Synthesizer Build part-33: DIGISOUND-80 ENVELOPE GENERATOR with AS3310.

A great ADSR with 3 different types of envelopes and extra outputs including an inverted one. Warning: This was a temperamental build because it didn't work perfectly when I first built it. However the problems have been identified and solved. See text below for a more in depth explanation. 

NB: Please don't attempt to build this if you're a beginner in need of a simple reliable workhorse ADSR. This can be a bit of a temperamental build because of the many options this ADSR offers. I recommend the 7555 ADSR if you want an easy to build, good, reliable ADSR. Regard this one as an experimental or advanced project. 

This Envelope Generator or ADSR is a very luxurious one because it produces three different types of envelopes. The following description is from the original text for this module:
First there's the 'Damped' mode. The object of this mode is to more closely simulate the piano envelope which has a sharp attack, a brief initial decay, a long release and finally a very short release as the damper is applied to the string. So it's an ADRR response and in this mode the end of the gate pulse causes the final short release to occur. In other words releasing the note has the same action as applying the damper on a piano.
In 'Normal' mode the ADSR functions as any ADSR would with the duration of the Sustain period being equal to the duration of gate signal being present and the key being pressed down.
The 'Automatic' mode is particularly beneficial when envelopes are being initiated from non-keyboard sources like an LFO or from a clock signal. A short pulse will now generate a complete ADR envelope and, by adjustment of the time constants, this type of envelope can be made to approximate the ADSR type envelope. Usually these external sources would only generate a limited AD type of envelope.
    When I first built this ADSR I had my problems with it and so did many others so please treat this project as experimental. However the layouts are 100% verified. Mine is working fine in the normal and damp settings, and for a long time I thought automatic mode was faulty but that is meant for external trigger sources so it's behaviour is normal although useless for normal use. Read the comments below to see what problems people run into. If you want a reliable ADSR without any bells and whistles then build the 7555 ADSR

Further features of this envelope generator are:
- Independent trigger input for re-triggering and generating multiple peak envelopes in the Damped and Auto modes.
- Gate and Trigger pulses within a range of +3V to +15V are acceptable.
- Wide range of time constants. Typically 2 milliseconds to 20 seconds. If longer times are needed you can increase the value of C9.
- 0 to +10V peak attack output
- 0 to 100% Sustain level.
- Low control voltage feedthrough which means low residual voltage when the envelope cycle is completed thus ensuring that the VCA is off.
- Manual gating facility.

Features I added:
- Extra buffered envelope output.
- Extra inverted envelope output (0V to -10V).

Dual 12 Volt operation:
This envelope generator is designed to run on a dual 15V powersupply but I tested it on a dual 12V supply and it works just as well with only a very small loss in envelope voltage. On 12V the envelope is about +9V so no problem running this on +/-12V. One change should be made however; the current limiting resistor R25 should be changed from a 750 Ohm to a 470 Ohm according to the datasheet of the AS3310. However I test ran it on dual 12 Volt without changing the resistor and it worked perfectly fine.

PROBLEMS I ENCOUNTERED:

I had build this envelope generator some time ago and I've been using it in my synthesizer for all that time but I didn't write an article about it until now because there was something wrong with it. In the 'Normal' mode, which is the one you'll be using most I think, the Decay was oscillating. It kept on being triggered for as long as a gate signal was present. The only way to stop it was to turn up the Sustain level so it matched the Attack level and then you wouldn't hear the constant up and down oscillation of the volume level. This is mentioned above in the features, that it has an option for multiple peaks in the Damped and Auto mode but that's not supposed to happen in 'Normal' mode.

