Dual ADSR Module Design

[ve3wwg]

While my MIDI module has one digital ADSR output, I realised later that I needed a second ADSR output to control the VCF (filter). Rather than just build a single additional ADSR unit, I've decided to build a dual unit so that both are functionally identical.

So after reviewing several designs, I have decided upon a custom design of my own, documented here:

http://www.experimentalistsanonymous.com/ve3wwg/doku.php?id=synth_dual_adsr

This design has the following general features:

• Has hysteresis on the gate input to allow analog signals to trigger it
• Can be switched into "ganged" mode, since you'll often want both units triggered from the same Gate signal (this saves an extra patch cord)
• Fairly simple design

If I left out the Schmitt trigger circuit, this would be a very simple design. But I felt that the Schmitt trigger is a useful addition, allowing flexibility to use LFOs or other analog mixed signals to trigger a unit.

The Schmitt trigger gave me the most difficulty. I had done several iterations of using discrete transistors for this purpose. But after doing some Monte-Carlo analysis realized that the resistor values were quite critical, requiring 1% components. For me, this meant "special order" and potential for failure. 

Using a normal split supply opamp for the Schmitt trigger presented other problems, since this requires work to keep the low output near ground potential, instead of going negative.  The best compromise seems to be to use the LM3900 opamp, which is a Current Differencing Amplifier (CDA). The main attraction is the fact that it likes to operate from a single supply and swing to ground rather than negative. Creating a Schmitt trigger from these CDAs seemed to require the fewest number of components. Since you get four CDA units to a package, this worked out just right for a dual design.

The only other special innovation of this design is my emitter follower after the output TL072 opamp.  Once you review the circuit, you'll see that the best that the charged capacitor C6 can be discharged is to within 0.6 volts due to the diode involved in the Release circuit. The emitter follower circuit corrects for this and reduces the ADSR output voltage (by 0.6 volts) to within about 100mV or so, depending upon the output level you've chosen.

If you wanted to economize, this dual design could accomplish that simply by sharing a single Gate input circuit. You'd of course lose some flexibility but it would eliminate the whole 2nd Schmitt trigger and trigger transistor circuits.

[expanoncolin]

That's a great design.  The simplicity of the 555 ADSR is really nice.  I added schmitt triggers to my ADSRs a while back using CD4000 series chips but as I recall it didn't work great.  I may try your triggers.

-Colin

[ve3wwg]

Thanks Colin. I can't take credit for the 555 design but I did see it used in a number of different forms. It's really hard to beat the 555 for an ADSR circuit since it has most of the required parts - particularly comparator and flip flop - all in a cheap convenient package.

I plan to start building the input circuit shortly, so I should be able to report back on that shortly. I have to order parts for the potentiometers etc. for the rest.

[ve3wwg]

Update: I have the Schmitt trigger section wired using the LM3900s and it appears to work nicely on a sine wave input signal.

http://www.experimentalistsanonymous.com/ve3wwg/doku.php?id=synth_dual_adsr&#construction

I'll need to wire up Q1/Q11's circuits to see how sensitive the gate inputs are to a sloppy sine wave.

[ve3wwg]

I had a few moments this morning to solder in the 2N3904 transistors (Q1/Q11). I then performed another test and it worked beautifully.

Using a signal generator, I fed in a sine wave at a low frequency and was able to see nicely squared approx 30 usec pulses. I was a little worried that the first unit might not have the same sensitivity as the 2nd, based upon testing without the transistors wired in place. But the high Beta (421) for the 2N3904's that I used, seemed to really do a nice job of squaring things up.

The pulse is nice and solid and should trigger any 555.  Obviously this also works very well with normal gate signal.

[expanoncolin]

Glad to hear the triggering is working sharply.  I know that can be a little fussy to get right.

-Colin

[ve3wwg]

Indeed. I have the NE556 partially wired now and after correcting some wiring errors, I can see the timer triggering reliably (better than hoped actually).

I don't have all the pot and jack hardware, so there is only a little more I can do at the moment.  I plan to wire up the buffer amplifier and emitter follower tomorrow. After that it will be a wait for ebay (jacks) and Mouser (pots).

[ve3wwg]

I spent most of the day today, just building the front panel. It is such tedious work!  It took me quite a while just to figure out how to wrap the 0 to 10 labels around each knob.  The OpenOffice has plenty of room for improvement. Now that Oracle owns it, I wonder what if anything will become of it.

I had to move things around quite a bit to make room for the knob dials. However, I think this will be helpful if you make notes about special patches and settings.

With this out of the way, I can finally get back to doing the "fun stuff".

[ve3wwg]

Finished the wiring for unit 1 today (unit 2 needs parts). Due to the slow speed that events move at here, it was difficult to capture a scope trace with a camera. But I've attached what I was able to capture. It was enhanced by Gimp so that you could see the trace more visibly.

For this image, the ADSR output level was set to "7" (out of 10), while the scope was set to 2 volts per graticule.  In theory the output level set to 10, should provide about 10 volt peaks, while at 5, it should peak about the ADSR normal 5 volt level.

Many EGs don't have an output level control. Call me a control freak but I kind of like the idea of being able to tweak that. It remains to be seen if this will be a useful customization or not.

[expanoncolin]

Looks very ADSR to me!  I think most modules don't have output level controls because you're expected to have to use an attenuator/mixer/gain/etc module for that, for better or for worse.

-Colin

[ve3wwg]

That approach certainly works and is sensible (not to mention flexible). But it seemed like a waste of a module (section) if all you wanted was a gain adjustment.  For me, it gives me the added flexibility I need before any such a module gets built -- which by that time, will require another cabinet and power supply!

The price you pay for the extra pair of controls of course is the added real estate that is dedicated to them.  The panel would have laid out nicer and more symmetrically if I didn't have to make so much room for the extra controls.

[ve3wwg]

The Dual ADSR module is finally completed and installed. Whew!

Building this one was like building a stereo. By the time you get working on the 2nd half of the unit, you get pretty sick of it. This build was complicated by too many wires going all over the place and to the front panel.  I think I'd do a few things differently next time.

But the main thing is that it all works. I monkeyed around with the 2nd half with it patched to skew the VCO etc. I can't wait to get the VCF built for it now.