Tuesday, 14 January 2020

Synthesizer Build part-18: A REALLY GOOD AS3340 VCO DESIGN!!

This is the Digisound-80 VCO. The answer to my DIY VCO prayers. Easy to build, easy to tune and all the extra's like Synchronization, Frequency-Modulation and PWM.

Not only is this VCO easy to build, it can actually be tuned easily too. Before I found this, I used the datasheet VCO schematic for the AS3340 using the stripboard design from the LookMumNoComputer website. I could never get that VCO in tune over a wide range of octaves and I couldn't get it to play really deep tones either. I think the fact that he left out the HF tracking had something to do with that. This 'new' design however changed all that! After looking through all sorts of VCO schematics, like the ones on the Electric Druid site, I decided to go for the Digisound 80 design but I left out the triangle- to sinewave converter because I already have one of those build in as a separate module. I can tell you, these VCO's (I built five so far) sound soooo much better than the Datasheet VCO. Of course it's the same waveforms but the range is so much bigger (0.1Hz to 50kHz!) and tuning this VCO is a breeze! And this Digisound design isn't even that different from the Datasheet design. Except for the extra trimmer, the Hard Sync options and a few resistor value changes, but this makes all the difference in the world. As a first time synth builder and having been into modular synths for only 6 months (at the time of writing this article) this VCO was a real revelation for me. You can even use this VCO as an LFO, a Low Frequency Oscillator, because it goes down to 0.1Hz. If you're looking for a good AS3340 VCO to build, I think this is it. It certainly is perfect for my synthesizer DIY project.

Here's the new version of the schematic with all the opamp buffer stages drawn in. All the outputs are buffered and the same with the PWM input. So a total of 4 buffer stages are used here, all housed in one IC, the TL074. You can also use a TL084. (There's a link at the bottom of this article to the original text and schematic.) I did not include any de-coupling caps in this schematic because I don't use them. If you have a normal linear dual power supply there should be no need for de-coupling but I have included them in the stripboard layout.


(Last revised: 14-Feb.-2020 Added numbering to opamps and extra text. 23-March-2020 Emphasized connections with black dots and added remark about R4. 18-Oct.-2020 Changed position of Pulse Width Resistor R18.)

The picture below shows the stripboard layout that I made and used for my build(s). This layout is verified and the placing of the buffer stages follows the numbering on the schematic. For the Octaves control I recommend you use a 100K potmeter with a center detent. You know the ones they use for balance control on amplifiers with a little click you can feel when you reach the center point. This is very useful to easily re-tune the VCO after you've been using the Octaves control knob. Use a normal 100K potmeter for the Frequency Fine control though, not one with a center detent because you need accuracy around the center settings.
I put in a 47K resistor for R21 which is the pull down resistor for the squarewave output. It originally was a 10K resistor in the Digisound-80 circuit, because the CEM3340 chip was used, but that will work fine too. (Use 10K if you're using a CEM chip) It's stated in the datasheet for the AS3340 that it needs to be 51K but in practice it doesn't matter at all so I use 47K. For the current limiting resistor (R23) I put in a 1K. This is necessary because we will connect it to negative 15 Volt. It says to use a 910 Ohm in the schematics but I always play it safe and use a 1K resistor. Use a good quality polystyrene or polyester or silver mica type capacitor for C7 (1nF). This is the frequency determining capacitor and must be stable with temperature changes. So do not use a ceramic capacitor for C7.

Running this VCO on a dual 12V powersupply:
If you're going to use this VCO with a dual 12 Volt power supply (Eurorack) then use a 680 Ohm resistor for R23. (On the stripboard layout R23 is the 1K resistor going from pin 3 of the AS3340 to the negative 15 Volt rail.) It doesn't matter if you are using the CEM3340 or the AS3340 ICs. According to the datasheets for both the CEM3340 and the AS3340, the chip can be fed with a voltage between +10 and +18 Volt and -4.5 to -18V. I wouldn't run it on any lower voltage than +12 Volt though.
Further down this article, in the 'Tuning' section, I mention that if you experience problems with tuning while running this VCO at 12V, you can make R4 bigger. Use a 270K or even a 300K resistor instead of the 200K. Some people experienced problems because trimpot-A was at its end before the VCO was in tune. Making R4 bigger will prevent that. (It's not always necessary to make this change though.)
When running this on dual 12V you must also change C7 from 1nF to 0.5nF or 500pF. Otherwise you'll only get very low frequencies out of this VCO. This is not in the Datasheet but it has been established by feedback from many users who built this VCO for Eurorack.
Lastly change R18, the resistor connected to the PWM potmeter to 18K or 20K or 21K to make full use of the throw of the pulse width modulation potmeter. 

So to sum up the changes you need to make for +/-12V operation:
- Change R23 to 680 Ohm.
- Change the Timing Capacitor C7 from 1nF to 500pF (0.5nF).
- Change resistor R4 for a 270K or 300K resistor to assist tuning.
- Change R18 to 20K or 21K.

Further hints and tips:
The potmeter for High Frequency Tracking or Linearity can be a normal trimpot, not a multi-turn one. The influence it has is minimal. But you must use multi-turn trimpots for A and B on the layout, otherwise tuning will become very difficult. I used metal film resistors with 1% tolerance in those places where it matters and this is good enough. In fact, I used cheap 1% resistors from China and they are not 1% but more like 3% but this is still good enough. But the 100K CV input resistors should all be measured and matched so they all have the same resistance value. This makes it easier when you connect different CV sources to those inputs, they will be in tune straight away. I was surprised that the two 100K potmeters I used in the panel for Octaves and for Fine Tune give exactly the range that is stated in the original description although the Octaves control is not linear, at least mine wasn't but maybe that is due to the potmeters with center dent I used. Octaves is plus and minus 5 octaves and Fine is plus and minus half an octave. I'm not used to things actually working out as originally described in DIY projects. It's usually either a bit off or way off but the Digisound 80 designs are really good and spot on.

