>
YAMAHA
STRINGS
SS-30
RACK-MOUNTED WITH MIDI
MIDI STRINGS

Saturday, July 25, 2020

Fun With Filters (Pt.5) - Old Crow CS-80 module edition

This is the final Fun With Filters post. I started with the ConcertMates post, showing how the SS-30 could be used with an external filter. Then the fun really started with three parts documenting my Yamawah filter module based on the Yamaha IG02611. That was fun but the conclusion was that it wasn't good enough for the SS-30 as a companion filter.

In the fourth part I took a look at how the envelope of Yamaha CS synth filters works and invested in the ALM Busy Circuits Quaid Megaslope to be able to replicate that behaviour.

Old Crow 480


During this time I was clearly trying to find a Yamaha CS type of solution and that led me to considering a few options. These were limited to the discontinued Gotharman CS filter based on actual IG00156 ICs, the Studio Electronics SE88 VCF and the Old Crow 480 filter. The SE88 was a strong contender, but in the end I decided I wasn't happy with the price and I wasn't really sure about the tone of it. I don't like the design of it either. There, I said it!
That leaves the Old Crow DIY option. This won for several reasons: price (around £200); DIY is satisfying and lock-down means I have the time (and the need) for projects; Old Crow is a expert on the CS-80 and his design is not only very well respected in it's 5U modular form - the MOTM480 - but is used in the Deckard's Dream CS-80 (don't call it a clone) inspired synth. The Eurorack form was adapted from the 480 and is actually an impoved revision.

Old Crow's Synth Shop was one of the first synth site's I remember exploring in the early days of the web. His exploration of the CS-80 in particular was fascnating and had lots of photos of what was going inside, which was interesting for someone working on a contemporaneous product.
He really know's the CS range and has done more than most to explain it's mysteries. The following quotes are  from Old crow posting on the Muff Wiggler's forum.

The thing to remember is, the filter is only one part of a much larger signal chain in the configuration of a CS-50/60/80. The voice chain is a *minor* portion of the "CS sound." The greater amount of instrument character exists within the user performance controls. The touch response and rate scaling, as well as the timbre changes imparted by the wah pedal and chorus circuits are the trademarks of the instrument. If you were to construct a CS80 filter from IG00156s as a standalone module, it would sound just like my modules. I know, because I did exactly this in 2000~2001.
In fact Gotharman did create a module from IG00156s, but both the module and IC are very hard to find.
I think I was aware of him and Paul Scheiber (design engineer of the Moog MG-1) reverse engineering of the filter for Synthesis Technology, but the 5U world is something I've never paid much notice to. The process of cloning was mostly a matter of trial and error.

The MOTM-480 and RR-480 aka MOTM 480 mkII are the result of me spending a long time hand-selecting parts to test as a CS filter in place of the Yamaha parts in my CS-80. In the end, I got it as close as can be brought into this world.

The Eurorrack version of the 480 filter came out in 2015 and was created at the request of modular synth artist Robert Rich, it seems, hence that version is called the RR 480. Deckard's Dream is inspired by the CS-80 and apparently they licensed the '480 design from Old Crow for their filter.
The Deckard's Dream uses my RR480 circuit. I calibrated this for Roman about this time last year. The DD is not a CS-80, but the CS-80 is likewise not a DD. With a bit of configuration and sound tests, you can get one to sound like the other.

Black Corporation, who make the DD are now poducing a Eurorack mono-synth version called Deckard's Voice. This includes not only the '480 filter but also the real deal CS envelopes, as well as the oscilators and LFO. It would actually be a pretty good solution for me, but the price of $699 is above what I want to pay and and the RR480 has more options for the filter, overall.

The PCBs and a panel are currently available from Old Crow's CS-80.com shop and Modular Addict. I got mine from Modular Addict (I can't remember why) and started work on the BOM. The build was quite easy, but the BOM is wrong and the design overall wasn't quite to my liking. Below is a guide to the build which I wrote up and posted to the Muff Wigglers build thread.

Anyway, it wasn't hard to construct and it worked first time with no issues at all. Compared to my slightly over wrought drum modules this was a breeze.






RR480

The RR480 is a module consisting of two state-variable (SVF) filters. Both filters have low-pass, band-pass and high-pass outputs. The are two modes which are selected by a toggle switch.
HPLP is a preset serial HP to LP configuration where the HP1 Out is routed to SVF2's input, just like the CS. In this mode the LP2 Out is assumed to be the output, although HP2 and BP2 outputs are still active.
In the Split mode there is no routing between the two filters and they can be patched either completely seperately, in parallel, or together, in series.

