Friday 22 May 2020

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.

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 might try this solution later in the year and I'll publish any results I get here in the article when I do.

One little thing you need to be aware of with this ADSR is that I 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.

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.

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).:

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:

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:

Here's the original Digisound article in PDF form, about this ADSR:

In the original Digisound modular synthesizer this is actually a dual ADSR:

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!

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!

Sunday 10 May 2020

Synthesizer Build part-32: ELECTRIC DRUID VCLFO-10 with extras.

A truly awesome LFO with 16 waveforms and 8 different controls. I added 2 extra outputs for 10Vpp and a frequency indicator LED.

This Low Frequency Oscillator was one I had on my wish list for a long time. Last month I decided to buy the chip, it was only 5 Pounds anyway, and it arrived precisely a week later from Tom Whiltshire of Electric Druid in Portugal.
I downloaded the Datasheet PDF with all the schematics etc on it but I found parts of the schematic a bit confusing. The Frequency and the Level controls have their own opamps and they both have two potmeters connected which looked a bit weird to me. So I asked on the Synth DIY Facebook Group what the deal was with those two potmeters. It turns out both the Level and the Frequency controls can be connected to an external control voltage so one potmeter is connected to an input jack and serves as the input level control or attenuator and the other potmeter is for manual setting of the Frequency and the same for Level. So after I had that straight I set about making a stripboard layout. Now, I noticed there was no rate indication LED in the schematic. I always find it handy to have a frequency indicator LED on the panel, so I designed a second print and included a rate indicator LED on it, together with two extra outputs that have a DC offset of +5V so the outputs are 0V to +10V. That is the level I use most on my synthesizer so I needed to have that included. This does mean that the Noise level through these outputs has a +5V DC offset so use the +/-5V output for noise.
I made a second stripboard layout and made a mounting hole in it, on the upper right corner, so the second print can be stacked on top of the main circuit board, using a 3 cm M-3 bolt  and a bit of plastic tubing as a spacer to keep the two prints from touching eachother. The layouts worked like a charm and everything worked fine when I tested it.
If you find that you wired up the potmeters the wrong way around, you can easily correct that by connecting pin 2 of the VCLFO chip to ground. That reverses the working of the potmeters. Make sure the two level potmeters are wired like on the diagram though. They are not influenced by pin 2 of the chip.
The main print has its own plus and minus 5V power supply included, so everything can be powered from a single dual 12 Volt power supply. I used the big 7805 and 7905 in TO-220 package because that's the only ones I had available but you can use the smaller L versions. This circuit hardly draws any current at all so they won't run warm and don't need heatsinks. I did not include any de-coupling capacitors or electrolytic caps on the power rails (except for the -5V because that was indicated in the schematic). You can put those in though, if you think you need it. Use two 100nF ceramic caps, one from +12V to ground and one from ground to -12V.

There is an extra 100K trimmer in the layout with which you can set the Synchronization mode between Sync Off, sync-ing the LFO, the Sample and Hold or both. Instead of a trimmer, as seen on the layout,  you can also make this a feature on the front panel and connect a potmeter to the same points as where the trimmer now sits, and of course you then leave out the trimmer. That way you can change the sync setting on the panel itself. This is what I later did. Beware these extra's are not listed in the Bill of Materials. I used a 100K potmeter but you can use any value potmeter or trimmer for this function because it is just a voltage divider connected between +5V and ground. (Use a value of 10K or higher.)
Looking at the panel-potmeter front-on with pins pointing downwards, connect the left pin to ground, the middle pin (wiper) to pin 8 of the chip and the right pin to +5 Volt.
It took me 5 and a half hours to solder the stripboard components in place and to wire it all up. The whole proces of designing the layout, designing and making a panel etc. took a whole weekend so it was a nice project to do because at the time I, and everyone else, was stuck at home in Covid Lockdown anyway.

