Page 1 of 1

Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 17:02
by trad3mark
Hey all,

so here's some tube theory. I'm learning about triodes at the moment, in lieu to building an awesome boosterdrive. Anyway, i've read a few articles about using triodes instead of MOSFETS, but at 9V, i can't seem to get any levels of gain/clipping/overdrive/whatever close to the likes of a J201 or a 2N2222 etc etc. So first question is, can this even be done with 1 triode at 9V???

Second question is, can you use a triode as other types of transistor? whatabout a normal NPN like a 2n3904 or something?

cheers,
tm

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 17:30
by Dr Tony Balls
I asked a similar version of this question once before but really got no answer.

I became curious after looking at the various preamp modeling circuits on sites like runoffgroove where they take the circuit from, say, a tweed champ, and replace the dual triode 12A*7 tubes with fets, running at some voltage set by a bias pot, with a supply of 9V to the system. If you look at both of those circuits, they're like dead-on identical save for the valves. So, my thought was why cant i go the other way, replacing a FET or two in a circuit with a dual triode tube like that? I was gonna test the idea by making a tube SHO, but my 12V power adapter has been on backorder and ive got other things to keep me busy the meantime.

Now....as for what happens if you try the same thing with circuits meant for bipolars instead of FETs? I'm not sure. I suppose the first logical question is can you replace a FET with a bipolar tansistor in a circuit? You *can*. I know this from my first pedal build where I didnt know the difference and tried a radioshack 2n2222 in a SHO circuit before I could get my hands on a BS170. It worked, although didnt sound quite the same (ill revisit the sound of it later if i can dig up that box), but with tweaking you could maybe get it sounding good. So I guess the answer here is to just try it? Build something FET-based with a socketed FET and swap it with an NPN and see what happens! If it works, then you could probably make the same swap using tubes, too.

In regards to 9V operation, i think thats pretty much a no-go. I always thought the "12" in a 12ax7 meant that the heater voltage was 12V. And i think that if you arent running the heaters hot enough, the thing basically wont function right or wont have the proper gain to do what you want.

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 18:21
by frequencycentral
Maybe this is want you're looking for: http://circuitworkshop.com/forum/index.php?topic=207.0

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 18:22
by trad3mark
yeah i've been trying it on 9V for the last few days with no such luck. :(

So, i'm going to get something like a MAX1044 and bump up the voltage, so it can still technically run on 9v.... sortof...

I'm trying to make a tube version of a super duper 2 in one. The only thing i'm curious about right now is, if, for example, i run this at say 18V. Do i just use the same circuit but put 18V into the fillament and 9v into the same places it was (via a resistor into the plate i think it is?)

Edit:
thanks FC. Unfortunately, it's not that i want to replace the tubes with transistors, it's the other way around. I'm basically looking for an easy conversion from using a FET in a SHO to using a triode (half a 12AX7).

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 18:53
by frequencycentral
trad3mark wrote:Edit:
thanks FC. Unfortunately, it's not that i want to replace the tubes with transistors, it's the other way around.


Which is what I'm showing you, ie how to successfully run a 12AX7 at 12 volts by using a 2n5457 as a constant current source. Read the article again.

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 19:56
by trad3mark
there isn't a facepalm big enough for my blindness there lol!

So as a guide line/rule of thumb then, once the 12A#7 has 12V going to the cathode, it can more or less replace fets straight away? does it need much tweaking to the resistors going to the plate or anything??

Re: Tube Theory: Triodes to... ?

PostPosted: 12 Dec 2009, 20:29
by JiM
I get a decent overdrive out of a ValveCaster, just like this one : http://www.beavisaudio.com/projects/ValveCaster/ but with a NoS russian tube. Usually, the signal straight from a guitar (without booster) cannot overdrive a single tube stage unless it's badly biased. But that first stage can overdrive a second stage. The ValveCaster uses both triodes of a preamp tube for that reason, and a 12V power supply.
That 12V is convenient for the 12A?7 tubes that can have series heaters. Mine doesn't, the heaters are always in parallel, so i've added a 7806 regulator and its heatsink to provide 6V, 300mA to them.

One fundamental difference between transistors and tube is the voltage they need to work well. One can also say that tubes are high-impedance devices, whereas transistors are low-(output)-impedance devices. Let's have a look at the basic properties of each device :
  • Bipolar Junction Transistor : the current at the base commands the current going from collector to emitter.
  • Junction Field Effect Transistor : the voltage at the gate commands the current going from drain to source. That voltage sort of "pinches" the channel where the current flows.
  • Metal-Oxyde-Semiconductor Field Effect Transistor : the voltage at the gate commands the current going from drain to source. The gate has a big input impedance (causing sensibility to electrostatic discharge), but some capacitance also. The same "pinching" effect occurs, but within a capacitor.
  • Triode: the voltage at the grid commands the current going from plate to cathode. The grid has a huge input impedance, because it's basically alone in vacuum. The electrons that jump from the heated cathode are attracted by the high-positive-voltage plate. The grid can reject the electrons (when negative) or attract them (when positive, but that's usually not recommanded).
So what happens when a tube is used with a low anode voltage, aka "starved plate" ? The electron cloud around the cathode is not much attracted by the anode, because it's far away (a couple of millimeters) and the electric field decreases with the square of the distance. The tube will have a very low gain, if any. It may need also to be driven into positive grid, to accelerate the electrons towards the plate : this requires providing substancial current into the grid, which passive guitar pickups cannot do.

The trend in "cloning" tube circuits with JFETs is cool, but it's a bit of over-simplification. It doesn't take the Miller capacitance of the tube into accound, the drain trimpot is not the best way to bias, etc. Just look at these ROG article for more details (and acknoledgement that pin-for-pin replacement is not faithful enough) : http://www.runoffgroove.com/fetzervalve.html
http://www.runoffgroove.com/sd.html
But the most important thing is often overlooked : a ~300V plate voltage is translated into a ~9V drain voltage. When doing the opposite, shouldn't you also scale to a high voltage ?
A starved plate design, how convenient and safe it is, does not react the same way.
On the other hand, there were "space charge tubes" designed especially for that mode of operation. I suppose that's the case of most of the submini, "hearing aid" tubes used in Frequency Central designs.

trad3mark wrote:Do i just use the same circuit but put 18V into the fillament and 9v into the same places it was (via a resistor into the plate i think it is?)

That's the best way to fry your tube ! It's the other way around : you'll want the heaters to get 6.3 V (in parallel) or 12.6V (in series), and the plate to get the highest voltage you can get (up to 350V).

trad3mark wrote:I'm basically looking for an easy conversion from using a FET in a SHO to using a triode (half a 12AX7).

Basically, you won't get any boost from a single starved plate triode alone. You'll need either a higher voltage, a second triode gain stage, or some transistors/opamps to do the actual amplification.

I recommand reading this : http://www.freewebs.com/valvewizard1/gainstage.html

Re: Tube Theory: Triodes to... ?

PostPosted: 13 Dec 2009, 01:31
by trad3mark
ok. that actually is starting to make things clear.

It's not crazy complicated, it's just maths and physics. And i love those! :D

ok, awesome. So, new plan. I'm going to build some sort of cool tube overdrive from scratch. As with many of my college architecture projects, the theory is great! Taking inspiration from a Super Duper 2-in-1, it is possible to have 2 cascading overdrive/boosters from 1 x 12AX7, to produce awesome levels of overdrive. The thing to do now is:
1. Understand the theory; do the research, and get a good knowledge of what i'm aiming for, and how best to achieve it.
2. Start off with a basic, simple model
3. Tweak, add and subtract to it as needed.

I'm definitely going to start this over the christmas. I'll set up a workbench thread for people to follow / help my progress.

Once again, thanks for all the help guys. :)

Re: Tube Theory: Triodes to... ?

PostPosted: 07 Oct 2010, 08:53
by DrNomis
You can think of a FET as the Solid-State equivalent of a Triode,in fact the FET was invented because there was a need for a Solid-State Semiconductor device that had electrical characteristics that resembled valves,but if you look at the Transfer Curves of each device,the FET's Curves are more like a Pentode Valve than a Triode Valve.... :)

Valves and FETs are Voltage Controlled Devices,and they tend to have a very high input impedance,something like tens of mega-ohms,whereas Bipolar PNP and NPN Transistors,are Current Controlled Devices,when a Bipolar Transistor is operating as a Common-Emitter Amplifier,the input impedance is the two base bias resistors in parallel,depending on what resistor values are used,the input impedance can be something like tens of kilo-ohms.... :)

Re: Tube Theory: Triodes to... ?

PostPosted: 07 Oct 2010, 21:44
by HydrozeenElectronics
I've always wondered too if you could build a Tube version of a Fuzz Face? that would be sweet!

Re: Tube Theory: Triodes to... ?

PostPosted: 15 Oct 2010, 18:14
by DrNomis
I actually tried making a Tube version of the Fuzz Face a few years ago,I used a 12AX7 Valve,and just wired it up same like a Fuzz Face,and all I got out of it was just a clean signal,in actual fact,the way the active devices(bipolar transistors) in the circuit are biased so as to turn them more or less "on",is the opposite to how Tubes are biased,which is to turn them more or less "off",so,if we put a Tube in the Fuzz Face circuit,there is no way for the circuit to bias the Tube,and so,the tube just saturates,that is,conducts the maximum current from the cathode to the anodes,that the anode resistors will allow..... :)

Re: Tube Theory: Triodes to... ?

PostPosted: 15 Oct 2010, 18:49
by DrNomis
Just for interest's sake,here is a pic of one of my only two 6SN7GT Twin Triode Valves,the heater is powered by 6.3Vac..... :)

Re: Tube Theory: Triodes to... ?

PostPosted: 19 Oct 2010, 00:40
by HydrozeenElectronics
That's good to know. thanks :)


DrNomis wrote:I actually tried making a Tube version of the Fuzz Face a few years ago,I used a 12AX7 Valve,and just wired it up same like a Fuzz Face,and all I got out of it was just a clean signal,in actual fact,the way the active devices(bipolar transistors) in the circuit are biased so as to turn them more or less "on",is the opposite to how Tubes are biased,which is to turn them more or less "off",so,if we put a Tube in the Fuzz Face circuit,there is no way for the circuit to bias the Tube,and so,the tube just saturates,that is,conducts the maximum current from the cathode to the anodes,that the anode resistors will allow..... :)

Re: Tube Theory: Triodes to... ?

PostPosted: 25 Oct 2012, 12:03
by DrNomis
frequencycentral wrote:Maybe this is want you're looking for: http://circuitworkshop.com/forum/index.php?topic=207.0



Hi Rick,

I've actually tried connecting a single 12AX7 Valve up like one Mu-Amp stage in a BSIAB circuit, it does work but you need a fairly high plate/anode voltage, when I built the circuit on my breadboard, my plate/anode voltage was about +300V, but I think you could get it to work at something like +45V if you tweaked the biasing, I remember that I got quite a bit of voltage-gain when I built the circuit.... :thumbsup


To generate the plate/anode voltage from say +9V you can make a voltage-multiplier from a CD4049, some 1N4007 diodes, 2X 33k resistors, a 330pF ceramic cap, and some 1uF/100V electrolytic Caps, I'll post a hand-drawn schematic here shortly.... :thumbsup

Just by adding extra 1N4007 Diodes and 1uF/100V caps, I was able to get up to about +90V from this circuit.... :thumbsup