Negative Tuning on the Prophet VS

One of the items on the VS wish-list from Jason Proctor (Vector Surgeon VS Editor/Librarian) is negative tuning values. In the stock OS, Coarse values range from 0 to +24, and Fine values range from 0 to +99. This ultimately makes all de-tuned patches slightly sharp. [Edit: Or, like SCI did, you can do a "negative" tuning by setting Coarse: 11, Fine: 90, for example]

I figured out a way to do this, but it's not a simple change.

You can see that the positive values are still stock, which gives 5 octaves of tuning range compared to the original 3 octaves. I did this expecting to keep factory patches compatible (eg if the original patch was set to +20, it would still be +20) but that was a dumb assumption. As the video shows, this works fine in the edit mode, but it is unsaveable as a patch.

The VS sets the bit-depth of each parameter as it saves & loads based on an array in the code. Here is the MIDI chart compared with the stock bit array:

In order for the VS to save a negative tuning value the # of bits has to be 8, for a signed parameter. The issue is that once any of these values change all internal patches and even sysex becomes corrupt since the bits are being unpacked incorrectly.

There are a few options to make this work though:

1. Rebuild the factory set by hand to be compatible with the new scheme. This would make it harder for people with their own libraries & patches.
2. Include a power-on diagnostic-style routine that loads whatever is currently in parameter memory with the correct bit-unpacking, then re-save with the new bit scheme. Users would load their sysex (it would look corrupt), restart into this mode, then resave. I am not even sure this would work, but I'm gonna give it a shot.

Prophet VS v1.4 bug

I've been informed of a bug in my modified VS OS that causes wave 126 to not be silent anymore. Looking at it now I understand the problem, and it was a dumb mistake on my part.

Not knowing quite where things were in the ROM and space being an issue, I put all my code at the very end of the ROM. That just happens to be where Wave 126 (silence) is.

I have since (unrelated to the bug) made the code relocatable and started optimizing. The original assembler Sequential Circuits used seemed to not optimize forward branches, so all were word length when sometimes byte length would have worked. This freed up tons of space in the main part of the code so this bug has already been unintentionally fixed for the next iteration.

Anybody who has already purchased v1.4 can get a free new copy of the corrected version.

PPG Wave 2.2/2.3 Rack Project - Pt. 1

This PPG has a strange history. It's recent description led you to believe that this unit has always had mismatched voice cards and PROZ board... someone's botched attempt at mixing parts from a Wave 2.2 and 2.3.

However, I found the original (or as old as I could find) listing on the Spheremusic VEMIA site. When it was sold at that time all the parts were matching according to the photos. Whether purposeful or accidental, the PROZ board got swapped sometime between VEMIA and the owner I purchased it from.

It is very easy to identify PROZ card versions, even from the back. Here is a photo of the rack PPG from the VEMIA page in 2011:

VEMIA Photo from 2011 showing matching boards

The sections I circled clearly show this is a V2.3 PROZ board. The white square is two chips that are a noticeable difference at the top of the board. The magenta square is the PPG-BUS connector for the ribbon cable and is probably the most obvious difference. Here is a photo from PPG Synth Net showing three versions of the 2.3 board:

Three versions of the Wave 2.3 PROZ board

There are other obvious identifying parts as well, like the yellow capacitor on the far top corner.

On V2.2 PROZ boards the bus connector is in the top middle of the board as you'll see in following photo. This one is taken from MatrixSynth, Reverb, and the Ebay auction where I purchased the unit. Perhaps unrelated, but the user unregistered his account shortly after selling it to me.

Recent photo of the 2.2 PROZ board now living in the rack

Notice the very different layout.

In the next post I'll talk about certain missing components and getting the 8-bit 2.2 PROZ talking to the 12-bit 2.3 voice cards.