Experimenting with tubes - Do's and Don'ts?

[timd]

Hey everyone -

So I built the Matsumin Valvecaster, and I fear I might be getting hooked on building with tubes. I have more tubes of different types and sizes, and I located the pinouts/datasheets for each one. The amount of information I know about tubes is limited to my one build (really no knowledge....).

What I really am looking for is this:

1. Is there any hard and fast rules to get a given tube to work properly? ie - are there any connections that have to be made among plates, grids, cathode, etc? This would be useful for V+ and ground, plus input and output. I was planning on using some of the valvecasters connections as a starting point.

2. Is there any danger of doing damage? I'm planning on running my possible builds on 9 volts. I have a feeling that some tubes may not run on that low of a voltage, but I'm concerned that the output of one of the builds could be more that the next pedal in line (or amp) could handle and ruin things. Is this a real concern? I'm planning on having fun with this stuff and doing alot of experimenting, but I don't want to have tragic results. For the builds, I won't be focused on only overdrives, etc. Maybe something that just adds some color would be fun.

[R.G.]

1. Is there any hard and fast rules to get a given tube to work properly? ie - are there any connections that have to be made among plates, grids, cathode, etc? This would be useful for V+ and ground, plus input and output. I was planning on using some of the valvecasters connections as a starting point.

1. The heaters have to be provided the nominal heater voltages, +/- a little bit. Cold heaters mean no electrons to work with, and too-hot heaters wear out FAST.
2. The plate must be positive with respect to the cathode or nothing happens. Tubes run entirely on electrons running from the cathode to the more-positive plate.
3. The grid can be either more positive or more negative than the cathode. But more positive is essentially useless, as when the grid hits 0V with respect to the grid, most of the useful stuff is already over.
4. Tubes conduct full current like a vacuum tube diode (which they are) unless you make the grid more negative than the cathode some way. A negative grid repels the electrons trying to get to the positive plate and cuts down the current. More negative grid compared to the cathode = less current flow.  So in general, grids most negative, cathodes next and plates most positive.
5. Tubes can't conduct much current. A 12AX7 is really hurting to provide more than 1-2ma to the plate. This is partly because every stinking electron has to be boiled off the cathode by heaters and then sucked to the plate by a positive voltage. The higher the positive voltage on the plate, the more the electrons want to get there.

2. Is there any danger of doing damage? I'm planning on running my possible builds on 9 volts. I have a feeling that some tubes may not run on that low of a voltage, but I'm concerned that the output of one of the builds could be more that the next pedal in line (or amp) could handle and ruin things. Is this a real concern?

Not really. Many 9V pedals can put out nearly 9V of signal. If that's all your tube has to work with, it can't really hurt any more than one of the 12V or 18V powered distortion pedals.

However, I think you're going to be disappointed in the amount of gain you get from a 9V plate. It can be done, and there were some special tubes made for running from automotive 12V ( as an aside, "Motorola" started as a company making car radios; they made some low voltage tube stuff, I think).  If you can find some of the low voltage tube types, you'll get better results. Otherwise,  you're in for making the plate supply higher. There are some tube pedals that work from lowish voltages - 12 to 50V - but normal tubes don't really get fired up below 80 or 90V. The gain is very low down in the low plate voltage region.

[mac]

If you are experiment with a tube pedal at 300v to be used ONLY with a tube amp, I guess it is not necessary to reduce the output voltage to 9v.
A resistor bigger than 100k in series with the out cap, and a volume pot are enough?

mac

[Earthscum]

1. The heaters have to be provided the nominal heater voltages, +/- a little bit. Cold heaters mean no electrons to work with, and too-hot heaters wear out FAST.

(and)

5. Tubes can't conduct much current. A 12AX7 is really hurting to provide more than 1-2ma to the plate. This is partly because every stinking electron has to be boiled off the cathode by heaters and then sucked to the plate by a positive voltage. The higher the positive voltage on the plate, the more the electrons want to get there.