Here is the (verified) layout. Wiring diagram:

The +5V and -5V points on the left side of the print are simply noted by me so you know that voltage is available at that point. You could say they are test points you can check if you are troubleshooting this print (which of course I hope won't be necessary. ;) But nothing needs to be connected to those points. Sorry if that's a bit confusing.
This circuit will run fine on +/-15V too. The voltage only feeds the IC's and the regulators and they can all take it without problem or without any changes needed.

Here's the layout for the main print. Beware that some stripboards are sold with 56 instead of 55 holes horizontally. The layout is 55 holes wide!!
In this layout the synchronization mode is set with a 100K trimmer at the top of the stripboard. As I mentioned before I myself exchanged that for a panel mounted potmeter later. The sync mode depends on the voltage on pin 8 and the trimmer or potmeter sets that voltage.
There is no capacitor on the input of the 7905 voltage regulator, it doesn't need one to work but if you want you can put a 1µF electrolytic cap over the input to ground. Easiest way to do that is to connect it to holes B-11 and C-11 with the negative pole in position C-11 (-12V)

The Zero Adjust 100K trimpot is used to set the symmetry of the output signal. You must use an oscilloscope to adjust this. Set it so the LFO output signal's positive amplitude is the same as the negative amplitude.
Make sure you get the cuts right in the stripboard. Especially those in the power rails at the top otherwise you'll have a direct short circuit between ground and -12V. Always measure continuïty over the power connections to rule out short circuits before you connect it to power for testing.

Here's the Bill of Materials for the main print. Note: component numbering does NOT follow the numbering in the datasheet schematic. 

Here's a close-up of the second print with the extra outputs and rate indicator LED. You can use this print for other projects too, if you need to add a DC offset voltage to a certain output. I chose 150K resistors for around the opamps (R3,4,5 and 6) because I have a lot of them but you can use any value from 47K to 500K instead of the 150K's as long as you use the same value for all four resistors:

The DC Offset trimpot must be set to a +5VDC Offset to get a 0V to 10Vpp output signal. Use an oscilloscope connected to the output socket to set this. I don't have a schematic drawing for this part but it's a really straightforward opamp offset circuit with the offset trimpot connected to the non-inverting (positive) input of the opamp. The signal goes in at the inverting input and then it goes through a second opamp stage to invert it back to normal again. Then the second output (at the bottom) is simply fed the signal of the first output via an opamp buffer. The second opamp of the chip on the right is not used and properly connected to ground to 'park' it so to speak.

Here's the Bill of Materials for the extra print:

You don't have to use this extra board for 0 to 10 Volts outputs of course. If you find you need more Bi-Polar outputs of +/-5V then set the trimmer to 0V offset and it's done. 

This VCLFO produces 16 waveforms in 2 sets of 8. I have made a little compilation image of oscilloscope images I took of the waveforms and some sample and hold results. The blue line is the original +/- 5V output and the purple line is the one I put in myself with 0 to 10Vpp. You can see that the noise has a +5V DC Offset on the purple line. When you start testing this circuit after completing the build, it's possible you don't see a waveform but just a flat line. That means your offset voltage is too high or too low, so all you need to do then is set the offset voltage with the trimmer on the main stripboard. Then check the 10V outputs and set that offset with the trimmer on the small print. I advise to use multiturn trimmers for those, but you don't have to. Make sure your oscilloscope is set to DC mode for measuring these waveforms.

Each of the waveforms produced can be sent through a sample and hold unit which is built into the chip and as the chip can also produce noise you can also get random tones produced by this LFO if you connect it to the CV-2 input of one of your VCO's. The sample rate of the S&H can be set with a 10K panel potmeter and if you turn it to zero the S&H switches off automatically.
The VCLFO has a synchronization input and it can be frequency modulated by means of a Frequency CV input with attenuation potmeter. There's even a separate input for the Level control which is a volume control changing the amplitude of the waves.
There's also a control on the panel for 'Distortion' which bends the bottom or top part of the wave with the middle setting being the clean, undistorted wave.
The LFO has 4 frequency ranges and they are:
8 seconds per wave to 12,5 Hz
4,6 sec/wave to 25 Hz
2,6 sec/wave to 50 Hz
1,2 sec/wave to 100 Hz
You set the frequency range with the LFO Range potmeter and then you can set the Frequency within that range with the Frequency potmeter. There's a smoothing switch included in the circuit which rounds off the corners of the waves and makes them smoother (obviously, LOL). This is to prevent the sharp edges of some waveforms from causing clicking sounds when you're using the LFO as a Tremolo.
The possibilities are endless with this LFO and with the chip only costing 5 UK Pounds, like I mentioned, you should really get this one.

Here are some pictures of the finished panel and of the stripboard and wiring. I admit the panel is a mess but it works for me:

In the picture below you can see I made a change by adding an extra potmeter (the one with the yellow knob) with which you can set the Synchronization mode between synchronizing the LFO, the Sample and Hold or both.

And here's a little video I shot using the LFO in a reasonably complicated patch. I've got 3 VCO's feeding squarewaves into 3 filters and a triangle wave into the wave folder. Each filter receives an LFO signal from a different LFO. The Electric Druid VCLFO-10 is feeding a quad pulse into the Steiner Parker filter. All LFO's are synced from the main LFO which is the Music From Outer Space LFO. You can also see the Mixer/Passive Attenuator in action with the bright blue clipping LED coming on occasionally and the Digisound 80.6 LPF sounding really good!

If you're interested in recreating this patch then here is the basic set-up I made. The eventual sound is, of course, dependant on the settings of all the potmeters and little changes can make a big difference but this at least is the foundation of this patch:

Here is a link to the product page of the Electric Druid VCLFO-10 from where you can order the chip:

If you have any questions about the chip or simply want to say thanks to Tom Wiltshire, drop him a line on his website. He's a really nice guy and he'll appreciate your feedback.

Okay, that's it for now. I have now finished the second stage of my synthesizer and so I have no more room to put new modules unless I build a third case. That will no doubt happen but not right away, what with summer coming it's going to be too hot in the attic to spend all day in there wood-working or soldering. I also built up a Eurorack system in early 2022 which took a big chunk out of my budget, which wasn't/isn't too big anyway, but that's all in the game right?

As always, if you have any questions please post them on the EB Projects Discussion and Help Facebook group, or in the comments below or contact me directly via Facebook.

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!

Thursday 7 May 2020

Synthesizer Build part-31: NOISE MODULE with 5 TYPES OF NOISE + Random Gates.

A very easy to build noise module with 5 different sorts of noise, two of them being 'Grainy' noise with adjustable graininess.  Works on dual 12V so Eurorack friendly.

(The Random Gates section is located half way down the article)
This is a module I adapted from the MFOS Noise Cornucopia schematic by Ray Wilson. It's turning out to be quite a popular project because I'm getting lots of feedback from people who built it and are really happy with it. Especially the addition of the Grainy Noise.
So, I needed a good noise source in my synth and this one seemed perfect. The original schematic has a random gates section which I didn't need but which you can easily add on if you want it. But I left that out. (I made a separate layout for the Random Gates Generator section which you can find further down the article) I also changed the transistor used to generate the noise and I changed the way the transistor is integrated in the circuit. My way is simpler and generates 200mV worth of noise right at the emitter of the BC547. The transistor's Emitter-Base breakdown voltage is exceeded thus the transistor is operating in avalanche mode, creating nothing but pure noise.
This was a one day build for me. I spent the morning adapting the design and making a stripboard layout. Then I built it in the afternoon and by 8pm that same day I had a good functioning noise module built into my synthesizer. The layout I made worked right from the start. No troubleshooting needed.
In the layout below the transistor is shown as a schematic symbol, and not as it's normally shown in the TO-92 package, to make it clear that the collector is not connected. In fact, you need to cut off the collector leg completely to stop it working as an antenna. I've put the pin-out of the BC547 in the layout to make this extra clear. You might need to choose a BC547 that gives you the best noise results. I heared through feedback comments that there can be differences between transistors but you should get noise with any transistor. It's just that some transistors produce more noise then others. I myself put in the first transistor that I had, and didn't choose between them. It worked fine as you can se in the video. Should you experience hum or something, from the power supply, then you should resort back to the transistor arragement in the original design as shown in the original Noise Cornucopia Schematic

