three O three mod

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

three O three mod

Post by nologin »

I've done some 2nd VCO experiments for TB303 and x0b0x. It is largely inspired by x0xi0. I published some videos in French about it:





Sorry, the mod wasn't perfect, i made a mistake on my PCB that i have since rectified... :doh:
But the original idea was rather something for TB303, not x0xb0x. i wanted to do something a bit devilfish, but different. :cloud: And i remember reading what B. Castro wrote in the past on the ladyada forum, TB303 devilfish is great, but it has been done with mods that the power supply can handle, more mods would not be possible because the power supply would weaken. I had read something like that.
I know that the TB303 power supply can handle a 2nd VCO and the xmod between those 2 VCOs. But I don't know how much more mods it can handle. I am not an engineer. :oops:
I found a picture of the inside of the TB303 devilfish, if anyone has opened one, do you know what the 16-lead IC is?
I'm thinking maybe a TL084? I can't read, the image is too small... would allow me to estimate the limit to not exceed, so that the power supply doesn't fail.

Image

I had a bad experience with a mod kit for x0xb0x, it was called atomic mode 2, i think all those extra ICs were too much for the power supply. Logic would have been to add a 2nd 12V source... In the case of the x0xb0x, the TB303 power supply is different, less flexible.

I want to avoid making this kind of mistake, i know that with TB303, i have to make do with the existing power supply and that i can't exceed a certain limit.
Last edited by nologin on Thu Nov 24, 2022 10:21 am, edited 1 time in total.
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Since i didn't get any answers, I'll try to answer myself.
In the end, as i went with the 2nd oscillator option, i am limited to make more modifications.
I rectified the error made on my pcb, so i was able to get a very stable 2nd oscillator, nothing to do with my old youtube demos.
But adding a second oscillator, i'm already very close to the maximum that the power supply can provide!
So i tried to add an FM mod and filter tracking through a 4558. This works but as i am asking for more than the power supply can provide, as soon as the play button is pressed, the 12 volt becomes unstable, resulting in a slight instability of osc 1 and a more obvious instability of osc 2 creating a portamento on osc 2, this creates a cool and eliminatory effect. :omg:
That said, on the 6562 oscillator 2 , have half of it free, so in theory i can either do the FM mod or the filter tracking here, which is annoying because i had planned to do an x mod between the 2 oscillators here. :doh:
I conclude that either way, I'm going for something very different from devilfish and because of the presence of a 2nd oscillator, I'll have to make a narrow choice on the mods adopted. :bang:

But i could do all the "basic" and known mods, those where you simply replace a resistor by a potentiometer. :cloud:
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

I found the right compromise, i could add filter and tracking and FM mod on the an6562 of the 2nd oscillator. I found the limit that the power supply can assume. I can't do the xmod because it would have to add another operational amplifier, it would result in a malfunction of the synthesizer, the oscillators in the first place.

Now it works perfectly; sorry it's in French, i don't speak English well, but I can read and write it a bit.

User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

good work, it sounds pretty wild there at the end.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Thanks
On the other hand, the FM mode and filter tracking, i made with a common pot for these 2, so that i could switch from one to the other via a switch. I did this so as not to overload the power supply, which was already under enough strain. Anyway, when use FM mod, don't hear much filter tracking, so it's not must be necessary to have 2 separate pots.

I find that with FM, when i use the tune control pot on osc 2, it saturates and modulates the VCF in such a way that it sounds like a bitcrusher effect, probably the same as with x0xi0, but I've never been lucky enough to own an x0xi0. :doh:


There are some differences with my RE-303, VCO1 has less volume and filter tracking is less efficient with the TB303. My TB303 is old anymore, but i like it like this, it has a different sound signature than the others. I prefer not to recap it, because my TB303 really sounds amazing when plugged into a distortion. I did this test with it just to compare with the RE-303. I will only modify my RE-303. :mrgreen:
charkachak
Common Wiggler
Posts: 174
Joined: Mon May 20, 2019 3:38 pm
Location: France

Re: Have you seen the inside of devilfish?

