Spring Chickens

In the golden age of analog synthesizers, everybody had a spring reverb in their system. In fact, sounding like the Radiophonic workshop is often as easy as adding spring reverb to simple sine and square wave sounds. Buchla was no exception including the 190, 275 and part of the 212, 208 and 227. The 275 in particular is a rare and interesting beast, offering voltage control of the wet/dry mix as well as simple EQ on the dry signal. 

 

The module itself sends out and receives a line level signal to an external reverb unit. Most likely, the signal is sent to a 17" x 7" x 3" box that has the reverb drivers and recovery amplifiers built inside, along with 2 15" dual spring tanks.  Any line level in/out unit could be used with the module though. The brochure reads, "The reverberation electronics and delay elements are remotely mounted. Interconnections are at line level, permitting the use of the control module with reverb units of alternate origin (E.M.T.'s, for example)." In the case of the module in my hand above, the "remotely mounted" electronics were not included. Rather than connect an "E.M.T.", the owner and I decided to whip up some electronics to use a couple of reverb tanks pulled from a 190 unit.

The electronics inside the remote unit for a 275 have a discrete amplifier, envelope follower and a variable gain recovery amplifier. In other words,  a compander is built in.  The slowed down attack makes a bit of a pre-delay on the reverb and the gain reducing as it decays takes out some of the noise from the spring reverb. Sadly, the amplifier uses +/-24 volts, which I wanted to stay away from. I created a similar circuit, using the same amplifier IC that is used in the 208 and 227, LM380, with a 275-like compander. I attached it to the back of the module and ran pigtails to the tanks, mounted on the back of the cabinet. The trims control how much action the compander has. I just trim them to have unity gain out when the signal is full strength. I wish I had made a sound clip once it was installed.

 

This unit could be built inside a box with the springs like the original one was. Since it has line level in and line level, 100% wet output, it could be put on a send of an analog mixer or on an insert of a DAW, like the vintage Roland and RCL units I have right now on my ProTools rig. I use them in my sessions and since they don't have any controls on the front to change, everything comes up the way I left it. Kinda cool I think. 

Harmonic Oscillators

Buchla made the 148 Harmonic Generator in 1969. It is a sawtooth oscillator core, the same as the 158, with waveshapers creating 9 harmonics above the fundamental. The even harmonics are created with full wave rectifiers doubling lower frequency triangles. The odd harmonics are shaped using a series of diode clippers that mix the triangle with a clipped version to make a higher frequency triangle. The effect is a wavefolder, the precursor to the Timbre circuit in the 259.





I built a modified version of that design 10 years ago. The original has no -15volt rail, so it AC couples the signal in many places. I eliminated all these caps and powered the circuit from +/-15 volts. I never got the quality of the waveforms to where I wanted them. I have since worked on a Buchla 148 and found the waveshapes to be about the same as my clone. They sound a little wavetabley, especially the higher odd harmonics.

Legend tells that Buchla made a version of the Harmonic Generator for the 200 series. It was based on the design of the 148, but had a few extra parts added. It had reversing attenuators on the CV ins and an output mixer with sliders and even/odd outs. I have never seen this module and don't even know the model number.

In 2009 I started designing my take on this forgotten concept. I made up my own waveshapers using distorted CMOS opamps, like the 259 Timbre circuit, instead of diode clippers. This does not sound good. I have since thrown the design out.

Finally, I have my finalized design. The Harmonic Oscillator. The analog oscillator core puts out saw, square, triangle, and spike (208 style). The waveshaper puts out 10 decent quality sine waves. The voltage controlled mixer has sliders and CV ins for all channels as well as scanning circuitry (sort of freq. and bandwidth for the harmonics) and tilt (to favor the high harmonics or the low harmonics from voltage).

Prototype Music Easel!?

The Music Easel is the street name for the combination of the Buchla 208 and 218 touchplate keyboard. I was recently called upon to repair the prototype unit, built in 1973. This unit was resurrected in 1987 and had since been in a basement, used from time to time without ever being moved. He built a very cool blue wood enclosure for it with a built in linear power supply.

When it changed hands, several problems appeared. The 208 module is built from a large motherboard with 12 perpendicular daughter cards, attached with .156" Molex connectors. The biggest problem was that the connectors had oxidized and the cards had to be jiggled around to get several of them to work. The 20 years that this thing sat in the same spot it was fine, but I guess shipping did a number on them. I replaced the female side on all 12 cards and things started to get better.


The Oscillator and some other parts were still not working properly, so I set out to find the problems. Being the prototype, there are several things that are not the same as other units. The panel has a sine wave in the modulation oscillator, when it actually puts out a triangle, the memory card edgecard connector is smaller, the Envelope and Pulser sliders are labeled backwards and the modulation oscillator's banana output jack covers some text. The boards are filled with cut traces and flying resistors. It was a lot of fun.


Finally, I got to the 218. It was only triggering notes while you touched one of the grounded frets. The original owner thought this was normal, but I have used these before so I had to figure out what was wrong. I found that the resistors controlling the keyboard signal's gain were different on the schematic, the parts overlay AND the picture that I had taken of a working unit when I serviced it. I tweaked these values until the keys triggered normally.

Good Ol' Filters

The Lopass Gate is widely recognized as a sweet sounding filter with a Vactrol-induced slew to it. Set into the top switch position, it is a Sallen-Key filter very similar to the Korg MS-10/MS-20 filter. The LED/photocell elements, known as Vactrols, add something around 10ms of slew on the attack and 100ms of slew on the decay. This is more than an OTA filter for sure, but what if there was some REAL slew on the CV?

Then, there's the Buchla 192. A simple, nonresonant, 2 pole lopass filter, featuring an all discrete transistor signal path with the control elements made from photocells and incandescent lamps in a little bit of heat shrink tubing! The original run (it's quite rare, but I don't know how many are out there) of this module didn't even have a CV input. It has a slew of several seconds when the cutoff is changed.


The one that landed on my workbench was living behind a Dual Reverb front panel with the labels scratched off. It was not functioning. Since I discovered at least one dead "optical element" and this device was such a hack anyway, the owner and I decided to mod it for actual VTL5C3/2 Vactrols. After all was said and done, it is a very interesting sounding lopass. Sort of like a 292 in lopass mode, but with an all discrete signal path.

Just because

I didn't know Vactrols could do that!

I recently repaired a 227. It has several problems: A dead channel, reverb wasn't working, I forget. It seemed like a quick turn around repair, until I sent it home and got the dreaded, "this thing's broken again!" phone call. Only now, it was completely shutting off the main output whenever it got hot. So, as long as I kept it on my bench, with air flowing around it, no problem showed up. After 5 minutes inside the boat, the main output went dead. This problem drove me nuts for a while, and in the end it was a Vactrol! Since the main output buss is quad, the panel volume control is actually a 4 channel Vactrol VCA. I guess diodes sometimes go open with heat if they're old enough. In this case the diode was a vintage Vactrol.

Good Times In Anaheim.

It was great to meet everyone at NAMM. I'm glad I decided to go.