Don’t bother trying to make a PCB of this one, you could stripboard it in half-an-hour or less, I’d imagine. As on the original it’s tacked onto the end of the sample and hold PCB.
Notice no screws for the slide potentiometers. The slide potentiometers are bunged on a bit of veroboard (at a jaunty angle to the strips), with acrylic spacers screwed to the board and glued to the front panel using JBWeld, which is as shit as it sounds.
I had a couple of failures before I got it to stick. I’ve also tried using some special/horrific acrylic cement, which totally failed. There would be a photo of the back of it here but it’s just embarrassing.
Since then I’ve reverted to making PCBs for the slide pots, with screws in the middle of the pot throw, which works much better, and helps reinforce the front panel.
Argh, PCB is dated Dec 2015, it’s been a while… three simple circuits for the price of one on this board, which you should easily be able to stripboard. Here’s my version of the schematic – R3 has no value specified on the schematic, but the board overview in the service manual (and Yves Usson’s photo of the board) has it as 1K.
Here’s a video demo of the envelope follower, using the drum loop from “being boiled” as the followed signal, with the env follower output going to the System 100 envelope input socket. See how the red light goes on and off!
I’ve never found much use for envelope followers but maybe I’m not just creative enough with them. If the signal volume on the input gets boosted enough it bleeds through to the output, but at a normal sort of level there’s no problem.
The integrator works fine, but with the specified values when using it as a CV slew for a VCO it needs a fine touch to dial in the right amount. I used a CA3140 instead of the specified (ICL?) 8007.
The amplifier does the amplification thing as you might expect, and could be really useful (although I regret not putting a switch on the panel for DC coupling).
Also made a mistake on the panel – it should be E.F. on the switched input for the integrator. Had to correct the spelling of “integrater” too, couldn’t let that stand.
…and round the back – aluminium bent up at the edges to try and stiffen the panel, though it’s not so bad at this smaller size.
I’ve got better at arranging the board-to-panel wires since I did this, it looks a bit higgledy-piggledy.
I’ve had this board hanging around for…about three years? That’s how excited I am by it. This is an old photo, ‘cos it currently has about an inch of dust on it.
It’s all a bit kind of just OK – the phase shifter is alright, the spring reverb is nothing special (nowhere near as nice as the Tombola spring board) and the test oscillator is my own version and that sounds… acceptable, as far as a test oscillator goes.
For the spring, I replaced the TA7136AP with a stripboard adaptor using a TL071 and kept everything else the same, as far as possible.
There’s no specification for the reverb tank as far as I can see, so I can only guess Roland used the Z-3F tank they used in the Space Echo and System 100 103 mixer. Somewhere I’d seen suggested that the Accutronics 8AB2D1A would do the trick, so that’s what I’m using.
I did some underwhelming demos at the time – here’s a familiar-sounding bassline going through the spring:
And a blippy-ish loop, starting from dry:
For the phaser I roughly matched the 2SK30A FETs as far as I was able to using the method outlined by RG Keen in JFET Matching for Effects. They’re not super-cheap, so it’s expensive to buy loads just to match eight.
These demos might be an even greater test of your patience – hold on: here’s some twiddling with white noise:
And here’s an annoying loopy bassline, first with the resonance turned down, then turned up:
For the 1KHz and the A octave test oscillators I couldn’t find any of the large variable inductors, so I went with a couple of phase shift oscillators with a bit of extra simple RC filtering afterwards.
The octave divider chip is an LM3216 – at the time I designed the board, this was fairly rare (£12 from Little Diode), so to avoid using a sought-after IC for such an unexciting purpose I designed the octave divider around the CD4024.
I say “designed”, I probably nicked it from somewhere.
Here’s a sample of the 1KHz test tone – three seconds is as much as I can take
Looking at the spectrum analysis the odd harmonics are accentuated, so I guess it’s a bit squarer than it should be, but it’s OK.
The board still needs a front-panel, and I’ve spent way too long trying to work out whether it’s worth doing a PCB or just a stripboard for the mixer section.
Either way, it’s the summer and while it’s not piddling down or throwing a gale I’ll be out riding my bike.