Plate characteristic of triode.

[Alan0354]

I am looking a the graph of Ip vs Vp at various given Vg of the triode, the graph almost looks like a non linear resistor when the operating point is away from the cutoff. It is obvious that as you increase the Vp, the current increase kind of linearly. The Vg almost like controlling the resistance of the triode. The triode is almost like a voltage control resistor!!!!

If so, the output voltage is just like a resistor divider with the bottom resistor as voltage control resistor!!! Am I missing something? This is not a trick question. For the first time, I stop and got a tube book and study the theory and just trying to understand the tubes.

[Enzo]

DO you have the RCA tube manual? Or at least find it online. The 6C4 is a small power triode in a 7-pin mini. It has curves in its section. The 12AU7 signal triode has curves, and the 12AX7 signal triode curves are found on the 6AV6 page. Plate voltage and current at various grid voltages. I think it is what you are looking for.

Oh wait, you are looking AT it, not FOR it. Oh well.

The grid voltage controls the current through the tube. Puting a load resistor in series with the tube then will have a controlled voltage across it due to the varying current. I myself don't think in terms of output voltage unless I have a circuit around it. You can take an output from various points. And by bottom resistor, do you mean the cathode resistor? IN a typical gain stage like a 12AX7 uses, the voltage across the cathode resistor also changes - the current through the tube is also the same current through the resistors in the plate and cathode legs. BUt if we have a 100k plate resistor and a 1k cathode resistor, the voltage change across the 1k will only be 1/100 as much as across the plate resistor's 100k.

[Alan0354]

I have the data sheets in front of me, but I think I was wrong. I was looking at the Vp vs Ip of each of the curve and try to think of it as a resistor. It is wrong because each of the curve of different grid voltage has a horizontal displacement and it cannot be a voltage control resistor. Sorry.

I am for once trying to study the book on tubes instead of monkeying around.

[Jazz P Bass]

I love this book.
http://www.antiqueradios.com/archive...acuum_tube.pdf

[Alan0354]

I love this book.
http://www.antiqueradios.com/archive...acuum_tube.pdf

Thanks for the book, I downloaded it. Now I have two. I just received the Guitar Power Amps by Richard Kuehnel three days ago and been studying it. I have been faking it by pretending the tubes are just transistors that I am familiar with. It is really hit and miss because they are different enough even though the very basic circuit are kind of similar. But they do have their share of difference. Particular the pentodes which have the screen that has special behavior.

[Enzo]

Well here, download this thing and read it, at least the parts that pertain to audio and power supply.

The Radiotron Designers Handbook.

RDH4 mirror

[Steve Conner]

You're dead right. "Plate resistance" (rp) is one of the fundamental parameters of the triode.

What's slightly counter-intuitive is that the grid voltage doesn't actually control the plate resistance, except in the non-linear region near cutoff. If you study a family of triode curves parameterised by grid voltage, you'll see that they all have a similar slope: changing the grid voltage just shunts the curve left or right. So, to a first approximation, plate resistance is constant. For a 12AX7 it's about 56k, IIRC.

The physical reason for this is that the electric field from the plate affects the electron flow just the same as the electric field of the control grid, only less so. The amplification factor (mu) quantifies this: if a triode has a mu of 50, then 1 volt change in grid voltage has the same effect on plate current as 50V change in plate voltage.

Tetrodes and pentodes have mu and rp of almost infinity, because the screen grid prevents the plate from affecting the electron flow. They behave very similarly to JFETs. There is no semiconductor that behaves quite like a triode tube, though.

There is one more parameter, the transconductance (gm), which means exactly the same as in solid-state work. But there are only two degrees of freedom in triode design. If you know any two of mu, gm and rp, you can calculate the third: for example mu = gm*rp.

[loudthud]

If you study a family of triode curves parameterised by grid voltage, you'll see that they all have a similar slope: changing the grid voltage just shunts the curve left or right. So, to a first approximation, plate resistance is constant. For a 12AX7 it's about 56k, IIRC.

The plate current curves diverge slightly as plate current increases. That's where the tube gets it's compression when B+ sags. The load line always has the same slope. As it moves to the left when B+ sags, the grid voltage lines are closer together and therfore the gain is less. See the first two attachments.

Tetrodes and pentodes have mu and rp of almost infinity, because the screen grid prevents the plate from affecting the electron flow. They behave very similarly to JFETs. There is no semiconductor that behaves quite like a triode tube, though.

I'm finding that JFETs don't always behave like pentodes. The J105 is typically used as a switch. It's on resistance is less than 3 ohms, Idss is 500mA. But there is a large region where it behaves more like a triode. See the third attachment. Today I purchased a bunch of JFETs that I'll be looking at. That J105 is too linear!

[Alan0354]

Well here, download this thing and read it, at least the parts that pertain to audio and power supply.

The Radiotron Designers Handbook.

RDH4 mirror

Thanks for the link. Chapter 2 is exactly what I needed. That's what I've been looking for on the change of parameters not published in the normal data sheet. I have been searching on the web and bought the book by Richard Kuehnel that is supposed to be more detail, but still it is quite cook book approach. Now my printer is working overtime printing all the 52 pages!!!

[Alan0354]

I'm finding that JFETs don't always behave like pentodes. The J105 is typically used as a switch. It's on resistance is less than 3 ohms, Idss is 500mA. But there is a large region where it behaves more like a triode. See the third attachment. Today I purchased a bunch of JFETs that I'll be looking at. That J105 is too linear!

This is particular true when the plate voltage gets lower and lower, the screen dissipation becomes a problem. Also, FETs' on channel resistance is quite low and it can pull down hard where tubes cannot. And then when the grid become possitive, it starts to draw current where FETs don't. FETs are much simpler to deal with. When you drive a N JFET or NMOSFET positive, it literally become a resistor with resistance of the on channel which in power MOSFET can be as low as a 1/100 ohm easily for a power FET. Even with 12AX7, you can't exactly pull down all the way like FETs.

Only in the FETs saturation region ( not the same as saturation in tubes) when the Drain/Plate acts like a constant current source ( or voltage controlled current source ) where the characteristics become similar.

That's the reason I decided to bite the bullet and spend a little time studying the tubes. If you work in the most linear region, you can get away treating tube like a transistor. But when you deal with Guitar amp where you drive they all the way, then they are quite different, Much more so than the difference between BJT and FET.

[Alan0354]

Thanks for all your help and support. Got a lot of reading to do rather than talking!!!!

Alan

[Steve Conner]

Loudthud, a change in gain from 57.5 to 55 is only about 0.5dB, I don't believe it is an audible effect.

If you really want to understand the fundamentals of tube circuits, old textbooks like the RDH4 are hard to beat. Could well be the same books that were on Leo Fender's bookshelf as he made his first amps.

I haven't seen much information on tube non-linear behaviour in a musical context. The classic work in this field is (IMO) Russell O. Hamm's 1972 AES paper "Tubes vs transistors: is there an audible difference?" which can be found online.

[Alan0354]

I just studied about 6 of 7 pages of cpt 2 of the RDH4, it already answer a lot of questions I have and give me so much more insight than the other books I read. I guess you have to go back to the old old books to get it. I give me new appreciation of those graphs to figure out gm, rp and \mu.

One question looking at the plate voltage vs plate current graph of 12AX7. Take the common example of Fender front end. Rp=100K, Rk=1.5K. The Vk usually about 1.6V which give say 1.2mA of plate current. B+=400V and therefore Vp=280V.

You look at the graph, draw horizontal line at 1.2mA all the way to Vp=280V, you'll see the Vg is between -2 to -2.5V instead of 1.6V as in the real circuit. Why?

[frus]

I've never seen a fender with that high a B+ in the first gain stage. It's more like B+ is 280V, so the plate voltage at idle is much lower (180V or so?)

[Alan0354]

I've never seen a fender with that high a B+ in the first gain stage. It's more like B+ is 280V, so the plate voltage at idle is much lower (180V or so?)

My memory is a little off. the Vp is around 250V, the spec called for +1.9V at the cathode, I guess it is closer.

[loudthud]

Loudthud, a change in gain from 57.5 to 55 is only about 0.5dB, I don't believe it is an audible effect.

I calculated 0.39dB, but remember that's only for a single stage. The B+ and amount of sag is just something I pulled out of the air. The experiment I want to perform is to measure one input channel of a 5F6A with the volume control all the way down and play through the other channel. The re-issue amps have more filtering than the vintage amps, that will probably change the results.

Other factors that contribute to the compression effect are the B+ sag effect on clipping level and bias shift throughout the amp.