Explanation needed

[Leevi]

I have wondered why the wrong wiring of tube input (picture below) works now and then
and some times the tube is totally blocked. What is the reason for that?
/Leevi

[JayB]

I have wondered why the wrong wiring of tube input (picture below) works now and then
and some times the tube is totally blocked. What is the reason for that?
/Leevi

The un-scientific explanation.  The grid needs a ac/dc ground reference to properly balance the triode. The grid will build up a dc charge that will eventually throw the bias off until it no longer works. The grid needs to be at a lower voltage potential in respect to cathode. At first you may have less than a volt on the grid while the cathode maybe at 1.5v until the grid builds up a charge that's higher than the cathode. Then it shuts off. If you need that cap, put your resistor after the cap.

Maybe PRR will give you the gritty details.

[Dave]

That sounded pretty gritty to me. 

Dave

[tubeswell]

And the grid builds up a charge in that case from two sources, namely:

1) the longer time-constant of lower-frequency notes causing the coupling cap to 'overcharge'* and build up a surplus of charge which then feeds the following grid, causing the bias of that stage to be affected, causing grid current limiting and ultimately blocking distortion. This is why lowering the capacitance value of a coupling cap helps reduce the tendency to blocking distortion; and

* i.e.: it can't get rid of its charge fast enough

2) the grid is a lot closer to the cathode than the plate is, and so the grid has a natural tendency to heat up and emit a few electrons of its own (which then 'escape' to the plate), thereby causing the grid to become +vely charged, and without a grid leak resistor there to replenish the electrons that are being coaxed off the grid the +ve charge then affects the bias of the grid in the cathode biased stage, causing grid current limiting and ultimately blocking distortion.

That's my take on it anyway. Someone with more nouse than me might have a better explanation.

[sluckey]

Your circuit does not work, period. You must have a dc path between the grid and cathode for a tube to bias up and work properly.

[RicharD]

>works now and then and some times the tube is totally blocked

When it's working, it shouldn't be, and is operating without input control.  You have not finished setting the DC bias, ie. Since the grid leak leak is blocked by a cap, there is no set DC condition on the grid.  In a grounded cathode amplifier, the grid must be more negative than the cathode for it to work properly.  Counting on thumbs, we know that a 12AX7 is happy around 1mA of current.  Current times the cathode resistor = the cathode voltage, usually about a volt.  So that means we need the grid to be somewhere between 0 and 1 volt.  Tieing a 1M between the grid and ground achieves this without creating so much load that the previous stage or geetar isn't loaded down.  Now the tube wants to bias up and something less than 1V is very little in the grand scheme of thangs.  With enough AC signal, you need only touch the grid momentarily with your finger ( not recommended ) and the tube will start passing signal, but the bias is fluctuating and it probably sounds funky pumpy.  If you stop passing signal, the tube loses bias and quits working.  I suppose with a leaky cap or some sorta gremlin it might work intermittently.  Perhaps a big AC spike my look enough like DC for the tube to try to bias up.         

[Leevi]

Actually the circuit was exactly the following and the guy had used the amp
about one month without problems.
/Leevi

[sluckey]

Actually the circuit was exactly the following and the guy had used the amp
about one month without problems.
/Leevi

Surely there's a mistake in the drawing or the actual circuit. It cannot work 'without problems' as drawn.  There's no way to establish bias conditions for that circuit. The grid cannot control the current flowing from cathode to plate without a DC path from grid to cathode. That's about as basic as Ohm's law. I'd love to be shown how it can work.

(pic cropped to a more usable size.)

[PRR]

Some tubes will work, UN-reliably, with no grid current path.

Some tubes, with reasonable plate conditions, have an S-shape grid current/voltage curve. There is a _ZERO_ current condition at a few tents volt negative of cathode. There's even some benefit to working here.

However. Any drift or dirt in tube or on socket tends to throw the bias WAY off, the tube blocks-up. Also guitar signals are large enough to throw the tube off the zero-current point on the first large pluck, or some string of transients.

[sluckey]

PRR, that's an interesting paper. Just goes to show there's always an exception, I guess. It's the only written material I've ever seen making that statement. I'd love to see a real world manufactured circuit that operates with an open grid. Got any useful examples?

[PRR]

Perpetually open, no.

Electrometers will typically have a shunt when idle, which is opened just long enough to take a measurement.

Actually.... doesn't RDH 3rd have an acorn-tube VTVM where the input probe is the naked grid-pin? (Comes out the end of an acorn.) Obviously this is not expected to ready anything unless connected to "something", and that something would bleed grid current.