Here is the verified layout.

(Last revised 16-May-2020: Added grounding wires to output jacks and pinout to noise transistor.)

Stripboard only. 
For extra clarity: connect the 'Base [B]' of the transistor to copper strip 'I' and the 'Emitter [E]' to copper strip 'G'.
The cut on position D17 in the layout above, has been moved to position D21 to make it more visible.

Below is an overview of the cuts and the wirebridges alone. This is seen from the component side! As ever, mark the cuts on the component side with a black waterproof marker and then stick a pin through the marked holes and mark them again on the copper side. Now you can cut the copper at the marks with a sharp hand held 6 or 7mm drill bit.

Bill of Materials. Instead of the TL084 and TL082 you can also use the TL074 and TL072 opamps:

Here is the altered schematic, made from the original Noise Cornucopia design:

The noise output from the transistor goes through a highpass filter consisting of the 100nF capacitor and the 2 MegaOhm resistor. This creates a filter cutoff frequency of 0.8Hz letting through all the frequencies and rejecting any offset voltage. Should you experience an offset voltage after the filter then lower the resistor value from 2M to 1M. That will make the cutoff frequency 1.5Hz.
I changed the 500K trimmer, used to set the amplitude of the noise, for a 200K panel potmeter so you can use it as a level control on the front panel. In my panel I used a 500K panel potmeter but that really is too high a value. When I turn the potmeter 1/3rd open, the amplitude reaches it's maximum at 10V peak-to-peak and the rest is just maximum volume and starting to clip, so I think it's better to use a 200K potmeter. (However I haven't tested it with a 200K potmeter) .
If you only have a 100K potmeter you can try changing R5 from 10K to 4K7 to get the gain right, and then it should work with a 100K potmeter. I've had confirmation that this solution works just fine.
To be honest, you don't need a gain option in a noise module like this, so you can just as easily forget about the Gain potmeter and put in a trimmer, set it so the output of pin 7 gives +/-5Vpp noise level and leave it at that. That's also how it was intended in the first place. The gain option was just my own idea.
The opamps used here are not critical. The schematic says to use TL074 and TL072 but I used the TL084 and TL082. I think you could even use an LM324 instead of the TL074. The pinouts are all the same. 
This module is designed to work on a dual 12 Volt powersupply (so ideal for Eurorack systems) but it will work equally well on 15 V.

How Grainy Noise works:
The Grainy Noise consists of very short pulses with an amplitude of plus and minus 5V. The Opamps IC2 a and b are set up here as voltage comparators which are being fed on the non-inverting inputs with white noise and on the inverting input with a voltage that can be set with the Graininess Potmeter to between 0V to + or - 8.25V (roughly). Each comparitor has a diode on the output so handles only one part of the voltage phase (either positive or negative). So if the voltage on the negative inputs is very high, the noise will only occasionally go over it and we'll get only a few Grainy Noise pulses. When the opamps are not producing a pulse they are at rest in either full positive or full negative voltage on the output pins but those voltages are being blocked by the diodes. So only the Grainy Noise pulses are being fed to the output. As you lower the voltage, the threshold will become lower and the noise will tip over the boundary more often creating more and more Grainy Noise pulses.

The 'Grainy' noise is a real asset to have. It's very useful because of its harsh sound. It sounds a bit like the noise you get from old TV sets. If you look at the scope image in the video you can see that most pulses from the Grainy Noise go into negative voltage. The more you turn up the Graininess, the more pulses you get that go positive and that's what is used to create random gate pulses. In fact only the positive noise pulses are used in the Random Gates generator and the negative ones aren't used because that diode has been left out. See the original Noise Cornucopia Schematic for that. The Highpass grainy output is a bit low in amplitude. That's not just in my build but other example videos show the same thing. Maybe a different type of capacitor would make it better but to really change it you should put it through an extra opamp and give it some extra gain but I didn't bother with that. I don't think I will be using that output much anyway. If you change the resistor to ground you also change the highpass filter so I don't think that is advisable to make it louder.
Btw, the LowPass noise is not exactly the same as Pink Noise in my opinion. The LP noise has more rumble (bass) in it I think but you may have a different opinion on that. I leave that open :) I think to call it LoPass and HighPass etc is more intuïtive than to assign different colour-names to the noise. It's also a useful type of noise to have because a lot of people prefer mixing LoPass noise into the signal path instead of white noise because the LP noise sounds less muddy.

I've made a separate layout for the random gates section of the MFOS Noise Cornucopia design, for those interested in adding this on. The two 7 pin headers are there to provide a choise in randomness of the Gate signal. See the original Noise Cornucopia article for the schematic drawing. The signal has the most randomness if you place the jumper on the lower settings. The higher up you go the less random the Gates get. 
NB.: Place only one jumper on the pinheaders!! 
If you have a rotary switch with 7 positions you could use that, instead of the pinheaders, and make a feature of it by placing it on the front panel. That's up to you. Let me just say also that I have not built this random gates module myself but I've gone over the noise cornucopia circuit schematic with great precision and it's a very simple layout so it should work fine. If you have built this layout please send me some feedback about how it's working. I've had some feedback saying if it doesn't work like it should to put a 2,7MOhm resistor in parallel over the 10pF capacitor.

Personally I don't find this type of random gate circuit very useful. It's not synchronized in any way and you only get one output with a random pulse train on it. I would prefer the Yusynth 8 Random Gates project where each pulse has it's own output. What I would prefer even more is to pair a Sample and Hold circuit with a noise generator like the one above. But it's all up to you of course. 
(Check the comments below to read about Tim's findings when he breadboarded this circuit to test if he could build it with 7 outputs instead of one.)

Here's the layout I made for this section: 

Here's a demo video I made with sound samples of the different types of noise:

And finally some pictures of the finished product. As you can see the finished module is very small. In fact it is only 3 centimeters wide so it won't take up much space in your modular set-up:

To finish I want to direct your attention to a great video by Moritz Klein about building noise modules where he explains the theory behind it very well.  Click here to see the video on YouTube.

Okay, that's number 31 done. A very satisfying build because everything worked right from the get go. Any questions or remarks? Please put them in the comments below or post your questions on the EB Projects Discussion and Help Facebook Group.

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!

Sunday 3 May 2020

Synthesizer Build part-30: LFO with SYNC and FM INPUT (Yusynth).

A very useful LFO with synchronization and Frequency Modulation input, using the ICM7555 IC. This is an other Yusynth design.