Post by charkachak »

well done buddy ;) Bravo pour ces améliorations !
And as Guest mentioned it sounds "pretty wild" at the end of your video. A 303 on LSD!
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

I transferred the board to my RE-303 dinsync, big disappointment there is a beug, while it works perfectly on the TB303! :doh:
I don't understand, i checked everywhere on my RE-303, which is identical TB even with the 650C CPU...

The malfunction is a detuning from the env decay pot... It's unbelievable. I'll explain.
If the decay pot is at its lowest, i probe on a single long C note, i have a kind of slight portamento in the attack of the note, then the note stabilizes at 65.4 hz, it's barely perceptible to the human ear. If i increase the decay, the oscillator two this detune, actually i think it lengthens the portamento and turns the note into a 1.8hz higher note. I'm at 65.4 hz decay low and 67.2 hz with the decay at maximum. :omg:
I susppected CD4066 was the culprit, i tried TC4066 and HD14066: no better.
As i have several boards, i installed today on both, indeed it works correctly only with the TB303.
I have not noticed any other problems with the RE-303 except for the detune problem related to decay. :hmm:
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

can you check if the 5.333V is staying steady through all of this?
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Yes, there is no problem with 5.333V.
However, when i built my RE-303 years ago, the 5.333V was totally unstable, i replaced the TM6 trim, that fixed the problem. But it was impossible to get to 5.33V if i remember correctly i was 5.29, i replaced R180 with a 33K to get there. But these last weeks i found the reason of this small problem(33K instead of 39K) , i had been one of the alphas builders, so i had built without BOM.(or partial bom...)

Made a mistake. I had soldered D43 a 6,8V zener diode. But 2 weeks ago i rectified with a C6V2PH at D43.(zener 6,2V) Normally the substitute is a BZX85 C6V2TR. I think it's OK with C6V2PH, they are diodes that my neighbour had in stock, it's philips diodes that date from the end of the 80's, from what my neighbour said.

I guess they are equivalent to the vishay BZX85 C6V2TR, that the name is different because it's not the same manufacturer and not the same era. To note that the bom TB303 says D43 is 1SZ46A, I checked with a microscope in my TB303, the diode is red + silver and is badged 24, i did not find information about this reference, i do not know of which zener it is. But I'm starting to wonder if there might be a little detail there, because to build my RE-303 i copied my TB303 and the power supply of my TB303 has some small differences from the original schematic , D43 is not a 1SZ46A, if i remember well there is also a different value of resistor, and capacitor from the schematic. It seems that Roland had made some differences in the power supply according to the different batches.


I probed 12V and 5V of the two 303s.

TB303
stop sequencer: 11.89V
Play: oscillates from 11.88V to 11.89V

It's OK, i understand that as long as the oscillation does not exceed 0.10V, the synthesizer will behave normally without altering the tune or anything else.

5V oscillates a bit in play, but that is normal.

RE-303
Stop sequencer: 11,93VDC
Play: oscillates from 11.93 to 11.75V :foul:

5V oscillates a bit in play mode.

In fact when i press play the sequencer, which is powered by 5V, seems to oscillate by 12V. :despair:

Next week a friend will lend me another RE303, i have to look at it, it has some cracking pots. I'll get another one to compare.
But i think it's only mine that has a problem, it's crazy because unmodified it was working pretty well, with modifications the poor thing is running out of steam. :doh:
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

might be that there is some slight difference in the power supply so it cant source the extra current, or for some reason the extra current being drawn is more on the re303 mods.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Yes, it must be a close call. On the other hand, my RE-303 is very close to a real TB.
Ram, cpu and 662 are the original Roland ones.
I even have an original sumida in the power supply.
I still have some used parts, 2SB596 and 2SD667C and 2B647C. I'm going to try with them because I don't totally trust the last two, the ones I have are laser engraved and come from China. In doubt.

