I bought a Midiverb II a while back, and when it arrived, it was dead.

It was advertised as working so I should’ve sent it back, but it was thirty quid and I thought I’m a super–DIY-electronics-pillock who totally knows what he’s doing – I’ll have this fixed before bedtime.

Four years later, and the Midiverb had fallen into a tiny gap between the guitar amp I got when I was fifteen and an Ikea shelving unit, still waiting for me to resurrect it.

Here it is in its dead state:

The dust! You’d think I’d have given it a bit of a wipe before taking a photo.

And inside – the board takes up about three-quarters of the rack.

The big square chip with the sticker on it is the magic Alesis reverb chip, and the chip to the left is the DAC.

The ICs above are for memory, and on the left-hand side there are some old-fashioned silver-topped TL084 quad op-amps, some of which might be used for output.

I’m guessing because there’s no schematic available, which makes it more exciting/fruitless, although the Midiverb III service manual might be similar in some ways. Anyway, for the Midiverb II this page on wolzow.com gives a far better impression of the insides than I ever could.

Here’s a random picture of the power regulation bit – the Alesis external PSU is a 9V AC adaptor, which gets rectified into +12V, -12V and +5V.

I optimistically replaced the capacitors in this section a while back, plugged it in, and when it still didn’t work I thought, oh well, that’s me fresh out of ideas.

Looking at the date codes on the ICs this seems like it was probably a good thing to do anyway.

I rechecked the +5V power supply after reading this post on anlage-e – seems like some of the decoupling capacitors go bad after a while – no shorts, seemed fine.

Recently I plugged it in again and boringly got the same result, all LEDs on, no LED display. Maybe something to do with the microprocessor? I wondered if the reset line on the 80C31 wasn’t going high – measured it on the oscilloscope, it seemed to be blipping high when plugging the MVII in.

Alright, next thing – clock on the 80C31, it should be ticking along at 12MHz, what have we got?

FLAT LINE. (I should’ve taken a photo, although it wouldn’t have been very interesting). At this point I dragged the probe slightly, LEDs flickered, and the front panel 7-segment display burst into life.

Tears of joy.

I worked out that if I flicked the resonator next to the 80C31 hard enough with my finger it would jump-start the clock – restarting the Midiverb would produce the reset signal and get it going.

This wasn’t much of a fix, so I replaced the resonator with a brand new 12MHz crystal.

And it’s been working happily ever since. Here’s a picture of it under my Quadraverb, on the infinitely large patch 29.

Here’s an edit of the second thing I did with it, which might be an actual track eventually. Soz about the quality, it was onto cassette tape…

The main high-ish sound is the System 100 102 expander through patch 29 on the Midiverb, the bonky bass is the 100M clone, drums are a 606, and the pad-ish synth at the end is the Poly-61M. All running off the MC-4.

Anyway, what have I learned from this? The main thing is that digital things aren’t totally unfixable. Also, the lack of a service manual isn’t the end of the world.

It looks good, but it’s not very good.

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.