[MrFixr55]

Edited 12/21/22
The P-T156 transformers offered by AES *Antique Electronic Supply) do seem to work well for this app. I have a Radiola 16, 17 and 18, along with a 100A loudspeaker. I used the P-T156s on the 16 and 18 successfully. The cabinet for the 100A speaker is pot metal which swells. This is possibly the reason for your not being able to remove the screws. The speaker is a magnetic type very common for the era and not known for being high fidelity. There is a "PI" Filter that acts as a low pass filter. However, they were much less expensive than the GE "Kellog - Rice" Electrodynamic speaker, the archetype of all dynamic speakers that followed. For best fidelity, any good 8" speaker with a 2K Ohm to 8 Ohm output transformer will do. (A transformer from an AA5 AC DC Radio is just what you need.) Your Jensen is just the trick. PA Speakers with a 70V Matching transformer work well. Use the 0.5W 0.75 W or 1.0 W tap

The Radiola 17 and 18 used the new "AC" Tubes, the '26, '27 and '71A. The existing 01A tubes, used on the Radiola 16 and many battery operated radios dating back to 1921, would hum horribly if the filaments were powered by AC.

What are you using as a transformer for the power supply that you are building? The 26s operate on 1.5VAC, the 27 operates on 2.5V AC and the 71A runs on 5V AC or DC.

Do expect hum! The '26 and '71A are touted as an "AC" Tube by the "P T Barnums" at RCA at the time (Read their ads that described reception by RCA radios of this era as being indistinguishable from live performances). However, being that they were both direct heated tubes, there will be hum. The 26 hum was mitigated by the slow responding low voltage high current 1.5V 1A filament fed by a supply with a variable rheostat for hum balance. the 71A Filament differed from the 71 by also having a more heater-like thoriated filament whose emission varied less during the peaks and valleys of the AC than the original 71 tungsten filament did. The real AC Tube in the set is the '27, the first true AC Tube with separate heater and cathode. This is the archetype of every voltage amplifier triode with a heater and sleeve cathode ever since. There was a big push to make what RCA referred to as a "Socket Powered" radio. The first radio in this series was the battery powered Radiola 16 which used 201As ('01As) for the 3 RF amps, detector and 1st AF Amp. The detector used a spring mounted 4 pin socket to minimize "microphonics. The "Power" Amp was a 112A, emitting 0.35 Watts at 135V Plate voltage. The Radiola 17 used the same chassis and almost all parts, except the detector socket and tubes as the Radiola 16. The 27 was used as a detector instead of the 26 as, for some reason, the 26, which is functionally similar to an 01A except for filament, was deemed as unsuitable for use as a detector. Possibly, the AC on the filament would raise the hum level, which was already objectionable by today's standards. The Radiola 18 was similar to the Radiola 17 except for the IF transformers which were "neutrodyne-like" without violating the Hazeltine patent, a slightly different power supply design and the use of "choke and capacitor" coupling instead of the transformer coupling of the Radiola 17 and the direct coupling of the Radiola 16. Note that the Atwater Kent 36 and 40s used the same circuit as the Radiola, while a Zenith of the same era used 27s for all tubes except the output tube. The RCA Radiola 60 Superhet, introduced the same year as the Radiola 18 also used 27s for all tubes except the output tube. This contributed to better fidelity through lesser hum. In addition, the Radiola 60 used a biased "Power Detector" instead of the grid leak detector used by RCA
BTW, Westinghouse and GE built almost all RCA Radios from its founding in 1919 till the acquisition of Victor, (who had started producing radios in 1928) and the introduction of the Radiola 80. Then the tables were turned and RCA built almost all GE and Westinghouse radios from 1930 till between 1934 and 1936.

The 71A has an output power of approximately 0.7W.

I built an RC Coupler with a switch and volume control that subbed for the non-existent radio phono switch. I was not impressed with the level of distortion using this as an AF Amp. The hum likely contributed to that. There are several discussions on the internet on single ended amplifiers and preamps using constant current power supplies, 01A and 71A tubes. (Other than the filament voltage, the '26 and '01A have very similar characteristics and the same base diagram.

Another source of distortion is the "grid leak" detector, which had higher distortion than biased detectors or diode detectors.