My old "light-socket" AC radio set was designed for nominal 110 volts AC line. Modern line voltage is typically much higher, up to 120 volts or more. This higher voltage unnecessarily stresses the tube filaments and may shorten precious tube life in these old sets. A simple rheostat solves the problem. Also, strong noise and high-power local AM signals may enter the set on the AC line. I used a high quality RFI filter to address this problem. I built a line filter and voltage control unit for my Atwater Kent Model 40 set; http://www.youtube.com/watch?v=qa7-h9FIqWo
The writeup with drawings is in my SkyDrive folder; https://skydrive.live.com/#cid=A7EE38C4 ... 2053CA!274

What is the wattage of the rheostat?

A Good question, thanks. The rheostat I used is an Ohmite #RJS50R, 50 ohm, rated 1 amp maximum (50 watts). As used here, it dissipates about 3 watts after adjustment for a 10 volt drop. I then determined by measurement that my AK Model 40 set uses about 34 watts at 110 volts AC. See also the schematic in the write-up I posted in my SkyDrive folder. Thank you for asking!

Is the lamp socket meant for sets that had Edison-type plugs on their power cords rather than standard 2-prong plugs?

I don't know the answer. The earliest AC-tube sets were sometimes referred to as "light-socket" radios in their advertising. Some photos in the Atwater Kent Parts List Manual show their AC power cords with attached light socket adapters. Also, see below an excerpt from http://www.philcoradio.com/history/hist2.htm

"...However, a new development from the Radio Corporation of America (RCA) in August 1927 would finally make it practical to build a radio that could be operated from the light socket, without batteries or battery eliminators."

You don't know why you put the lamp socket on your box?

RRM

I had three reasons for using lamp socket;
1) Just for fun, since my radio was called "light socket" radio (so a two-prong adapter is screwed into the light socket, and then the radio into the adapter, just like the old days).
2) When I first breadboarded the circuit it was convenient to use a 40-watt light bulb as a dummy load so I could test the range of output voltages that the rheostat would provide. My AK40 consumes similar power (about 35 watts) to 40 watt bulb when the rheostat is set for 110 volts. A 40-watt bulb might blow the 1-amp fuse I am using, due to its low turn-on resistance. I haven't tested with a 40-watt bulb since I added the fuse.
3) The third reason cropped up by accident one day, when my wife plugged in her sewing machine to the "110 volt" outlet of my breadboard (I wasn't around when she did it). The sewing machine still worked but not quite normal, and she complained about it. The lamp socket helps her remember that she should use one of the provided 120VAC outlets instead.

I believe that the light socket radios were the early models when the only use for electricity was lighting. The common (now) duplex outlet didn't exists. Why should it, there were no electrical appliance with plugs to plug in. Lighting was "hard wired" in, not lamps with plugs, We forget how primitive it was back then. The radios were screwed in where a bulb would have been. PL

By the time AC radios came around, 'light socket' plugs were on their way out. By 1930, most manufacturers did not sell actual light socket plugs on their appliances. It was up to the consumer to buy the adapter that converted prong to socket.

Adapters are in the front row...

The change from the Edison-size screw-threaded plug/receptacle over to the Hubbell (original patent for the bladed receptacles was to Harvey Hubbell Co., competitors like Bryant and Crouse paid royalties to Hubbell) bladed receptacles was a major safety innovation through the mid-20's to early '30's, according to old player piano service manuals I have. Electric-run pianos usually came with a very long cord w/ Edison-male. Since it's easy for a child to insert their fingers or tongues into the Edison socket (or splashed mop water), the national code did not allow Edison sockets any lower on the wall than about 6 ft high, and in some NE States (such as Maryland) only ceiling sockets were allowed. It took a while for the now-common 4" (in baseboard) to 18" high (in plaster-lath) receptacles off the floor to become standard.

Have seen this posted before about today's line voltage lowering the life of old tubes. I for one believe it is a bunch of malarkey. I for one have never had that kind of problem. Granted i do not run any of my radios 24/7. Even then how would you be able to tell that 120 volts made a tube burn out faster then 110 volts. Lets say you had two radios on the shelve. Both of them the exact same down to the smallest resistor. Turn them both on and leave them on. Would any one want to bet as to which radio would die first. I don,t think so. To many variables involved here
Stan Ski

I don't know if it is malarkey or not for receiving type tubes. However, the Eimac Corp. addressed the subject of filament voltage management for extending life of transmitting tubes; http://www.bext.com/filament.htm.

For example, they said "The key to extending the life of a thoriated tungsten filament emitter is to control operating temperature. Emitter temperature is a function of the total RMS power applied to the filament. Thus, filament voltage control is temperature control; temperature varies directly with voltage. As the emitter temperature rises, the de-carburizing process is accelerated and tube life shortened. Figure 2 shows that the useful tube life can vary significantly with only a 5% change in filament voltage. "

The 1928-era AK manuals state 110VAC as the nominal line voltage for their AC sets, not 120VAC. The model 42 and 44 sets have line-regulating resistor mounted on side of the power supply enclosure. I have measured our line voltage as high as 125 volts, occasionally. The Globe type tubes in my model 40 set are all 84 years old now. A NOS UX-280 is now $100 from AES (or $66 used). A NOS UX-171A is $86. Given these considerations vs. the risks, I am more comfortable to operate my set at 110VAC. Thank you for your interest.

The question I have is what type of voltmeter was used to check line voltage back when these radios were made. Was it a 1000 ohms/volt, 20,000 ohms/volt, or a vacuum tube volt meter. Depending on what was used back then, 110v on a modern DMM might be considerably lower than what the voltage actually was back then.

VOMs back them were usually 1000 ohms per volt. I have one just for that reason - they were used used to generate a radio's table of voltages from radio manufacturers, etc..

Chuck

The ac line voltage can be measured with any type of meter. Meter resistance is only of concern when measuring circuit voltages that have a high source impedance.

RRM

When I re-capped my Atwater Kent 40, I used three .47uF, 600 volt film caps and two 10uF 450 volt e-caps, which drops things down to a safe level when plugged into "modern" house current.

No difference using a 1000 ohm /volt or 1 Mohm meter readind the AC line. The source impedance is such that meter loading is next to nothing.

To Retired Radio Man and Unusualdesigner , thanks for that info.
Bob

You're welcome.

The purpose of providing AC line voltage reduction has more to do with temperature vs. life of the tube filaments, rather than B+ voltages. Another possible benefit of having the additional resistance (about 35 ohms) is a slight reduction in the cold turn-on surge current in the filaments. This can't hurt, and may help.

In response to several good feedbacks above, I have posted Revision A to my writeup in my SkyDrive folder. See also the excerpt from the Atwater Kent Service Manual attached below.