I noticed that my leds are "brighter" than the TB303's, I checked the problem is not from too greedy leds.

I noticed this by chance this afternoon, it's a sign, it will bring luck next week:

Image
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

another thing to compare between your boxes, is what is the 15V rail like? and what are the opamps youre using for the 12V rail generation? you can check the 15V rail on pin8 of the 12V regulator opamp.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Thank you, i think that's where the problem lies.
I'm wondering why i didn't check 15V before!
On the 2 synths I use AN6562, anyway as they are mounted on a stand, i had already tried to compare with others.

Indeed at the TB303 pin 8, i have 14.14Vdc, RE303 is lower, 13.91Vdc. If I press play the same RE303 is lower!
If I probe further upstream, at the cathode of D41 TB303 15,36V while RE-303 15,11V. I think it's this small difference that makes it go off further.
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

does the TB303 15V rail droop when the mod is engaged? if so, does it droop by the same amount as the RE303? is R172 the right value? the votlage is set by D42, so maybe try putting another diode in series with it so you get an extra 0.6V.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Osc2 is always on, as on the x0xi0 the oscillator volumes work with the 5.33. So it's always on, only the oscillator volumes determine whether it's on or not. I haven't tried putting a diode in series with D41, I'll try that by tonight. I haven't had time this weekend. I will also check my 12V with osc2 totally disconnected.
But about the 15V rail drop, i have something similar on RE and TB.
TB stop: 15,37V TB play: 15,36V/15,37V.
RE stop: 15,11V RE play: 15,11V/15,10V

Only difference RE has 15V about 200mV lower, i think that's the problem.

But i did try something, just to see, i replaced the 12V from the RE303 routed to my VCO2 PCB with the 12V from an old I/O board that you'll recognize. Indeed it works very well with this second 12V source. So that confirms that it is indeed the 12V of the RE-303 that is too low. The limit is something like 200 or 300 mV.
By the way, it reminds me of the original question, although the subject has drifted a bit, i think Robin the devilfish guy must have done a lot of experimentation, to find a good compromise of modifications without weakening the power supply. Even if the mods are different, i may have to find the right compromise.

Image
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

do you have an oscilloscope? if so, put the probe across R172 and compare the TB to the RE. this is the resistor on the powerline of the AN6562, and the drop across it tells you how much current is flowing. this will let you know if the RE is drawing more current than the TB. if they match, then its just a headroom issue. but, it might be that the RE circuit is drawing more current for some reason. the reason i suggest using a scope, is that the current draw will be spikey, and this will allow you to see those spikes.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Yes, thanks.
I'll check that tomorrow. But last night i tried to solder diodes in series to D41, indeed it raised the 15V rail and therefore the "12V" is also raised, it went to 11.93V. So the RE now has a "12V" higher than the TB303.
And yet as soon as i press play, 11,93V still oscillates, unstable. :foul:
On the other hand i found that if i electrically disconnect the VCO2 PCB, the 12V remains stable, stuck on 11.93V. As soon as i connect it electrically it becomes unstable. (In play mode) It's all the more frustrating since it's the PCB i connected last week with the TB303. Since last week i tried with several VCO2 PCBs same result, already i know 100% that the problem is not in the VCO2 PCB.
I'm starting to believe it's not from the power supply, that my RE303 actually draws more power, i probably made a mistake somewhere while building it, and it wasn't affecting how it worked without mods. I think that tomorrow morning when i go to probe with my oscilloscope, i will find my RE-303 consumes too much. :bang:
In any case i think Friday a friend will lend me a 2nd RE-303, I'll see what will happen with it. :hmm:
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

You mean at the junction of R172/D41?
I probed here i have this kind of unstable sine, it is a bit higher on the TB303.



