Synth-DIY -- how to start?

[alanp]

Doing it one module at a time (FreqCentral) is hugely more manageable, especially if you have board mounted pots and not the MFOS spaghetti.

[peAk]

Yeah, I think I'll take your word for it. I just need some type of path because I will be starting from scratch.

[alanp]

Who wants some sequencer?



It's using the Fonik Baby 10 schematic (fixed reset problem), going to solder the LEDs in when I mount it to the faceplate. (Ordered one from Front Panel Design, too damn pricey to order from them again!)

Selectable 1 to 10 steps, Gate is toggle-able. Clock in, reset in, reset and step have pushbuttons, and CV and Gate out

[karter2000]

Awesome! Where did you get the pcb for the Fonik?  I've been looking for a good sequencer pcb that's fairly reasonably priced.


Sent from my iPad using Tapatalk

[alanp]

Laid it out myself. I'm still not entirely happy, might send in another version where the pots and switches aren't so damn close.

[karter2000]

Looks great! If you have any extras, I would be willing to take one off your hands.


Sent from my iPhone using Tapatalk

[alanp]

Well. I got something for laying out a pcb, so I used it to get more PCB's fabbed.

In a few weeks, I should have boards for the yusynth Steiner-Parker VCF, and the EFM CA3046 VCO -- with boardmounted 3.5mm jacks! I only have the one System X vco (and two minimoog VCF's and a Electro-Cardiogram VCF that can all fake being a sine-out vco) at the moment, so in a way this is kinda sad on my part.

[alanp]

So, I was having fun noodling on the synth earlier (this is hugely fun, although it can annoy the rest of the flat pretty badly at times.)



But hang on... what the heck is a guitar pedal doing there?!

The clue is in where the other end of the 6.35mm jacks are plugged into



Cat Girl Synth's Stomp Box Adaptor (CGS60V1), laid out by me for Euro power.

The fact that you've got the signal chain going through stomp boxes introduces noise (it's inevitable, synths run at like 5V peak to peak, stomp boxes are maybe 200mV ?) but it's fun!

If you're wondering, I had the minimoog VCF oscillating (it sounds nicely gritty), tremolo'ing through the LFO controlled VCA into the multiFX (to give the De Profundis delay something to chew on), with the output going into the Electro Cardiogram filter (red panel.) Obviously while titu'ing with settings on both stomp box (turning fx on and off) and titu'ing with synth knobs as well.

Got 9 more boards for this, going to have to order more jacks. (They're Tayda PCB mount ones.)

[lincolnic]

Dude. Alan. Enough with the teasing. When are you going to record us a demo of all your fun synth toys?

[alanp]

Who wants a really, really simple VCO? Saw/square out, freq and PWM controls (either CV or keyboard CV (this one has a trimmer to scale it to v/oct.)) The timing cap (the 10n one) could do with adjustment, but other than that I'm happy with this one.

It's Tom Gamble's EFM 3046 VCO. With board mounted jacks and pots, no more wiring for ME!

[RobA]

I'll be interested to hear how this one goes for you. I've been playing with the LM3046 and CA3046 quite a bit and there are two issues with it related to the schematic I found for what I think is this circuit. The first is that pin 13 of the IC is supposed to be attached to the most negative point in the external circuit for stability of the transistors (see note 4 in the "Absolute Maximum Ratings" section of http://www.ti.com/lit/ds/symlink/lm3046.pdf). The second is that in testing of bunches of IC's, I found that the second set of transistors 9,10,11 and 6,7,8 aren't well matched on most of them.

I did a circuit that used the LM3046 before I saw the note about the substrate connection on pin 13. It seemed to work fine on two of the prototypes but then on a third it didn't sound right. I don't know for sure that it was the issue though, but it made me suspicious enough that I scraped that version and went to using discrete transistors for that design. It's a distortion circuit though, so using the discrete transistors has a couple of other benefits too and the added effort of matching is worth it. I'm still playing with the LM3046 though and as long as I plant pin 13 to ground or -Vcc and don't use the unmatched pair as a matched pair, the IC has been really good. It could easily be that I misinterpreted the note in the spec sheet too, but I found a warning about it in some notes for an EE lab that was more explicit about the warning.

[alanp]

It doesn't have the flexibility and range of a fancier VCO, but for what it is (ultra-simple) it does what it's supposed to. The substrate pin 13 is connected to -12VDC through a resistor in this circuit.

[RobA]

It doesn't have the flexibility and range of a fancier VCO, but for what it is (ultra-simple) it does what it's supposed to. The substrate pin 13 is connected to -12VDC through a resistor in this circuit.

My interpretation of the note was that the substrate needs to be held at a lower potential than any of the other pins. The warning from the EE class was that you needed to just tie it straight to the lowest voltage rail. What I don't know, and I couldn't find any good info on it, was how strict this requirement really is.

I like simple. I'll have to try this one on the breadboard for sure and do some playing with it. Did you use a tempco for this? If so, where do you get them?

[alanp]

There's a 1K resistor that is supposed to be tempco, but I didn't bother (it means a special order for it), and the layout has it at right angles to the IC.

Oh, yeah, the schematic calls for a J112 fet, and I used a 2n5457. Seem to have got away with it.

I'm thinking of ordering panels from FPD for it -- if I get 10, then it's 20% off.

[alanp]

Waiting on caps from chrome, and switches from china, but they look so cool even not working