Tube Pedal Voltage

[tomstuder]

I recently experimented with a 12at7 tube running at a plate voltage of 33 volts. I was positively surprised to see that you can get some unclipped gain with a very linear frequency response from the tube. At the same time, it's easy to drive the tube into clipping - just put a linear booster in front of it.

I decided to use this idea for a tube-based overdrive pedal design. The 33 volts can be generated from 9 volts by means of an internal DC/DC converter. Enough gain to drive the tube into clipping can easily be generated with discrete transistors. Having 33 volt available for the transistors helps here too in terms of gain and headroom - no opamps required. The design doesn't need transformers and fits in a typical stomp box enclosure.

Before I started with this project, I looked around for tube pedal designs on the web. Most of what I found either operated at typical tube voltages of 200 and more volts or else was run straight from the 9 volts pedal power.

My question here: does anybody know of tube pedals that run the tube at voltages in the, say, 20 to 80 volt range (i.e. voltages higher than standard pedal power but less than tube spec voltages)?  Any particular experiences with the 12at7 in those designs? Alternative tubes? Any feedback is greatly appreciated!

More info on my particular design can be found here:
http://tomsguitarprojects.blogspot.ch/2014/10/RealTube-One.html

[antonis]

Sorry for off-topic but...

Haven't you noticed any problem with ripple voltage from DC-DC converter..??

[tomstuder]

It took some experimenting. The boost converter I ended up using is the LT1613 from LinearTechnology. It's a fast switching converter. If I remember correctly, the highest noise component I saw on the oscilloscope in one of my test setups was 2mV at 800KHz. It's not an issue when used as plate voltage or collector voltage in my design.

I'm planning on putting together more info and a PCB layout for the converter in a separate post.

[antonis]

As long as I can see, you use "low" voltage only for input jfet and output bjt (and LED indicator ..)

Shouldn't be more practical to feed them with 33 Volts also and eliminate the need for DC-DC converter..??

[tomstuder]

The pedal runs at 9 volts. The DC/DC converter is used to boost convert the 9 volts from the input jack up to 33 volts. The converter is required in any case unless you want to power the pedal with 33 volts.

[antonis]

unless you want to power the pedal with 33 volts.

That's exactly what I mean..
(unless you want to connect it in a chain with other effects with common PSU..)

[tomstuder]

That was a personal choice. I didn't want to rely on external 33V. 9 Volts are easily available on pedal boards. But back to my original question:

Does anybody know of tube pedals that run the tube at voltages in the, say, 20 to 80 volt range (i.e. voltages higher than standard pedal power but less than tube spec voltages)?  Any particular experiences with the 12at7 in those designs? Alternative tubes?

[amptramp]

There were a number of tubes designed for battery portable radios using a 67.5 volt B battery and the 1U4 pentode might be good, but these are filament-type tubes with no separate cathode, so some redesign may be necessary.  There were some subminiature tubes with inline pins like the 1AD4 that were used for hearing aids that are smaller than the normal 7- and 9-pin miniature series.

There was also a run of 12 volt tubes used for a short period of history when car radios eliminated the vibrator power supplies and used a TO-3 germanium transistor for the output.  The 12K5 space-charge grid triode provided the 50 mW power drive to the transistor base but there were a lot of more normal tubes, pentodes, triodes, converter tubes etc. that rounded out the tube complement.  I have an old Buick radio like this.

But as you have found, some tubes normally specified for higher voltages work well at lower voltages but you have to be willing to experiment.

[GibsonGM]

12AU7  *cough cough*

[mac]

What about a 220v to 24v transformer.
It gives 33v DC and with a heatsinked voltage regulator you can have 9v.
Heaters also from the 9v output with a series resistor.

And if you need more plate voltage, a 220v to 12v transformer.
Heaters connected directly to the secondary, 9v with a voltage regulator, and that diodes and caps network also from the secondary to multiply the output voltage by N times 12v.

mac

[tomstuder]

There were a number of tubes designed for battery portable radios using a 67.5 volt B battery and the 1U4 pentode might be good, but these are filament-type tubes with no separate cathode, so some redesign may be necessary.  There were some subminiature tubes with inline pins like the 1AD4 that were used for hearing aids that are smaller than the normal 7- and 9-pin miniature series.

There was also a run of 12 volt tubes used for a short period of history when car radios eliminated the vibrator power supplies and used a TO-3 germanium transistor for the output.  The 12K5 space-charge grid triode provided the 50 mW power drive to the transistor base but there were a lot of more normal tubes, pentodes, triodes, converter tubes etc. that rounded out the tube complement.  I have an old Buick radio like this.

But as you have found, some tubes normally specified for higher voltages work well at lower voltages but you have to be willing to experiment.

Thanks! I was mostly focusing on tubes found in modern guitar amps. But looking at some older types of tubes that were actually designed to run at lower voltages might be pretty interesting. Designing a guitar pedal based on one of those older tubes would make a cool project.

