The Moog 905 reverb module is one of those parts of the modular system that typically never got as much attention as some of its flashier counterparts like the filters or oscillators. For those in the know however, it went on to play an essential role in helping to create the patches that have become a part of the popular canon of sounds and effects that the instrument has become known for.
Back in the 60's, reverb springs were a pretty new and exciting way to spice up the sound of an electronic instrument. Guitar amplifiers, home organs and even PA systems utilized them to add spaciousness to an otherwise dry sort of electronic sound.
This was a relatively cheap and easily portable solution to mimic the sort of reverberation effects available in a professional recording studio, which usually used a large hard-surfaced reverberant room or specially treated acoustic chamber with a speaker and microphones inside to create their sound.
Reverb springs work by using one or more actual metal springs suspended between two electrical transducers. The input transducer is hooked to circuitry that actually shakes or twists the springs in response to an electrical signal...the bigger the signal, the greater the movement of the the spring(s). The signal travels down the springs, and is picked up by another transducer that converts this motion into a small proportional electrical signal that is then amplified and mixed with the original signal. The delay in this signal caused by the time it takes to travel down the springs, plus the fact that the signal "bounces" up and down the springs several times as it decays gives it its reverberant quality. Shake or smack the springs, and you get a great explosion of sound as they clang around!
Bob Moog realized that this technology would be valuable in creating space and realism (or even unrealism) for electronic music, and so the 905 Reverberation Unit was born.
Using it was simplicity itself...it had an input and output jack, and a single knob. When turned completely to the left, only the dry sound of the instrument was heard, along with some coloring from the electronics involved (more on this later). As the knob was rotated clockwise, the dry sound was slowly lowered, and the reverberation effect from the springs was slowly added to the signal. At the full clockwise rotation of the knob, only the output from the reverb springs was heard. This allowed for a wide range of effects that added the reverberated sounds to the original dry signal in various amounts.
There has always been some confusion about the actual circuitry involved in this module.
This early 905 schematic, dated from mid-1966, details most of the circuitry fairly well, but the value of the resistor connected to the collector of transistor Q3 is missing from the document. The mix knob is either drawn incorrectly, or missing the necessary info to show that the dual-ganged potentiometer elements must be connected to operate opposite from each other for proper blending of the two signals to occur.
This module's schematic was also included in the Norlin Modular Service Manual, but it was reproduced with a number of nasty mistakes (as were some other schematics In that collection as well) that have continued to circulate around the web. Some resistors were shown improperly connected, some resistor values were not representative of what was actually being used in later production models, and the dual element blend pot was again drawn incorrectly connected so that no blending of the two different sounds would occur. The resistor connected to Q3 is still shown here with no listed value. The input to Q2 is shorted!
So, not much help.
Trying to build a working unit from this schematic would be frustrating, to say the least...
Over time, the schematic has had some much-needed corrections made (shown here in red) and was reposted by various individuals, but there are still errors present in these modern day versions, including a resistor on these newer models that is not shown in any of these schematics.
After a lot of sleuthing on the web, examining various circuit board photos of these modules, and more lately seeing some excellent photos at the Moog Forum posted by an individual called Rezin, I've been able to piece together what I believe is the most accurate version of this schematic to date.
For those like myself who are planning to recreate one of these modules for their own system, this may be very helpful in capturing the magic that this module brings to the party!!!
After examining the various photos available, I created this stuffing guide using a picture of the copper side of the board to enable me to see just how the various components were really connected together "under the hood" in an actual production model unit.
At this point, I began to notice some differences.
In all of the schematics up to this point, the circuitry connected to the output of the springs was shown as using the +12 volt buss as the positive supply, and the spring chassis and the return of the associated components was connected to the -6V buss. All other electronics in the module used only the +12 volt buss and ground for their connections.
Looking at the 2.2k resistor and 75uf capacitor connected to the emitter of Q3, you can see that they are actually connected to the ground trace on the PC board, not to the -6 volt supply rail.
