Twitter   Patreon WIGGLING 'LITE' IN GUEST MODE

 Goto page 1, 2, 3  Next [all]
Author Model 1011 Discrete VCO
SlightlyNasty
 Hi all, Before I get into details, just a little bit of backstory to explain this whole thing. I've been messing around with electronics for years, mostly with digital logic type stuff (74HC etc) and old computers, and although I always wanted to get into analogue electronics and had built a few kits/PCBs like the ASM-2, I could never quite get a foothold on the theory of it. Every tutorial page seemed to just want to throw masses of equations at me (I'm useless at mathematics) or was fixated on aspects of design that weren't really relevant to the stuff I wanted to make. It's only very recently that I decided to try learning it using the system I've used to learn everything else in my life and career - ignore all the tutorials and just force myself to make something, figuring it out as I go along. Also, at this same time I'd decided that I really wanted a true modular system, so in deference to my life of poverty I made a deal with myself that I could have one as long as I designed and built the whole thing myself. So I set a few ground rules to make sure I'd get the most educational value out of the process. Things like "No peeking at other peoples' designs unless you're absolutely stuck", and "use jellybean parts wherever possible", and most importantly "finishing something is more important than making it better". With those rules in place, I set out from the obvious starting place for someone who's never designed a synth module before and has only a vague grasp of how transistors work - a fully discrete VCO. And here's the first prototype: It's a sawtooth-core unit with three outputs - Ouput A is sweepable from sawtooth to square with another pot adding suboctave square. Output B sweeps from sine to triangle, and there's a separate output for raw suboctave. There's linear and logarithmic FM inputs, sync, and PWM along with a static PW pot input. Tracking is good, temperature stability is... reasonable. I've only got a 3000ppm tempco in there, so it's hard to say. The PCB looks pretty intimidating, but it's easily built in a day - I laid it out in subcircuits so that it can be built and tested a piece at a time. Transistors are all BC550 and BC560, except for two lonely 2n7000 FETs. Now the bits I'm embarassed about. For a start there's way too many components in there - because of my "no peeking" rule in designing this thing, a lot of my solutions were a bit, um... "inelegant", let's say. Stuff like using positive feedback to implement hysteresis in the sawtooth comparator simply didn't occur to me, neither did the notion of virtual-ground summing on inputs. Also, because I was making it all up as I went along, there ended up being AC coupling and buffer stages that really wouldn't need to be there if I'd blocked it all out properly from the beginning. But I didn't want to get in stuck in the graveyard spiral of perpetual improvement, so if something worked I left it in and moved on. Obviously some bits I had to take from other people's designs - the expo converter is based around the designs from RenĂ© Schmitz's page, just flipped upside down for PNP and reimplemented using discrete transistors and my own half-arsed temperature compensation. And I obviously didn't reinvent Franco compensation either. Oh, hang on - sound. Here's a demo with it roughly calibrated by ear, being played on the Monotribe's ribbon keyboard through a Pro-Solo (Didn't have the PW pot or log FM input hooked up for this recording, so they're absent): [s]https://soundcloud.com/raynor-pettge/m1011-demo[/s] Eventually this thing is going in a Loudest Warning style 4U system, using the eurorack power standards. The front panel for this guy will look something like this: Now the job is to lay up the second version of the PCB, removing a few components, making some tweaks and generally tidying things up. And that's it really. Make of all that what you will.
mush
 Looks great. Any chance that the next revision of the PCB will be under 10 cm in length (making euro-adaptation a lot cleaner)?
SlightlyNasty
 mush wrote: Looks great. Any chance that the next revision of the PCB will be under 10 cm in length (making euro-adaptation a lot cleaner)?

Hi Mush.

