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 altitude on Sun Aug 15, 2010 12:16 pm

antto wrote:
Luap wrote:
antto wrote:
but i was thinking we're talking about the clicking of the notes, not DC problems


The thread is discussing both. However, the solution to the DC offset thing is now known, obviously. Because Altitude kindly posted the solution in the first post here.
But the note/gate click issue currently has no proven fix yet..


i suggested adding a LP filter wherever the Gate signal is (unfortunately i can't tell "where" )



The gate goes through R146 to the base of Q37

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

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

by antto on Sun Aug 15, 2010 2:26 pm

nope, this is the gate signal right after it comes out of the CPU
i didn't meant that one

first there is this GateCV from the CPU which triggers a RC-decay envelope somewhere
then this is used to amplify the filter output
a long Gate coming out of the CPU will result in a decaying note
the LP filter should filter this decaying envelope

i just tryied to hook a 1000uF capacitor from R146 to ground and the result is the notes don't retrigger and actually become a single long note (which decays to silence)
so, this is not the right spot..
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 Mon Aug 16, 2010 12:43 pm

I'm just trying to finish off my 4th x0xb0x and i'm having a nightmare with this clicking. The first two i built were both adafruit kits and one was virtually silent while the other had what up until now i thought was too much clicking. The third machine was self sourced and about the same as the second adafruit one clickwise, but this one i'm working on now is just ridiculous.

I've removed the DC offset clicking using the trimmer pot and tried every solution suggested on this thread, but theres still an annoying amount of pop and click at the start of notes with the filter closed. What is even more annoying is that i know the x0xb0x CAN be silent, as i've built a silent one before with exactly the same transistors and a BA6110. The only mild improvement i've got is by changing Q32 for a 945 rather than a 536, but that just meant the clicking now happens more in waves like its controlled by an LFO rather than happening on every single note as it was before.

I'm starting to suspect i've just got an unlucky line up of tolerances somewhere.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

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

by antto on Mon Aug 16, 2010 1:32 pm

crustypaul wrote:...
The only mild improvement i've got is by changing Q32 for a 945 rather than a 536, but that just meant the clicking now happens more in waves like its controlled by an LFO rather than happening on every single note as it was before.

NO, if it sounds like the clicking is LFO'ed - this is normal, because it clicks depending on the VCO
depending on where the note starts opposed to the phase of the oscillator, if the waveform there is high (erm, away from zero) the click will be more pronounced, while when the waveform is somewhere near zero - it will be less noticable
but the point is - it always clicks because of the VCA Envelope
it doesn't matter if it sounds like it clicks on every note or just in "some" notes (aka like LFO'ed) or just when the filter is closed - it is there all the time, it's because of the Envelope, it's too sharp!
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 Mon Aug 16, 2010 1:45 pm

antto wrote:if the waveform there is high (erm, away from zero) the click will be more pronounced, while when the waveform is somewhere near zero - it will be less noticable


Yeah, i got that much.

Its weird that some x0xb0x's click while other don't when using exactly the same components though. That would suggest to me that the VCA attack going too fast on some machines is down to a bad lineup of resistor and cap tolerances or something to do with the varying beta of the transisors. There doesn't seem to be anything else that could be that random.

On one of my x0xb0x's i've got a switch to enable a variable VCA attack knob taken from the ADSR mod in the wiki. This can remove the VCA clicking completely, but also changes the shape of the envelope decay in an annoying way.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

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

by Luap on Mon Aug 16, 2010 2:04 pm

crustypaul, Do the clicking x0xb0xes you made all use 536 transistors? The non clicking ones use 945's or something else perhaps? (There are 4 or more types or transistor that Roland used here).
Im still wondering if there is something about 536's which exaggerates this problem, because b0xes built with 945's don't appear to suffer from it so much.

I also wonder if when Roland built 303's and used these different transistors, if maybe they changed the spec of other components to accommodate them. Im sure I read there was 2 or more revisions of the PCB layout. But i've not seen any documentation of what the changes actually were. But for all we know, it could relate to this..
Luap
 
Posts: 363
Joined: Wed Jul 08, 2009 7:10 pm
Location: London'ish. UK

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

by crustypaul on Mon Aug 16, 2010 2:26 pm

Every x0xb0x i've made has used 536's but as i say, one of those was more or less silent so i'm thinking the actual type of transistor doesn't appear to make much difference. Unfortunately the only silent one i've made was built for a friend so i don't have it here to take any measurements.