The mounting plate of the reverb assembly and the chassis of the springs are connected to the uppermost trace on the board... you can get a better look at how the plate is mounted in these pictures of the unit. Notice especially the soldering lug that connects one of the plate mounting screws to that uppermost trace...
So, at this point I was thinking, "Perhaps this trace is the -6 volt rail shown in the schematics?!??"
Looking at some better pictures of the various modules available, it looks like that trace is connected to ground as well, through the shield of the cables attached to the reverb springs.
OK, so just where is the -6 volt supply rail on this unit???
In fact, as you can see here, pin 3 of the power connector (-6 Volts, 3rd tab from the top) is not connected to the circuitry at all in these modern examples of the 905 module. You can however see a red wire connected to pin 1 (+12 Volts, top tab) and a green wire connected to pin 2 (ground, 2nd Tab from top) if you look closely.
This makes sense to me because if -6 volts was actually connected to the main chassis and springs, it would be fairly easy to short this supply to ground, as the springs tend to poke out quite a bit inside of the module.
Yikes!
This might be one of the reasons this was changed in ongoing versions of the module.
Having the extra headroom that connecting this circuitry to both supplies provides doesn't make a lot of sense here anyhow, as the signals that are generated by the spring transducers are fairly small, even if you shake the springs vigorously. It would probably make more sense to do this at the front of the circuit, as the single transistor is driving a fairly low impedance load and could use the extra "oomph" from a larger supply, and the input to many spring assemblies is often isolated from chassis ground and left floating for just such a purpose.
I've read accounts from various individuals that this module adds a subtle but pleasant amount of coloring to the oscillator signals run through it, even if no reverberation is applied, and it is often used in this way as a "thickener" to the overall sound. This is probably due to the nature of the electronics being powered from a single ended supply, and the characteristics and biasing of the transistors used in the input and output stages. This is, in some ways, much like the acclaimed discrete transistor mixers used in in the earlier modular systems, which added their own pleasing element of distortion to the sound, delighting modern listeners. (Cha cha cha)
The electronics in the 905 module also invert the phase of the dry signal passing through it, so if it is mixed with the original signal source this will cause various cancellations as well, depending on the settings.
So, after all this investigation, intimation, instigation and implication, this is the schematic that I believe now represents the connections and component values of most production models of the 905 reverb.
Using the pictures of various production models seen on the web as a guide, there have been some component value changes from that early 1966 schematic, and one addition.
The input resistor R1 was changed from 33k to 15k.
The resistor R8 was changed from 100k to 220k.
The resistor R9 was changed from 220k to 330k.
The resistor R15 was changed a number of times during production. On some units I've seen it was a 330k, or a 470k. I show it as a 150k in these schematics because that seemed to be the most common, and that's what was used in the unit photographed recently by Rezin. Any other changes might have been dictated by the transistors used in that particular example, which are usually not readable in most of the photos I've seen.
A 100k resistor R17 (not shown in any previous schematics) was added from the module output to ground.
In the unit photographed and examined by Rezin, all transistors are NPN types 2N3391A.
The reverb springs available back when this was designed typically had an 8 ohm input impedance, and a 2-3k ohm output impedance. There's lots more choices nowadays, including multi-spring units with varying decay times, so some experimentation with different units might help you create your own perfectly reverberated customized listening experience as you roam this nation's highways and byways....
So, for those intrepid individuals who would like to build their own unit (or repair a janky one) I hope that this will help you in your efforts. It helps lessen any hum pickup if you keep the reverb springs as far away from the power supply and transformers as possible, and use good quality shielded cables to make connections to and from the unit.
Many, many thanks to Rezin and Tom from the Moog forum for their most excellent photos and insights into this mysterious unit (as the dry ice fog slowly clears).
So, join us again next time as we go ...
FORWARD INTO THE PAST!!!!!!
Posted with Blogsy