What's the lowest voltage (say, 12AX7 and maybe 12AU7, if it's different) that you can run the heater at? I remember reading that people were having no problems on 9V. However, is that still the same amount of current being supplied, or is it cut down by battery? I haven't been able to figure out if some of these people are really running off a battery or not, or if it's just "battery capable", which is kind of why I am asking.

I know you can parallel a 12ax7 heater and run 6V, more current, could it run off 5V, 1A supply with enough juice to step up for plate supply (for a single tube)? Just thinking oddball power supplies, here.

[timd]

R.G. - You have just made made my day! Thank you.

[timd]

What's the lowest voltage (say, 12AX7 and maybe 12AU7, if it's different) that you can run the heater at? I remember reading that people were having no problems on 9V. However, is that still the same amount of current being supplied, or is it cut down by battery? I haven't been able to figure out if some of these people are really running off a battery or not, or if it's just "battery capable", which is kind of why I am asking.

I know you can parallel a 12ax7 heater and run 6V, more current, could it run off 5V, 1A supply with enough juice to step up for plate supply (for a single tube)? Just thinking oddball power supplies, here.

I'm not sure what the lowest voltage would be, but when I made the Matsumi Valvecaster, I didn't bother with a battery clip, as I read that this thing sucks batteries like no tomorrow.

[Earthscum]

Well, 12AX7 needs 150mA @12V, which would make the total resistance 80 Ohms. 9V across 80R is about 113mA. That's alot of suckage!

I guess I should try complete thoughts... obviously I know how to figure up this stuff, I'm wondering how little current you could supply to a heater and still have it do it's job?

[R.G.]

The tube specs say "rated voltage +/-10%". For a 12??7, that comes out to two 6.3V sections. You can parallel them or series them. If they're series, that's 9Vdc into a nominal 12.6V heater, for 71.4% of nominal. That's pretty weak. I suspect that what happens is that since the filaments count on the increase in resistance with temperature, it may be pulling more current at low voltage than you'd expect, which helps with the too-low temp thing a bit.

There are some studies on running the filaments of similar tubes at 5.0 to 5.5V or so. That's -17% to -13%. A "12v" filament would be running at 10V to 11V under these conditions. Still awfully light.

The problem with such low heater voltages is the complex electrochemistry of oxide coated cathodes. All indirectly heated cathodes - and this covers all the tube you're likely to use - have barium and oxides coating the cathodes. This lowers the temp at which electrons boil off with the strict proviso that they not get so hot that it degrades the coating and that gas ions don't poison the chemicals. Coated cathodes produce a huge number of electrons at proper temps and this "electron cloud" shields the cathode from some of the ugliness. At lower cathode temps, the electron cloud is much thinner and more easily depleted. That's counteracted by the low ability of the low-voltage plate to suck them away. Which effect wins?? Who knows?

I've seen reports of a "starved" sound of tubes run at very low currents and voltages. Noise could be worse, because each electron moving charge to the plate carries one packet of charge, and as the sheer numbers go down, the averaging of electrons into smooth current flow doesn't cover up the quantized charge movement as well, so there's a "rain on the roof" effect.

It's worth experimenting, as long as you remember the possible gotchas. Try it! Hook up a 12AX7 with 9.5V on the heater (I personally would not try this on batteries; I'd use a 1Spot, but then I get free 1Spots - which cost about $20 normally) and the same on the plate resistor. I'd make the cathode resistor about 100 ohms and the plate 100K, grid leak about 1M. Then see what voltage happens across the plate resistor. If it gets down from 9.5V by even a volt or two, run some signal through it. See what you get. You may have to tinker plate voltages or whatever.

There's another way to do this. Parallel the two halves of the 12AX7 heater and put a series resistor from 9.5V to the heaters. This costs you 300ma for the heaters, and a 9 (...or 10, or 11) ohm 2W resistor, but runs the heaters at full temp. That only leaves the gain from cathode/grid/plate to worry about. Another 150 ma is not much of a stretch for a power adapter. See if you get usable gain. If not, raise the plate voltage and re-tinker.