The frequency of this oscillating Decay could be changed by changing the Decay time. Short time equals fast oscillations, long time equals slow oscillations so you could almost think this was meant to be but I can not believe it was meant to work like this in the 'Normal' mode.
So I was using this ADSR with Sustain turned up but it annoyed me that is wasn't functioning quite right because this is an awesome ADSR and I wanted to do an article on it. So I asked on the Synth DIY Facebook group what could be causing this. I was told it was due to capacitor C7 and that I should remove it. They were absolutely right. Removing C7 did the trick, at least in the 'Normal' mode but when I switched to 'Damp' or 'Auto' mode the ADSR was hanging. It wouldn't go into the  Release state. So for these two modes capacitor C7 needed to be in place.
By happy coïncidence I used vintage double pole 3-way switch to switch between the different modes and I had one pole left unused. So I connected the capacitor to that unused part of the  switch in such a way that it was connected in 'Damped' and 'Auto' mode and disconnected in 'Normal' mode. This worked fantastically and now it behaves just as it should do. Should you want an oscillating Decay in 'Normal' mode you could easily add a switch to connect C7 again. Now you have the choice between the two. (This option I leave to you. It is not documented anywhere in this article).
All these changes have been drawn into the layout and into the new schematic that I made.
I used a SPDT toggle switch to go between manual triggering (with a momentary switch) and inputting gate signals. You could use the internal socket switch of the Gate input socket for this too, that's up to you but then you can not press trigger when a cable is connected to the Gate input.

NB: In 'Dampened Mode' the Decay control determins the length of your envelope. 

EDIT 30th of JULY 2023: I was made aware of an article addressing the decay oscillation issue and it offers a solution for the multiple triggering in Normal Mode: It advises to use a schmitt trigger on the trigger input so the trigger level is always at the highest possible voltage. The cause of this retriggering namely, is an impedance issue and the fact that the trigger pulse isn't high enough in voltage. I'm posting the original article below here, so you can read it yourself. I'm not going to try this because it says with this modification it will trigger fine with Gate's higher than +9 Volt. With signals lower than 9V the schmitt trigger doesn't trigger. Not much use then.

I think my own solution is a much better one.

 
NOTE ABOUT AUTO-MODE:

One little thing you need to be aware of with this ADSR is that you need to switch to Auto mode whilst holding down a key on the keyboard. If you don't do that, then the ADSR only gets triggered (in Auto mode) if you push the manual trigger button but not by the keyboard. I think that's meant to be though because Auto mode is for external sources so that would make sense. If however you switch to Auto mode whilst holding down a key then it will work with the keyboard. Any key you press after switching it on will keep sounding until you press an other key and it will keep sounding until you switch back to Normal mode. Once you get used to this it's actually not a problem at all. Just something to be aware of.

IMPORTANT CONSIDERATION:

If you plan on building this ADSR you might just build it first like it was intended with C7 connected to pin 7 of IC1-B and without connecting C7 to the second pole of the 3 Way switch. In the stripboard layout it's simply a matter of connecting the 10nF cap between pin 2 of the LM358 and the strip directly underneath the LM358 which connects it to pin 7 via a wire bridge. Then it's back to how it was originally. Should you encounter the same problems I had then you can make the same alterations I did and have it function perfectly that way. Instead of a double pole 3-way (rotary) switch you can use a single pole one and if you need C7 to be disconnected in Normal mode, just use a little toggle switch for that. Double pole 3-way switches can be expensive unless, like me, you have some lying about in your junk box.

SCHEMATICS:

Here's the new schematic drawing that I made and used for my build with C7 connected to switch S1-B. (That's the only difference to the original schematic) :

This is a re-drawn version of the original Digisound-80 schematic, without any changes. You can click on the picture and then use the "J" and "K" keys on your keyboard to quickly switch from one picture to the other so you can easily see the changes (only on a Mac or PC).:

THE LAYOUTS:

Here's the verified stripboard layout. The changes I made are implemented in the layout but if you connect the lower pin of C7 one strip higher, you can do away with switch S1-B and everything is back to how it originally was, so the changes (if needed) are very easy to make.
BEWARE! All IC's are mounted with pin 1 to the lower right!
The layouts were rivised to make them easier to read in Nov. 2023.

Wiring diagram:

Stripboard only. Don't forget to cut the copper strips at holes H-32, K-32 and P-42 (under the capacitors):

Beware that some stripboards are sold with 56 instead of 55 holes horizontally. The layout is 55 holes wide:

Cuts and wirebridges as seen from the COMPONENT SIDE!

As always, mark the cuts on the component side first with a Sharpie or Edding pen and then stick a pin through the marked holes and mark them again on the copper side. Then cut the strips using a sharp 6 or 7mm hand held drill bit. Then solder in all the wirebridges before you get on with soldering in the components.

Bill of Materials:

CALIBRATION:

There are two trimmers in this circuit, RV2 and RV6.