Inputs

There are three audio inputs and each has a level control. In the HPLP mode 1, 2 & 3 are all mixed together. In Split mode they are, yes, split, 1 & 2 to SVF1 and 3 to SVF2. These could be labelled more clearly, but once the layout is clear in your mind it's not a problem.

Control Voltages

A single set CV inputs for cutoff frequency and resonance control both SVFs. Cutoff frequency can be   controlled 'V/Oct' with positive going CV signals and 'FM' for bi-polar for modulation CVs. Resonace CV is also bipolar.

Tracking The Spread

Tracking by the Volts/Octave input is not the same for both filters though. SVF1 tracks 2V/Octave and SVF2 tracks 1V/Oct. This is how Yamaha did it. The HPF tracks differently to the LPF and . Apparently this can be changed to 1V/Octave with a resistor. There is also a resonance CV input.

Irwin VCA

Handily, the RR480 also has a VCA, which uses up some of the spare OTA capacity left over from the filter design.
The VCA is: linear; inverting; has two inputs, VCA IN1 has a level control, VCA IN2 can be switched between AC/DC input; and a CV input with 'Initial Gain' level control.

This is the so-called Irwin design of VCA, although that doesn't matter.

Drop the BOM

There are two notes on the BOM
Note 1
R33, R33, R34 - 150K - DNP (do not place) on the mainboard, instead place on the Host Board
This has no effect on the BOM as the parts are still located on one of the boards

Note 2
R75, 76 - 150K - Only place if input signal (audio) is >5vpp
Eurorack is 10Vpp for audio, so these should be fitted.


Issues with the BOM

Missing/Wrong

These are the errors on the BOM which should be updated to match the PCB, or you will end up with missing parts when you order (unless you have them in your own stock):

C47 - 100nF 0805
C37,C40 - 180pF not 100pF - I used some large old radials which I had in stock.
C49,C50 - 10uF 25V 0805
R35 - 68K not 100K
R101 - 3M3 not 4M7
R55, R58 - 130K 5% not 200K - I used 120K

JP30 - is SW1 and the jumper is not needed.
On Schematic Filter (fig 1) it is next to VCA IN 2 JP21 and JP23A
Annotation is 'AC (open) DC (closed)'.
On Host board schematic (fig 4.) it relabelled SW1 with the same annotation.

Obsolete:
Regulators LM78/7910 are not available anymore.
Equivalents are available though and I just grabbed whatever was rated right and fitted okay.
Issues with panel
Jack sockets
The 3.5mm jack sockets specified are CUI PJ-301B (Erthenvar PJ-301B)
http://erthenvar.com/files/pj-301b.pdf
The height of this part is 10.5 - board to panel.

I bought Thonk PJ301BM, with the same height.

However, the actual panel to board height must include the 'collar' between the top of the base and start of the thread. The Old Crow panel does not cut out for this collar and so the panel to board height is actually 11.5mm.

Potentiometers

The 9mm pots in the BOM are Smallbear 1012A 1203 http://smallbear-electronics.mybigcomme ... -pc-mount/

That is Alpha RD901F-40-15R1-Val. https://www.mouser.co.uk/datasheet/2/13 ... -40-15R1-B(resistance_-1627810.pdf

These are the same as the parts I got from Thonk and are 10mm board to panel height.

I had to buy some M7 1.6mm nylon washers to get the heights to match.
x3 of the washers or x1 nuts that come with the Alpha pots also work, if you have plenty to spare.




Switches

There's a similar issue with the switches. I used x2 spare jack socket nuts to fill the space, although they are too large and something else more suitable could be found.

Additional Parts

Nuts and bolts for the regulators. I used M3 from stock.

11.5mm stand offs for joing boards.
Old crow said " I also went with 12.7mm nylon board locks, much faster to assemble vs. screws, pillars, etc.", but that is too high and you can see there's a gap in connectors. I ended up using M3 10mm brass threaded pillars and some locks washers. It went together quite snug and whilst not essential I'm glad it's all secured.



Curses

I ordered the wrong 3u3 caps. They should have been bipolar and I ordered polarised. Grr.
Weirdly I orderd bipolar parts for the DC section, which should have just been cheaper polarised caps. Go figure. I probably miserad the BOM as they are next to each other.

Panel Scales

And finally, the scales on the pots are 270 degrees but the Alpha pots are 300 degrees. Modular Addict don't have the .fpd file but if you order from CS80.com they say you can get that and, if it offends you enough, make your own.

No comments :