Music and Synth DIY

Circuit bending, Effects, Eurorack, Schematics, Synth DIY

Picoverb Euro conversion with LTC1799

I bought a couple of Alesis Picoverbs on Ebay for peanuts with the intention of using them in my modular.

Of course, anything I can think of has probably been done already, and surely enough, there’s a simple and well documented guide at Pulplogic – – which is what I followed.

Couple of drawbacks – one, it’s not skiff friendly (not an issue as I’ve only got one skiff case and it’s not going in that) and two – there’s no control of any of the parameters for the effects, only mix and level. Not much good for synced delays etc.

Control over delay times/reverb length

I got the idea of trying out replacing the Picoverb’s CPU clock crystal with an LTC1799 precision oscillator- the old circuit bending pitch control staple.

Usually, and often in 80’s/90’s hardware, this also affects the DAC, and so there’s a pitch drop and a sample rate reduction. However, in the Picoverb, the DAC must be separately clocked (seems to be in the Nanoverb service manual, couldn’t find one for the PV and couldn’t see more than one resonator) and changing the clock freqency doesn’t seem to affect the sample rate – meaning it works well as, say, a way of modifying the delay time or lengthening the reverb, without the usual artifacts.

As it’s going into the modular, I also wanted a way to modulate the clock speed with a typical +/- 5v, 10v p-p LFO.

To do that, the mod would need to:

1: Voltage clamp to 10v p-p

2: Convert the bipolar 10v p-p Euro CV to 0-5v unipolar

3: Mix this CV with a CV from a manual frequency control (0-5v pot as voltage divider): A summing mixer (inverting) adds a -5v reference voltage
with the Euro CV at .5 gain and the CV from Manual
Clock Frequency pot at .5 gain. 10k resistor for the -5v, 20k for Euro CV (0-2.5v), 20k for Manual frequency CV (0-2.5v) gives a range of 5v-0v.

4. 5v-0v would give 51mhz to 1 khz or so – the effective range of the Picoverb is 11.25mhz to around 3-4mhz. The 5v-0v needs to be scaled.

A). Attenuate to maximum desired clock frequency
when CV = 5v (eg. 5v attenutated to 4v=11.25mhz)

B). Add offset to set minimum desired clock frequency
when CV=0v

The most difficult thing was effective voltage scaling. The solution I went with is a bodge, everything else I tried (biased op amp, etc) had unintended side effects like a small window of effective range with large dead spots either side. This isn’t elegant but it works in this application. The Picoverb’s functioning clock range is relatively narrow compared to some of the old ’80s and ’90s processors.

Pros and cons

It does add a lot to the Picoverb to have control over delay times, etc, and there are some useful stereo effects from mono signals to be had. The modulation idea was partially successful: it really needs something like a slew limiter, as while sine waves, triangle waves etc work well, anything that jumps values like S&H or sawtooth waves cause loud artifacts when applied to the clock speed.

As I mentioned previously, I couldn’t find a service manual for the Picoverb, but it looks like it shares a similar “logic buffer” circuit with the Nanoverb/Nanoverb II – rather than a crystal straight into the chip, it goes through a couple of transistors. Need to investigate further whether this is the problem, or whether the chip itself just can’t handle the slewless jump to zero in sawtooth waves. If it can’t, an LFO fixed to sine waves/triangle waves would be better than a CV input for the clock speed. Haven’t tried voltage control of an LTC1799 without a vactrol replacing the pot before, which would naturally slew the CV changes to the reaction times of the LDR, so I don’t have a reference point.

There is a small problem with mains hum/ground loop, which is almost certainly due to my lack of breadboard planning and shielded cable for ins/outs. I aim to make a small SMD PCB for both Picoverbs (it’s basically the LTC, a couple of trimpots, Euro power filtering and header, and a single TL074). In fact, a TL072 might be better using a transistor for the LED buffer as there’s one of the four op amps unused. I’ll retrofit this one, hopefully adding something to stop drastic jumps in clock speed.

Needs to look prettier as well…

View schematic