Ok so now the power supply is built its time to move on. I figure out the next thing is building a synthesiser voltage controlled oscillator. I’ve got the power supply to operate it and I guess its one of the things that’s more of a standalone item and easy to test. I was going to try and come up with a design of my own but that could take years and there are so many circuits floating around the web it seems a waste of time and there’s also the fact that the guys that have designed these circuits are way better than me with better designs than I could ever hope to achieve so that s my mind made up. I just have to decide which is going to be the best one for the job.
I need it to be good meaning there’s no point in building something that’s known to be problematic with tuning etc. I’ve read on one forum from a guy trying to get something going that he’s built and an expert has commented on the design saying that it’s never going to be able to be tuned correctly more than a couple of octaves and given the current generated by the component values it would never go higher than a 1khz.
No difficult to get parts. There seems no point in trying to build designs that are using hard to find parts I want to be able to get components for years to come so if I’m struggling now they will be impossible to find soon things like….
If I can I want to be able to lay the circuit out on Vero board so I’ve not got to try and track down a PCB. That also precludes designs that are overly complicated with loads of components.
I want to be able to get something up and running but also be able to upgrade it in the future. It seems that the best way to build a VCO is to use matched transistors and a thermal resister touching the two in thermal contact. There doesn’t seem to be any other way for it to reliably stay in tune. So a least I can use two un-matched transistors and a normal resistor to check its working and then match transistors and try to get a thermal resistor later.
After weeks of research I have found the perfect answer. The Thomas henry 555 VCO. He’s well respected in synth module designs and even he states he is particularly proud of this design as he was trying to come up with something similar to a CEM3340 type of oscillator. I always thought the oscillators on my Electronics and Music Maker Spectrum synthesizer to be among the best analogue voltage controlled oscillators so I’m looking forward to building and testing these. Also the component list looks great with only commonly availably parts called for.
The VCO board with printed layout, with wire links and IC sockets mounted.
After playing around with layouts I’ve decide that the best way is to mount the exponential generator and matched transistors on separate small Vero boards as I’ve already been playing around with this part of the circuit and got a few already working on the smallest sized Vero board and I’ve got loads more of these to use up so each voltage input will be built on its own bard. I can then fit two VCOs on one large Vero board that I’ve been using. I bought a pack of 10 of them extremely cheaply and I mean extremely, I paid less than three pounds for eight boards.
I’m going to build 8 VCOs and so use 8 small Vero boards and four of the large veroboard. So that will mean 8 VCO’s without having to spend anything on the actually board material.
Im building 8 because I’ve noticed in some of the comments that you can switch in a capacitor and make it into a voltage controlled low frequency oscillator. So no messing around with a separate design for that. It will lead to a very flexible configuration, 8 VCO’s or 1 VCO and 7 VC LFO’s and everything between. Making the modular even more versatile.
The layout for the exponential converter is fairly simple there are a few cuts to make on the Vero board and only a few components to solder on.
The exponential generator built on a small veroboard. Notice the thermal resistor mounted over the top of the two matched resistors.
The board that the two oscillators are built on is quite large. One thing I noticed was that there were a few track shorts where there shouldn’t have been! You need to get a magnifier on it and check and it’s always easier to do this before you start soldering parts on to it.
To make things easier still you can partially build each oscillator and check it works before you fully populate the board. I’ve looked at the circuit and it can be stripped down to the triangle wave oscillator. The other waves are derived from it so you can just get that working first which is how I did the first board the others I just fully populated in the normal order of wire links IC sockets and then resistors. Finally the IC’s were inserted into their sockets and the variable resistors that would be panel mounted.
The picture above shows the VCO board fully populated with all the board mounted components.
The component overlay of the VCO.
This is the print out I used to stick on the veroboard.
Each one was powered up and tested. Seven of the eight worked first time which was a bit of a result, the eighth had an op amp inserted the wrong way round.
After I had made the front panels (you can read about that here) I fitted the variable resistors, sockets and switches. I then did the wiring between the panel components and then between them and the circuit boards. The next stage was to wire everything up to the power supply.
The boards mounted and wired up to the front panel controls. I must admit to not realising how complex and messy this would all become.
The final bit was to tune the VCO’s. You can read about that here.
While at this stage I noticed that the stability wasn’t as good as I was expecting. I thought this may be due to several things.
4 VCO's being tested with an oscilloscope.
These were all the reasons I thought of to investigate so firstly I’d look at the power supply. I disconnected all the VCO’s and just powered one board. To my surprise this proved to be the problem. With one board the VCO’s pitch was rock solid. If I connected another board up the pitch varied, in fact it was all over the place. I experimented by moving and adding decoupling capacitors and increasing the values and putting twenty ohm resistors on the supply of each oscillator. Nothing improved the stability. In the end I decided to just use a separate power supply for each oscillator board.
I noticed that the Yusynth VCO power supply has separate 78L15 and 79L15 regulators on the board so maybe that’s just the best way to do it. In any event it had taken be loads of time experimenting without any real progress so it was easy to just make three more power supplies.
So that’s how they are going to end up. Eight VCO’s switchable to LFO’s with four separate power supplies. Now it’s time to think about the filters.