Vacuum Tube Synth

Disclaimer
This project includes high voltages which can be dangerous if not lethal. Only attempt this build if you are experienced and well trained in electronics. I take no responsibilty for people who get injured while working on this project. Build it at your own risk.

This project was insane. I had the idea of building the old Metasonix TS modules in my mind since two years. Since these include the work with higher voltages I wanted to gather a solid understanding of electronics and schematics before I attempted this project. In January 2022 I finished my solid-state synth and now I thought I just need to go for it. In this article I want to talk about the experiences I made, how I did the things and what lessons I’ve learned. I’m not going to share my protoboard layouts for three reasons. The first one is that you need to be able to read schematics and to work with those when attempting a project like this. The second is that I’d strongly recommend to use turret boards instead of protoboards for the circuitry or to take the elegant route by designing you own pcb’s. This will make it way easier to assemble this device. The third one is that I’ve made some changes to the layout while working on them.

This article will be a bit longer. Here is the table of contents if you want to jump to specific part:

• The Idea
• Power Supply and Case
• Metasonix TS-21
• Metasonix TS-22
• Metasonix TS-23
• Conclusion

The Idea

I was always interested in vacuum tubes. Coming from the guitar and bass guitar world it was always the best choice to use a tube amp – eventhough they are pretty expensive. Over the time I was wondering why tubes are so rare in the synth universe. Of course you got some pioneers like Friedrich Trautwein who has created the Trautonium for example. Nowadays most synth companies use tubes with low-voltage, which is nice too! And there is one exotic folk, Eric Barbour, who just designs super interesting synth modules – following his principles of making unique and non-standard audio devices with vacuum tubes.

Since this project includes to work with high voltages it was important to me to gather enough experience with electronics. I’m also coming from an engeneering background with a masster’s degree in renewable energies. A big portion of it was electrical engeneering. I’m also used to work with mains voltage in my job. It is important to know what you do when working with vacuum tubes. After I successfully built some complex modules by translating their schematics to stripboard and did enough reading about vacuum tube electronics it was time to attempt this project.

Power Supply and Case

The first thing was to make a case and a power supply. Fortunately I already got a wooden box lieing around which was used for my first diy modular synth. I cut it a bit smaller and modified it a little.

I’ve made the panel like always using sheet metal from the hardware store, drilled all the holes, spray-painted it and put all the potentiometers, jack sockets and switches in. Then I soldered all the wires to it. As you can see, I also equipted the panel with an analogue amp meter. The idea was to visualize the tubes heating process. When the tubes are cold and get switched on you get a pretty high inrush current which is getting smaller when the temperature of the filaments increases. To reduce the current draw in that phase I also decided to build in a stand-by switch. When it’s in standby, only the tubes filaments get voltage (6.3 Vac). When the switch gets flipped over the rest of the circuitry is set under voltage.

After this was done it was time to build the power supply. In this project we got 13 tubes in total. Most of the power consumption is drawn by the filaments. To know which transformer to choose I had a look at the datasheets of all the tubes which will be used. In the datasheets you get the info of the current draw of the filaments in the state of operation – when they are heated up. In total all tubes consume around 4.8 A at 6.3 Vac. Make sure to use thick enough wires here. Since the inrush current can be 3 times as big while they heat up after the device is getting switched on, the transformer needs to be able to deliver at least 15 A at the 6.3 V secondaries. The transoformer I got is the Hammond185F12. It is dimensioned for 130 VA and got two 6.3 V secondery windings (20.6 A). This is more than enough to power the tubes heaters and the circuitry. The second secondary goes to a second tranformer which has two 115 V primary windings and two 6.3 V secondaries (only one of those is used). The second secondary of the first transformer is wired to one of the secondaries of the second one, which is wired in backwards so to say. For the second transformer I’ve choosen a Hammond183K12. It is super important to install fuses before the first and between the transformers. If you skip this you get the risk of burning your house down when something goes wrong.

This is the layout of the power supply:

The rectification is done with an E91AA tube. You can also use rectifier diodes configured as a full bridge rectifier here. After the RC filter we get 113 Vdc for the circuitry. For the circuitry all resistors are rated for 2 W and all capacitors for 450 V (250 V is also enough).

The circuitry is done on protoboards which are mounted to an internal metal chassis. Apart from just holding the circuit bords it also shields the circuitry from the electromagnetic fields of the transformers and provides the central star-point of all the ground connections which finally go to mains earth. Here is a good read about proper grounding techniques with vacuum tube circuits. In retrospective I’d use turret boards instead of proto boards. This would have made the installation a lot easier.

What’s also important is to twist the wires which are carrying AC voltage to reduce the risk of getting the 50 Hz mains hum into the signal path. I just put the wires into my drilling machine, twisted them up and put some heat shrink sleeves around them. Here you get some good information addressing this subject.

The next step was to install the tube sockets to a piece of sheet metal and to solder in all the wires.

The case also got three 6V light bulbs to make it a bit more shiny:

I know it can also be difficult to source tubes and such. This is the retailer I used here in Germany. Here I got the transformers and all the tubes.

Metasonix TS-21

It’s hard to describe what the TS-21 Hellfire Modulator is and what it does. So I just summarize the information from the original source. Definitelly a pretty gnarly sounding thing!

The Hellfire Modulator consists of two independent effects units. The first one is a PWM circuit acting as a monostable multivibrator – a metastable circuit. The switch point of this circuit can be changed with the PWM panel potentiometer. Best practice is to feed it with sharp waveforms like saw or square.

The second one is the beam modulator with a tube origanilly intended for applications in TV’s. By electrostatic deflection the signal switches between two anodes, which results in a stereo output. You can also plug in a mono cable – getting a mono output. Last but not least it also has a built in LFO which can be switched off.

What’s also important to mention is that the labeling of V3 and V4 is reversed in the schematic! So V4 and V5 should be the same instead, not V3 and V5. Another useful tip I found in a post on wiggler is to put a 100k resistor between pin 6 of V4 and V5 and ground. Here is the schematic. I also figured that there was only a tiny useful bit on the L2 potmeter. I swapped it for a 500k instead of 2M5 and put a 2M resistor between point W of the schematic and the potmeter. I would also recommend to use an even smaller potmeter like 100k. You can experiment here, just keep in mind that you get 2M5 in total. This device needs also some time to understand and some dead spots are normal. It needs waveforms with discontinuities like saw or square waves. If you put a square signal into the signal and PWM input you can have a lot of fun with it. Hint: experiment with the pulse width and the frequency of your incoming signals. The Hellfire Modulator really lives up to its name!

Okay, enough talking, here is how it sounds:

Here are two other demos, not mine, which I found on Youtube:

Metasonix TS-22

The TS-22 Pentode Filterbank was a total secret to me. I wasn’t able to find any information, like demos, for that device.

It consists of four independent bandpass filters with separate resonance and tune controls each. They mix down to a VCA. Each filter is based on a pentode tube with a simple Twin-T circuit. The CV is driven by a little opamp circuit. The schematic does not explicitly tell the pin numbering of it. So I asked my friend Eddy Bergman and he just passed over this little sketch. Thanks! I used a TL071 for it, since this is what I got in stock already.

The tricky part here is that you have to make a quad vactrol consisting of one LED and four LDR’s. I’ve choosen a white LED and the specs of my LDR’s are 50k-160k/20M. I matched them by switching on my desk lamp to get a consistant light source and took the four with the most similar readings. Then I crammed them together with some heat shrink sleeves and a little hot glue. I was lucky and it worked out from the get go!

I couldn’t find a single demo of it, but here is how mine sounds. Maybe it’s the only demo in the world wide web. Happy how it turned out!

Metasonix TS-23

The TS-23 Dual Thyratron VCO sounds unlike any other VCO I’ve come across. Each oscillator is based on a 2D21 tube which acts like an instable switch allowing a capacitor to charge and discharge. The cycle of this process can be adjusted with the tune potentiometer, thus changing the pitch of the sound you’re listening to. It delivers a saw wave that isn’t perfect. This results in an unique and pretty raw tone.

Another feature of it is the beam modulation. This results in a squishing effect, based on a very nonlinear waveform clipper. But I couldn’t get this portion to properly work yet.

Again, this circuit is driven by a little opamp circuit and the schematic doesn’t tell the pinout. Kindly Eddy sent me another quick sketch to make it clear. I used a TL071 like with the TS-22. It worked also right away but I haven’t really tuned it yet.

And there is still another problem. The level of oscillator 2 is way lower than of oscillator 1. This can be solved by tweaking some resistor values, but I leave it for now.

Here is how my TS-23 sounds:

Here is another demo from someone else:

Conclusion

All in all it was by far the most complex diy project I’ve ever done. I also gathered a lot general knowledge in electronics like proper grounding techniques and electromagnetic fields resulting in hum or noise. In total it draws about 55 W – 80% of it caused be the tubes filaments.

Also there was way more crafting work involved than I expected before – especially the wiring. One quote of one of the wiggler posts is definitelly on point:

“Tubes suck, okay? They do amazing things but hate you along the way.”

Again a shoutout to Eric Barbour of Metasonix for publishing his old schematics!

I’m super happy that it is in my arsenal now – a super unique piece of gear! It’s so much fun to play around with it’s gnarly sounds and it’s great that it is also compatible with solid-state synths. The combination of both is where it really starts to shine!

That’s it for know. Thanks for passing by! If you got any questions or feedback just leave a comment below or come over to my Facebook group. Cheers!