I seem to be building a lot of Yusynth designed circuits lately but that's because I know they work so well. This LFO is no exception. This is a medium difficulty project. I wouldn't advise it for beginners. Just take a look at the layout and you'll know what I mean.
This LFO circuit uses the well known ICM7555 chip as main oscillator and two TL074's (or TL084's or any other equivalent) to produce the different waveforms. The 7555 is the CMOS version of the NE555, Do NOT use an NE555 in this circuit! 
The LFO has 4 outputs, one for Sine-, Triangle-, Squarewave and Ramp wave. It has a switch for two frequency ranges. The normal setting (x1.0) goes from about one cycle per 14 seconds to about 100Hz. Then there's a x0.1 setting that divides this roughly by ten so you get (in my case) one cycle per 60 seconds to 18Hz but this can be set with a trimmer on the print so you can set it to your own liking. 
Because the layout is pretty chaotic looking, you need to go about this build very methodically. Mark out all the cuts you need to make first. I've made a special layout with just the cuts on it, to make it easier for you to do this accurately.
I must say I absolutely love this LFO. It has quickly become my goto LFO for modulation duties. It's particularly hand for modulating the LowPass Gate because the speed can be modulated with an ADSR for instance so a sound can start off sounding continuous with the LFO driven into audio range by the Envelope Generator and then lowering in frequency, fading out into a pulsating beat created by the Lowpass Gate. It's awesome :)

Here's the stripboard layout I made for this LFO. I built mine using this layout so it's verified. All wire bridges connecting to ground are coloured green. Btw, you can use other values for the 50K panel potmeter. It's just a voltage divider level pot. You can use 10K or 100K or 1M, whatever you have available.
Naturally, instead of having a switch to go between Saw and Inverted Saw (Rampwave) you can install two output sockets and have both available at once. That's up to you.
Instead of the 50K resistor at the top right, you can use a 47K one.
Wiring diagram:

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

Here's the overview of where the cuts need to be made. I usually mark them with a black Sharpie on the component side, because that way they are easier to identify from the layout, and then I stick a pin through the marked holes and mark them again on the copper side. (That's why I'm showing both sides here). Then I cut the copper side with a 6mm or 7mm drill bit (or a Dremel-tool) in the marked places.

Bill of Materials:

Here's the schematic I used for the layout:

You can see in the schematic that there's a fifth output, underneath the saw output. This is an inverted version of the sawtooth wave and I installed an extra switch to give you the choise between Saw or Ramp. (The un-inverted version is actually a Ramp (rising voltage) and not a Saw, but whatever.)
All waveforms are bi-polar, they have the zero volt line as their mid point so they have a negative and positive phase.
Here is the result of some measurements I took from the LFO:

In the x1.0 setting:
Frequency Range = 1 cycle per 14 seconds to 100Hz
Squarewave amplitude = +5 to -5 V.  Duty Cycle = 26% to 86%
Sinewave amplitude = +5.3 to -5.3 V
Triangle wave = +7 to -7 V
Sawtooth wave = +7 to -8 V

In the x0.1 setting:
Frequency Range = 1 cycle per 60 seconds to 18,7Hz
Amplitudes are the same.
Squarewave duty cycle = 18% to 98%

The synchronization pulse threshold = +2,9V.

As you can see, a fantastically broad range of options and synchronization works very well. When you put a high amplitude sawtooth wave on the CV input the resulting frequency sweep can reach well in to the 400Hz (in x1.0 setting). The LED indicates the frequency rate and is connected to the squarewave output so it will react to changes in duty cycle by being on longer or shorter.

Calibrating the circuit:
You can set the Frequency range by turning the Rate panel potmeter all the way counter clockwise and then use trimmer T1 to set the lowest rate.
Trimmer T2a and T2b are used to set the sawtooth wave in such a way that the positive phase has the same amplitude as the negative phase. In other words you set it so the zero volt line runs right through the middle of the wave. There are two of them because one is used in the x1.0 setting and the other in the x0.1 setting, so only one of those trimmers is active at any one time. Therefore you need to set this twice.
Trimmer T3 is used to set the Sine symmetry. Turn it so that the top of the wave has the same curve as the bottom of the sinewave. This potmeter also influences the duty cycle of the square wave, so you need to set the duty cycle panel potmeter in the middle position and trim the Sinewave so it looks good and then look at the Squarewave and make sure the panel potmeter for duty cycle can be used over its full throw. To make things even more complicated, this trimmer also has an effect on the shape of the Triangle wave so it's a bit fiddly but you need to go between all of these three parameters and find the right setting. You'll get the hang of this soon enough though. It sounds more difficult than it really is. You just have to find the setting that looks the best for all three waveforms. A multi channel oscilloscope will be of great use here.
If you can not get the waveforms right you need to change the 1µF and 10µF capacitors for some other ones with the same value. Yusynth says to use Tantalum caps here but I tried those and it only made things worse. But you may have a different experience. You need to be able to experiment, an other reason why this is not a beginners project.
One other thing which I became aware of through reader feedback; if your output levels are very low and transistor Q1 gets hot then you might be using fake chips. I've had feedback where this problem turned up and changing the chips for ones from a reputable source fixed the problem. So once again, make sure your chips aren't fakes from China.