About Pulse Width Modulation:
Pulse Width Modulation is now also spot on. Before the 18th of October 2020 I had the PWM connected as is show in the original schematics in the PDF file linked below but that didn't work perfectly. There was a significant amount of throw left on the potmeter when you reached the 100% mark. However, I got a suggestion in the comments below to move R18 from the wiper of the PWM potmeter to pin 1, the +15V connection to the potmeter, and that did the trick.
I should have realized this myself it's so obvious. The 47K resistor R18 now forms a voltage devider with the 100K potmeter that takes off 5V from the +15V supply and leaves the potmeter with +10V on pin 1. This is then halved by the voltage devider made up of R19 and R20 (both 47K) to feed the chip with 0 to +5V, which is exactly what it needs for the correct pulse width modulation.
This Pulse Width problem was really buggin' me because it was the only thing that was not working right in this design but now that is solved too.
The results I get are as follows: With the PWM potmeter fully counter clockwise I get 0% pulse width, meaning that there is no signal, just a flat line. Then as I turn it clockwise the pulse appears and goes through the percentages to stop at 99% pulse width when the potmeter is turned fully clockwise. So fully clockwise there's a very thin pulse left over. This is absolutely perfect. Of course your results can differ a tiny bit because of resistor tolerances but I got the same results with all 4 of my VCO's. 
So if you are using this VCO with a Eurorack powersupply of +/-12V you need resistor R18 to be as near to 20K as possible. 21K or 18K will work fine I think. (The schematic and layouts have all been updated with the new R18 position.)
For external Pulse Width Modulation you need a signal that goes from 0 to +10V on the PWM input jack. This can be a problem if you use this VCO in a Eurorack setting where the signals are usually -5/+5V.  Just so you know. But there are LFO designs on my website that will give you the 0V to +10V output option you need.

Temperature Compensation:
Don't place this VCO directly over the power supply in your modular set-up. If it gets influenced by the heat from the voltage regulators too much it can de-tune a bit but I think this is true of almost all VCO's. The AS3340 has internal temperature compensation but this only really works for changes in room temperature. If you put it over a heat source like a power supply it will most definitely de-tune. Of course other components around the chip will also warm up and add to the de-tuning of the VCO when influenced by the heat from the powersupply.

Here's the layout. I didn't put in the input jacks for the sync inputs or the output jacks for the wave forms and CV-OUT. It's already spaghetti junction and that would make it even worse. I assume you know how to hook up jack sockets. All potmeters are frontal view with shaft facing you. I have recently added 100nF decoupling capacitors directly between the IC's and 22µF electrolytic capacitors on the power rails, because this came up on Facebook. These are not included in the schematic drawing but they are in the original schematic in the PDF linked below. (You can use any value for the electrolytic caps between 10µF and a 100µF as long as they are 25V or over.) As of the 18th of October 2020 I deleted the version that had a switch to go between external Pulse Width Modulation and the potmeter. This to avoid confusion. Here's the wiring diagram:


(Last revised: 18-Oct.-2020: Changed position of R18 from the wiper of the PWM potmeter to between the +15V and pin 1 of the PWM potmeter. I also changed the colour of the capacitors to be in line with other layouts and I made the wirebridges that connect to ground a green colour for clarity.)

And here's a close-up of the print. Don't forget to cut the copper strip underneath the 1M resistor above trimpot A. The cut is difficult to see on the layout but it's there of course, otherwise the resistor wouldn't work:


Bill of Materials.


(07-June-2020 Revised version. Numbering now follows numbering on schematic.)

Here's a look at the finished stripboard. I soldered on a little copper eye to make mounting the print on the particular panel I made easier, but there's room enough left on the print to drill a few holes to mount it however you like. Make sure the copper traces are cut so no contact is made with the bolt and nut etc. In this picture you can also see the annoying little circle at the bottom of the AS3340 chip. Do NOT mistake this for the pin-1 indicator, and put the chip in the wrong way as I once did!! I had the chip mounted in the socket the wrong way around and had it switched on for about 20 seconds. It got so hot that I could smell it, that's what allerted me, and I switched it off immediately thinking the chip would be waisted but no, it survived! (They call that 'burning in the chip', LOL.  DON'T TRY IT!)




Here are the two identical VCOs side by side in my synth. You can see a 'Tuner' and 'Sync Out' output, which I added lately. More on that in the 'Synchronizing' section below:



About the CV-OUT connector:
You can see in the picture that I have CV-OUT jacks on the VCO's. These are simply in parallel over the 1V/Oct. input jack so I can daisy-chain more VCO's to one 1V/Oct. signal so that all VCO's receive the 1V/Oct signal from the Doepfer A-190-3 MIDI to CV converter. This is not included in the stripboard layouts but you can see it in the schematic drawing. Do not use the CV-OUT as a CV-IN because it shares a 100K resistor with the 1V/Oct in and will detune your VCO. If you use the Dual Buffered Multiple described on this website, then you don't have to include this CV OUT and you can spread the 1V/Oct. signal over all VCO's with the Multiple. But I do advise to include it. If you daisy-chain your oscillators like this you keep the Buffered Multiple free for other functions and you can daisy-chain upto 8 oscillators of this design before you'll get a slight drop in voltage in the 1V/Octave signal.

Synchronizing multiple VCO's:
I recently added two more outputs. One is parallel over the squarewave output socket and is used to connect the VCO to the negative Hard Sync of one of the other VCO's, so I can keep the Squarewave output free for normal use. The other output is in parallel over the triangle wave output and is used to connect the hacked Joyo tuner to the VCO, also to keep the Triangle output free for normal use. These outputs are not on any of the pictures or on any of the layouts but you can easily add them if you feel you need them. I find them very useful. If you want to synchronize multiple VCO's then just take a squarewave out from the main VCO into the Negative Hard Sync input of the second VCO. Take a squarewave out from the second VCO and connect it to the Negative Hard Sync of the third VCO etc. Now if you turn the Frequency Control of VCO-1 all the other VCO's will stay synchronized with the first one. Now you see why the extra outputs are so handy to have.