RV2 is used to set the maximum Sustain voltage to the same value as the peak Attack voltage so no sudden voltage change occurs when the attack cycle is finished or so that the Sustain voltage can never be higher than the peak Attack voltage. The best way to set this is to use an oscilloscope but you can do it with a voltmeter too. I advise to check out the original text (second link below) and read the calibration instructions there. They are on page 4.
RV6 is more for polyphonic systems and for normal use it can be left in the middle position.

So, that's all the calibration you need to do ^__^

Here's a screenshot of the oscilloscope that illustrates the oscillating Decay problem I had in the beginning:

Here are some screenshots of the different modes of this ADSR:
This is the Damped mode with short and continuous key pressing You can see that every time you let go of a key an almost instantaneous release kicks in and kills off the note:

Here's the 'Automatic' mode with the same quick key presses.

Here you can see that letting go of the key will not stop the envelope. It will go through its complete cycle even if no gate signal is present. If you press a key before the cycle is finished it will start at the beginning again as you can see at the right side of the waveform in the screenshot above. This way you can create multiple peaked envelopes by re-triggering the ADSR.:

Finally here's a shot of the normal ADSR mode:

Here's a look at the response time of this ADSR. It's not the fastest response but still, 1.36mSec is pretty fast I suppose. The yellow line is the Gate signal and the blue is the ADSR output with Attack set to zero:

I'm really glad I was able, with the help of the Synth DIY group, to get this envelope generator working like it should at least in Normal and Damped mode. I do have one little quirck with mine. I can only use Auto mode if I switch from Normal to Auto while holding down a key on the keyboard and then the envelope is constantly retriggered so it functions as an LFO. Personally I find this very useful so I'm keeping it like this but let me know in the comments if yours does the same and/or if you found a solution for this. Or maybe this is just how it should be. I really don't know.

Here are some pictures of the module and print. The first one was taken after I installed it in the synth and the second one after I just finished the build. You can see that I put in a lot of output jacks for the envelope. It's always useful to have a few extra I think. The top two outputs are switched in parallel over the ADSR output and the bottom two are switched in parallel over the extra output on the stripboard. Below the inputs for Gate and Trigger there are two more sockets. They are Gate and Trigger outputs. They are each switched in parallel over their respective input sockets. I later added a yellow LED to have a visual indication of the envelope. The LED is soldered over one of the extra ADSR output sockets using a 15K resistor as current limiter so as to not influence the envelope voltage and to make sure the LED doesn't shine too bright:

Here's a link to the Electro-Music Engineer PDF article by Charles Blakey about this module:
http://www.digisound80.co.uk/digisound/other_documents/doc_files/1981-12_EM_Eng_CEM3310.pdf

Here's the original Digisound article in PDF form, about this ADSR:
http://www.digisound80.co.uk/digisound/modules/80-18_files/80-18.pdf

In the original Digisound modular synthesizer this is actually a dual ADSR:
http://www.digisound80.co.uk/digisound/modules/80-18.htm

Okay, that's number 33 done. If you have any questions please post them on the Eddy Bergman Projects Discussion and help Facebook Group, or the comments below or contact me directly.

See you on the next one!

Synthesizer Build part-23: DIGISOUND 80.6 LOWPASS FILTER.

A very cool AS3320 design that sounds amazing! With verified stripboard layout and new schematics.

After having taken out the Sequential Pro One lowpass filter to make room for the Korg filter, I needed a new use for the AS3320 chip that was inside it. I found the Digisound 80 point 6 lowpass filter module on this awesome website that has all the schematics for the entire Digisound 80 modular synthesizer.
You can configure the filter for any type you want (it's all in the original text) but we are going to build the lowpass filter because for subtractive synthesis the lowpass is the best sounding and most useful of all the filters in my opinion.
I first made a new schematic drawing because the original had those zigzag lines for resistors and I find the rectangular way of drawing resistors easier and you can put the value of the resistor inside the box. Makes it less complicated to look at imho.
Anyway, here's the new schematic drawing:

So after that was finished I made a stripboard layout. It is verified because I used this twice and both times the filter worked perfectly. Furthermore it has been used successfully by others in their builds. Make sure you work accurately though because I wouldn't consider this a beginners project. The layout includes a second audio output with 3 times the gain of the original output. This is of my own design and is not included in the schematic drawing. It is this output that is wired up to the output jack-socket in the layout below. The original output is marked on the layout too. More about this further down the article:

(Last revised: 19-March-2020. Added a second audio output with 3 times gain compared to the normal output. 25-5-2023: Removed colour codes from resistors, added colouring to wirebridges.)

