Getting a more 303 sounding box..

Discuss x0x construction and related issues

Moderators: altitude, adafruit_support_bill, adafruit, phono, hamburgers

Please be positive and constructive with your questions and comments.
rarara
 
Posts: 169
Joined: Wed Apr 05, 2006 9:47 am

Re: Getting a more 303 sounding box..

Post by rarara »

very interesting - thanks aminoacid 8)

i have 100 733s on the way from HK but suspect they might be a bit duff as you also experienced. not much money i guess...

have reordered some 945's but this time made sure they have the P suffix

just need to perfect my desoldering :roll:

roxxx303
 
Posts: 93
Joined: Tue Nov 08, 2011 11:33 am

Re: Getting a more 303 sounding box..

Post by roxxx303 »

Hi,

in another thread I posted a link to my YouTube-Comparison-video
of tb-303 and x0xb0x:

http://forums.adafruit.com/viewtopic.php?f=7&t=37205

-roxxx303

rarara
 
Posts: 169
Joined: Wed Apr 05, 2006 9:47 am

Re: Getting a more 303 sounding box..

Post by rarara »

OK made an interesting discovery with my meter - C13 would appear to give no reading.

are capacitors often duff or could it be too much heat at build? also, should I check every capacitor in the box?

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

Re: Getting a more 303 sounding box..

Post by antto »

uhm, what reading are you trying to measure?
C13 is connected to ground, and acts as a LPF, it creates the accent "wow"

rarara
 
Posts: 169
Joined: Wed Apr 05, 2006 9:47 am

Re: Getting a more 303 sounding box..

Post by rarara »

dang it - i was getting readings on other caps but think you are right. may swap c13 though as no wow :cry:

rarara
 
Posts: 169
Joined: Wed Apr 05, 2006 9:47 am

Re: Getting a more 303 sounding box..

Post by rarara »

OK, my quest for perfection continues, and still at this stage I am looking for a "wow" sound....

Originally I thought there was a problem with C13 but now think its OK but my voltages for "IC12A thru D24 - R46 - C13" are:

D24 - 0v, 2.76v
R46 - 2.77v, 2.89v
C13 - 2.89v, 0v
IC12A - seems mostly along the lines of the 'analog voltages' posting/eagle file which is on another thread from quango. I can't measure mV's but the only differences are where quango reported 11.73v on a couple of pins, I am getting a bit lower - more in the region of 10.8v

i tried following some traces around these components and VR4 and noticed a few things....

R47 - getting 3.2v one side and 3.79v the other - a bit low?
C15 - 0v and 4.17v (R67 also has the 4.17 - does this feed C15 or vice versa?)
J5 - is this important? it looks like quango gave the voltages for this the wrong way round (?) but ignoring that, where he noted 34mV/0v/5.96v I am seeing 1.91v/0v/5.96v (with J5 on the I/O being the same)
C43 is also showing 2.4v on the negative side and 2.97V on the other

I am fairly positive that with enough effort I can get this thing as I want it now that i am pursuing voltages, but really would appreciate anyones help. (....and have checked that i am getting 5.3v and 11.8v)

3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Getting a more 303 sounding box..

Post by 3phase »

Interestig question..and antos nonxox actually froces me right now to attack the topic once more..

Just removing all the mods and bring the xox back to a clean state..

Some mods that got it more 303 alike i keep or intensivate.

Following problems got eliminated since i got it years ago..

1)
Roland 662 VCA ..selected..they all sound different.. but make a more mushy sound as we know it from the 303.. the newer chip is clearer and snappier. I like it actually..but dont works in facor of the anyway harder xoxbox sound

2) original foil caps in the filter.. The original ones are transparent green drops and these give a finer sound that the ceramic tile style green drops.. these sound indeed harder..
I also checked various other types.. even much better ones than the originals..theese deliver aswell a very fine tone..but extended high frequency response…what dont works in favor of the xox hardness..
So the original transparent green was the best i could find.. i robbed other old roland machines to get them..

3) Old semiconductors.. I got my parts from a trustworthy guy. but..how good are modern batches of discontinued transistors? Are they remakes? fakes? .. I got a big step smoother by sourcing old transistors in old VCR and Hifi equipment. In 100% of the cases the older parts was giving a warmer and smoother sound… Maybe i really had remade versions? I know fake chips.. theese usually dont work very good at all..but on the single transistor it maybe works but sounds harsher?

4) Transistor type? What is the best sounding one? I am still uncertain..except again that old versions sound different than new batches.. that is especially visible on transistors still produced ..sinnce they are to available to get faked… A topic i actually would like to discuss.. My xox is 945 based …but i ve a handfull old 1815/1015 available..according to what i just wrote the old 1815 are brobably better than new 945..even when new 945 are better than new 1815..

That the question i maybe try to answer now…since i really hate that my xox sounds not 303 alike even after all this changes..

