Sunday, November 10, 2013

Envelope Response Comparisons

While developing the Andromeda A6, we tested many vintage synths to gain insight into their particular magic. Here are graphed responses to the envelopes in the Roland System 700 and the Oberheim SEM. I believe they are both readings of the output amplitude curve. The 700 was set to minimum time settings to see how fast it was (much faster than the Minimoog, for example). The SEM shows an amplitude decay curve which is interesting...the ear prefers exponential, as this is how sounds in nature decay, but its opposite, starting slowly then speeding up approaching the end of the cycle, slightly mimics what you hear after compressing or limiting a sound.

Sys700_Amp_Env

The Roland System 700 ADSR set to all minimum times. Total duration is less than 1millisecond. Note that the Minimoog's minimum attack time is about 1 millisecond alone; this is important as hard voltage changes of less than 750 microseconds cause the Minimoog VCA to click.

SEM_Amp_Env

The vintage Oberheim SEM.

Not all envelopes are the same. We noted that the Roland SH-5 ADSR's decay slope is steeper (begins more quickly) than the Moog's.

The Buchla modular envelopes are typically exponential. The 281 envelope is linear, in that it begins and ends at the same rate, a straight line across a time graph. As with any voltage-controllable envelope however, its slope is thus completely variable from linear to steep regular exponential to its reverse as seen in the Oberheim graph, through routing the voltage output of the envelope back to its VC inputs through inverting attenuators (attenuverters).

Nods to Grant Richter for explaining the last part in his SDIY pages.

Alesis A6 Andromeda Voice Architecture

Click on the images for larger versions. These are the contents of the analog oscillator and analog filter ICs, with two voices per chip.

A6_ASO_Large

A6_ASF_Large

Thursday, November 7, 2013

Buchla 190 Dual Reverberator DIY


The 190 dual reverberator is called "the" Buchla 100 sound. It's entirely discrete using only transistors and transformers with a single Amount pot crossfading between in input signal and reverb return. Note that it uses a +24V supply but that's provided in the 100 power supply. If you don't have one of those, small +24V supplies are pretty inexpensive in the form of a wallwart from Mouser, etc.
 
I have been told that this is a single-ended Class "A" 1W amplifier (into the reverb tanks). That may partially explain why the 190 is so very good-sounding...


SCHEMATIC

Buchla 190 Schematic

Sowter 1427a transformer wiring:

Yellow- +24V
White- Collector pin of 2N3053
Green- Ground of reverb send to tank
Brown- Audio of reverb send to tank

Red and grey are unused.

Matching to the Triad TY30X transformer colors listed in the schematic:

Triad:    Sowter:
Brown   Yellow     to +24V
Blue      White      to Collector pin of 2N3053
Black     Green     Ground of send wire to reverb
Green    Brown     Audio of send wire to reverb


ETCH ARTWORK AND PARTS LAYOUT

Download HERE


COST:

2) Accutronics tanks: average $30.00 each; "MOD" versions are less but sound different.
2) Sowter transformers: $110/pair, which is significantly less than vendors charge for the original TY30X/Magnetek.
2) Quality heat sinks: about $30.00. You can use less-expensive sinks but this should extend the life of the 2N3053s.

This build will likely cost over $200.


CABINET COMPATIBILITY:

The transformers are 30mm tall by themselves, so figuring with 15mm standoffs for the PCB plus PCB and transformers, 50mm or so deep behind the front panel.


Clone using original Triad transformers:
Buchla 190 clone


The main issue with doing a 190 build used to be the Triad TY-30X 2W audio transformers. These were also used in the 410 module as well as the 212 Dodecamodule (for the headphone output).

TY30X

Here it is in a 1961 Triad catalogue:

http://www.triadtransformers.com/images/bigArchive/TR-61%20pg-010.jpg

And in 1978:

http://www.triadtransformers.com/images/bigArchive/TR-77~78%20pg-030.jpg

"Output and driver transistor transformers: High level

Primary matching impedance: 100CT
Primary current: 100ma DC Balance
Maximum level: 2W
DC resistance Ohms DC: Secondary: 8/4 Ohms; Primary: 14.4 Secondary: 1.1 Overall Turn Ratio: 3.54:1
Freq Response +/-3dB: 50-10,000"

MageneTek purchased Triad a while back and have an equivalent listed but I've yet to manage to source it in small quantities / initial order expense. They no longer manufacture transformers.

Triad tell me if they still have the technical papers, it would cost $2500.00 to re-manufacture a run of them. I'd contacted both Hammond and Edcor looking for equivalents to no avail. Wiggler Mike Silver contacted Sowter and Brian quickly came up with a working clone!

Buy a pair of Sowter clones of the TY30X HERE. It is of course the 1427 Transistor output transformer. It is far less expensive than originals from vendors if you can even find them. I paid more than twice as much for a pair of originals and have seen them at far higher prices than even that! I've run a Kraftwerk track through the circuit with both transformers wired in and switched between them; they are really close, from this test on headphones. I haven't checked to see if they begin to distort at the same level.


The 100 series apparently used the now-rare Mallory "white" axial electrolytics (100uf 15V; 10uF 35V), which =might= play a tiny part in the overall sound.


Mallory 200uF tedssdotcom

For illustrative purposes, a 200uF Mallory of the type used in the vintage module. Image from tedss.com, who do not appear to have these in 100uF 15V or 10uF 35V.

