cutoff/rez CV in Mods, Gate in to x0x?

[controlvoltage]

OK, I almost feel lazy for askin questions before I try a bunch of stuff...
But I am working on this at the same time and have some questions based on the schematic. A lot of my questions relate to *reading* the schematic, which uses some different conventions than I am used to. F'rinstance, the dual-ganged pots are a little difficult to follow.

I am used to 0 to +5V CV signals where the 0 to +5V comes directly from the pot, which is dividing the 5V across it into a control voltage seen at the wiper.

On the x0x, I measured ~7.15V across the cutoff pot, and 3.2V across the Env. Amt. pot! This is a whole new world... a strange, Rolandy world full funky circuits and accidental greatness...

can anyone give me some pointers?

ideally I would like to add CV ins for most knobs... cutoff, res, env amount primarily... I guess decay would be useful too... this way one could drive the x0x from an analog sequencer, which would be most excellent.

I am assuming a GATE IN would be simple enough... I am expecting the Atmel to go 0V/+5V Lo/Hi on its GATE OUT pin, which goes into R146... a switching jack inserted at this point should allow an external +5V Gate signal to do the same job... correct? I don't have a (working) scope at home right now to look at this pin for confirmation.

So, any pointers would be great... these mods will of course be tested immediately and added to the x0xd0x wiki...

cheers!

-Amos

[controlvoltage]

Here are some really vague pictures of CV in mods to the original 303. It would have been really nice if component labels were used instead of measurements from the edge of the board... these are about too un-technical to follow.

VCF input:



Gate input:



any ideas as to what this VCF input is connected?

[Jonnay]

I tried to figure out what was going on with the Gate in with a copy of the tb303 circuitboard, but I had a hard time making heads or tails of it. Maybe i'll make a second crack at it.

As far as i know about the VCF in tho, it is for VCF audio in, not CV in.

[Guest]

i think the gate in mechanism could be done more easily
by just connecting a 22k resistor to the base of q37
the other end of the resistor going to the 5v signal
this way the x0x could respond to both
the internal sequencer and the external source
but perhaps this isnt preferable

as far as the vcf is concerned
i think that only the cutoff is voltage controllable
the rest are set with the potentiometers which must
carry audio signals
the pots could be replaced with digital pots though

to put the cutoff control voltage in
connect a 100k resistor to the base of q10
the other side of the resistor goes to your cv

the main issue with this is dealing with
the offset voltage in the vcf
the signal does not reference ground
but rather the emitter of q9
maybe it will make for some interesting sounds

[controlvoltage]

Thanks! (br0x?)

That explains the dual pots then, I guess...

I think it might work out OK with the cutoff input- a 0V to +5V CV would act as a bias voltage on Q10, regardless of the reference voltage... I think?
Nothing for it but to give it a try.

as for the gate in, my only concern re: allowing both the sequencer and ext. gate input to work simultaniously,
is the possibility that the voltages would be summed at Q37... would it mind a +10V signal (with both sequencer and ext. in turned on)? If not then there is no problem, I guess... if so, then I would want to use a switched jack to interrupt the gate signal from the sequencer when an external gate signal was plugged in.

[controlvoltage]

Ok, mods are installed!

The Gate in mod didn't work as specified... here's what I did
And there is room for improvement

I disconnected R146 from the CPU
and connected the anode of a 1n4148 diode to the trace on the CPU side.
I connected the kathode to R146
I then wired the gate input (tip connection of the Gate in jack) to the junction of R146 and my diode

This may have been a pointless use of a diode, and it may even be the cause of the slight clicking sound on note-on
which occurs when the internal sequencer is running and there is a plug inserted in the Gate IN jack but no signal on the plug.

Anyway, it works...

As for the filter-CV in, it rocks
The apparent effect is that with a cable plugged into the cutoff-CV input,
the range of the filter cutoff is extended
from absurdly subsonically low, up to the normal maximum.

I may want to add slide and accent inputs also, to my own x0xb0x... this one is done, I think. It rocks. I played the x0x from a CV keyboard and it was totally slick. Mod wheel hooked up to the filter cutoff... aww yeah.
It was cool controlling the cutoff with a cv touch surface also. Good times.

-Amos

[Guest]

glad to hear the vcf mod worked

i was suprised to hear that the gate mod did not
so i tried it on my x0x
and it worked fine
i connected a 22k resistor to the q37 side of r146
and a function generator doing 5v peak to peak
on the other side of that 22k resistor
the function generator was then grounded
to the x0x digital ground

your diode connection setup has one possible problem
if the x0x is putting 5v on the microcontoller gate out
and the gate input jack is trying to assert 0v
then the microcontroller is shorting its pin to ground
through the diode

this could be fixed by adding another diode
from the gate input jack to that connection

[controlvoltage]

Aha!