I think the next revision will just be a cleaned up version of this 4U board, I would definitely like to do a Euro version (you probably remember that was part of my original plan), but I'll have to think about how to go about it. The current board is 105cm2 and very dense, so it'd have to be something like a 21HP module to fit.
search64
 You could go SMD, of course. Or use ICs instead of some of those transistors. As far as temperature stability goes, you can also go the other direction: using a tranny as an oven which keeps the core at temperatures far above ambient temperatures.
PWM
 But If there's ICs it's not discrete anymore.. I don't see the problem, the module is 105cm a eurorack panel is 128,40cm..! Great project by the way. I love discrete VCOs!
search64
 You'd lose the bragging rights, sure.
PWM
 To my ears most discrete VCOs sound different (I think better) then non discrete VCOs. Plus, I like to work discrete and the look of a stuffed with trannies PCB!
mush
 Hi Slightly Nasty. Yes, I thought you've dropped the euro-version as the PCB posted looks higher than 105mm. FWIW. I'm be in for a couple as I can see some myself layout a panel for a pair version in eurorack, maybe with a discrete amplified transformer based ring mod on protoboard. Let me know if you have any extra PCBs on the next prototype run.
SlightlyNasty
 PWM wrote: I don't see the problem, the module is 105cm a eurorack panel is 128,40cm..!

 mush wrote: Hi Slightly Nasty. Yes, I thought you've dropped the euro-version as the PCB posted looks higher than 105mm.

Sorry for the confusion guys, I was referring to the area - 105cm2. The current boards is indeed too tall at 70x150mm. It's made to fit neatly behind a 3" wide 4U panel.

Mush - You're right, I did drop the euro version once it became apparent that I couldn't really make a board small enough to work neatly for both 4U and euro. I was originally thinking it might need to split over two boards anyway, in which case I could have done a setup that would work for both, but once I figured out I could squeeze it onto a single board behind the 4U panel I decided to run with that, for simplicity.
search64
 PWM wrote: To my ears most discrete VCOs sound different (I think better) then non discrete VCOs. Plus, I like to work discrete and the look of a stuffed with trannies PCB!

Well I haven't been able to compare them side to side, but I find it slightly hard to believe there would be much difference in sound. It's just a different package for the same part... Unless you know something I don't?
SlightlyNasty
 search64 wrote: You could go SMD, of course. Or use ICs instead of some of those transistors.

I did want to keep this one all discrete just because I think it makes it just a bit more fun and interesting as a build. I didn't really do it for the sonic qualities, more for the challenge and because I've always liked engineering that puts the working of the device proudly on display.
search64
 You could still do it SMD though Once you go SMD, you never... go back... or something.
PWM
search64 wrote:
 PWM wrote: To my ears most discrete VCOs sound different (I think better) then non discrete VCOs. Plus, I like to work discrete and the look of a stuffed with trannies PCB!

Well I haven't been able to compare them side to side, but I find it slightly hard to believe there would be much difference in sound. It's just a different package for the same part... Unless you know something I don't?

I didn't realize that you ment (matched) transistors in IC form.. I thought of OpAmps, silly of me..
And probably no.
ashleym
 Always good to see a different flavour of module. Interesting take on the outputs. I agree with the others, look at an IC transistor package to help with matching and temperature.
Isaiah
 SlightlyNasty Impressive! I'd be keen to see the schematic you came up with if you're happy to share it. You might enjoy studying the schematics of the Steiner Synthasystem; the circuits are mostly discrete transistors and a few op-amps. Are you likely to organise a PCB run for this (in 4U/CGS format)?
SlightlyNasty
 ashleym wrote: Always good to see a different flavour of module. Interesting take on the outputs.

Thanks ashleym, that was partly to reduce the number of output buffer/amp stages and partly because I always found myself wanting some sort of basic wave mixing functionality when I was using my ASM-2. I figured that it made sense to split the waves up so that one output would give you the "hard" high-frequency sounds like saw and square, and the other the "soft" sine and triangle.

