EML Electrocomp 101

Electronic Music Labs (EML), based in Vernon, CT, was a rather strange but interesting outfit that, for a brief period in the early 1970s, had some success in the commercial synthesizer market. The company was founded and largely run by electrical engineers rather than musicians, an attribute with both strengths and weaknesses.

EML 101

The EML 101 was designed to compete against the Arp 2600 and the Minimoog, and in several ways it was a superior instrument. For one, it cost considerably less. For another, the instrument was built almost entirely with mil-spec parts, and housed in a very sturdy (if not particularly stylish) wood case with a heavy-gauge aluminum front panel.

Architecture

The 101 has four oscillators, one more than any of its competitors. Two of the oscillators can be switched to an LFO range. However, to economize on the design, the EML engineers observed that many players use multiple oscillators primarily to reinforce each other, tuning them either in unison (for phase cancellation effects), or in fixed intervals. So, although four oscillators are provided, only two expo converters (circuits that make the oscillators respond in a volts/octave mode instead of a linear mode) exist and must be shared by those four oscillators. For this purpose, the 101's designers created a rather complicated control bus architecture.

The duophonic keyboard can generate two control voltages. Each of these is fed to one of the two control busses, which are called KB1 and KB2. Modulation comes from the oscillator control mixer. (Oscillator 1 also has a dedicated external modulation input.) A rotary switch on oscillator 2 allows selection of either the unmodified KB1 control voltage, or the KB1 bus added to the oscillator control mixer output, as its input. Oscillator 3 can select whatever is selected for oscillator 2 (this is designated "CM2"), or the unmodified KB2 control voltage, or the KB2 voltage added to the oscillator control mixer output. Oscillator 4 can select whatever is selected for oscillator 3 ("CM3"), or the oscillator control mixer output (not added to any keyboard voltage), or strictly manual control usng the oscillator's octave and fine tune knobs.

One thing the 101, straight from the factory, wasn't very good at was interfacing with other synths. EML was the only manufacturer that ever used the 1.2V/octave scaling standard for its VCOs. (It sort of makes sense if you realize that this equates to 0.1V/half step.) What was worse, though, was that the only external control voltage inputs available were the CM3 input, which doesn't scale the same as the rest of the synth, and an external input for oscillator 1 which is linear. I have modified mine so that the CM2 output jack at the upper left is now an input, and it takes 1.2V/oct scaling. A JKJ Electronics CV5 MIDI-CV converter, tuned to the 1.2V/oct scaling, drives it when I'm not using the built-in keyboard. I have identifed in the schematics how to modify the CM3 input so it scales properly (it's just a resistor substitution), but I haven't been motivated to do it yet since I seldom use the duophonic mode.

Servicing the 101: One nice thing about the 101 is that it was obviously designed to be serviced. The case is easy to open -- if you know which of the dozens of screws go to what. The trick is: ignore all of the screws in the case. The only screws you need to undo are five Allen screws on the front panel: the two large ones at the left edge, the two large ones at the right edge, and the one at top center (above the FILTER CONTROL jacks). Do this, and the panel pops out. And the nice thing is: all of the electronics (except for the keyboard) are attached to the panel. Nothing is attached to the case. Take a couple of lengths of 2x4 scrap lumber, saw a kerf in each about 3/4-inch (2 cm) deep, and place the panel in these slots. Now, the panel will stand up on your workbench in the 2x4s, and you can operate the synth completely out of the case. See it at the left edge of this not-very-good photo:

Back of 101 in Homemade Fixture

The circuitry is contained in three boards, plus the power supply module. As seen from the rear, going from right to left, board #1 contains the keyboard interface circuitry, control voltage exponential converters, oscillator control mixer, the control voltage bus switching, and the envelope generators. Board #2, the middle board, contains the four VCOs, the noise generator, and the filter control mixer. Board #3, the leftmost board (behind the Panavise, resting on the bench top) , contains the VCF, the VCA, and the sample-and-hold circuit (which the schematics refer to as the "effect sampler").

Connections between boards are made with point-to-point wiring, with the wiring connected using push-on terminals. Many colors of wire are used. (Other than black being used for ground and return, there doesn't seem to be any specific color code.) This makes it fairly easy to trace connections between boards. All of the front panel controls are individually mounted to the panel itself; there are no board-mounted controls. With the boards mounted on standoffs behind the panel, it is fairly easy to access boards, jacks, and panel controls.

One thing that is odd about the 101's design is the configuration of the sample-and-hold circuit (which is labled "sampler" on the front panel). The S/H is provided with its own clock oscillator, adjustable from extremely slow to about 30 Hz on the front panel. An external trigger input, which overrides the internal clock, is also available. However, for some reason the designers chose not to provide a patch point for the signal input; it's hardwired -- and not to the noise source, but to oscillator 4. Now, this configuration can do some interesting effects, but one thing it cannot do is traditional random-note or random-filter effects, since the input isn't random. The photos below shows what I did about this:

Board #3 with S/H Input from VCO4 Disconnected

This photo shows board #3, which contains the sample-and-hold circuit. A pink wire, which has been disconnected, can be seen at left center. This is the wire that brings the signal over from VCO 4 on board #2. It's a fairly simple matter to splice a switching jack into this connection.

Jack Spliced into S/H Input

The two pieces of light blue shrink tubing shows where the pink wire has been cut into for the modification. The half coming over from VCO 4 has been spliced into a dark green wire that goes to the jack mounted in the panel at top center (with the empty hole to its left). A red wire comes from the jack and is spliced into the other half of the pink wire, which has been reconnected to its normal termination point. The jack is configured as a normalling-type jack. With nothing plugged in, it allows the signal from VCO 4 to flow to the sample and hold signal input as usual. Plugging into the jack interrupts this connection, and sends whatever signal is coming in through the jack to the sample and hold.

Stealing Jacks from the Mults

I didn't want to punch holes in the front panel; it's in good shape and I have a general aversion to altering the appearance of vintage gear. So what I did is steal a jack from one of the blocks of multiples. The stock 101 has two blocks of multiples, each containing four jacks that are connected to each other. It allows an output signal to be "fanned out" to multiple destinations. However, there are plenty of other ways to create multiples. Taking a block away creates four locations for additional input or output jacks, without messing up the panel. Here I have removed all four jacks from the rightmost multiple. The one installed jack is my new signal input jack for the sample and hold. At the time I shot this photo, I had reserved the other three jack locations for tapping into the multimode filter...

The 101's multimode filter with its blending control is neat, but I thought it would be even more useful if it had individual outputs for lowpass, bandpass, and highpass. Easily enough done; the 101's filter does in fact produce all of these signals. The factory-wired output is produced by a three-tap pot with the lowpass fed into one end of the pot's track, the highpass fed into the other end, and the bandpass fed into the track's center tap. The position of the wiper determines the mix of these three signals that is output. See the photo below:

Filter Output Jacks taken from mult block

This is board #2 again; it's been removed from its standoffs and dropped down a couple of inches. The pot at right is the filter mode blend pot; the terminal at top with the gray wire is the wiper terminal. The other three terminals are at the bottom of the pot, and are unfortunately hidden by the top edge of the board. However, you can see where the trio of red, green, and black wires extend out from this point and sweep to the left and up, to connect to three of the block of four jacks at the top right. This is our hijacked multiple block, completely converted to its new functions. In the block of four, the highpass output is wired to the upper right, the bandpass to the upper left, and the lowpass to the lower left. (The lower right is the sample-and-hold input mentioned previously.)

A lot of analog gearheads diss the EML 101's filter because it doesn't self-oscillate and isn't "phat". I like it, though. It's an incredibly versatile filter, 12 dB/octave multimode. And the mode is continuously variable -- something that is not common even today, much less in '70s designs. I like to use a mix of lowpass and bandpass to emphasize sweeps and peak a certain frequency while retaining the bass. It does a nice job with highpass, too, and without hurting one's ears like some highpass filters do.

In looking at the filter schematic, I discovered that the variable multimode is actually implemented by mixing outputs from separate lowpass, bandpass, and highpass sections. It gets combined into one output by a three-tap potentionometer, having a tap at each end plus a tap at the center of the resistor track (plus the wiper tap). I soldered additional leads to each of the three taps on the pot, and took each one of them to an output jack. Now I have individual access to the separate sections. I stole the three remaining jacks from the rightmost multiple on the front panel for the purpose, so now there is no multiple there any more. Additionally, I think I know how to make the filter resonate. It appears that the circuit has an anti-oscillation feature that reduces the internal gain as the resonance is increased. (Actually, you can make the filter oscillate if you use external gain and feedback. But it sounds awful that way.) It appears that by just removing a resistor, or substituting a trimmer, I can alter the internal gain profile and make the filter oscillate. Thoughts for a future bench session. If I do this, I will look for a way to add a switch to turn it on and off, because I'm afraid that changing the anti-resonance circuit will change the sonic characteristics of the filter. Because I don't want to drill a hole in the panel, this is a bit harder than it first appears. (There are other things I want to add to the panel, so what I might wind up doing is removing one of the remaining multiple jacks and using the hole to extend a cable out to an external panel.)