No space is provided here for them if you can source them; instead, common types may be used although it in the interest of "fidelity" it is recommended to use Elna Silmic II "audio" type capacitors, being the budget choice for higher performance over common caps. Most of these caps used in this build should be radial, not axial types.

The 190 also used several huge Mallory film capacitors, .1 and .0022 values. This also may play a part in its character, as well as that there are several tantalum caps involved in the audio path (1uF, 4.7uF).

There are two types of Tantalum capacitors in the original module. The 1.0uF are mainly Sprague

190 Parts 1uF Sprague Tantalum

1uF (C

190 Parts 1uF Tantalum version 2

1uF (C

190 Parts 4.7uF Tantalum

4.7uF Tantalum

The same is true of the three film capacitors, 0.1uF and 0.0022, which are C8, C9, and C10. Originally they were the very large Mallory polyester types, a very deep red coloration. The 0.1uF were 100V and the 0.0022uF I believe was 400V or so. You can certainly use Orange Drop equivalents (specifically polyester types) but you'll have to crimp the leads as adequately pitched (spaced) solder pads are not provided.

Several of the resistors are carbon composition types: 470R, 4K7, 47K, the 33R 1/2Ws, 1K, 10K, and 220R. I do not know if this was due to any particular strategy or finding on Don's part. His 258B schematic indicates specifically that certain values were to be carbon composition types; again, it is unknown "why". There are only a couple CC types in some of the later 258Cs I've seen.


190 Parts 2N3055 TO5 Can

TO5/TO39 can 2N3053 mounted on the 190 PCB. 

Edit: I am told that this is a 2N3053, which is also NPN, and is easily available in the above form, which is TO-39, from Mouser. The hand-written schematics are for the most part not a problem from which to extract parts numbers and values; in this case, the kicker is that there probably were never TO5 2N3055, which should have alerted me immediately. You can just barely tell that the schematic says 3053 and not 3055. Very sharp eyes...Thank you!

2N3053 from Mouser (Temporarily out-of-stock and it's the only one available from Mouser.)


The 3053 run exceptionally hot so use a good heatsink such as this: 532-1130B or 774-TXBF032025B.


TEST THE ETCHED PCB FOR CONTINUITY

After etching, confirm continuity with a VOM set to diode (hopefully one which beeps so you won't have to constantly refer to its display).

190 Continuity V2


COMPONENTS:


Buchla 190 Clone V2 Parts


BUILD: 

The 33r 1/2W resistor mounts on the trace side if you are using the large Elna Silmic II audio capacitors. Nichicon FG or KZ types are perhaps even more desirable and they're smaller. 

It is important to attach the heatsinks to the 2N3053 prior to soldering them to the PCB. Also stand them about 5mm height above the PCB so they can breathe a bit.


REVERB TANKS

In the 100, the tanks were mounted inside the cabinet against the inside left and right sides. Terminal strips mounted inside the cabinet to the wood assist in the wiring task. I am told that the closest, if not exact modern equivalent are the Accutronics Type 4 with the following descriptor / part number:

4AB3C1B

http://www.accutronicsreverb.com/?PHPSESSID=b9752031617a105db2e4ad3e826ee4f0

They are easily available through eBay vendors and online guitar stores. Note that these are larger and are different than the tanks used in the 208r clone/group buy/etc.

Accutronics tanks are available and there is a newcomer called MOD who offer one quite similar, by the same name and spec, but it is typically less expensive. Here is a video I found of a comparison. The comments section indicates that the mounting of the new Accutronics might make a difference sonically...I find the MOD to be less blurry like you'd expect from classic reverb, and it's a touch metallic and meaty. The Accutronics to my ear has the blur you'd expect from a spring reverb or acoustic space.




WIRING

A wiring doc will be published shortly. Not much is a mystery but I'll provide visuals.

It is recommended to use existing long RCA male to RCA male stereo cords for each circuit's send and return to the reverb tanks. The tanks are color-coded with White input jacks and Red output jacks so maintain this continuity for ease of setup and tear-down.

The transformers can be mounted to the PCBs to save space if needed, but they're noisier that way, than if you mount them away from the amplification circuitry. If you do mount them on the PCB, I suggest using mu metal to seal them and to ground it as well.

A terminal strip can be mounted at the rear of the PCB. Tie the two +24V wires to one terminal input. The reverb send wiring may be done in the same way. Obviously, you'll want at minimum a 5 position terminal. In the original 15 module space wooden 100 Series cabinets, the audio transformers were mounted on the inside top, and used terminal strips to connect to both the PCB and reverb tanks, each of which were mounted inside of the left and right cabinet sides/panels.


POWER

The 190 uses Ground, +15VDC, and +24VDC.

Get an inexpensive 24V supply here. 1A might be too much but I believe in headroom.


SUGGESTED FRONT PANEL ARTWORK




REFERENCES / FURTHER READING:

Check out Mark Verbos' installment on the 275 dual reverb for a pic of the later mounting style: naked springs suspended in an aluminium boat atop each other along with the driver electronics.

http://buchlatech.blogspot.com/2012/11/spring-chickens.html

Errors in the CBS Buchla Schematics?

Carbon Composition Resistors in Buchla Gear

Huge thanks to JonDent, Benge and Jason Butcher for information and guidance! And "blame" Opsysbug on Muffwiggler.com for believing I could do this, which is why it now exists. ;) Thanks to Mike Silver for finding the right company to clone the transformers!


Boat