RCA PA-103 / PA103A1

[emrr]

RCA PA-103 (PA103A1). 1932ish mic preamp from the sound reinforcement catalog. Fixed gain with unusually high output and B+ current requirements, almost qualifies as a line amp. I had the opportunity to make a restoration plan for a pile of these several years ago, so observations are based on multiple units. These are super scarce units, I’ve not encountered any others.

250Ω in/out, 58dB, 300V @ 27mA, 2.5V @ 3A
Output looks pretty clean to at least +16dBu, definitely distorting by +22dBu.

Lots of quirks here.

- Early telco style design with things protruding off the front and back of the panel, somewhat helps with rack weight distribution. The tube deck is shock iso’d from the rest of the chassis. 0.5 inch thick transformer case shields.

- mechanical layout DEFINITELY NOT with service in mind. The caps are all in a sealed pack (5 caps, 1 inductor) with wire leads which looks like a transformer case, have to disconnect and isolate those wires and find new spots for caps. It's a very tight sandwich between the resistor board and the shield plate.

- not being broadcast line, it’s carbon resistors. “Dog bone resistors” as the old timers call them. All new modern carbon comp in keeping with the clients wishes, except for the 1W R.

- 2.5VAC tube types, not a lot of obvious 2.5V/3A supply options.

- input low pass cap to limit upper bandwidth, also tames a big transformer resonance. You can take it out and live with the inaudible resonant peak for a surprisingly modern sound. Larger cathode caps also afford more bottom, an octave. Response is best with a 250R load resistance, mainly treble improvement.

- input center tap is grounded, and must stay that way, out of control resonance if lifted, typical of early RCA input transformers.

- does not run quietly or with expected frequency response without the tube cover box in place. In fact you can reseat the cover on these multiple times and watch the response plot settle differently each time! There’s a shield plate between the tubes and the resistors too, which also affects response and noise.

- you could put an interstage volume control in, tests showed no response changes across the full throw of a gain pot. Where to put a gain pot is another matter, practically and non-invasively.

The tube current meter points were someone's DIY addition.

Restored they sound really good, and a bunch of channels would heat a room and run the power bill up in no time. Full info can be found in the Rider manual, lucky break there.

The associated PSU is a crazy brute force CLCLC unit that can handle 4 of these, if you use less than 4 PRE there are (3) 100W wire wound resistors in a cage to be strapped as loads replacing any less preamps, so a PSU with one PRE is set up to burn 78ish mA of B+ as heat!

https://live.staticflickr.com/65535/53646345124_a1d7245687_b.jpg

https://live.staticflickr.com/65535/54183832088_65bfb0045c_b.jpg

https://live.staticflickr.com/65535/54182676347_bd0711aa8b_b.jpg

https://live.staticflickr.com/65535/54184001655_694f7f3237_b.jpg

https://live.staticflickr.com/65535/54183846094_34128d6568_b.jpg

https://live.staticflickr.com/65535/53646464855_3291903d07_b.jpg

https://live.staticflickr.com/65535/53646351284_6ee20d9530_b.jpg

https://live.staticflickr.com/65535/53646469685_75a6b66507_b.jpg

https://live.staticflickr.com/65535/53646470840_89938779ff_b.jpg

https://live.staticflickr.com/65535/53646022581_1ff9f5ec55_b.jpg

https://live.staticflickr.com/65535/54184031070_39af76c1d7_b.jpg

 

[mjrippe]

Very cool, Doug! I guess if you didn't have the original PSU, you could put the filaments in series and run them at 5v.

 

[emrr]

Very cool, Doug! I guess if you didn't have the original PSU, you could put the filaments in series and run them at 5v.

One's a 2A tube, other's a 1A tube.

 

[mjrippe]

One's a 2A tube, other's a 1A tube.

So they need to be balanced and won't each just pull what they need? Or does that only work in parallel filaments?

 

[emrr]

the 1A tube would have 3.3V across it, the 2A tube 1.7V. Voltage divider, resistance equivalent of current draw, filaments likely much more non-linear than the equiv resistor, so results vary.

 

[Brian Roth]

Thia convo reminds me of the sets of valves intentionally designed to "play nice" with series filaments for the "All American Five" radios sold in the 40/50/60's. Individually 35, 50, whatever Volt filaments so the heater string added up to 110 VAC-ish.

35L6, 50L6 beam power tubes anyone? lol

Bri

 

[mjrippe]

the 1A tube would have 3.3V across it, the 2A tube 1.7V. Voltage divider, resistance equivalent of current draw, filaments likely much more non-linear than the equiv resistor, so results vary.

Ah, that makes sense. Thanks.

 

[mjrippe]

Thia convo reminds me of the sets of valves intentionally designed to "play nice" with series filaments for the "All American Five" radios sold in the 40/50/60's. Individually 35, 50, whatever Volt filaments so the heater string added up to 110 VAC-ish.

35L6, 50L6 beam power tubes anyone? lol

Bri

Yes, I was thinking of the AA5 but too lazy to look them all up! II don't think you could give away a 35L6 these days

 

[vintage-electric]

Emrr, Great work! As a fellow restorer, seeing someone do a clean job is a real pleasure. Curious, were the dog-bone resistors badly out of spec? In my experience, carbon comps seem to drift more out of spec than the dog bones. Admittedly, there are many variables.

Again, a clean job!

-Kent

 

[emrr]

Emrr, Great work! As a fellow restorer, seeing someone do a clean job is a real pleasure. Curious, were the dog-bone resistors badly out of spec? In my experience, carbon comps seem to drift more out of spec than the dog bones. Admittedly, there are many variables.

Again, a clean job!

-Kent

Thanks, I do as much as possible to keep things original in every way, mods non-invasive and reversible, etc.

The resistors were all over the place, and the before pic was one of the few units that hadn’t already had most replaced with carbon 50-60 years ago. Then they lived in a barn for 40 years, no temp or humidity control.

I feel like I find plenty of carbon that’s in spec, but noisy from decades of humidity.

 

[MidnightArrakis]

Thanks, I do as much as possible to keep things original in every way, mods non-invasive and reversible, etc.

[I do as much as possible to keep things original in every way] -- Then.....my response here is probably both unneeded and unwanted, but.....as I was reviewing your photos, it seemed to me that perhaps the metalwork used in this old piece of gear could use some "sprucing" up. If so, I'm game to mechanically recreating the rack-panel, brackets and enclosures so new and identical metalwork could be fabricated. Then, you could place all of the old circuitry into new metalwork have this preamp looking just "spiffy"!!! Just sayin'.....

Also.....the rack-panel "looks" to be slightly thicker than the standard 1/8th-inch thick (0.125") rack-panel. Do you know its thickness? Was the old wiring "brittle" and ready to crack at the slightest touch? I can remember working on old audio gear and radios back in the 60's and having the insulation on wires "just fall off" when they were touched!!! The wiring I see in your photos looks they would be very stiff and resistant to be moved around. GOOD JOB!!!

/

 

[emrr]

You wouldn't dare try to spiff this up. The 80+ year old wiring feeds directly into the transformer cans and will break if you look it funny. If you took it apart to paint you'd kill something for sure, or have to replace sections of wiring, making it look non-original. This metalwork package is complex, and would certainly cost a lot to replicate correctly. I can't see it. Almost all older panels are 3/16" copper plated steel.