[zambo]

I love making 9 volt tube stuff. Its pretty fun. In space charge mode tubes like to develop way to much treble and bass and have an auto mid scoop. Its a challenge to compensate for this with smaller caps and treble roll off etc. Grid stop resistors Are a good thing. I run 9 volts through 4 1n4007 diodes to drop it to 6 volts for the heaters. I also use a 1spot power supply. I dont get them free though

@R.G. - Awesome explanations. Thanks!

[R.G.]

I dont get them free though

That was actually kind of a joke aimed at me. I get the dead ones returned for warranty repairs. One of my jobs is to find out *why* they failed and track it to be sure it's within the bounds of random component failure and not a design problem or a manufacturing issue. Some of these are actually still working correctly, but were returned for other reasons.

So it's kind of "Here's a box of 20 1Spots we've collected over a couple of months. Why did they fail?" and I get to keep the ones which are actually good but erroneously returned. 

I suspect that paying money for them would actually be cheaper overall. 

[DougH]

Here's a tip for decent sound out of a low voltage tube circuit- low impedance.

Check the BK Butler tube driver schematic and you will see what I mean. Of course, this means low gain too. So you will need to drive it with something to get some overdrive going. Something like, oh, say...an op amp maybe? The guy knew what he was doing.

Seriously though, the low voltage tube circuits with the big 100k plate resistors and so forth don't sound so hot, IMO. I've heard the clips, and experimented plenty on the breadboard. No dynamics, no headroom, too much gain. Think about it. You're setting it up like it's running at 250V but chopping the supply down to 9V- think what that will do to the sound. Scale down the grid/plate resistor networks and it starts to make more sense, esp when driven with a hot signal from a gain stage like an op amp. And you can always limit the gain of the op amp circuit so that you don't exhaust headroom and get distortion from the op amp itself.

[Perrow]

So it's kind of "Here's a box of 20 1Spots we've collected over a couple of months. Why did they fail?" and I get to keep the ones which are actually good but erroneously returned.

Can you combine parts from different non working units to make more of them work? Or would that be too much work?

[iccaros]

Some helpful tips using non Space Charged Tubes at low voltage.. This is for Pentodes....
from http://www.junkbox.com/electronics/lowvoltagetubes.shtml


Avoid grid confusion! The first grid (the grid physically closest to the cathode; what in conventional tubing is called the control grid) should be tied to 12V. If you're using a conventional tetrode or pentode in space-charge mode, make sure you put 12V on the control grid, and use the screen grid as the control grid. Pentode suppressor grids are almost always tied internally to the tube cathode, but if the suppressor grid comes out separately to a pin, try putting 12V on the suppressor as well. This may allow greater gain in some circuits, as you have two grids accelerating electrons toward the weakly charged plate.

[R.G.]

Can you combine parts from different non working units to make more of them work? Or would that be too much work?

I could. However, there is live 120Vac (or 240Vac, depending) in there, and then about 350Vdc, also hot to the AC power line as well. Make one mistake and it becomes a Light Emitting Power Supply ( or LEPS~!    ). I used to do this in my technical infancy, but it's not worth the trouble any more. I'd rather pay the cost of a new 1Spot (about US$20) to NOT have to do that. I have learned over the years that I am, in fact, mortal.

[zambo]

I was thinking man....you gotta own and play a lot of pedals to be endorsed by 1spot power supplies...

@ Dough - Thanks for the wisdom, I am currently building a twin tube with a SHO driving them. Will post results 

[Cliff Schecht]

I dont get them free though

That was actually kind of a joke aimed at me. I get the dead ones returned for warranty repairs. One of my jobs is to find out *why* they failed and track it to be sure it's within the bounds of random component failure and not a design problem or a manufacturing issue. Some of these are actually still working correctly, but were returned for other reasons.

So it's kind of "Here's a box of 20 1Spots we've collected over a couple of months. Why did they fail?" and I get to keep the ones which are actually good but erroneously returned. 

I suspect that paying money for them would actually be cheaper overall. 

Can you not crack them open and adjust the feedback voltage divider to increase the output to 9V from 12V? The design can't be that finicky .