The x1.0 and x0.1 frequency range settings.
Calibrate the LFO in the frequency setting that you think you will be using most. If you get the waveforms right in the x1.0 setting then the sinewave may not look ok in the x0.1 setting.  That's a little quirck of this LFO and difficult to get right but I usually only use an LFO in the 10 second to 10Hz range, so if all is well in the x1.0 setting, then that's good enough for me. The duty cycle range of the squarewave varies too, according to how the frequency range switch is set. It's really only the sinewave that I personally can not get right in the lower frequency setting. It rises normally and then drops off so it's more like a sine version of the ramp wave. But that's the only thing I can't get right. I found that adding a 0,1µF electrolithic capacitor in parallel over the 1µF cap helps in getting it all looking good. This however will vary from build to build with component tolerances etc.

12V vs 15V:
This LFO will work on a dual 12V powersupply but the frequency will go down by a large amount but you can turn that up again with the trimmer T1 on the print. The amplitudes of the waveforms will go down to between 2 and 5 Volt so that is significantly lower. The LFO is not really meant to work on +/-12V but it will work. However, if you need to address this problem I advise to make an extra print with a TL074 quad opamp chip and set these opamps to a gain of 2 and have all the waveforms go through it. That will double their amplitudes. You can also give them a DC offset voltage to keep them all at a positive voltage if that's what you need. However, if you're a beginner and don't know how to do the above mentioned extra's then don't worry. Don't bother with it for now. Just build the LFO and run it on 12V. LFO outputs are usually attenuated anyway so the lower amplitude signals will still be very useable. This will be a module you will use a lot! I guarantee it.

Here are some screenshots of the waveforms. You will need to try and trim the negative spike in the top of the Triangle wave away while keeping the sinewave looking good. I don't think it's possible to get rid of it completely but you won't hear it in normal use.
As you can see from the screenshots this is a bi-polar LFO. Meaning the output voltages go both positive and negative.

The result of introducing the synchronization pulse. The waveform resets at the rising edge of the sync pulse and will remain high until the pulse falls away. Short trigger pulses will work best here:

Here's what happens when you put an inverted ramp wave (from high to low) on the FM Modulation input (CV IN). You get a frequency sweep that can be quite high in frequency, but you can set the level, and with it the maximum frequency, with the FM Level potmeter. You can see that the amplitude drops a bit in the higher frequencies for some of the waveforms:

Some pictures of the finished module:

I am thinking of adding a second print, like I mentioned earlier, with just a single TL074 on it to use the 4 opamps to give the 4 waveforms a +5V DC offset so they go from 0 to 10V and stay in the positive voltage range. Edit: There's now a Dual Voltage Processor project on this website that can be used for this purpose too.

To conclude this article I made a little test video showing off the 'Synchronization' feature of this LFO, which was the main reason I wanted to include it in my modular synth. As you can see it works very well:

Here's a Falstad simulation of this circuit which I drew myself. It's not working quite like it should but it gives a good indication of how the circuit works: -- CLICK HERE --

Okay that's article number 30 done! Quite a milestone for me I must say, to write 30 articles in so short a time. As per usual, please put any remarks or questions in the comments below, or post them in the Facebook Group for this website.

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!