5) Resistor type … you need to use the resistors specified by ada fruit.. metal film gives a harder sound.

6) other factors?

internal gate timing?

power supply circuit?

internal HF polution by usb chip and higher takt..or the relative absence of hf polution and noise by the missing dc/dc converter and its 70khz interference?

stray capacitys? the original board is way smaller and has not as much shielding as the xox board..
that is pico farads here and there thru the circuit and areas of potential low level crosstalk fx..even crosstalk on the modulation is thinkable..resulting in slightly different envelope shapes.


This are the 4 suspects i ve left..or what else could harden the sound of the xox so much?

I actually fear that it is a combination of all this aspects and therefore a 100% 303 is reserved for the 303..

But its thinkable that further optimisations are possible.. See the TT bassbot that has so many sgit sounding component choices that it actualy shoudl sound even more shrill and hard as a xox.and it actually does..but has a certain elasticity i miss on the xox..i wonder if this is the impact of the power supple..and a xox with bassbot style original power supply would give a decend result..

Problem..where to get the parts for an original power supply? or shall i slaughter my 303 to isolate its power supply from its circuit to feed the xox with it? actually i dont love ,y xox enough to do this.. maybe i should lend a 606? ;)

User avatar
rv0
 
Posts: 395
Joined: Tue Jul 14, 2009 4:50 pm

Re: Getting a more 303 sounding box..

Post by rv0 »

3phase wrote: 5) Resistor type … you need to use the resistors specified by ada fruit.. metal film gives a harder sound.
I really doubt the sound character is influenced by the type of resistor, that would be getting close to voodoo & magic.. Resistor value, probably yes.
I'm considering temporarily rerouting the PSU of a 303 to a x0x to see if there is any truth in the power supply myth. Not right away though.

User avatar
mmm
 
Posts: 517
Joined: Tue Apr 24, 2012 1:51 am

Re: Getting a more 303 sounding box..

Post by mmm »

darffader wrote: I really doubt the sound character is influenced by the type of resistor, that would be getting close to voodoo & magic.. Resistor value, probably yes.
It's also helpful to keep in mind that a large number of resistors isn't even in the signal path. I don't see how for example the type of a bias resistor influences audio quality.

3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Getting a more 303 sounding box..

Post by 3phase »

cole resitors have more noise..that is a factor that even cuts thru on modulation cv´s..

I ve changed the resistors back in the relevant areas to the specified 5% cole types.. that was a good portion of work

but a huge sound difference.. Ok..you can say thats the 5% difference.. the metal films just had 2%..

but..the sound difference was way to big IMO to just be explained by that..and the xox became much more mushy and smoother sounding.. more like a 303 but more noisy.. a little.. but with the metalfilms there was indeed less noise.. There is an audible factor..and at the time adafruit published the xox cole resitors was already difficult to get..but are explicit requested in the parts list..

As i know now for a reason.

However..the basic tone of my xox is much closer to a real 303 with the resistor change..original vca, filter caps and NOS transistors ( only 50% yet ) but..

it still dont sounds like a 303.. mainly the enveloping i stiff.. and the output still a little sterile..

i wonder if i have an error in the circuit somewhere.. or maybe gate timing effects the elasticity of a sequence. or some original 303 crosstalk is missing?

all 3 factors might be adjustable by hard or software changes..

And the power supply question? The originally 303 sound s actually best on low batteries.. In the 90´s we often intentionally used old batteries for important recordings..

So the condition of the powersupply does fx the sound output of a 303 ..but how?

the batteries get transformed up by a discrete dc/dc converter.. this dc/dc converter is allways in tthe circuit..but has to do less on the external powersuply that provides stabillized 9 V DC..

the batteries just provide 6V DC..and old ones even less..

And this makes the 303 sound better..

So question..what is happening when such a discrete dc dc converter gets a 2 low voltage? is it dropping in voltage? or just in current delivery? or booth?

It sounds like it kepps the voltage..because the tuning stays intact..a least the scaling..

But that is something an electronics expert might comment..

I tried reducing the current delivery of the xox power supply by a resistor ..dont gives the desired fx..

i used a little lamp as resistor..that got a little closer..what gives me the idea that the current limitation of a weak dc dc converter is somehow a dynamical process and not a staic on like current limitation thru a resistor,

But.. i am sound engineer and no electronics engineer. And thats pretty deep stuff.

The logical way would be to run a xox from a original 303 power supply..maybe i do that to my 303 for a test to get this validated once and for all..

cut all traces that connect the 303 power supply with its circuit. cut all traces for that on the xox ..and connect the 303 supply with the xox.. its a bit like brain surgery and a lot of potential ground issues make this a complicated task.

However.. i might do that..but only after i brought my xox to the maximum 303 closeness that is possible without getting to the power supply..

So question..what else could effect the elasticity of the enveleope action of the xox?

transistor choices? or different caps in key positions?

