Close
0%
0%

Tube Synth

Moog Modular style Synthesizer using Vacuum Tubes (and diodes)

Similar projects worth following
I've been working on various tube synthesizer ideas on and off for a few years. The main problem has been keeping the scale in tune when using tubes, and keeping the frequency in tune when using silicon. For example, using transistors in the same cabinet as the tube oscillators etc causes the transistors to heat up and go well out of tune. I'm now using a triode to do the linear to exponential conversion, which gives me about 4 octaves of correct tune, and more on a good day. It's very temperature stable as well. Mains power is a problem with a small change in the mains voltage causing filament temperature to drift, which causes it to detune.

The Oscillator

The oscillator is a regular Phantastron oscillator (6AU6) configured to generate a sawtooth waveform. The output on the anode is wired directly to a cathode follower (right 6BQ7A) to buffer it, with the integrating capacitor wired from the cathode back to the pentode's grid. The summing node on the pentode is where current is applied, which generates a frequency proportional to the amount of current flowing into the node. The voltage does move around a lot because there isn't a lot of gain in a single pentode.

The voltage to current conversion is done by the first triode. I found the 6BQ7A dual triode to be the best fit but a 12AT7 will also work. To get a constant current out of the triode you need to have a constant voltage across the cathode and grid. Because the cathode is connected directly to the summing node, which moves around, the control voltage must be referenced to the summing node, i.e. it must float.

The floating control voltage is generated with a switched capacitor network using reed relays. The tuning voltage comes from a simple resistor divider circuit (labelled coarse and fine) and also gives an optional AC modulation input. This voltage is sent to the lower part of the capacitor network. The upper part is connected to the CV input itself, with a trim pot to tune it to 1V/oct.

When the first two relays are energised by the AC filament voltage, the difference between the two control voltages is dumped into the 100nF capacitor - the tuning voltage subtracts from the CV input. On the second half of the AC cycle, the charge in the 100nF capacitor is dumped into the 10nF capacitor, which sets the voltage across the cathode and grid of the triode. It then stores the control voltage between cycles. The 100nF capacitor is bigger so it can charge the 10nF capacitor more quickly, which reduces portamento.

This means there's a sample rate for the control voltage running at mains frequency, but it means I can avoid op-amps. This means there might be a short blip at the start of each note when the frequency changes, caused by the slight delay from the sample frequency.

Voltage Controlled Amplifier

The objective of the VCA is to use as few tubes as possible. Therefore I used a transformer to convert the balanced output of the differential pair (12AU7) to unbalanced. This saved a tube.

Those familiar with VCA designs may recognise the topology of this circuit. There's a differential pair at the top (12AU7's), with the current through it being controlled by a constant-current sink made from a pentode (6AU6). The current through the pair determines how much gain the circuit has. The current is controlled by the grid voltage of the pentode.

There is the problem of getting the 0V to 10V control voltage down to -10V to 0V, or thereabouts. I found a balanced resistor network works quite well. The idea behind this is to have an equal 'pull' between the top 150k resistor and the bottom 150k resistor. The value of the 10k resistor determines how far down (or up) the signal is shifted (with some attenuation). Replacing the 10k resistor with a pot arranged as shown gives a manual control over the DC offset. The control voltage is applied over the top of the manual control.

Envelope Generator

The envelope generator is a convoluted mess, so let me explain.

The trigger is a Moog style S-Trig input to keep the control voltages low. Normally the input floats at 20V and the resistor network biases V1A to conduct. This also switches V4B on and V4A off in the flip-flop. 

When the trigger input is grounded, the voltage on V1A's grid swings negative, and consequently the plate voltage will swing positive to about 120V. In response, the cathode of V1B will follow the voltage and current will flow through the attack pot into the 4.7uF capacitor.

V3A and V3B make a Schmitt trigger. When the capacitor voltage gets to its threshold voltage through the 100k resistor, it will sharply switch the plate voltage...

Read more »

LittleFugue_TubeVCA.mp3

More Bach. Little Fugue in G Minor, multi-tracked. Watch out for the bad tuning. I'm sure Bach would not approve! I've got a different type of tube on order that might fix it.

mp3 - 3.63 MB - 11/17/2019 at 10:41

Download

Tube_BWV1006_VCA2.mp3

A MIDI version of BWV1006, a solo for violin by Bach. https://en.wikipedia.org/wiki/Partita_for_Violin_No._3_(Bach)

mp3 - 4.32 MB - 11/11/2019 at 05:33

Download

  • Paint!

    256byteram11/18/2019 at 09:50 0 comments

    I've not updated about the paint situation.

    The first thing to do is stamp in the labels for the various controls. I used a lettering kit available on eBay.

    The areas where I didn't want black paint were masked off with tape - the flanges and screw holes. A couple of coats of satin black spray paint were applied, then white acrylic paint was applied to the lettering. I found the letters had to be stamped quite deeply for the paint to hold. I could also try an oil paint.

View project log

Enjoy this project?

Share

Discussions

Michael Wessel wrote 11/06/2019 at 17:55 point

Can we hear it please?

  Are you sure? yes | no

zzeroo wrote 11/07/2019 at 09:16 point

Yes please, upload some sound examples. I would also like to hear this thing!

  Are you sure? yes | no

256byteram wrote 11/11/2019 at 05:38 point

There's an MP3 in the files list. Sorry for the delay. I didn't see your comment until just now.

  Are you sure? yes | no

Dan Maloney wrote 11/05/2019 at 00:55 point

Moog-Tube! Cool!

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates