Tuesday, July 17, 2018

Reducing grid current by reducing cathode temperature


All potentials of electrodes of electron tubes are referenced by definition with respect to the cathode [Gray].


The open grid or space potential of the control grid can be of the order of -700mV (even before applying anode voltage). Its value can be increased by increasing the heater voltage or decreased by decreasing the heater voltage. If careful connection with heaters only (using a fuse in series for safety) of a heater voltage the grid to cathode potential can immediately be measured with any voltmeter, even one of not a very high input impedance, typically 100KΩ.


Where does this voltage come from? It should be the due to cathode emission. Escaped electrons by "boiling" leave the cathode positive as the total number of positive protons are still there. So the cathode is positive with respect to grid or the grid negative with respect to cathode. This is negative grid bias even before connecting anything else to an electron tube.


If we connect the usual grid to cathode resistor in fact we are forcing this grid potential to change (like connecting a load across the terminals of a real source of voltage with non 0 output impedance). Current flows and we might call this grid current although it is really part of cathode emission current. The mechanism should be as follows. When we connect an external resitor from grid to cathode, electrons find an external path to return to the cathode finding back the lonely positive protons. It should be as simple as this. The positive protons and negative electrons find a path to meet by electrostatic force of opposite charges.


Curves of this current vs voltage as the resistance of the path is varied can be seen in [Blencowe] and [Whitlock]. On Whitlock the curves are called variation of initial velocity grid current with grid potential and heater voltage as a parameter. So it is a family of 3 curves with heater voltage of 5.7V, 6.3V and 7V. One can see clearly the decrease of negative grid bias as the cathode temperature is reduced. And of course the coresponding decrease in grid current or rather cathode emission current. The slopes remain the same indicating the same input impedance or rather the same source impedance as the electron tube clearly becomes a source of voltage the origin of which is the thermal energy externally supplied to the cathode.


As early as 1914 Armstrong found a substantial decrease of anode current even by the presence of an open grid [Armstrong].


While experimenting a few years ago with low anode voltage, even before reading the excellent information on the above publications, it was found by accident that an electron tube can normally operate with almost any low anode potential as long as a high Megohm resistor is connected to a positive potential or even the anode itself. The connection to anode was prefered on grounds of more natural sound quality of a microphone monitored in real time. See previous euroelectron posts.


The other chance discovery was that when the heater circuit was opened, the anode current immediately increased a little and the signal to noise ratio gradually improved at a lower anode current until it got worse by a too cold cathode.


The first microphone head amplifier to operate with just an AAA 1.2V battery supplying both filament and heater is the Pleiades V1. Was it possibly there that the grid was first connected to a positive source? By accident or curiosity or chance the 60MΩ grid resistor was returned to the positive filament terminal instead of the negative and a substantial increase in anode current had resulted.


The following clip is possibly the first recording ever produced with an electron tube microphone headamp powered by an AAA battery for both heaters and anode. A small aluminum box housing the Freed input transformer, the CV2269 electrometer electron tube with Pleiades bias by 60MΩ from Vb to grid, (RC coupled anode follower configuration) and the AAA 1.2V rechargeable battery:


Tezcatlipoca - Ira, George - YouTube renewablemusic channel:
https://m.youtube.com/watch?v=2RD02ATDPAk


The positive effect of reducing cathode temperature was observed by chance on the Pleiades 6, a few years later while listening in real time to the low output Grampian ribbon GR2/L microphone.


Pleiades V6 schematic


Pleiades V6 circuit, Pleiades bias resistor is usually 6MΩ, Cc=22nF, the Nuvistor 7586 electron tube can also be connected and Ia is approximately half.


4V for both anode and heaters was found to give an exception low noise circuit while the gain is somewhat reduced.


Anode current is typically 50A for the EF183, 25μA for the 7586 electron tubes. Output impedance is typically 30KΩ for the EF183 so it becomes 300Ω balanced by a 10:1 output transformer.


Input impedance is typically 100KΩ. So it becomes 1000Ω balanced by a 1:10 input transformer. So almost any 200Ω or even 25Ω microphone can be connected (see Grampian DP4/L).


A typical value of grid bias is -70mV. Without the Pleiades bias resitor the grid bias is typically -700mV so there is cutoff of anode current.


Measurements as to whether the input impedance changes by further underheating have not been made yet.


It cannot be yet concluded if the ear-brain monitored increase in signal to noise ratio comes from a reduction of grid current.


Underheating of the cathode was well known at least since 1930s and was used succesfuly by Georg Neumann in 1947 on the U47 Neumann condenser microphone head amplifier.


A reduced cathode temperature reduces the internal anode resistance thermal noise [Llewellyn], [Pearson]. Reduced electron tube electrode temperatures and potentials result in reduction of secondary emission, photo emission by the glowing cathode, and other sources of noise or grid current [Pearson]


Even if electron tubes turn out to be the lowest noise devices on the planet with absence of shot noise due to space charge (Llewellyn) why using electron tubes on 2018? Electron tubes are possibly the most linear electronic devices without feedback need, distortion tending to 0 as the signal tends to zero. A music signal is most of the time around the 0 crossing point area. Distortion is increasing in the right way as the suply rails are approached by higher signals or peaks. The electron tube artfully succeeds in fitting something tremendously big such as the dynamic range of music to something small which can safely drive op amps or analog to digital converter circuits before they run out of 1111111...1's. Electron tubes do objective signal instantenious limiting while preserving subjective loudness [Hamm]. So that on 2018 we can enjoy When I Fall in Love - Nat King Cole on YouTube...


Nat King Cole When I Fall in Love (HQ)
https://m.youtube.com/watch?v=91bQyER32GY


...as fresh as if it has been recorded right now or in the future.


Another reason why this track sounds state of the art and so natural is possibly the fact that electronic engineers clearly hundred of hers as ahead of their time ere thinking in terms of flat frequency response not just from the mic to lids pear but from producer's brain to listener's brain [Lowe, Morgan]. Inductors in parallel with the mic with inductance index itself varying with frequency might have been the trick, see also Pleiades (R,L) filters.


Further investigation and playful experimentation is needed.


References:


Applied Electronics - T. S. Gray - MIT


Triodes at Low Voltages - Merlin Blencowe


Techniques for application of electron tubes in military equipment - Rex Whitlock


Operating Fratures of the Audion - E. H. Armstrong


A study of Noise in Vacuum tubes and Attached Vircuits - F. B. Llewellyn


Fluctuation Noise in Vacuum Tubes - G. L. Pearson


BBC research department report on the Neumann U47 microphone


Open-grid tubes in low-level amplifiers - Robert J. Meyer - electronics - October 1944


The use of multi grid tubes as electrometers - J. R. Prescott - Jan 25 1949


Tubes vs transistors (vs op amps) is there an audible difference? - Russel O. Hamm - JAES


On preserving transconductance of electron tubes at anode potential as low as 3V - euroelectron


Flat frequency response from producer's brain to listener's brain, Sound picture recording and reproducing characteristics - D. P. Lowe, K. F. Morgan - JSMPE



No comments:

Post a Comment