Image
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

i was thinking of clipping the ground clip on one side of R172, and the probe tip on the other, that way you are measuring the voltage drop across the resistor, which is proportional to the current through it. be sure that the 303 isnt connected to anything else while doing this, as the scope ground is earth ground, and might mess things up if there is a loop someplace. the power transformers should isolate that section just fine. you can then compare the current draw on the RE versus the TB. im suspecting they will be the same, in which case the only hope is to increase the supply to the opamp. its also possible that the opamp has higher output resistance or something, you could try swapping it.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Thanks, i compared it by probing R172. This waveform is quite unstable, it's not something you can probe with 2 channels and try to overlay. So I set a single channel and probed separately, used the stop function on the oscilloscope to compare. And I got almost the same with both synths:

Image

Image

As the waves move all the time it is difficult to compare them, i have the impression that at certain times the RE-303 peaks are 1/2 tile higher, it is not obvious. I will try to swap the 6562 with these 2.
I have the impression more and more that there is a problem elsewhere, I didn't have the time to go and get a 2nd RE-303, I'll go next week.
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

youre probably right that the problem is elsewhere, but youre also zoomed way in timewise on that waveform, so its hard to tell. youre at 10ns/div, and the current draws that are causing issues are probably happening on the 100ms/div level, as thats the rate notes will be firing at.
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Yes, thanks.
I haven't had time to work on it this week.
I have just probed, the rotary scale of my oscilloscope, does not allow me to set to 100ms, but i have probed with 1000ms, i observe that i do not have the same thing at all. The peaks are higher and more irregular with the RE-303.
However, I have a second RE-303, i will check what happens with it.

Image
Image
User avatar
guest
Super Deluxe Wiggler
Posts: 8304
Joined: Mon Aug 19, 2013 11:49 am

Re: Have you seen the inside of devilfish?

Post by guest »

ok, thats looking a bit better - or at least its showing a difference that makes some sense. can you zoom out any more? the top shot looks like its at 2us/div and the bottom at 1us/div (microseconds per division). if the sound warbles around only when notes are played, you might see even larger voltage drops at the note rate, so maybe 10ms/div (10 milliseconds per division) or longer. on the TB303 you can clearly see the switching regulator switch rate (~150kHz).
openmusiclabs.com
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

Over the course of this week i will look at this in detail. Especially as I have a 3rd one to probe.
I have a 2nd RE303, i installed my kit and i could see, that there was the same problem, when i turn the decay potentiometer, the osc2 doesn't hold the tune, i even noticed that on this 2nd RE303, it is even worse than mine. :foul: If i hold a long note, there is a very obvious portamento, the osc2 fails to hold a long note, this behaviour i know is caused by asking more from the power supply than it can provide. This 2nd RE303 is not exactly like mine, it has a sonic potions, mine is equipped with a µpd650C like a TB303, i have the impression that it fucks up even worse with the sonic potions, maybe it consumes a bit more.
I'm already glad to see that the same problem exists on a RE303 built by someone else, in the end it's not due to an assembly error. :hihi:
On the other hand it still works fine on the real TB303 and also on x0xb0x. I'll keep trying to find out why. :lol:

Image
User avatar
nologin
Wiggling with Experience
Posts: 299
Joined: Thu Jan 21, 2016 2:53 am
Location: France

Re: Have you seen the inside of devilfish?

Post by nologin »

I haven't had time to work on it lately. But this morning i tried something, i desoldered IC16 and IC11 from the TB303. Then i installed them on the RE-303, still the same problem, so the op-amps are not physically at fault.
I'll probe again later.
I made a friend listen to this problem of a difference of about 1.5hz when i turn the decay potentiometer, he didn't find it eliminatory. Personally, it bothers me, because this small difference on a C at 65.4Hz, is not the same if i play higher notes, for example with a C of the next octave higher the difference increases to 3 Hz.
Post Reply

Return to “Music Tech DIY”