Stripboard only. Beware that some stripboards are sold with 56 instead of 55 holes horizontally. The layout is 55 holes wide:

Cuts and wirebridges seen from the COMPONENT SIDE! As always, mark the cuts on the component side with a Sharpie and then stick a pin through the marked holes and mark them again on the copper side and then you can cut them with a sharp hand held 6 or 7mm drill bit. 

Bill of Materials:

The panel potmeters used are all 100K linear types but the value isn't that important. Since they are all connected to either a powersupply voltage or an audio signal you can use any value you like from 10K upwards. I myself used three 100K potmeters for the Coarse, Fine and Resonance and I used three 10K potmeters for the audio and CV level controls. This works just fine. 

You can choose to include the Frequency Fine control potmeter or leave it out to save more room on the panel. I personally never use it but it is there if you want to play the filter as an oscillator when it is in full self-oscillation mode with the Resonance turned up full. You'll need to tune the self oscillation pitch to the chromatic scale of notes so in that case a fine tune knob will be very useful. But I personally never tried this so I don't know how well this filter responds to that. If you have any experience with that then please put it in the comments below so I can share it in this article.

This is a 24dB/Octave, 4-pole LPF and it is self oscillating unlike the Prophet One filter I used this chip in earlier. That one refused to self oscillate. I used simple ceramic capacitors for the 220pF caps and this works fine. There's no need for fancy polystyrene caps ^___^.
The 1µF electrolytic cap C7 at the input may seem to have the wrong polarity. Usually a cap like that would have the positive pole connected to the point where the signal comes from and negative to where the signal needs to go. In this case it is mounted correctly because the input opamp is an inverting buffer with a negative gain reducing the amplitude of the 10Vpp input signal by a third to an amplitude the chip can handle. In the output buffer the signal is then inverted again to a positive signal with a gain of 3 to give us the original amplitude.

Calibrating the filter:
There are three trimmer potmeters on this print and you can set them as follows:
RV8, the 100K trimmer, is used to trim away the DC voltage on the audio output. Measure the DC output voltage with nothing connected to the input and turn RV8 until it reads zero.
RV7, the 20K trimmer is an interesting one. It's used to have the filter track 1V/Octave oscillators correctly but I simply tune it for best sound. If you have a squarewave on the input and you turn this trimmer you can clearly hear the over-tones, the harmonics, change in pitch. You should be able to hear the frequency beating effect of the note from the VCO against the tone of the resonance. Trim until there's no frequency beating but also listen to the tone while changing the cut-off frequency and trim until it sounds right to you. There's a full description of the proper way to calibrate this filter in the original text, which is in the link I mentioned earlier in this article.
The last trimmer is the one in series with the current limiting resistor for the AS3320. Simply measure the resistance and set it so the total resistance of the trimmer with the 1K resistor equals 1,5K. You could also just put in a 1K5 resistor but turning this trimmer does have a little influence on the sound but you'll have to try it to know what I mean. I just set it to 1K5 and left it at that. Turning this trimmer all the way to zero resistance won't damage the chip though, eventhough it needs a 1K5 current-limiting resistor, 1K won't hurt it. The one thing I learned building this filter is that the AS3320 is quite a robust chip. I made a few mistakes building it the first time and the chip has had voltages (through resistors) placed on the wrong pins, short circuits and all sorts of other mishaps but it survived all that without a scratch. Thank goodness because I only have one of them at the moment :) Luckily I was able to test if the chip still worked by placing it back inside the old Prophet One filter and seeing if that still worked. That was very useful.
Anyway, you can use this filter with a dual 12 Volt power supply, but in that case the current limiting resistor should be 1,2K in total. But it's really not that important. Simply connect it to +/-12V and it should work fine.