Having said that, i guess the next logical step would be to swap out every 536 in the VCA of my latest machine for a 945 and see if anything changes after each swap. Doesn't sound like much fun :roll:
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

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

by Luap on Mon Aug 16, 2010 2:45 pm

crustypaul wrote:Having said that, i guess the next logical step would be to swap out every 536 in the VCA of my latest machine for a 945 and see if anything changes after each swap. Doesn't sound like much fun :roll:


Yes, that is basically where I am at with my b0x. There are 24 of those transistors! :shock: But I expect rather less than half of them play any possible part here. I did try changing Q32 and Q37, and didn't really notice any change. I only swapped them 1 at a time though, then put the 536 back in. They are socketed now, so i'll try both at once, but im not expecting miracles yet :|

Ironically, it was largely little problems like this that made me socket every single transistor in my second x0xb0x. To make life easier if there were any issues. But typically that x0xb0x worked and sounded perfect first time and didn't need any components changing at all!
Luap
 
Posts: 363
Joined: Wed Jul 08, 2009 7:10 pm
Location: London'ish. UK

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

by altitude on Mon Aug 16, 2010 3:49 pm

Has any1 tried what guest suggested?

guest wrote:there are a number of places
this could be coming from
and beta differences in transistors might be it
i dont have my x0x currently
but when next i do i will compare betas
with my 303 and see whats up

the main culprit would be q32
although q38 and q31 also have an effect

you can try increasing r134 or c41 to see if it helps
which is actually a 72hz lowpass filter

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

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

by crustypaul on Mon Aug 16, 2010 5:49 pm

Changing Q32 had a very slight effect but nothing that really made much difference. Changing Q31 and Q38 did nothing.

Increasing R134 and/or C41 also had no effect.

I've just swapped the 536's at Q3,Q4,Q32,Q35,Q37,Q40 and Q41 with 945's, which is all the 536's in the VCA and envelope sections, and theres no difference whatsoever.

I don't think this is down to transistors
Last edited by crustypaul on Mon Aug 16, 2010 7:12 pm, edited 1 time in total.
crustypaul
 
Posts: 27
Joined: Fri Sep 26, 2008 5:16 pm

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

by rv0 on Mon Aug 16, 2010 6:44 pm

one of my 303's has the clicking problem too, it has the "C1685" transistors which arent used in a lot of models
another one with 536 transistors doesnt click
it never really bothered me, although i always prefered the one with 536 in it

rv0
 
Posts: 394
Joined: Tue Jul 14, 2009 4:50 pm

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

by altitude on Mon Aug 16, 2010 9:11 pm

I'm going over the x0x and TB-303 schems (love how he laid the x0x out identically :D ) and one thing that immediately jumps out at me is the fact that the x0x does not have a R147 which connects to Q36. What does it do and why was it omitted? No clue.

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. Yes they do the same thing but different brands can have different results. This was especially true with my 9090 where a 4013B made by Fairchild was causing weird distortion on the sound of an instrument, this was the case for ALL Fairchild 4013s that I tried. I replaced it with a TI part and the problem was solved. Both are 4013B CMOS parts. Same thing for the MB808 where the brand of the 4548 hex schmitt trigger had a impact on the sound. So seeing that the 4066 does a fair amount of work here and that it is also a Fairchild part, next time I'll order some ON parts which are incidentally still produced and called out in the original 303 parts list (MC14066B)

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

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

by rv0 on Mon Aug 16, 2010 9:36 pm

Altitude wrote:I'm going over the x0x and TB-303 schems (love how he laid the x0x out identically :D ) and one thing that immediately jumps out at me is the fact that the x0x does not have a R147 which connects to Q36. What does it do and why was it omitted? No clue.

error in 303 schematic
r147 does not exist

rv0
 
Posts: 394
Joined: Tue Jul 14, 2009 4:50 pm

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

by Luap on Mon Aug 16, 2010 9:54 pm

darffader wrote:error in 303 schematic
r147 does not exist


Did you google it too? :lol:
viewtopic.php?f=7&t=2568
If I read Guest correctly, it is marked on the board, but replaced with a 'short' (link?).

This is what I was suggesting earlier though. Could it be that some 303's did have it, and some didn't?

As for the 4066.. No idea.. but somehow I ended up with a pile of the bloody things that I have no idea what i'll do with!
Luap
 
Posts: 363
Joined: Wed Jul 08, 2009 7:10 pm
Location: London'ish. UK

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

by rv0 on Mon Aug 16, 2010 10:31 pm

Luap wrote:
darffader wrote:error in 303 schematic
r147 does not exist


Did you google it too? :lol:
viewtopic.php?f=7&t=2568
If I read Guest correctly, it is marked on the board, but replaced with a 'short' (link?).


heh
actually i had copypasted that info in a txt this morning while looking for something else involving the tuning circuit. i was googling around for IC10, as I didn't know exactly what it did.
after asking around a bit, it's apparently a buffer: in original 303 design, ic9 cant deliver enough current for the upper part of the R-2R network and would collapse giving an error on the cv.
x0xb0x doesn't have this problem as the part is omited (due to a better part for ic9 being used?)

rv0
 
Posts: 394
Joined: Tue Jul 14, 2009 4:50 pm

Please be positive and constructive with your questions and comments.