The other very common method of bias is via a grid leak resistor. This resistor (on the grid of, say, a 6Q7, 12AV6, etc.) is typically in the neighborhood of 5 to 10 Meg, and connects from the grid to B-, or the common connection point of the cathode of the same tube.
A capacitor isolates DC on the grid from any DC that may be riding on the incoming signal
Electrons from the cathode hit the grid and build up on the grid, establishing a negative charge. These electrons further charge the isolation capacitor that feeds the grid with the incoming signal. The grid leak resistor drains off excess electrons and prevents total blockage by the grid.
The bias is further established by the signal itself. The signal actually establishes the majority of the bias. When the positive portion of the signal wave strikes the grid for the first time, grid current flows between the grid and cathode, since the grid will be positive. This effectively shorts out the first positive wave. Then the wave swings negative, and brings the grid negative. Grid current does not flow in this case. Some of this negative charge is maintained by the isolation capacitor on the grid, and the grid leak capacitor leaks off the excess. If the grid leak resistor were removed, the capacitor would charge up more and more negatively, and eventually the tube would cut out.
A small amount of distortion and fuzziness occurs due to the initial grid current that is drawn, and subsequent non-linear operation of the tube. The positive portions of all successive waves can also be flattened a bit, depending on amplitude variations, which also causes distortion.
All-in-all, however, a great majority of radios use this method of grid bias, and they sound very good.
You should probably try rejuvenating your bias cell first, before converting your radio. You can also substitute the bias cell with an appropriate voltage watch battery, if the bias cell won't rejuvenate.
If you do convert the radio to grid-leak operation, be sure that there is an isolation capacitor of from .002-.01 MFD between the grid and the signal source, and be sure to run the grid leak resistor from the grid of the tube to the cathode of the tube.
You could also add a cathode bias resistor, and tie the grid leak on the more negative side of this resistor for increased negative bias. Cathode self-bias is also self-adjusting, and is less likely to allow grid current to be drawn, if the bias is set at the right point.
You can also eliminate the grid leak bias with cathode self-bias by decreasing the grid leak resistor to a valve of from .5 to 1 Meg. Increase the value of the isolation capacitor to from .01 to .03 MFD, since the decrease in resistance will decrease the time constant of the circuit, which will decrease bass response if the capacitance is not increased.
You should likely be able to rejuvenate the bias cell by immersing it in boiling water and then sealing with silicone caulk.
If not, you can use a watch battery to serve the purpose. You could also use a higher voltage for the filament, with a dropping resistor on either side, to establish a bias. However, this would waste current to only achieve the same result as the watch battery, which would last much longer due to the absence of current draw.
I have come across at least one radio with grid-leak bias on the output tube. I forgot the make of the radio, but it used something like a 3 Meg resistor on the grid, and had no cathode bias resistor. The cathode was terminated to B-, the same location as the grid leak resistor.
When a radio like this is converted for AC/DC operation, all of the tubes are strung in series. By nature of the voltage drop across the filaments, negative bias can be developed for biasing the output tube. To rearrange the filament string in such a way, use something like a Zenith 6-G-601 radio as an example. The grid biasing resistors will have to be terminated as shown in its schematic in order to establish proper bias. Also note various small-value filter capacitors around the oscillator filament. Furthermore note the resistor shunting one half of the output tube filament. This is because the more negative portion of the filament will conduct more cathode current, and will be run brighter. This current will also be drawn through the rest of the filaments in the string, and run them brighter than appropriate. The shunting resistor reduces the current drawn through the filament string, and keeps the filaments at proper voltage. It keeps the two filament halves of the 3Q5 output tube balanced, too.
The addition of small 25-35 volt 40 MFD capacitors from various points in the filament string to B- will also help reduce fluctuations in cathode current from straying from one tube to the next.
If you wanted to convert a 'farm' radio for AC use, it would merely be a matter of converting the filament string as explained above, and then build a simple B supply. The simplest design would be the power supply of the Zenith 6G601. Otherwise isolation transformers could be used for the A and B supplies to isolate the set from the AC line. The AC/DC design does not isolate the set, but works just fine so long as the set is protected within its cabinet. The AC/DC design also uses warm-running dropping resistors for the filament string, but is easier to filter with smaller value capacitors.
Converting the 3Q5 to grid-leak operation, or rejuvenating the bias cell might be the easier route to go, however, with the filaments being powered by a transformer-fed supply.
I now just empty the cell, and use it as a battery holder. I install the watch battery inside the cell.
Vianney:
That sounds like the best way to do it, IMHO. Why try to use an antique cell when there are modern, better shelf life and everything, components that you can hide in the original and make it look real and work better?
Lewis
:
: