Spark Gap 2: New Low-Voltage Tube Overdrive

[merlinb]

A few years ago I experimented with tube diodes in a feedback-clipper type circuit. That project seemed to generate some interest, although it was a primitive circuit and the level of overdrive was kinda lame. I decided to re-visit the concept, and have come up with something far better. Meet Spark Gap 2, a low-voltage tube distortion pedal that can deliver much higher levels of flame and a crisp tone, unlike the fart-under-a-blanket you get from most starved tube circuits!
https://valvewizard.co.uk/sparkgap.html

I hope to create some sound clips at some point, watch this space!


 

[printer2]

Can't wait to hear the first sample.

[frequencycentral]

This looks great Merlin! Love those tubes sticking out too.

Quiet thread! This would have been on fire 10 or 15 years back.

[FleshOnGear]

Very cool! How did you find out that the threshold of the 6AL5 increases with less voltage on the heater? That seems like a real game changer, and quite lucky since pedals typically run on 9V. The Fender Roc Pro amps from maybe 20 years ago used tubes as clipping diodes, but I think they had silicon diodes in series with the tubes to deal with the threshold issue, IIRC. I've never played one. I really want to try the Spark Gap 2!

[stallik]

This one's on my build list even before hearing it. Valves can still be had for less than the cost of the valve base. I have a sneaky feeling that might change if this circuit gains traction

[amptramp]

Are you counting on conduction in the 6AL5's to set the bias for the second stage?  What differences in bias do you see with SW1 on or off?

IIRC there was a graph in the original spark gap thread showing different conduction levels for various heater voltages.  Maybe running the two heaters in parallel with a pot to control the heater voltage could bring you to a more decisive clipping threshold.

[PRR]

less than the cost of the valve base.

A valve with no DC bias and reduced heater does not need a base/socket. 6AL5/12AL5 tend to last 10,000 hours at rated heater and much-much longer underheated.

[stallik]

Are you suggesting just directly soldering the tube pins directly to the PCB? Would save some room

[mozz]

I don't know if you can get solder to stick to the pins. I mean if you are not trying of course it will stick but on purpose i don't think it will flow unless you brighten them up.

[merlinb]

Ugh I've had a nightmare trying to make a sound clip. I'm not really set up for recording at home anymore, so I used a looper pedal to record a riff, then brought it to my work computer. But I ended up recording a click over it somewhere, and then half way through trying to figure out how to use Reaper it somehow cut off the start of my riff! Yeesh. Anyway, here's a crappy sound clip at various settings. I'll try again at some point.
https://valvewizard.co.uk/SparkGap22.mp3

[fryingpan]

I'd suggest two things:

1. Use an IR loader to simulate what the circuit would sound like through a guitar speaker. There are several free ones, including one in Reaper. Free IRs are plentiful too. Just use the IR you feel sounds most like what you hear through your speaker.

2. If you're not comfortable using Reaper, Audacity is more than good enough for these things. It can also apply VSTs non-linearly now, so you don't need to render the file every time you want to sample effects.

I would also imagine that the circuit might benefit, for the solid state part, from higher operating voltages (so that the opamp never actually clips, although it is debatable whether the opamp clipping would be audible through the valve diode clipping). A simple zener regulator on the valve heater should do the trick?

[fryingpan]

FWIW, I passed the file through a cabsim (a 4x12" loaded with 2 V30s and 2 G12T75s; this IR is of one V30 cone mic'ed with an AT4050 on axis).

Here it is:

https://whyp.it/tracks/281943/sparkgap2-cabsim?token=1r8aq

If I may, the pedal might benefit from further/stronger filtering in the low end especially. I suppose the sample was made with the tone flat and varying the gain level.

[amptramp]

As for the suggestion that you could solder the tubes directly to the board, the pins are made of a nickel steel called Kovar so that the pins will have the same temperature coefficient as the glass envelope of the tube.  It is not solderable and any attempt to solder it will likely result in cracking since the pins will be hot but not the glass.

Another idea for this would be to use the octal equivalent, the 6H6, in place of the 6AL5.  It is most commonly seen as a metal tube and might last longer on tour.  For a studio effect, it won't matter and the two tubes have identical characteristics.

[drdn0]

As for the suggestion that you could solder the tubes directly to the board, the pins are made of a nickel steel called Kovar so that the pins will have the same temperature coefficient as the glass envelope of the tube.  It is not solderable and any attempt to solder it will likely result in cracking since the pins will be hot but not the glass.

Another idea for this would be to use the octal equivalent, the 6H6, in place of the 6AL5.  It is most commonly seen as a metal tube and might last longer on tour.  For a studio effect, it won't matter and the two tubes have identical characteristics.

God the 6H6s look cool though!

[printer2]

As for the suggestion that you could solder the tubes directly to the board, the pins are made of a nickel steel called Kovar so that the pins will have the same temperature coefficient as the glass envelope of the tube.  It is not solderable and any attempt to solder it will likely result in cracking since the pins will be hot but not the glass.

"Kovar may be readily welded, brazed or soldered to other metals."

https://www.efineametals.com/wp-content/uploads/2020/06/Joining_Kovar_Alloy_to_Other_Metal-_Brazing_Welding_SolderingR.pdf

I have soldered to the pins before.

[FleshOnGear]

The 6H6 does look cool, and I like the thought of a more durable metal envelope. But does it have the property of increased threshold of conduction that the 6AL5 has? Maybe putting germanium diodes in series would raise the threshold to an acceptable level. You'd probably need a larger enclosure for two octals, though.

[mozz]

"Lowering the cathode temperature in a vacuum tube diode reduces the contact bias that occurs when electrons emitted from the cathode contact the plate. That bias must be overcome before rectification can take place. The effect of contact bias in a vacuum tube diode is similar to the forward drop in a silicon diode. Both National and Hallicrafters used filament dropping resistors in the 4 to 7 ohm range for the 6H6 and 6AL5 tubes used for detectors and noise limiters in many of their receivers."

Here's my understanding of how the full filament current and cathode higher temperature (filament dropping resistor not being used) causes a low level of hum in the tube noise limiter circuit. With no plate voltage applied to the limiter (a break in audio level (cathode to plate equals zero volts)), there is a very small amount of current flowing to the plate. Because the filament is being heated by an AC current, the electron cloud on the cathode (which causes this small current flow) is fluctuating at that frequency, so the very small current flow to the plate is also fluctuating at that frequency, 60 cycles. Hence, we have AC hum in the limiter plate circuit. This, of course, will go directly to the audio input. Normally we will not notice the hum because there is usually sufficient audio signal level to obscure it (an audio signal equals voltage on the NL plate). But, when there is no audio, then it is noticeable. Zero audio happens often enough that it is a problem.

The issue won't appear when powering the filament via DC, the hum problem occurs when a "parasitic" diode is formed and the filament/cathode forms an inefficient half wave rectifier. Slightly reducing the temperature of the cathode further reduces the activity of this undesired rectifier removing the hum issue.

Drake took another path using an 8BN8 instead of a 6BN8 in their 2 series receivers instead of using a dropping resistor.