This is my take on the legendary Serge Smooth and Stepped Generator (SSG). I received the SSG PC board as a compliment from Ken Stone, who designed the board. The circuit is of course by Serge Tcherepnin. Ken has an excellent writeup of the SSG.
The banana jack color coding used on this panel is as follows:
- Yellow: signal input
- Blue: control voltage input
- Black: gate or trigger input
- Green: signal output
- Red: gate or trigger output.
The SSG contains two almost identical voltage-controlled circuits that place a limit how fast the voltage can change on an incoming signal. The effect is a lag, that, when applied to a VCO pitch, will produce a portamento. That is just the beginning. Each slew has a Cycle output, which when patched to the input turns the module into a triangle wave VCO. There is also a Hold input, which takes a gate signal. When Hold is held high, the slew rate becomes so low as to stop the output from changing at all.
The board also contains two ‘one-shot’ generators that emit a short negative pulse when triggered at the input. The one-shot output normally sits high, around +15V. Patching it to a Hold input turns the Smooth function into a Stepped function, a type of sample and hold. The Serge panel did not provide the outputs of these one-shots, but internally jumpers one of them to the Hold input of one slew, thus making it into the Stepped generator. I put the inputs and outputs of both one-shots onto panel jacks, so I can have one or two Stepped Generators, by patching to Hold inputs. I added a 1K resistor in series with each one-shot output for current limiting.
Below is a picture of a one-shot on a scope. The top trace is the input square wave that is triggering it. The bottom trace shows the output of a one-shot. Each positive transition of the input results in a negative-going pulse of a short duration (about 1 ms). This transition is from about +15 to about -15 volts.
One-shot on the scope
The last circuit on the SSG board is called the Coupler. It’s a comparator that, in the original design, looked at outputs of the two slews and put out a gate indicating that one is higher than the other. I decided to make this into a separate module on the panel, and I modified the circuit just a little. I disconnected the inputs from the slews on the PC board, added 100K terminating resistors on the inputs, replaced the small feedback capacitor with a 4M7 resistor for DC hysteresis, and brought the 0 to 6V output to the panel.
I added a small board with three LED driver circuits, one for the Coupler and one for each of the Smooth generators. Here’s the board and its schematic, and also the new Coupler schematic.
The LED driver board is powered by a short cable coming over from the SSG board itself. As with the other CGS Serge boards in this panel, I used a combination of MTA 100 and flying wire connections to insure that the board can be completely removed from the panel without unsoldering. (Click this pic for more detail.)
Here’s a more detailed photo of the board. I found some beautiful old dipped mylar capacitors to use for the timing caps. The 220nf one is pretty big.
Tune Up
Each slew has a trimming pot that sets the initial slew rate. I somewhat arbitrarily set the faster slew (the one using the 100nf timing cap) to max out at about 1Khz when the knob is turned all the way up. I set the slower one for about 500Hz max. When oscillating, these circuits are not very stable. But they seem to settle down and stop wavering after a while.
I’m glad I’m not the only person who finds components to be beautiful! The panel looks tops too.
Hey Richard,
Great info as usual.
Love to see a picture fromy your banana system!
Cheers,
Dennis
Was curious were you got your connectors? I’d prefer to use those in my mods instead of permanent flying wires. Thanks -Sean
The flying wire connectors were made with Mouser part numbers 571-5-103958-3, 571-5-103946-3. These are 4-pin. They are available in other numbers of pins. They are meant for a crimp tool, but you can solder them, if you finesse it.
Hi,
Nice layout! I wan’t to make one like that but I saw on Ken’s schematic a difference between the smooth & the stepped: between u1 and u2 you have 100n/j201/820k/1M and between u4 and u5 you have 220n/j201/1M/820k … how did you manage to make 2 identicals circuits?! Thanks … (I know you’ve made that 5 years ago …)
The two circuits are not identical in my build, because I did use different values for the capacitors. I may have swapped the 820K/1M resistors in one of them, so they both look like the Smooth circuit. Both circuits have the same panel features, because the are both Smooth. I can make either or both of them into Stepped functions by patching a One Shot output into a Hold input.
By the way, the later SSG boards included LED drivers. I built one of those, too.
http://pugix.com/synth/cgs-serge-positive-slew-ssg-and-noise/
Thanks for the answer! Maybe I’ll do like your later version with the switch!
Hi Richard, please, could you describe what each “one shot” circuits return each time they receive a pulse high? The one normally wired for the lower section (stepped) makes a very short negative pulse and go back to high, to implement the “sample and hold” function, am I right?
But what about the usually unwired (F-T) “one shot” circuit? What’s it’s output when receiving a pulse high?
Hi Fernando. The unwired one shot circuit behaves similarly to the one normally wired internally to the hold input of the stepped section. I’m trying to remember (I can check later), but I think that the one used for the stepped section emits a short negative going pulse on BOTH transitions of the input, from low to high and then from high to low. So if you put a square wave into it, you’d get a sample on both rising and falling edges. The ‘unwired’ circuit is simpler and emits a negative pulse only on the positive transition of the input.
Interesting! thank you. If you ever check it with a scope and can post a picture it would be very illustrative.
I thought the stepped section would only need one neg. going pulse (sample once). With those two pulses one could imagine it will take two samples… (scratch scratch)
It needs a short, negative going pulse to do a sample. If you use a square wave to drive it, there will be two samples per cycle of the square wave. I’m going to look at this on the scope again and try and get some photos.
Hey, Fernando. I did check it with a scope. The post now has an added photo. I was wrong! The triggering happens only on a positive transition of the input, not on a negative transition. So you get one trigger per cycle when the input is a square wave. Both one-shot circuits look to behave the same, even though having slightly different designs. The only difference I saw was that the top one has a 1.25 ms pulse, but the bottom one was close to 1 ms.
Excellent, thank you very much! Great to know how it really works!
The more complicated one-shot also triggers for LARGE negative falls … the sort that the HOT coupler spits out. It is useful for the Random Voltage Source patch.