Doepfer A-190-3V MIDI to CV:
The Doepfer A-190-3 is the one module that I bought because I didn't trust myself to build one of these and I wanted the interface between the keyboard and the synthesizer to be absolutely fool-proof and reliable and it was certainly worth the €130 I paid for it. It adds a Portamento (or Glide) function to the synth and it has 3 extra outputs for the modulation and pitch-bend wheels on the keyboard that you can connect to CV-2 for instance to get pitch-bend. It also has a Velocity output and a 'Learn' option on CV-4. And it also has a USB input so I can connect the synth to my computer. Btw, I got the A-190-3V which is the Vintage edition which means the panel is black with white lettering to stay in keeping with the other panels in my synth. It's only 5 more euro's than the normal grey edition. I just made a 20cm high panel and cut a Eurorack sized hole in it and mounted the Doepfer in there. Then I made a special adapter cable to go from the Eurorack power system to the one I invented for my own synth. (See powersupply article).

OCTAVE SWITCH:
With VCO number four I changed the Octave potmeter for an octave switch, as an experiment. I used a 10 step rotary switch and I measured out a bunch of 10K resistors so I had 10 with the same resistance upto 10 Ohm accuracy. They were all 9K99. I soldered on the resistors in the way shown in the drawing below:


So you get 5 Octaves up and 4 Octaves down with a 10 step switch. If you want -5/+5 you'll need a 11 step switch and I think they are rare ^___^. This is more than enough though.
Now, this works but it is not the case that when you turn the switch you land on the exact same note as the previous Octave. To try and address that problem I exchanged the 10K resistor connected to -15V for a 10K multiturn potmeter, with a 2K resistor in series with it, going to -15V. Now it's not possible to tune it so it is spot on but I did manage to tune it so that each octave I go up, I can turn the Fine Tune one stripe up on the dial (decal) and I'm in tune. So you switch one Octave higher, you turn Finetune to the next stripe on the dial and you're bang on. And the same but backwards for switching down the octaves. This works well enough for me :) This will really only work well if you also have a hacked Joyo tuner connected to the VCO so you can see what you're doing.
To get this bang on the right note, you would need to experiment with the 3M3 resistor and try to buffer this potmeter and use really accurate resistors. So I wouldn't advise this switch solution, as it is presented here, for any serious project where everything needs to work perfectly. I'm just documenting it here because it is something I personally tried and want to keep a record of.

Here's a picture of the switch in the panel:



The output waves:
And finally a look at the waves this oscillator puts out. All nice clean waves as may be expected from the AS3340 chip but the ringing issue in the downward slope of the squarewave, which I mention in article 2 of this build series, is still there even in this design. Although it is significantly less prominent. This ringing must be common to this chip or something. Anyway, it's not audible so no real problem. I thought the zener diode over the squarewave output resistor might help to eliminate this problem but it has no influence but you can see from the pictures below that there are only a few spikes and only on the lowest notes. The picture below shows the ringing at note C1. Only 3 spikes! They only occur on the downward slope of the squarewave and they have a frequency of 28kHz so well above human hearing capabilities.


At note C3 there's only one spike left and after that it is completely clean. Maybe it adds to the character of the sound though. Who knows ^___^


Squarewave. You can see that the ringing is not even visible once you zoom out on the oscilloscope image:



Here's the ramp wave:


And this is the triangle wave:


Just for fun, here's a Triangle and a Ramp wave after being mixed together and after it's been through the dual Korg MS20 filter. You can see the high frequency resonance, produced by the filter, on parts of the wave form:



You can see that the output voltages are all around the 10 Volt except for the squarewave which is 13.4 Volt. I recently received a batch of 10 Volt Zener Diodes from China and I have soldered those in over the squarewave output and now all signal outputs are at the same 10Vpp level. Perfect!

TUNING THE VCO:
This VCO has 3 trimpots for tuning but we're only going to use 2 because the High Frequency Tracking or Linearity potmeter is not really effective for the lower octaves. So we leave that in the middle position. I have developed a tuning procedure of my own that is very simple and will get this VCO in tune over many octaves in less than 15 minutes.
If you don't have a useful tuner for calibration purposes but you do have a smart-phone then I recommend you download the 'Universal Tuner' app by Dmitry Pogrebnyak. It's available in the Google Play store for free. Of course any tuning app that displays frequency and notes will do. There's plenty to choose from.
Before we start tuning, turn the Coarse Tune or Octaves potmeter off with the switch on the panel and set the Fine-Tune potmeter in the middle position. In the original text the wire connected to the wiper of the Fine Tune potmeter is de-soldered, but I recommend just leaving it in the middle position. Take a little screwdriver and turn trimpot A up about 3 quarters of the way.  Now go to trimpot B and turn that down about 3 quarters of the way (it's not necessary to be accurate with this and it also doesn't matter which way you turn them. It's just for setting a start position.)

- Launch the Universal Tuner app. or the tuner of your preference.
- Open up the 'Gain' potmeter on your VCA so you get permanent sound.
- Now press key C5 on your keyboard and turn Trimpot A until note C5 is in tune on your tuner.
- Now press key C2 on your keyboard and turn Trimpot B until C2  is in tune.
- C5 will now be out of tune again so press key C5 and retune it with Trimpot A.
- Now C2 will be out of tune again so press C2 and retune it with Trimpot B.
- Repeat these steps over and over until the VCO is in tune.
- You'll notice that you will need to turn the potmeters less and less to reach the C notes. After a few cycles of tuning they will be spot on their respective C notes.
- If you find that you need to turn the trimmers more and more to reach the C notes then switch potmeters and use A for C2 and B for C5

We tune with the Octaves control switched off to prevent variations in resistance from de-tuning our VCO. This keeps the VCO nice and stable but if you want it to be in tune with both control potmeters in the middle position then you can re-tune the oscillator after the first tuning session but now with the Octaves control switched on. It's up to you. :)

Be precise with the final tuning. Check the exact frequencies of the C notes. The app I mentioned will display the note graphically and it shows the frequency. You can get it in tune to at least 1/10th of a Hertz although in my experience you don't have to go more than one figure behind the comma.

Extra info for tuning on +/-12 Volt and using the V3340 chip:
As mentioned earlier, if you're running this VCO on +/-12V and you have trouble tuning it, change R4 from 200K to 270K or 300K. Someone kindly reported having trouble tuning this VCO on 12V and reported this as the solution in the comments below, so I thought I'd include it here. That's why the comments are so useful. If you come across problems like this please report it down below and we'll find a solution.
I've also had a comment on Facebook about the V3340 chip not holding tracking when used in this circuit. I have no further details on that, but just so you know. It's recommended you use either the AS3340 or the CEM3340 chip.

It will usually be the case, when we start tuning, that the notes are too far apart rather than too close together and if you repeat the steps above and keep switching between C2 and C5 and using trimmer A for the high note and B for the low note, you will notice, as mentioned before,  that the notes get closer together and you'll have to turn the trimpots less and less to hit the right note. Eventually you will be spot on and the VCO will be in tune over at least 4 octaves. Be careful that you don't overshoot but you'll notice that soon enough if you have to turn the trimpots more and more to hit the right C note. In that case switch trimpots and use trimpot A for the low note and B for the high note.
You can of course use even higher octaves and other notes, like tune it between A2 and A7 for instance. I leave that up to you. I don't use C1 for tuning because it is so low my phone with the app has trouble tuning into it.
You'll get the hang of this tuning proces soon enough. It's really simple. It took me just 15 minutes after turning the VCO on to have it perfectly in tune, and when I say perfectly I mean perfectly! I was really chuffed about this ^____^

For those of you who want to go the scientific route and calibrate the VCO with a scope and an accurate voltage source and meter, I posted the original calibration procedure below here and it's also in the PDF I link to at the bottom of this article.
(click on text to enlarge)



Here's a little overview of features and technical data about this VCO:

Frequency range:                    0.1Hz - 50kHz
Most accurate freq range:          5Hz - 10kHz
Waveform amplitude:                0V to +10Vpp
Octave adjust control range:     +/- 5 Octaves
Frequency Fine control range: +/- 0.5 Octaves
+ and - Hard Sync
Soft Sync
FM input with level control
CV-2 input with level control
Pulse Width Modulation both internally controlled and externally controllable.
Extra CV inputs can easily be added by using 100K resistors connected to pin 15 of the VCO chip. Measure the resistance of the 100K CV input resistors and make sure they are all the same, that way anything you connect to the inputs will be in tune right away.
All outputs are protected and can be short-circuited continuously without damage to any components.

Synchronization and FM input:
The positive Hard Sync doesn't obviously do anything but I think it needs a pulsed signal. I'm getting excellent results with the other inputs. Negative Hard Sync and Soft Sync all work fine and FM also gives great results. I personally use the negative Hard Sync input for syncing up two or more VCO's. I input a square- or ramp-wave from an other VCO into Neg H.S. and then they both react to pitch changes of the main VCO.  The FM input is also very cool to use. I can't describe how it sounds but if you build two of these and input one into the other and you turn the Octaves control back half an octave on the oscillator connected to the inputs, you're gonna get some great results. You can also input a Control Voltage from an LFO to get Vibrato or Tremelo effects.

Finally and by request, here's a list of individual notes and their corresponding voltages, should you want to tune the VCO without a keyboard, using an accurate voltmeter. Ignore the 'Expo Output' column. It is not relevant to this VCO:



Here's a picture of VCO number three installed in the second case of the Bergman-Berlin synthesizer :)


Btw, I recently built a version with the sinewave section included but can't get the sinewave to work using the original Digisound schematic with the CA3080 chip. Maybe my chips are Chinese fakes though, I don't know but I resorted to using a discrete transistor based triangle to sinewave converter. This VCO gives you a choise in output voltage for the Triangle and the Sinewaves between +/-5Vpp or 0 to 10Vpeak-to-peak. Any future Digisound VCO's I might make will be made using this panel design and using the triangle to sinewave converter I've built into this one. I'll put up schematics for that in the near future because it is a slightly different design than the one in the Triangle to Sinewave Converter I describe on this website in chapter 3.

Here's a picture of VCO number four with the sinewave output. I have some more experimenting to do but there will be extra info on this website soon on adding the sinewave option:



The panel I made for VCO number four is 2 centimeter less wide than the other 3 which are 10 CM wide. I included an output labelled 'Tuner' to which you can connect the JOYO Tuner after you hacked it. That is connected in parallel over the 10V Triangle output, soldered to the 5V/10V switch directly. (This is not included in any layout or schematic because it was a last minute addition to keep the normal outputs free.) As I mentioned earlier I later also added an extra output in parallel over the Squarewave output to connect the VCO to the negative Hard Sync of an other VCO so as to keep the regular output free for normal use.

Okay, that's it for this one. I hope this is useful to you. After searching for a good VCO design using the AS3340 chip it was a real relief to see that this VCO performed so well and was so easy to build and tune too. I wish I found an article like this one when I first set out to build my first VCO but now I've written one of my own I really hope it will help out all those of you who are building their own synthesizer, maybe for the first time like me, and are looking for a good VCO design.
If you found this article useful or if you have any questions, please leave your feedback in the comments below. I would really love to hear from you. Share this article with your friends and follow this blog to keep informed of new posts. There are buttons for sharing on social media right below here. It would help to maintain this website. If you want to support my work you can buy me a coffee. There's a button for that if you're on a PC or MAC under the menu in the sidebar.
Thank you for checking my website out and see you on the next one! :)

Here is a link to the original PDF file with all the text and schematics and tuning procedures for the Digisound 80 VCO: <click here to download or read the PDF file>

73 comments:

  1. Thank you for this project.
    Today was my first try with this VCO but I have failed :P
    I think I have some short on the circuit because my PSU gets hot. Can I use this schematics to power it with 12V? I am not sure which resistors should be smaller because is different to "stripboard schem"... Well, I will play with it tomorrow!
    Best for you!

