siemens: 11GW; Supply and Negative bias generation for ou
I have run the 11GW on 115V and 220V. Now that restoration is finished, I left it in the my local 115V mains setting, because it draws less power/heat.
One interesting bit about this radio is that the B+ voltage follows the mains directly, so it is twice with 220V operation. The ballast resistor is used only for the heaters. I did notice somewhat more sensitivity/gain with 220V.
The RV12P2000/LD1 audio module replacement for the UCL11 works well in either setting because the grid bias also follows the B+, but the 220V setting delivers higher power before clipping at the preamp input.
I downloaded two different versions of the 11GW schematic from RMORG. The older version modified with the audio module, is at d_siemens_11gw_audio_module.pdf The later version d_siemens_11gw_sch.pdf shows separate negative terminals for the two supply bypass caps. This is a very important difference for hum injection into the power tube grid bias.
Having two separate electrolytic caps eliminates the need for the low voltage electrolytic bypass at the power tube cathode, and also makes it possible for the combination preamp/output tube to share the cathode at ground potential. The current drop is obtained from the current flowing between the two electrolytic negative legs, which still has the hum of the first cap.
In the original wiring, both electroytics returned to ground, the bias was obtained from unfiltered B- pulsating current. On first inspection, it would seem that this could not work at all, given that the pulsating current was zero most of the time, and produced -20V negative peaks accross the 125 Ohm bias resistor.
But the large 700k grid bias resistor and 10nF grid coupling cap formed a 2Hz low pass filter to filter the pulsating -20V into a ripply -5V. So the stage coupling cap serves two purposes in this circuit. The remainder of the hum clean up was done with negative feedback from transformer output to the bottom of the volume control. The feedback network is a wein bridge type of RC-CR network with a resonance around 15Hz, for maximum feedback at this frequency.
If the negative feedback lowered gain by 4x at 60Hz, that would mean a 4x reduction in hum for a given level of output. I would guess that if negative feedback is used there it should drop the gain by at least 2x, and probably 4x, otherwise there is not much point in NFB. ( I did not measure how much NFB dropped the gain)
The negative feedback is reduced at higher volume levels or for very weak stations as the volume pot is cranked up but is not eliminated.
My 11GW came wired as an early version with common negative bypass cap terminals. Separating the negative terminals of the electrolytics as shown in the later schematic, improved hum further. In my restoration, I splurged a bit and added a 1000uF bypass accross the 125 Ohm negative bias resistor because I wanted to clean up the ripple for use of this negative bias source with the preamp stage. I think I kept the spirit of the audio module design of gently improving the radio.
It appears that the economics of electrolytic caps ultimatelly resulted in two capacitor bodies in later desings: One low voltage cap for the cathode bypass, and two high voltage caps in the same can for B+. This solution should cost about the same as the solution in this radio, but having a total of three caps in two cans produced the least hum.