[merlinb]

People have built any number of pedals and circuits running at these sort of voltages. Here's some running at 24V:
http://www.diystompboxes.com/pedals/schems/mini_boogee_v12.gif
http://www.tubecad.com/2008/01/blog0132.htm

What exactly do you want to know? These's very little to say. Pick a plate resistor, tweak the bias resistor, job done... No different from high voltage designs (but a little different from really starved designs).
http://valvewizard.co.uk/Triodes_at_low_voltages_Blencowe.pdf

[tomstuder]

Thanks a lot for the links.

I'm just trying to learn more. http://tomsguitarprojects.blogspot.ch/2014/10/RealTube-One.html is my first design (apart from an earlier prototype). It took me a lot of time and research. I'm just curious to compare it to similar designs.

[sajy_ho]

Actually there are some tube pedals in that voltage range; I remember Greg (zamboo) built one using 1054 charge pump and 12at7 with around 60V on plates; his video demo was mind blowing! Just search for it on this forum. I don't exactly remember its name but it was something like "Thunder struck".

[PRR]

> plate voltage of 33 volts.

Plate *supply* voltage. Your plate voltage is more like 17V.

> surprised to see that you can get some unclipped gain

Tubes are nearly linear devices. Voltage can vary wildly and they "work". Supply voltage may be varied over a wide range and gain will vary slightly.

See the common "R-C Amplifier" tables for tubes. They show Vbb of 90V, 180V, and 300V (a 3:1 range) and gain varying less than 10%.

As we go to smaller voltages, both voltage and current fall-off, so maximum power output falls off. At least as Square of the voltage, actually somewhat faster (about the 2.3 power). Even when not designing a "power amp", any load needs power, even if only microWatts. At low-low voltages you can't get the power you may desire. Again to the R-C charts: max output voltage at specified load is very much smaller with 90V battery than with 300V.

However guitar-level signals are quite small, so this may not be our limit.

The other problem is that tubes act like linear resistors in series with small stray voltages, often lumped as "contact potential" (though cathode work-function also plays a part). As you get toward 10V supply these "small" voltages become not-small. And in general they are in a direction to "shut-off" the tube. At very low plate voltage you may find that you have to take the control grid +positive+ to get a useful plate current. Positive grid operation is usually avoided because grid impedance plummets from infinity toward 1K. The preceding stage may not be able to drive 1K worth a darn; or may distort badly as signal swings negative and positive. However the impedance plummets over a range of a few tenths of a Volt, and we can often sneak small signals through without great harm. Also crystal devices (BJTs, FETs, chips) can drive 1K loads better than vacuum devices, so a mix/match technology can do what the old guys couldn't do.

BTW: if you find the R-C Amplifier table for 12AT7 (I don't have one handy but I will attach some AX AU data), pick a more-likely plate resistor (Rp=100K Rs=220K is happy in many g-amp circuits), then extrapolate the data from 300V to 90V then to 30V, you may get even "better" performance than you have with your 5K Rp resistors. ("Better" may be too much for this application, just saying.) For 12AU7, gain falls from 12 to 11, so you'd predict about 10 at 30V. Eo drops from 41V to 14V so you'd pencil 3V or 4V (still large by guitar-cord standards). Suggested Rk changes from 4400 to 5000 so a trial-value for 30V might be 5.6K or 6.8K.



Note also that going from 5K Rp to 100K Rp is likely 1/20th of the supply current.

Note also that for these values, if Eo(max) is 4V and gain is 10, then the maximum clean input is 4V/10 or 0.4V. Your tube driver only needs like 1V supply. OK, for FUZZ we want 2 to 10 times that much to SMACK the tube grid, but the 9V supply could be ample. Now your 32V supply has MUCH less load. I do not know if that allows a simpler or cheaper voltage-booster (what you have is about as simple as it gets), but lower current in a switcher *may* be less HF hash to shield and filter, or less interference in cell-phones.

> with a very linear frequency response

First-approximation, I don't see why freq response would change much.

Freq resp is often more about current available to drive stray capacitance. And yes, a low volt tube will have low-low-current. However at "real" voltages and 5K load I would expect a 12AT7 to go past 10MHz. (12AT7 is really made for 100+MHz work, though with lower impedances and hot-hot current.) You would have to really lame it to lower the cutoff into the audio band. (And some treble-cut is often useful on heavy fuzz.)

[tomstuder]

Paul, that's great info! Thanks a lot. It'll take me some time to digest it completely.

However, as a quick feedback, here are the idle voltages that I measured:

Supply voltage: 32.3V
Voltage at anode: 31V
Voltage at cathode: 0.03V
Voltage at grid: -0.365V

The measurements for the sibling triode in the tube were similar (they varied by less than 10%).

And yes, the voltage booster in front of the tube is a very simple design. The combination of the gain from the booster and the headroom of the tube work really well together though. The sound is very dynamic, going from near-clean to crunch with little variation in picking strength or guitar volume knob - that was one of the things I was looking for.

[tomstuder]

Here's more info on the DC/DC converter I mentioned at the beginning of the thread:

http://tomsguitarprojects.blogspot.ch/2014/10/dc-dc-converter.html