    ReplyDelete
    Replies
    1. Oh sorry to hear it didn't work for you. I have had problems too with stripboards that had short circuits between the strips. That could be the problem. The circuit itself is good because I have build 3 of these now and they all work very well. If you want to use it with 12 Volt you need to change R23 to 680 Ohm. Good luck tomorrow. Let me know if you have any questions.

      Delete
    2. The resistor you need to change is the 1K on the stripboard layout that goes from pin 3 of the AS3340 to -15 volt rail.

      Delete
    3. Don't be sorry for my mistakes :) I am still learning and yours projects with schems are like a playground for me. I will let you know about my progress.

      Delete
  2. Eddy,

    First of all thank you very much for documenting your work like this. You and a few other folks have made getting into DIY synth stuff so much easier. I searched for quite awhile, but this is the first VCO build I found that looked like it hit the sweet spot between fully featured and too simple (I was aware of the Look Mum VCO, but the poor tracking is an automatic pass for me).

    I built this to be powered by +/-12V. The R23 resistor change was the only change I made, but I was unable to get the oscillator tuned properly. No matter which trimmer (A or B) I used for C5 vs C2, which trimmer I set first, or where I set the trimmers to start, I'd eventually run out of travel on trimmer A before getting it all the way in tune. It would be at 100k and I'd still need some additional resistance that I couldn't get from it.

    To fix this, I replaced R4 with a 270k resistor and got it tuned up on the first try after that. Sounds excellent! I'm not sure if it's because I'm running this off +/-12V or it was something else that caused that. I did make my own stripboard/perfboard layout for this in order to get its size down, so it's possible I messed something else up in the process. It got a little convoluted at the size I was aiming for (10HP eurorack). But besides that tuning issue, everything seems to work as expected, and swapping the resistor out worked just fine as far as I can tell. Just wanted to mention that here in case someone else has the same issue. Thanks again for posting this!


    Pics of my messy build:

    https://imgur.com/a/kuXPyI4

    ReplyDelete
    Replies
    1. Thank you so much for sharing your experience with this build. I'm sure is the 12V conversion that gave you trouble but loads of people ask me about running it on 12V. I'm glad you got it working in the end.

      Delete
    2. I've made a note of your tuning problems and the solution you found, in the article. I'm sure other people will experience this problem so I'm glad we can offer them a solution. Thanks again for commenting about it!

      Delete
  3. Hey
    I'm quite the beginner and building my own DIY synth as my High-school graduation project. I'm fascinated, by all the modules you actually built and want to build some of them on my own.
    But when it comes to capacitors I'm having some difficulties. Until now I've only been using cheap Ceramic caps but in your layouts some of them look different.
    I'm quite certain that the blue ones should be Electorlyte because they dont look symmetrical, indicating they are polarised. Also when you called for polystyrene ones I will try to get my hand on those. I know that they should be much more stable than other types and are used a lot in audio curcuits.
    But is it ok now for the other ones to use carbon disc ones? I heard that they can add a lot of noise, especially when they are under vibration.
    Also someone told me this:
    A pretty good rule of thumb that I think most people use and is reflected in vero layouts:
    Anything in pF-- ceramic disc
    Anything in nf-- Polyester film "greenies"
    1uF or over-- Electrolytics (although these will be specifically noted in vero layouts due to polarity).
    Should I go by this advice, and what did you use? And what colors stand for what type?
    Thanks for your help.
    Leon

    ReplyDelete
    Replies
    1. Hi Leon, that rule of thumb is okay for 99% of the cases. I use a lot of ceramic caps in my builds and I never had any noise issues with them. Many of them are vintage ones, I have a large stock of components I got out of other circuitboards. The one issue with ceramic caps is that they are temperature sensitive and they can be prone to 'microphoning', they can react to vibration, like you already mentioned. You can go-ahead and use them for almost anything but in oscillators there's always one cap that is responsable for the frequency and that should always be a stable polystyrene or silver mica type capacitor. Otherwise your VCO will go out of tune with temperature differences. Also in filters, the caps responsable for the actual filtering will almost always be polystyrene or silver mica types, but it's always stated what you should use. Most caps in the pF range are used for filtering of power supplies or to stop opamps from oscillating and they can be ceramics, no problem. The blue ones are indeed always electrolithic caps and the little light blue stripe indicates the negative pole. I use both axial and radial electrolithic caps in my designs so the elongated ones are axial, but is doesn't matter what type you use as long as the value is correct. Electrolithic Caps directly on the power rails can be any value from 10 to 100µF. That doesn't really matter. In the layout software (DIYLC) I use the yellow symbol for ceramic caps for everything that is not polarized, so if you see a yellow capacitor in my layout it can be any type that is not polarized but if it is anything other than ceramic it will always be noted, usually on the layout itself and otherwise in the accompanying text. Okay, I hope this helped you out and if you come across anything you don't understand, just ask me. I'm always here to help out ^____^ Good luck with the build. I hope you'll enjoy it as much as I did.

      Delete
    2. Hey
      Thank you very much. This really helps me a lot. I think I'm all prepared now.
      I will give you feedback, how the whole project turned out when I'm done.

      Delete
    3. Excellent! I look forward to finding out how you experienced the build and what, if any, problems you ran into. It should be pretty straight forward though. And contact me if you have any questions. Good luck!

      Delete
  4. Hey man and thank you for putting all this usefull information out there! I just finished building my first electronics project, a guitar pedal and it worked quite well! Now I am thinking about building synth. I already studied the original digisound 80 VCO design and made a pcb in kicad out of those. Now I am asking if it is really possible to swap the CE3080 with the TL074 since the CE3080 is really hard to get. Also could I use the original pcb design and just change the CEM3340 to the AS3340?

    ReplyDelete
    Replies
    1. Hi Nils, You can not swap it for the TL074. That is just an opamp. You can swap the CA3080 for the LM13700 or LM13600. Those two have two CA3080's in one chip and a few buffers. I did that with the Korg MS20 filter for instance. Btw, the CA3080 is not hard to get. I just ordered 10 of them on eBay for 4pound94.

      Delete
  5. Hello Eddy,
    really great site, very detailed !!!
    thank you for specifying which resistances have to be changed so that the vco also works with 12v eurorack!

    greetings

    ReplyDelete
  6. Great job on this, I'm definitely building this, thank you!

    Just a small tip to others building to avoid confusion: I think the BOM linked is for the original Digisound 80 VCO and not Eddy Bergman's updated design for the AS3340. (the reference designators in schematic does not match with BOM values).

    Other than that, happy building everyone!

    ReplyDelete
    Replies
    1. The BOM is for the stripboard design I made, only the component numbering does not follow that of the schematic. I'll make a note of that in the text. Thanks for pointing it out.

      Delete
    2. @ZnakeByte I made a new BOM and the numbering now corresponds with that of the schematic.

      Delete
  7. Any tips how you can adjust the VCO without midi keyboard ?

    ReplyDelete
    Replies
    1. If you don't have a MIDI keyboard you can use an accurate voltmeter and a bench powersupply that can go down to 1 Volt. There are tables online that can tell you the voltage for each note and that way you can tune the VCO without a keyboard. You can note down the voltage for C1 (1,0833V) and C5 (5,0833V) and feed that into the VCO, switching between the two voltages, and then measure the output frequency or tone with a tuner. Then you can use the tuning method I describe in this article to tune the VCO.

      Delete
  8. Maybe a bit more introduction, I am an electronic hobbyist, but without any musical knowledge. I built the VCO and it works. I tried the internet to find out which voltage corresponds to C5 or c2, but cannot find anything.

    ReplyDelete
    Replies
    1. I have a list of those voltages for you. You know what. I think this might be useful for others too so I will put the list in this article so you can easily refer to it. If you wait a moment the list will appear in the article above. =)

      Delete
  9. Wow, superthanks for the very quick reply

    ReplyDelete
    Replies
    1. You're welcome =) The list is up at the bottom of the article. It has the 1V/Octave scale and also the exponential scale and corresponding frequencies. I hope this helps you out.

      Delete
  10. Thanks, the AS3340 works nicely. That is it oscillates from something like 0.1 hz to 90 KHZ, but I cannot get the octaves per volt right on plus and minus 12 volt. So I have to experiment a little I think.

    ReplyDelete
    Replies
    1. Ah okay. I've had no complaints from any other readers on this subject yet but then this VCO was designed to work on +/-15V. If you manage to solve the problem, can you reply with your solution so I can add the info for other people to read. I hope you can solve the problem.

      Delete
  11. Hi I am still struggling with the calibration. On internet I find that C5 is 523 Hz, but in your table it must be way above 1 kHZ. If I follow your table a C4 is 4.08 volt and should produve a tone of 1046HZ

    ReplyDelete
    Replies
    1. I don't know where on the internet you found that but C5 is a really high note and you can easily see that it can never be 523Hz. That's actually C3. Are you sure you didn't mix up the Volt per Octave scale with the Hertz per Volt scale (as used by Korg for instance)?

      Delete
  12. Eddy, thanks for the reply. I found it here: http://www.sengpielaudio.com/calculator-notenames.htm

    ReplyDelete
    Replies
    1. This is all analog piano. You need to look at synthesizers specifically. The list I put on my website is the one you need to tune this VCO. So you don't need to look at anything else. ;-)

      Delete
  13. https://en.wikipedia.org/wiki/C_(musical_note), same here

    ReplyDelete
  14. I cannot recommend running this at +12 volt, just can't get it to tune and I tried modifying several components.
    Then tried 15 volt, no problems tuned in 5 minutes. So I will make a small step up converter from +12 to +15 volt.

    ReplyDelete
    Replies
    1. Sorry you couldn't get it to work. I'll make a note of it in the article.

      Delete
    2. Well so far I have not been very succesfull, the triangle and sawtooth signals are very clean. The square signal however below 1500 HZ is a mess, not stable. I think somehow something is wrong.

      Delete
    3. Yes, there must be a flaw in your build somewhere. I hope you manage to figure it out.

      Delete
    4. Found the error, at a certain moment I could not trim anymore with trimpot B? So I suspected it being faulty, soldered it out. Measured and indeed broken. New one in, working. It now works nicely and I can trim it on plus and minus 12 volt with R23 being 680 ohm. R4 I made 200k, I also looked into the datasheet and there is a function between frequency, power supply voltage and C7. I changed C7 into 0.5 nF, two 1nF in series. The only strange thing left is the square wave, out of the AS3340 comes a nice and clean signal but when I connect it to opamp 2-b it becomes a mess, the circuit I build is correct, I have checked it multiple times.

      Delete
    5. Well, I'm glad you figured it out. It's weird tho that you need to change C7 into a 500pF. Changing R4 to 270K should have done the trick but nevermind. If it works for you then it's okay. Strange that the squarewave is giving you trouble. There must be something wrong with the signal path if the squarewave on pin 4 looks good. Did you check the zener diode?

      Delete
    6. From the datasheet of the CEM3340 I took the following formula f=3*Ieg/((2*Vcc*CF). It also says Cf shall be 1000 pF at +15 volt. Since there was this clear relation I found it logical to change Cf for +12 volt.

      Delete
  15. 1 chip making my life difficult :)
    Last thing I built was this: https://www.reddit.com/r/beneater/comments/gtdr9x/i_already_had_an_analog_input_today_i_added_a/
    Almost 100 IC's.

    ReplyDelete
    Replies
    1. WOW! That's impressive!! Well, it must be the 12V that's making your life difficult with this build.

