chandler tube driver

[jeff]

My brother got me interested in the Chandler tube driver and I may build one. There's one thing I don't get.
Why do the 12AT7s have no cahtode resistor?
How is the tube biased?
Isn't the grid at the same potential as the cathode?

[SILVERGUN]

download Merlin's preamp first chapter free and look at page 12:
http://www.valvewizard.co.uk/gainstage.html
 

[PRR]

When you have 300 pounds water-pressure, you NEED some restriction in the pipes or your hose will burst.

When you have just 12.5 pounds pressure, you want all pipes and valves WIDE open to get even a slim stream.

Likewise at 12.5 Volts you may have to run a tube WIDE open to get enough current to do any good.

Yes, it distorts easy. But that's the point here.

[jeff]

Yes we want distortion but if the tube is full on, how can we get both positive and negative part of the wave?

To use your analogy if our "singal" is turning the water more off, more on, more off, more on,... etc. and the hose starts full open we can turn it down but we can't turn it up past full.

If the tube is pasing the most current it can when the grid and cathode are both at 0V(idle state) applying an input signal would reduce the current on it's negitave cycle, turning the tube more off, but what happens on the positive part of it's cycle? Does making the grid more positive than the cathode cause more current to flow or are we already maxed out and only getting half a wave? How can we turn the tube more on if it is starting state is full throttle?

Isn't that what "bias" is, the fact the grid voltage can go up to the point of turning the tube full on and down to the point of shutting it off. The closer you set it up to one extreme or the other will cause it to distort quicker but if you start at one extreme won't you only get half a wave?(isn't that how an octave pedal works?)

[PRR]

> how can we get both positive and negative part

Drive the grid positive.

In most all-tube circuits, we don't. It is HARD work. The tube input drops from 1Meg toward 1K. The tube before it won't drive a 1K load well. (When we "must" do it, we use a beefy triode and a step-down transformer.)

But. Here we start with an opamp which can easily drive 1K.

Also the grid input does not drop to 1K as soon as we cross 0.000 Volts. Usually have to go several Tenths Volt positive. We can go 0.1V and still have several K grid impedance. And if the tube gives gain near 10, 0.1V in gives 1V out which is a hot guitar signal.

I don't *know* that it works. Tube-data in this area is scarce. But it seems plausable. Even before considering that this plan has been around and I *think* I have seen reports of it working. (Also not-working when ANYthing is changed recklessly.)

[PRR]

> how can we get both positive and negative part

Drive the grid positive.

In most all-tube circuits, we don't. It is HARD work. The tube input drops from 1Meg toward 1K. The tube before it won't drive a 1K load well. (When we "must" do it, we use a beefy triode and a step-down transformer.)

But. Here we start with an opamp which can easily drive 1K.

Also the grid input does not drop to 1K as soon as we cross 0.000 Volts. Usually have to go several Tenths Volt positive. We can go 0.1V and still have several K grid impedance. And if the tube gives gain near 10, 0.1V in gives 1V out which is a hot guitar signal.

I don't *know* that it works. Tube-data in this area is scarce. But it seems plausible. Even before considering that this plan has been around and I *think* I have seen reports of it working. (Also not-working when ANYthing is changed recklessly.)

That's for THIS plan. If the grid resistor were 1Meg or more, grid current would shift it, generally negative (for small tubes working easy). Also the cap-coupling will tend to peak-catch large signals and drive the grid negative according to signal level. Both types bias are found in various input stages. Here I think the low 10K resistors are to *avoid* such effects.

===================================

Fuggit. I threw it to the Idiot Assistant (SPICE). Aside from being an idiot, the available tube-models are less trustworthy than two 'same' tubes from different eras/makers. However the numbers gotten agree generally with my expectations, and not very different from the DC number on the published plan (idiot says 9V on plate, plan says 14.3V; I know this model neglects Contact Effect which is small at 300V and large at 25.5V).

I rationalized the +12.5v and -13V opamp supplies to just a 25.5V for the tube alone. This also puts the grid at smack-Zero (actually part-microVolt due to grid current) for easy thinking.

First I swept the grid negative and positive (you can do this with a battery and pot) and watched the plate voltage. This is the basic transfer curve. The on-plan nominal bias is the green dot. You can see there is no sudden change of slope either side. And it does not go all to hell when it crosses 0.0V. Yes, it IS curved, and quickly gets more-curved. But for say the range of 0.4Vpp in 8Vpp out the THD will be far below 10% and pretty 2nd harmonic(*).

We do like to see our waves, not just transfer curves. I hit it with a big sine, 2V peak. Output is pretty bent, but at 12.5V peak output it is "only" 27% THD. The "clean" is still much bigger than the "crap". It's an OK fuzzbox, but many musical styles will want "MORE!!"

(*)This is all for one stage without coupling cap or load. When we line-up two stages and trimmings, bentness will increase. Interestingly it may sorta-cancel for small signals (be sorta-clean), but on large signals I expect it to be "MORE!!" bent. And more 3rd harmonic. Which is more what we want.