
Posts: 3155
Joined: Fri Feb 17, 2006 5:35 am

Re: BCbox's VCA "Click be Gone" Mod

by crustypaul on Tue Aug 17, 2010 4:24 pm

The accent setting has no effect on the click, neither does changing R132.

If we can't find whats causing the problem then I think what we need here is just a way of slowing down the attack by a few milliseconds without also changing the Decay. Sticking a 68nf cap between the juction of R131 & R133 and ground slows the attack enough to remove the real top end of the clicking, which makes it usable, but it also seems to lengthen the decay.

i'm thinking there must be something fundamentally wrong with the x0xb0x design that is vulnerable to component tolerances or transistor beta differences if so many people have complained of excessive clicking that can't be trimmed out.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

Re: BCbox's VCA "Click be Gone" Mod

by guest on Tue Aug 17, 2010 4:32 pm

it sounds like the component that is the culprit
is q36
i wonder if that mysterious 22ohm resistor that is missing
was put in for 303s that had too much click
is c72 well soldered

you could also try swapping out q36
for a low beta transistor

you could try the attack mod you mentioned
and short out d35
and see what happens
guest
 
Posts: 3155
Joined: Fri Feb 17, 2006 5:35 am

Re: BCbox's VCA "Click be Gone" Mod

by bcbox on Tue Aug 17, 2010 6:33 pm

I haven't noticed this difference in clicking between my 303's and the myriad of x0xb0x's I've worked on.
I'm reading with interest though, but quite, as I don't have anything to add at this point. I will pay attention to this though as well and if I can find anything while working on x0x's.
bcbox
 
Posts: 587
Joined: Thu May 25, 2006 1:30 am

Re: BCbox's VCA "Click be Gone" Mod

by altitude on Tue Aug 17, 2010 7:32 pm

It's a good discussion to have, my first one from the kit clicked pretty severely but all the new ones (self sourced) are not nearly as bad as that one was.. I am using 945s/733AP/All metal film resistors (if that counts for anything)

altitude
 
Posts: 995
Joined: Wed May 11, 2005 5:17 pm

Re: BCbox's VCA "Click be Gone" Mod

by crustypaul on Tue Aug 17, 2010 7:49 pm

Altitude wrote: I am using 945s/733AP/All metal film resistors (if that counts for anything)


That could well be the answer. Using 1% metal film resistors might well prevent a build up of 5% tolerances that could happen randomly using carbon ones.

i haven't managed to track down the fundamental issue so what i've done is more or less disguised it using 86nF between the juction of R131 & R133 and ground, and on 'guest's advice i've also shorted D35. Both of these mods are connected to a DPST switch so i can switch them both in or out at the same time. This doesn't remove the click completely but it does take off the horrible sharp top end of it so that it now sounds like its part of the attack of the note rather than something not working quite right. It does seem to affect the decay very very slightly but it could just be me hearing things that aren't there. 100nF also works alright but anything more than that and things don't sound quite right.

I'll try and get some samples up tomorrow of the mod switched in and out.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

Re: BCbox's VCA "Click be Gone" Mod

by antto on Wed Aug 18, 2010 12:43 am

yeah, the R147 (missing resistor) looks like it was involved with the Gate signal..

don't forget there are 303s with the exactly same VCA envelope attack problem, the same clicking
what i suggest is if someone can find where the envelope turns into what it is (Gate signal triggers an exponent, probably a capacitor that looks like a HP-filter with very long time, about 5 seconds and also it discharges when Gate signal goes Low)
so, my suggestion is - let's try to hook up an additional capacitor there (connected to ground) to take off some of the attack
this will eventually also make a small "tail" when the Gate goes Low and envelope drops, tho, i'm not sure how "bad" it will be

can anyone figure out where the VCA envelope is? it has to be some capacitor i think
We are here too: irc.freenode.net >>> #x0xb0x
..:: c0nb0x v1.00 ::.. http://forums.adafruit.com/viewtopic.php?f=7&t=29282 (new control app - win/osx)
VCO Tuning: http://forums.adafruit.com/viewtopic.php?f=12&t=24759 (do it the EASY way)

antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: BCbox's VCA "Click be Gone" Mod

by crustypaul on Wed Aug 18, 2010 8:39 am

Heres 2 patterns each with 4 bars of normal clicking, and then with 4 bars with the 86nF and shorted diode mod switched in. As you can hear, the difference isn't astounding but it does remove the nasty top end click and makes the pop sound more like its part of the note.

http://soundcloud.com/crustypaul/x0xb0x-clicks-test1

I can't really tell if it has any effect on the decay, which is probably a good thing. Its certainly a decent fix until someone uncovers the root of this issue.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

Re: BCbox's VCA "Click be Gone" Mod

by altitude on Wed Aug 18, 2010 9:37 am

crustypaul wrote:Heres 2 patterns each with 4 bars of normal clicking, and then with 4 bars with the 86nF and shorted diode mod switched in. As you can hear, the difference isn't astounding but it does remove the nasty top end click and makes the pop sound more like its part of the note.

http://soundcloud.com/crustypaul/x0xb0x-clicks-test1

I can't really tell if it has any effect on the decay, which is probably a good thing. Its certainly a decent fix until someone uncovers the root of this issue.



Very cool, looks like a step in the right direction.

Here is another observation/musing (which again may or may not mean anything). The 303 service docs call out 4 different types of diodes, one of which is a germanium type. I wonder if that can have a impact since it would have a lower forward voltage drop than the silicon based 4148. Since part of this patch is shorting out D35 (resulting in a 0 V drop) could the characteristics of the diode be coming into play?

altitude
 
Posts: 995
Joined: Wed May 11, 2005 5:17 pm

Re: BCbox's VCA "Click be Gone" Mod

by antto on Wed Aug 18, 2010 11:37 am

Altitude wrote:
crustypaul wrote:Heres 2 patterns each with 4 bars of normal clicking, and then with 4 bars with the 86nF and shorted diode mod switched in. As you can hear, the difference isn't astounding but it does remove the nasty top end click and makes the pop sound more like its part of the note.

http://soundcloud.com/crustypaul/x0xb0x-clicks-test1

I can't really tell if it has any effect on the decay, which is probably a good thing. Its certainly a decent fix until someone uncovers the root of this issue.



Very cool, looks like a step in the right direction.

Here is another observation/musing (which again may or may not mean anything). The 303 service docs call out 4 different types of diodes, one of which is a germanium type. I wonder if that can have a impact since it would have a lower forward voltage drop than the silicon based 4148. Since part of this patch is shorting out D35 (resulting in a 0 V drop) could the characteristics of the diode be coming into play?


yeah, the "curve" of the attack kinda reminds me of how a diode responses, as it looks like (x^2)
while it should look like a step filtered by a LP filter (exponential going up, smoother and smoother till it goes near the desired value)

crustypaul: maybe try slightly lower value for the cap (erm to decrease the filter freq more) ... maybe 47nF or something
We are here too: irc.freenode.net >>> #x0xb0x
..:: c0nb0x v1.00 ::.. http://forums.adafruit.com/viewtopic.php?f=7&t=29282 (new control app - win/osx)
VCO Tuning: http://forums.adafruit.com/viewtopic.php?f=12&t=24759 (do it the EASY way)

antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Please be positive and constructive with your questions and comments.