This filter sounds amazing! It has its own distinctive sound and I can not say it sounds like the Korg or the ARP or the Moog Ladder filter. It sounds like a Digisound 80 filter, although it comes veeeeeeery close to the ARP in sound. This one sounds a bit more well behaved, if you know what I mean. The sound of the Resonance is clearer than in the ARP which has a Rensonance that is sharper and rougher in sound. But that's the only  difference I could hear so it occupies a solid second place over the Korg-MS20 and the Moog Ladder filter in my personal top 5. The ARP filter is still number one because it's a real rebel and I love it. But hey, remember, this is all just my personal preference. You may judge it quite differently. Actually, I find myself using this filter more often than the ARP filter somehow.
The output from the Digisound 80 LPF is a bit more attenuated than the other filters I built and that's why I used the left-over opamp in IC-1 as an output buffer with a gain of 3. There's a 150K resistor from pin 6 to ground and double that value, 330K, as feedback resistor from pin 6 to pin 7. (You can use any value over 10K for both resistors, as long as the feedback resistor is twice the value of the resistor to ground.) This brings the volume up to the same level as the other filters I made. As I mentioned earlier, the amplitude is first divided by 3 and then multiplied by 3 again in the output opamp but, at least in my filter, I found the sound still lacking in volume compared to the other filters. That's why I wired up the left-over opamp as an amplifier with an extra gain of 3 to bring it up to normal. I'm not sure if it's just my filter or if this is normal, that's why I left the original output un-touched so you can use it if you think my solution is too loud. You could also install a potmeter of 500K instead of the feedback resistor so you can manually set the gain. (Put a 50K resistor in series with the potmeter so the feedback loop can't go to zero Ohm.)
The original audio output is marked on the layout so you can choose which one you want to use.
The first time I build this filter I had used a coupling capacitor of 4,7µF over the audio output because I measured a big DC offset voltage on the audio output, but then I read the original text and found out you can trim that away with trimmer RV8 so I took out the cap and trimmed the DC away and now it's all as it should be.
This filter has an input for 1V/Octave but unlike the ARP filter it's not necessary to use this. The filter will work fine without it but if you connect a 1V/Oct. source to it, the filter will track the octaves better. The sharp synthesizer sound we all love, will be more prominent if you use the 1V/Oct input. It actually makes the filter sound better.
Like I mentioned before, this filter has 2 potmeters for the Cut-Off Frequency but I advise to only use the 'Coarse' control. Fine is only for Polyphonic synths. I included it in my build so I could hear its effect and write about it here, but I normally don't use it. It stays in the middle position because turning it just changes the sound a tiny bit. It can be handy though to tune it into a certain harmonic frequency because this filter brings out the harmonics of a square wave really well, but all in all; leave it out.

Here's a picture of the finished panel built into the synth.:

As you can see in the picture I also included a bypass switch on my panel so I can put other filters in series with this one and if I only want to use one filter I can bypass this one and send the signal straight to the next one without having to change the patch cables. For instance, having this filter in series with the Korg MS-20 in High Pass mode sounds pretty amazing too! That way you have a Band-pass filter made up of two different filters. The bypass switch is only connected to the 'Audio-1' input though. If you want to see the wiring diagram for this switch, you can find it in the article about the Moog Ladder Filter, in which I also installed a switch like this.

Here's a little video with a demonstration of the sound of the Digisound 80.6 LPF (EDIT: at this stage in building my synth it still hadn't occurred to me that you need to connect an AD or ADSR to the CV input to get that typical synthesizer sound. #facepalm (We live and learn LOL):

Lastly I want to share with you the efforts of LookMumNoComputer Forum member Doolang who successfully built this filter using my layout. He made it so it fits the Eurorack standard by cutting the print in half and connecting the copper strips together with wire. This works like a charm and he did the same with the Steiner-Parker filter which also worked fine.

So that's the Digisound 80.6 lowpass filter done. I can really recommend building this. It has some very recognizable synthesizer sounds that you should really have available in your synth. Make sure you use good quality stripboard though. The first one I built had problems because strips of copper would become loose and break. So I rebuilt it with better quality stripboard. Make sure you use the filter with the 1V/Oct connected to get the best out of it. This will make resonance follow the notes you play (filter tracking). It'll also work with out 1V/Oct of course.

Okay, thanks for being here and if you have any comments or questions just put them in the comments below or in the special Facebook Group for this website.