 Isaiah wrote: SlightlyNasty Impressive! I'd be keen to see the schematic you came up with if you're happy to share it. You might enjoy studying the schematics of the Steiner Synthasystem; the circuits are mostly discrete transistors and a few op-amps. Are you likely to organise a PCB run for this (in 4U/CGS format)?

Thanks Isaiah, I'll have to release the schematics eventually I guess! I'll wait until I've gone through and made the various tweaks for the v2 PCB, and removed all my scattered development notes/ramblings of a madman though, I think.

I'm definitely up for doing a proper run of the next lot of PCBs, I'll probably put up an interest check when I'm closer to ordering those, and in the meantime I'll check to see how it goes on 15v (don't see why it shouldn't work) and write up a decent build guide. Also I've done up a panel PCB layout for the design above, so I'll see if anyone's interested in that.

And yes, I've started looking at other people's designs now that I'm mostly done with this one, it's mostly just very humbling to see just how minimal other designers can make their circuits. If I were to redesign this one from scratch I could certainly make it significantly simpler, but I'm happy to leave it for now as the interesting result of an unusual development process.
alain6870
SlightlyNasty

I didn't mean that I'd put up the schematics after ordering the next version, just that there are changes that need to be made to the schematics either way, and there's no point putting them up with errors or outdated information in them. I just haven't had the chance to go through them yet and tidy it all up.

Also, as to why I posted here - it's just something I've put a fair bit of time into, and I thought some people might find it interesting. I've always had a vague notion in the back of my mind that other people might want to buy boards for it, but first and foremost it's just a personal project to build something that I can use in my studio. I assumed that 'Music DIY' would be the place to show off the result and gauge if it would be worth my time thinking about a euro version etc. If it were to stay purely in my studio I'm completely happy with that.

Just to clarify - I'm happy to make boards for other people, but to me the whole point of this project was to set a bunch of limitations and see what I could come up with. I'm not that interested in making major changes to it, I'd rather design a new one from the ground up if I decide I want better performance and so on (and I definitely will do another VCO in the future, it's fun). I really have to move on to building the other parts of my own system, a box of VCOs might be fun, but it's a little musically limiting.

That said, when I put the schematics up I'd love to hear people's suggestions and ideas - even if I'm not making changes to this particular module it's all good learning material for both myself and others.
SlightlyNasty
 ashleym wrote: I agree with the others, look at an IC transistor package to help with matching and temperature.

Is there a particular monolithic PNP IC that people would prefer? Personally I'm happy to use the discrete parts, but I imagine it shouldn't be too difficult to lay out so that either option can be used. I haven't used monolithic transistors before, so I'm not sure what people have in their parts bins.
calaveras
 I really dig that it is discrete, Kind of prefer it. But then, I'm not starting a 4u rack anytime soon.
Don T
SlightlyNasty wrote:
 ashleym wrote: I agree with the others, look at an IC transistor package to help with matching and temperature.

Is there a particular monolithic PNP IC that people would prefer? Personally I'm happy to use the discrete parts, but I imagine it shouldn't be too difficult to lay out so that either option can be used. I haven't used monolithic transistors before, so I'm not sure what people have in their parts bins.

SSM2220
trcwm
 Great stuff! Is there a schematic of this?
Jarno
SlightlyNasty wrote:
 ashleym wrote: I agree with the others, look at an IC transistor package to help with matching and temperature.

Is there a particular monolithic PNP IC that people would prefer? Personally I'm happy to use the discrete parts, but I imagine it shouldn't be too difficult to lay out so that either option can be used. I haven't used monolithic transistors before, so I'm not sure what people have in their parts bins.

BCM857ds? Not supermatched, but very cheap, or were you looking for a TH part?
keninverse
 Hand-matching is fairly easy. Why not include an hand-matched transistor footprint for the expo converter as well as including an IC?
keninverse
 Oh and this thing has plenty of character...great sync sounds.
 Goto page 1, 2, 3  Next [all] Page 1 of 3