I once experimented with old caps from the 80´s--they are also part of the game..using panasonic BC´s for example gives a tighter sound..the type roland used sounds extremly mushy in relation to modern low esr types..

so getting to the envelope relevant caps and only going there for a mushy one might be an experiment.. and trying vintage transistors there.. In the moment my xox is mainly n new 945´s..only the double transistors are all vintage and the envelope fet is vintage.. I cant really say that the vintage fet in the envelope made any difference..but the vintage double transistors had a huge impact on all positions.

So still a riddle i actually believed to have given up on long ago..but now i am hooked again..since the nonxox os really brings some 303 feel back to the xox that makes it an interesting machine again.

Just.. how to get it more squelchy ? The smoothness on the filter resonance is not too bad on my xox now.. i ve heard real 303´s with a harsher sound ( they have worse sounding models too ).

User avatar
rv0
 
Posts: 395
Joined: Tue Jul 14, 2009 4:50 pm

Re: Getting a more 303 sounding box..

Post by rv0 »

3phase wrote: I ve changed the resistors back in the relevant areas to the specified 5% cole types.. that was a good portion of work

but a huge sound difference.. Ok..you can say thats the 5% difference.. the metal films just had 2%..
I guess you did not take any reference recordings of the BEFORE/AFTER situation

Whatever you do from now on, please make clean reference recordings before and after.
Otherwise it's just more words and opinions on top of the numerous claims already on this forum in the past decade.

Also for yourself it makes it easier to compare if you have recordings

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: Getting a more 303 sounding box..

Post by aminoacid »

I agree with most stuff you are saying and i think you can trust your ears.

I wouldnt swear on the metal film resistor thing. But tolerance does make units sound quite different.

The film/foil caps does the most difference.

The original psu is doing alot to the sound aswell. But this is prob not because of its construction but the fact that it feeds the opamp in the psu with ~15v insted of ~30v which is the case in the xox. The ml303 uses a 7815 to get a Similar vee voltage, you could prob make up something else.

Make sure that your 2sk30a(o) is authentic. Theres threads about this in the forum. And of course other transistors.

All the different npn transistors will sound different but i liked 536 and 945 the most.

In some places, the different measures of a transistor will give differences to the sound.

If you are experimenting with socketed transistors, dont forget to let everything settle for a while before you make your judgement.
And its a good thing to reference it with another machine, like a 606.

And about the el caps... ive tried xicons and panasonic fc. But i settled with che@p jamicons. Yes they are socketed in my vco , vca and vcf sektion.

Diodes seem to make small differencies, at least in the env and perhaps in the vco. Not sure but i think 1n4148 and 1ss133 sound different slightly or its the difference in voltage drop that made the differance. Id love to find some 1s2473 to try out.

The stuff about high hfe on q9-11 is not crucial key settings in my opinion.

The only transistor that i like a spec hfe is q18 which i like very low hfe so the bass doent dissapear to much when resonance is increased.

This dose of neardness will do for now :)

3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Getting a more 303 sounding box..

Post by 3phase »

before after recordings have theire difficulties since my workshop is not the studio.. and i definitely dont go back to metalfil, just for the sake of prooving something that is in passive filter design common sense.
Look up all the discussion regarding crossover design.. There of cause metal film is seen superior to cole based resistors..but nobody there doubts an impact on the sound. Thy call it a smearing of fine details.. what in our course of roundening the edges might be a good thing.

In any case i ve learned that component choices are crucial when you want to reproduce a specific sound and resistors are components too.

You mention the diodes.. intersting hint.. i lately experimented with diode patchcords.. and any diode gives a different sound on truncating a fast lfo modulation .. as faster the diode as cleaner the result..

When you look into the original 303 service manual you find a very wide varity of diodes beeing used in the 303.. Even a germanium one.. probably in the power supply..but still..why so many different types?

We have D25 in the oscillator that might have impact on the sound..
D24 on the accent/filter
D27 on the accent/Vca

the last ones rectify the envelope..and this is quite a similar thing than rectifying an lfo.. The attack phase can be effected by the diode speed..

hmm maybe really time to open the 303 and have a closer look what kind of diodes we find in this positions..all the same? or diferet types..what would indicate a design choice at this significant points..

3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Getting a more 303 sounding box..

Post by 3phase »

diods in my 303 are all the same type 1n4448 and similar to the ones used in the xox 1n4148

User avatar
aminoacid
 
Posts: 352
Joined: Tue Jun 20, 2006 5:27 am

Re: Getting a more 303 sounding box..

Post by aminoacid »

Im pretty sure The germanium diode is only present in the 606. They prob kind of shared their partslist...

Perhaps the diodes make changes for the accent. Mine has alot of wow but perhaps this is common with other xoxboxes...

Carbon resistors you say... that would really make differencies but how you know diss? I never heard about it...

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

Return to “Making x0x”