Thanks for the help

I think I was originally connecting to the wrong side of the resistor

I am now working on a crazy homebrew solution
to the problem of Accent gate input

more details to follow



DETAILS:

Is the accent circuit implemented differently on the x0x vs, the TB303?
Maybe I was just in a hurry
but I couldn't find IC13 as drawn in the TB303 schematic...
It looked more like there was a gate signal coming directly from the CPU
to the bottom of R141

The bottom leg of R141 shows +5v high whenever Accent is not active-
It goes low (0V) on any sequencer step where Accent is enabled.

So, here is my improvised untested plan for a solution
Connect an accent (gate) input jack to an inverting unity op-amp-
such that it puts out -5V if +5V is seen at the input.
Run the output of the op-amp to the bottom leg of R141
via the same diode arrangement as w/gate in above.
My hope is that the +5V normal voltage and the -5V from the accent input will sum to zero.
If I'm overlooking a fundamental principle that makes this impossible,
let me know... I will hook it up and see if it works.

[controlvoltage]

hmm... the inverting amplifier doesnt seem to be inverting... does this only work with AC signals? I didn't think so... anyway
I am using half of an RC4558P op-amp (running on +/- 12V), hooked up as the simplest inverting amp there is... R1 between source and inverting input, R2 as negative feedback between output and inverting input. I thought I should use a low value so r1 and r2 are both 33 ohms. This was a random and probably silly choice of value... and may relate to why it isn't working. I am a little confused anyway since I don't quite grasp how this circuit would be able to work with a connection (almost) directly from input to output... the current follower or non-inverting amp makes more sense to me visually.

Can anyone help me out on this one? seems like I am so close and yet my ignorance is keeping me from getting the rest of the way.

[controlvoltage]

Late-night brainstorm...

perhaps the Accent gate input could be accomplished with a CMOS switch
like 1/4 of a 4066.

This would eliminate the need for a bipolar power supply...

The signal at the Gate input would be applied to the Control input of the switch... when the CV in goes high, the switch will close

the switch would pull the voltage LOW at the bottom leg of R141 when closed.

Reckon this would work?

[Guest]

you should be able to implement an accent control
in the same way you did the gate control
just put a 100k resistor to the base of q35
the other end of the resistor goes to your source
when this source voltage is low
the accent is on
when this source voltage is high
the accent is off

and while your at it
you can do an external slide control too
just run a 100k resistor to the base of q29

[controlvoltage]

Slide input is already rocking...

here is the (almost) completed panel now


The only complication to the Accent input
is that usually you'd want it to be off
your usual gate signal is usually off/low
and then momentarily switches high...
which is why I wanted the voltage at the base of Q35
to be high when the source (gate input) is off/low
and to go low when the gate switches high.

If I had thought of it at the outset, I think I would have used the 4066
for all four gate inputs - I could take advantage of its built in buffering
and not worry as much about the source voltage going negative
or being otherwise out of spec... each gate input would simply go to a switch control.

Maybe that's how I'll mod my own x0xb0x...

[Guest]

the best way to invert the signal
is to use a digital inverter chip like the 74hc14
but if you want to save space
you can make a simple inverter out of
a transistor and three resistors
this is essentially what the circuits around
q35 q36 and q37 are doing
since youd need to pull up to high
you would use something like the circuit
around q36
except you wont need the capacitor

take a pnp transistor
and put the emitter to the 5v digital supply
then put a 100k between the base and the 5v supply
and then put another 100k from the base to your control voltage
finally put a 100k from the collector to the base of q35

when you put 5v on the line
the transistor will be off
so no current will flow to the base of q35
when you put 0v on the line
the transistor turns on
and puts current to q35 and turns it on

[controlvoltage]

wow is that easier than using a 7662 to power a dual op-amp which then doesn't work.

I'll have to try it later today. Although I'll look around for a 74hc14; you never know what might be kicking around the shop.

[controlvoltage]

Hmm... it's not going too well.

I was expecting the transistor to act like a switch in this arrangement, showing 5V at the collector when on
instead it is registering a puny 10th of a volt, or something...

also, when set up as directed
there is a constant 5V appearing at the CV input tip connection
Is this correct, and is it not a problem?
I am imagining that it would draw current from the CV-output device when connected and the gate-cv output was low... would this be ok?

I tried it just now with the connections suggested and it did something-
namely, when the cv in was low it would disable any accents triggered by the sequencer
but when the cv in was high it did not create an accented note... merely allowed the sequencer to do so.
My goal would be an accent in that could trigger accents directly
one way or another
I am having problems with the CD4016 analog switches I am trying to use also,
they don't seem to want to switch. I've used them for other applications and they are dead simple - I don't know what I could be doing differently to have so much trouble with them.

back to the bench...

[edit]
wait it's working now...
operator error seems to have been the culprit
now to proceed