      Delete
  16. Now that the circuit works, I have to get it on a eurorack front, https://github.com/rolf-electronics/Modular-Synthesizer/blob/master/DSCF8317.JPG
    Man, these things have loads of buttons. But it's fun to do. Next I have to copy your mixer :)

    ReplyDelete
    Replies
    1. Haha well, you can be proud of what you created. The panel looks great! You'll love that mixer. I built 2 of them and use them all the time. The clipping indicator works really well too. Are you going to include that?

      Delete
  17. Yes,I will include the clipping indicator seems like a usefull feature.

    ReplyDelete
  18. Ik lees nu ineens dit; Welcome to my website. - I'm located in The Netherlands.

    Groeten uit Hoogvliet

    ReplyDelete
    Replies
    1. Hahaha dus al die tijd hadden we nederlands kunnen praten. Oh well. Groeten uit Delfzijl in de top van Groningen. :)

      Delete
    2. Ik ben trouwens bezig met een nieuwe versie van deze VCO. Die wordt inclusief de sinewave uitgang mat de CA3080 chip zoals het originele ontwerp. Binnenkort in dit theater ^____^

      Delete
  19. Nou dan koop ik maar een kaartje :) Want die sinus converter moet ik nog bouwen.
    Heb een tweede VCO gebouwd, werkte in 1 keer. Met 200k voor R4 en 0.5 nF voor C7. De squre wave is qua kwaliteit slechter dan de driehoek en ramp, best apart. Zal wel aan het feit liggen dat de IC een "copy" is van de CEM3340.
    Next up een mixer, de mechanica zit al in elkaar.

    ReplyDelete
    Replies
    1. Raar dat je zoveel problemen hebt met de squarewave. Ik heb net vandaag de nieuwe VCO gebouwd en de drie golfvormen zijn allemaal perfect maar ik kan er nog geen sinus golf uit krijgen. Ik denk dat er een typ fout in het schema zit. Nog even verder expirimenteren :) Ook raar dat je een 0.5nF voor C7 moet gebruiken. Ik heb nu 5 van deze VCO's gebouwd en ze werken allemaal op 1nF. Nouja het gaat erom dat het voor jou werkt en ik ben ik elk geval blij dat ie direct werkte :) De mixer is ook een leuk projectje. Good luck!

      Delete
  20. Veel komt denk ik doordat ik op plus en min 12 volt werk.

    ReplyDelete
  21. Hi Eddy! And thank you so much for this amazing site! Being rather new (or at least uneducated/self taught) in electronics, even though I have a background as an electrician, I found very valuable use for the very pedagogic and well described builds on your page. In fact, this is where I mainly get most of my inspiration for my projects and the knowledge to make them happen. So again, thank you!
    Now, to the topic. Some short while ago I built two oscillators using two CEM3340 from a somewhat reworked version of the LMNC original design. They sure do sound really good, but as you also mentioned they are almost impossible to tune. And also like you wrote, I’m missing a bit of the low end in range. So now I’m building this “Digisound” version with your designs to it. However I’ve now come across a little problem regarding the 1nf polystyrene capacitor, I don’t have any of those and I might not be able to get some in time... I’m an artist working a lot with sound and I need this to be up and running for the next weekend since I have a show then. Could I replace that capacitor with a 1nf polyester or possibly a monolithic cap?

    All the best
    Simon
    (Sweden)

    ReplyDelete
    Replies
    1. Hi Simon. Thank you for the kind words! You can use a polyester cap. That's no problem. The only thing you shouldn't use is a ceramic capacitor because they drift with temperature but polyester will be fine. Good luck with the show. I would love to perform on stages. That sounds really cool. =)

      Delete
    2. Thanks for the quick reply! Perfect, then I’ll go ahead and use that instead.
      Of course you should perform! Your synthesiser is in itself such a beautiful piece of art! And also how it sounds and what it can do, well that would be more than enough for me at least to find a genuine interest in any show/appearance you might perform with it.
      Thank you! Actually it’s a performance in the woods in the north of Sweden. I’m a student at Umeå Academy of Fine Arts, so my practice is more like working with sound (or music if you want) within an art context than the opposite (doing art in a more musical context). For me this “difference” makes all the difference, so to speak. For example, the arts aren’t so focused in producing a product as such, it’s more about the investigation of an idea, if you get how I mean? For me this allows for a much bigger freedom of creativity not having to be restricted by already existing forms and rules.
      Now this got a little long, haha... sorry for that!
      Simon

      Delete
    3. Wow that is absolutely awesome. I come from a very artistic background myself so I know what you mean by not being restricted by conventional limits like art as only music. I that context I could indeed perform as a sound artist in the broad term or as a sound sculptor using the synth as the tool to do that. You're so lucky to be an arts student Simon. I really envy you. My father didn't trust my artistic capabilities even though he was a painter himself, and insisted I go the technical direction to be assured of a good job. That didn't turn out well for me but I was able to prove him wrong when I won a national competition with a portrait I made of queen Beatrix of the Netherlands, which hung next to the work of Andy Warhol in the 'Loo' palace in Apeldoorn in the Netherlands for 7 months. Anyway, I mustn't drag on. Good luck mate!

      Delete
    4. I definitely think you could! Make sure to announce it though so I wouldn’t miss it. :)
      With risk of sounding...i don’t know, spoiled maybe, but I really do consider myself lucky in that sense that I am able to study arts. But being in my late thirties now (having done 4 out of a total of 7 years) it hasn’t been neither a straight line or an easy thing to get here. A lot of sacrifices have been made along the way. I’m also one of the quite few (at least here in Sweden) who’s not from a somewhat wealthy background which can allow one to count on financial support should things not always work out. But still, I don’t regret deciding to risk it all pursuing this dream of a life as an artist. For me, as long as I manage to survive it’s always worth it. As mentioned before, I worked as an electrician for some 8 years or so, really trying to NOT put all my eggs in such an insecure future as the arts. But eventually I just couldn’t do it anymore. So here I am, with all the eggs in possibly one of the shakiest bucket imaginable... haha!
      What’s your background? Quite curious of where all your knowledge with this synthesiser building comes from?
      Thank you! I might need it, never performed with a modular set up before and frankly I’m quite scared to do it!

