Battery Eliminators

Battery Eliminators for 1920s Antique Radios

By Doug Criner

Tuned-radio-frequency (TRF) radios from the mid- and late-1920s were battery powered. Many required three DC voltages: a high-voltage B+ plate supply, the A filament voltage, and a C– grid bias voltage. Some sets used the A voltage for the grid bias, dispensing with the C–.

Nowadays, it is most practical to use a solid-state battery eliminator to supply these voltages. There are two popular choices: the ARBE-III and a kit from Antique Electronic Supply (AES).

ARBE-III

This is an elegant, well-designed device, and is highly recommended. The price, with shipping, is about $175, but is worth it: http://www.arbeiii.com/ Personally, I regard this battery eliminator as perfect for the repair bench. However, the price of this unit may rival the value of radios it is used with.

ARBE-III. The yellow tool is used to adjust the "A" voltage. The newer version has a pilot lamp and a silver-colored heat sink instead of black.

The ARBE-III is enclosed in a nice box, but has exposed terminals that have plate voltages, so the unit needs to be out of the reach of children or the unsuspecting.

The B+ and C– voltages, are brought out on separate terminals in the usual discrete increments. The A voltage is continuously adjustable with a potentiometer, which requires a voltmeter to adjust. In using the ARBE-III as a bench supply with different radios, it's important to remember to re-adjust the A voltage to avoid zapping an entire set of tubes!

AES K-101A Kit

This kit of parts from Antique Electronic Supply sells for less than half the cost of the ARBE-III, but lacks many of the ARBE-III’s niceties: http://www.tubesandmore.com/

If I were to recommend the first or only battery eliminator for purchase by a radio collector, I would suggest the ARBE-III. However, I like to have several old TRFs around the house in playing condition, and I enjoy giving away battery-powered TRFs to friends who appreciate them. For those sets, the cheaper AES kit makes sense.

Modified AES Kit, assembled.

The AES kit is assembled on a wooden “breadboard,” so all the plate voltages and 140-V AC (the transformer supply for the plate voltages) are exposed. Preferably, the completed kit should be mounted inside a suitable project box. Also, the AES kit is not fused and does not have an ON-OFF switch. The AES kit is a reasonable value, considering the prices one would pay for the individual parts, particularly the transformer.

The kit has three separate sections for the A, B+, and C– supplies. The A and C– voltages are continuously adjustable with potentiometers; a voltmeter is required to set those voltages. The B+ voltages are available at terminals in the usual increments from 22V to 135V.

Schematic of K-101A Kit (© Antique Electronic Supply)

The design of the plate-supply section is interesting. 140-V AC from the transformer is rectified by a bridge diode and capacitor filters. Then the DC voltage is regulated by two discrete transistors before it is supplied to a conventional string of zener diodes to provide the usual variety of plate voltages. It’s a modern, up-to-date, efficient circuit and will automatically adjust to variations in line voltage. I don’t like it.

The two transistors used to regulate the voltage to the zener string are, in my opinion, gilting the lily and susceptible to being blown. I have encountered some noise and other weirdness that might be attributed to the regulator circuit. In any case, I’m more comfortable with just feeding unregulated DC through a string of zeners. Therefore, I developed a modified design that does away with the regulation in favor of a simple zener diode string. It works fine, and most of the required components come with the kit.

With my modified B+ design, several components that come with the kit are surplus, namely the two transistors. Even so, I think that the AES kit is a reasonable value.

Modified B+ circuit for AES kit.

(The ARBE-III also has a voltage regulator ahead of the B+ zener string, but it’s in a sealed box where I can’t see it or get to it, so it doesn’t bother me quite so much. Also, the ARBE-III’s regulator is said to be an integrated circuit rather than discrete transistors with exposed terminals.)

The AES kit seems to put out more RF hash and hum than the ARBE-III. Most of my TRFs supplied by the ARBE-III perform well, even without a ground connection. For the AES kit, a ground connection to the radio seems to be more important.

Referring to my modified plate supply design, consider the sizing of resistors R1 and R2. The total, R1+R2, should allow the current through the zener string to be greater than the maximum current that will be supplied to the radio and less than the maximum current that can be safely carried by the zeners. The rated current of the 140-V winding of the AES transformer is 100mA, so we obviously need to limit the current below that. I suggest sizing R1+R2 to provide about 50mA. This will give plenty of power for virtually any 1920s radio – for example, a single ′01A tube will draw about 4-8mA plate current, depending upon the plate voltage, grid bias, and filament voltage.

(If '01A tubes are properly biased, the plate current will be toward the lower end of the 4-8mA range. Note: if '01As are run at 135V B+, the C– grid bias should be –9V rather than –4.5V prescribed for 90V operation. Radios with '01As that use the filament voltage for the grid bias, and have no separate C– connection, should not be operated at higher than 90V B+.)

This suggested current of 50mA will flow through R1 and R2 with or without a load connected to the B+ terminals. When a load is connected, it essentially “steals” current that would otherwise be going through the zener string. My suggested values of R1 and R2, 150 and 500 ohms, may not give 50mA due to variations in your line voltage or other tolerances. You can calculate the current by measuring the voltage across either R1 or R2, and adjust the values of either of the two resistors to achieve about 50mA.

If the B+ supply is limited to 50mA, the zener diodes will run quite warm, but well within their specified limits.

Additional information, May 2007: I have received comments from Norm Leal and Marv Nuce on my modified AES battery eliminator design.

Norm points out that with the modified circuit, accidentally shorting between two B voltage levels, one or more zener diodes can be blown. A fuse ahead of the zener string will not act fast enough to protect the zeners. Two possible protective schemes are available: addition of a crowbar circuit or adding regular silicon diodes to each output terminal, with the cathodes connected to the terminal and the anodes connected to the zener string. With or without such protection, I recommend that the power supply be fused in case one or more zeners fail shorted or any other overload/short condition occurs.

Marv theorizes that any noise coming from the battery eliminator, as originally designed, may be coming from the zeners, not the transistor regulators. He suggested shunting each zener with a small ceramic cap. My modified circuit produces no noticeable noise. Perhaps eliminating the transistors and feeding the zener string directly from a C-R-C-R filter network suppressed zener noise?

Shock Hazard

Either the ARBE-III or the AES kit present dangerous shock hazards due to exposed voltages. The high-voltage B+ connections on the terminal strip of the ARBE-III could be touched. The AES kit's circuitry is completely exposed.

Both battery eliminators have transformer supplies, which means that the exposed voltages are not referenced to ground, reducing the shock hazard somewhat since two separate points need to be touched simultaneously to complete the circuit through your body. However, most old radios call for a ground connection to improve reception, and this ground often is tied to B–. Then the B+ voltage is indeed referenced to ground, presenting a shock hazard between your body and ground.

Non-TRF Radios

Besides the TRFs from the 1920s, there are other tube-type radios that used batteries. Farm radios used a wet-cell battery and generated B+ plate voltage with a vibrator circuit. The battery eliminators discussed above are not suited for either of those types of radios.