A very interesting article was read yesterday. Triodes at Low Voltages - Merlin Blencowe.
http://www.valvewizard.co.uk/Triodes_at_low_voltages_Blencowe.pdf
Prayer was immediately answered on the thought of how would electron tube curves look at very low anode voltage. It was imagined that the shape is retained in a fractal way. The article demonstrates that the general shape is still there.
And it is a very well written and informative and imaginative article with great references too.
Below are some comments that may be relevant to the use of electron tubes at low voltage for very low noise amplifiers used for amplifying the tiniest signals as is done on the Pleiades V6, V5 and V4 pre preamplifiers:
Rather than reducing the grid current externally it may be reduced internally by a lower heater voltage. This leads to less cathode temperature and less space charge near the grid. A beautiful example of this concept is on one of the greatest microphones of all time, the tube Neumann U47.
Higher megohm resistors of the order of 4.7ΜΩ usually from anode to grid, or Vb to grid etc seem to work great in practice at low plate (anode) voltage as tested on the Pleiades battery electron tube microphone preamplifiers such as V4, V5 and V6. For an ECC82 amplifier with a military output transformer and a battery supply of less than 6V a 10MΩ is used at the moment with very good sound quality.
The effective input impedance of the stage is much smaller than this external resistor. For example on an EF183 triode connected and supplied by 3V-5V including heaters when Rag is 4.7MΩ, the input impedance is aproximately 100KΩ.
It is not necessarily a disadvantage to have a less than MΩ input impedance. This may be an advantage to damp the reasonance of an input transformer.
At the end of this post there is a reference of a 1930s patent titled Positive Grid Bias. It was found yesterday again by Googling. There the inventor shows how by heavy positive grid bias a region is reached where the slope of Ig Vg curve becomes flat meaning infinite input impedance and then downward meaning negative input impedance. Thus means that what is connected is magnified before it reaches the amplifier.
The Pleiades amplifiers currently use only subtle grid biasing with excellent sound and noise performance even at the smallest input signals.
On the Pleiades amplifiers grid bias is done by a resistor from anode to grid rather than from Vb to grid. This was found experimentally to give a sweeter sound. No explanation as yet, speed of light slap back positive feedback?
Out of many tubes that had been tried like ECC86. E88CC, ECC83, ECC82, ECC81, EF86, 6SK7, the ECC82 outperformed but it was with an EF183 that the sound took off in an unexpected miraculous way.
The EF183 is a wide? variably spaced grid vari mu pentode. It sings. It has a very high transconductance (gm). Naturally the screen grid is very close to the cathode and collects most of the electrons. So the tube is used as a triode with anode, suppressor and screen grid all connected together.
The EF183 gives a very high gain and signal to noise ratio compared to the ECC82 which is very good too.
It is very convenient to optimize a circuit in practice by treating it as an analog computer.
The only? 3 parameters that can be adjusted are the anode to grid resistor, the heater voltage and Vb.
It is a joy disconnecting a heater that was connected to 6V and hear the sound and detail increasing as temperature slowly drops!
By using a rheostat in series with heaters one can slowly and patiently find the sweet spot.
A Pleiades low voltage tube preamplifier is usually driving the input of an ordinary preamplifier. So everything concerning the performance is magnified many times.
It was in the way that it was found that a Pleiades pre preamplifier exhibits an exceptional low noise figure. This makes it suitable as a front end in amplifying the tiniest signal. This signal may be a microphone an MC cartridge or perhaps the front end of a radio telescope receiver.
The first stage is the most important noise wise as any noise developed there will be magnified by the rest of the stages. So the noise performance of the whole system is dictated by the first stage. So by adding the Pleiades pre booster detail increases while noise and hiss drops.
The small snode voltage is not a problem as the signal input is very small anyway.
A battery can supply both heaters and anode ensuring a hum free, noisy less and creamy smooth sound.
The battery can be inside the same enclosure so there is no RF pickup.
Tubes can be operated at spaces where radiation exists and can be power supplied by photovoltaics making their use invaluable and important.
Although it was found by chance while disconnecting the heaters and listening, the lower heater voltage increases signal to noise ratio. The use of lower heater voltage can be traced to the Neumann U47 (1947) tube microphone where the VF14 tube is underheated.
It may easily be verified by connecting a micro ammeter between cathode and grid that the grid current is reduced when the cathode temperature drops as expected.
We have to be careful by what we mean Vg=0V. We may take any tube and just supply the heater voltage. If connecting a 10MΩ input impedance voltmeter across cathode and grid we will read a fairly high negative voltage on grid! This just by heating only the cathode! So we have an EMF of a high internal resistance.
When we connect a micrometer between cathode and grid we force this potential to become zero as the meter has a very low shunt resistance. So all the voltage is dropped in the internal resistance of the equivalent circuit (grid resistance?). And according to Ohms law the only way this can happen is by having a grid current.
But the EMF is still negative as we can imagine. It is like short circuiting a (source of voltage in series with a high series resistance).
But where in reality does the grid current come from?
It may in fact be the current induced by the boiling steam off electrons. If we can imagine water boiling, the steam is a current in itself.
Reducing the cathode temperature reduces the grid current but also reduces what we observe by the voltmeter connected across the cathode and grid.
When the tube is cold this is a close to zero. The funny thing is that on a cold tube one can still read a potential different although very small between grid and cathode. It is very few millsvolts.
A great explanation on how a tube operates comes from the master Edwin Armstrong on his paper Operating Features of the Audion.
There, he demonstrates by experiment that even the physical existence of the grid (without it being connected to anything) reduces current flow, wing as he calls it or plate current.
On the very informative article by Merlin Blencowe one can see on the curves that even at 0V anode voltage there is a small anode current. This again may come from the boiling electron current at the cathode but it is smaller than the grid current as the anode is further. (It is fun measuring these currents on a tube just supplied on its heaters and also observing the variations as the cathode is let to cool).
On the excellent article by Merlin Blencowe there are some very interesting references, one example:
U.S. Patent 5022305. It is guitar distortion pedal employing a resistor from Vb to grid. By searching google one can find that it may be used by David Gilmour
Further work may be done to understand better and improve.
The Pleiades amplifiers would not have been possible where it not for Hliana insisting that the grid should not be negatively biased.
References:
Operating Features of the Audion - Edwin Armstrong
Positive Grid Bias - W Van B Roberts patent US1944574
Older entries of euroelectron on Pleiades V6, V5, V4 etc.
A very interesting site on space charge tubes and low anode voltage is Jeff Duntemann's www.junkbox.com. Great help and inspiration was received from this site for developing the Pleiades Electra 9V 12K5 tube and no other components headphone amplifier.