      Delete
    5. I think you made the right choise. You have to go with your heart because you only live once and you can't just work just for the income without having pleasure and fulfilment from your work. That leads to burnouts in the end. As long as you believe in your own capabilities you can succeed. As to my background. I was always interested in electronics. In the 80's I used to build pirate transmitters for the FM broadcast band. I played a bit of electric guitar too but I never really played in a band. The interest in music and synthesizers flared up again not so long ago. I have a mate who is friends with Tom Holkenborg (aka Junkie XL) and he has some vintage synths and so I got to thinking. I could never afford these things but I could build them so I did. I knew nothing of synthesizers when I started out in October last year, but it has been quite an eye opener and I learned a heck of a lot since then. I already had this website for a few years so I decided to use it to document my builds and to share with other people all the things I learned and all the schematics etc that I used and the layouts I made myself. It's taking of nicely. I get at least 200 visitors a day now on this website.

      Delete
    6. Yes, you’re absolutely right. That was the way everything was heading when I decided to change and follow my heart, or even my calling if you like...something like that.
      Impressive how you managed to build all that in such short time.
      Interesting with the FM radio! Actually my last exhibition was an installation with a crystal radio I built as the centrepiece. :)

      But back to the topic, I found some old caps on a circuit that used to belong to an old radio that I took apart some while ago. But I’m not really sure what they are. Could I send you a picture somehow if you’d like to take a look?

      Delete
    7. Yes, we can connect on Facebook if you're on Facebook. https://www.facebook.com/EddyBergman
      Message me there and you can send me the picture.

      Delete
    8. Perfect, thank you! I sent you message there now

      Delete
  22. And also, WOW! That’s beyond impressive! Is it still on display somewhere? Would be nice to see! Also happy for you that you could get that kind of revenge on his disbelief in your capability. Such a sweet kind of victory! :)

    ReplyDelete
    Replies
    1. Yeah it sure was. Here's a link to my Flickr acount. I just uploaded two pics. One of the drawing, which is a drawing I made in Photoshop consisting of close to 100 layers. Took me about a month to make it. The other is one of me with the portrait which was the official picture of 'the artist' which hung next to the portrait with some text about me.
      https://www.flickr.com/photos/ededitz/50093673986/in/dateposted-public/
      https://www.flickr.com/photos/ededitz/50093673851/in/dateposted-public/
      This was in 2013 and I was 43 in that picture. =) How time flies LOL

      Delete
    2. Really like the depth in the work. Also how it makes me feel as if her eyes is telling me something, there’s something organic to it and also I feel like there is something hidden in the layers that’s just waiting to be revealed. I like it a lot!

      Delete
    3. Thank you! I should explain about the competition that I was one of 65 winners out of almost 2500 entries. Most of the other winners were professional artists. I was one of only three amateur artists that won. Made me quite proud haha. Glad the you like the work :)

      Delete
  23. Hello I had another question. Im analyzing your circuit and i dont get why theres a zener diode in he voltage regulater part of the square wave out portion. With the 2 resistors you already get the voltage down to 10V so it should be unnecessary?
    Thanks in advance
    Leon

    ReplyDelete
    Replies
    1. So i just calculated it myself and realised that the voltage devider circuit only brings it down to 11.5 Volts. Why arent you using higher Resistance values tho to bring it down to 10Volts while not using a diode OR just hooking the Zener diode up directly, without the first part, it would bring the voltage down to 10V while the signal is high right?.

      Delete
    2. I can understand your thought process here but the resistors on the Squarewave output are not there to regulate the amplitude of the squarewave. The squarewave output (pin 4) is an open NPN emitter output and requires a pulldown resistor for it to work. Because you also want to be able to pull a little bit of current from that output (0,6mA) you don't want to high a resistance value in series with it. Also, resistors have tolerances on them which make them inaccurate. It's way easier just to clamp the output with a 10V zener diode. That way you always get a 10Vpp squarewave.
      And btw, this is not my circuit. It's the Digisound-80 VCO. The only thing I did was make a stripboard layout for it.
      Okay, I hope that answers your question. There's more info in the original document, link at the bottom of the article.

      Delete
  24. Hey Eddie I had a question regarding the two 100nf caps between the two IC's on the stripboard layout. What are these for exactly, I couln't find them on the schematic

    ReplyDelete
    Replies
    1. Oh sorry I replied so late. I didn't get a notification of your comment. The two 100nF caps are decoupling capacitors. They are there in case there's noise on the power supply voltage. I mention in the text that I did not include them in the schematic because I don't use them in my builds but a lot of people ask me about it and they want those capacitors included so that's why they are on the stripboard layout. You can leave them out if you wish at least if you use a linear powersupply. If you use a switching powersupply I would leave them in because they can be noisy.

      Delete
  25. For the correct control range at the input of the IC 3340 PWM from 0 to +5V, resistor R18 must be moved to the upper end of the 100K potentiometer and to +15V. Now the range PWM is from 0 to +7.5V.

    ReplyDelete
    Replies
    1. I adapted the article and implemented the changes you suggested. Thanks again for commenting about it.

      Delete
  26. For the correct control range at the input of the IC 3340 PWM from 0 to +5V, resistor R18 must be moved to the upper end of the 100K potentiometer and to +15V. Now the range PWM is from 0 to +7.5V.

    ReplyDelete
    Replies
    1. Ah yes. That would give 10 volt over the potmeter which will be devided in half by R19 and R20 to 0 to +5V. Why didn't I think of that ^___^. Thank you!

      Delete
    2. I tried your suggestion and it works perfectly!! The PWM now uses the full throw of the potmeter.

      Delete