Component C29 is backwards on all TB-303's... (more inside)

x0x0x0x0x0x

Moderators: altitude, adafruit_support_bill, adafruit, phono, hamburgers

Please be positive and constructive with your questions and comments.
Brassteacher
 
Posts: 248
Joined: Fri Oct 08, 2010 4:49 am

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by Brassteacher »

bcbox wrote:
And damn, that's quite a strip down job you are doing on that 303! :shock: What on earth is wrong with it that requires all that work?
I didn't accurately describe exactly what I was doing. It was a messed up unit, I was fixing the VCF and VCO and replacing all the electrolytic and tantalum caps. It was a good opportunity to remove all the components from the PCB and get a good scan of it so I could accurately clone the PCB. Much easier than working from the layouts in the service manual.
Hah! I'm glad to know I'm not the only one that does this! I've done the Moog MG-1, even though it had to be scanned in two sections. It's already been etched, and I'm doing a similar thing with it that I am with my x0x build, comparing more available modern components against the originals. At least in this case, I have the original synth to compare results. I'm also in the drawing phase of the Paia Fatman board, making it so that I can add features via a smaller daughterboard, as well as making some things a little larger and re-arranging them so that it'll be easier to etch at home. The Fatman also uses some components that are no longer available, so I'll probably re-route some things to use parts that are still current.

One of the cool things about the drawing software I use is that you can import a .pdf file as a background image, and draw directly over. Scan image, save as .pdf, import, and you're ready to clone, then hack, modify, etc. Another cool thing is that if you flood the vacant areas with "copper", it'll actually print the flood out also so that it'll be there when you photo or toner transfer a prototype (I have yet to find a way to get Eagle to do this). The coolest thing though is that's it's cheap! As in, free, unless you want to upgrade to be able to make a board that has more than 700 pins.

User avatar
computer controlled
 
Posts: 714
Joined: Fri May 06, 2005 5:48 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by computer controlled »

Are you cloning the 303 boards for personal use or is this something that will be available??

bcbox
 
Posts: 587
Joined: Thu May 25, 2006 1:30 am

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by bcbox »

Are you cloning the 303 boards for personal use or is this something that will be available??
Eventually it will be available, basically a x0x with 303 layout, and some other exciting things I won't mention until the documentation for the current x0xi0 is completed.

User avatar
computer controlled
 
Posts: 714
Joined: Fri May 06, 2005 5:48 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by computer controlled »

Cool =o]

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by antto »

antto wrote:it seems there is no (or too small) difference in the accent behaviour
the only thing is i needed to boost the resonance (R97) even more to get back to my previous resonance level
this doesn't really sound right
if i set my R97 pot to 10K - it's soo wrong.. i need at least 7.38K there to reach the normal resonance level with the reso knob up

EDIT: here are the recordings (wav files in .rar)
z_C29_original_90-130BPM.rar
z_C29_reversed_90-130BPM.rar
i noticed a few hours later that the resonance was too much, i had to tweak R97 again
i thought it's due to the CAP, it was still hot from the iron when i powered up the x0x (not much, i was touching it with bare hands)
now, today.. i had to tweak R97 again
i just measured the same reference "normal resonance level" as yesterday, and now it's 9.22K

nilsomat
 
Posts: 10
Joined: Wed Dec 01, 2010 12:38 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by nilsomat »

Hello everybody (my first post!)
This is really a very exciting find, i have to say. Funny thing it was only found now (and respect you did, bcbox!), since people have been cloning that little box since around 20 years. i wonder how many clones were built with the mistakes of the schematic...
I am quite happy you found it now, since i am just about to build my first x0xb0x, and i will surely install c29 backwards for the exact reproduction of the original!
In some of the audiofiles the effect is definitely noticeable. In 3saw, 1sqr and 3sqr it apeears to be most prominent, i could only listen to the files with my desktop speakers but i would describe it as an additional high pitched (hihat-like) chirping in the last 2 bars. I like the sound without the chirping more, so it appears it is supposed to be like in the original 303.
If someone has a wave editor you could easily subtract the 2 waveforms to show the difference.
I wonder if you exposed the last secret of the 303, or if there are more?

Greetings,
Nils

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by antto »

nilsomat: hi and welcome
the audiable difference in bcbox's samples is mainly the difference in the resonance level
you can't subtract the two versions in a wave edittor because it's a freerunning VCO.. and even if it wasn't you'd get the difference in the resonance levels plus some bits of drift

nilsomat
 
Posts: 10
Joined: Wed Dec 01, 2010 12:38 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by nilsomat »

Hey antto,
you are right :(
I tried it in a waveeditor, but couldn't sync the two versions closely enough. Well it was fun anyway :)
So the difference is really only more resonance? The waveforms looked quite different in the editor, i wonder if there is anything else being affected by the cap, there seemed to be more noisecomponents in the schematic version. Probably a spectrumplot with adjusted resonance would reveal that?
Would be worth a try.....

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by antto »

in my recording i compensated the resonance level so that it's the same as the reference recording
also, i recorded both samples at 96KHz and compared the envelope and accent curves on a sonogram
this is because i've tuned my filter cutoff as in the TB-303 manual (cutoff centered == ~500Hz)
thus, with cutoff and envmod both all the way up, the filter goes above 30KHz :roll: not to mention when accents hit it too :mrgreen:

in the past years i've been looking at TB-303 recordings alot and i know my job ;P~
in my eye, there isn't any noticable difference, which in your eye equals NO difference, trust me
i will keep C29 this way, i don't mind it

btw, maybe i gotta redo the test now with the cap cooled down

User avatar
tim stinchcombe
 
Posts: 11
Joined: Wed Dec 01, 2010 6:25 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by tim stinchcombe »

Greetings everybody!

What an interesting discovery bcbox! I have only spent an hour or so pondering this, but have run some simulations. It is clear that reversing a polarized electrolytic will not result in its instant destruction, as we are so often led to believe in Youtube videos, and indeed many cap manufacturers databooks and app notes. Coincidentally I recently came across a (more modern) synth module with a 10u bypass cap on the power supply fitted backwards because the silkscreen was wrong - I removed the cap and checked its value (it wasn't a million miles away from 10u), and had it running at 12V, reversed, on a bench power supply for several days (as it would have been in the module for heavens knows how long as it was second-hand) and it showed absolutely no inclination to explode or otherwise misbehave. Thus one might expect that the cap value itself is somewhat lower than its nominal value, but running some initial AC analyses in SPICE suggest that this will have no impact on the frequency response curves whatsoever - I'll try and do some transient analyses in the next evening or two, as they might show something up (because of the way SPICE does AC analysis, some small facet of the DC biasing could get overlooked...).

Thus I am kind of wondering just how we need to model a 'backward cap' in order to work out what its effect on that part of the circuit is, as even my indelicate ears could pick-up on the sound differences in some of the samples. My guess at the moment would be that it is subtly altering the DC biasing down the chain R116/R112/R109, and hence impacting the gain of that diff pair amplifier - it would be interesting to see some scope traces of the voltage at C29 when both ways round, in order to see if there is any material difference...? I don't own a TB-303, but I might blow the dust of my ASol TBX-303 clone, and try taking some readings from it, as it will hopefully exhibit similar characteristics. Sounds like a mini-project for the weekend!

Tim

bcbox
 
Posts: 587
Joined: Thu May 25, 2006 1:30 am

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by bcbox »

it would be interesting to see some scope traces of the voltage at C29 when both ways round
this can be done if I find a few minutes.

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by antto »

nice to see you here Tim :mrgreen:

guest
 
Posts: 3155
Joined: Fri Feb 17, 2006 5:35 am

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by guest »

the effects of using polarized caps backwards
has always been a question of mine
because sometimes you just cant get around it

so i took this recent discovery
as an excuse to do a full analysis

i took an electrolytic cap
and ran an ac signal through it at varying dc bias voltages
i measured its dc current draw and capacitance
along with noise and distortion measurements

there is a slight change in capacitance with voltage
about 2% or so
but this could also be due to heating effects
as there is a large increase in dc current draw

there are basically cutoff voltages
both forward and reverse
beyond which things start getting worse
and below these voltages everything is fine

for the 22uF/16v cap i tested
this was -8v and +28v
after these levels
noise increased and capacitance shifted
and with an increase of a few more volts
they basically ceased to function as capacitors

elmacaco
 
Posts: 31
Joined: Tue Mar 07, 2006 3:43 am

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by elmacaco »

I just reversed C29 and tested last night and it sounds much better, much more like a real 303 especially evident in the accented notes, but not limited to that.

Is this something that we could do to other synths?

User avatar
antto
 
Posts: 1636
Joined: Thu Apr 15, 2010 3:21 pm

Re: Component C29 is backwards on all TB-303's... (more inside)

Post by antto »

since C29 is backwards in all 303s, then it should be like so on all x0xb0xes
as for other synths: if it ain't broken.. don't fix it :wink:

Locked
Please be positive and constructive with your questions and comments.

Return to “General x0xing”