ERLA Radio Problem

Here is the rule of thumb I use for figuring audio impedances: Multiply the DC resistance of an audio transformer input or output winding or audio choke or in this case a 1920s horn or cone speaker by 10 to 20 to get the AC audio impedance.  So 1800 Ohms times 10 will be about 18000 to 20000 Ohms AC impedance which is just perfect for a 201A tube. This also works for 2000 Ohm headphones and the various audio transformers and reactors/chokes out there. Of course you need to take into account the turns ratio of transformers and their loads when you are trying to match them.

Rod  WB6FBF

I was thinking that the A01 tube plate resistance is either 10,000 or 11,000 ohms and the load is supposed to be matched to the output. So even though the earphones are less they would be more sensitive than the speaker. Since I have never worked on any really old radios of this nature It's just a shot in the dark. Just for the fun of it insert a matching audio output transformer with a primary of around 10,000 ohms and a matching speaker of 8 ohms on the voice coil.

I cannot find anything on matching tube impedance to to audio transformers.
But I understand that the two have to match.

1800 ohms

What is the resistance of the speaker voice coil?

Tubes are good and correct type Frank.

Crazy thought, have you swapped tubes around? Possibly a problem with the audio output tube.
Are you 100% certain that the output tube is an 01A? 
Some of those sets used a different tube as the output, and some had production changes that were not documented.

I use 2K Baldwins almost all the time when I need to use a headset Frank.

Gotta ask- What impedance headphones are you using?
I had a similar problem with a set until I realized I was using 8 ohm instead of 2K or better.

Touche!

One of the first things you do when restoring a battery radio is check the transformers since they often go bad. Both of the transformers are good.

Ok. Roger on all that. Good performance on headphones. That means the Tuned RF stage is good. Untuned RF stage is good. Reflexed 1st Audio stage is good. Fixed crystal Detector stage is good. Second Audio stage is not good. Have you checked the windings for continuity on that 2nd ERLA audio transformer? I know this is a very basic question to ask. But it is an important one to know if the 2nd audio transformer is good. A transformer with open windings will not perform up to expectations. Good work on all. Very interesting project.
Rod

I put in a diode and it made no difference. Sound is loud and clear through headphones. I'm starting to think the amp is just weak by design, similar to a Radiola Balanced amplifier.

One other thing to consider is how good is that fixed crystal detector? You could check it out in one of your crystal radios or substitute a modern germanium diode temporarily for the fixed detector in the set and give it another try to see what happens.
Rod

Well, that is interesting. I have to admit that I haven't operated a set based on untuned RF transformers but I gather that they were employed early on (1922) to get around the tendency of tuned RF amplifiers to self-oscillate. Hazeltine's work of 1923 brought out the method of neutralization for stabilizing TRF amplifiers that soon took over and superseded the untuned multi-stage approach. I mention this because I can't speak from experience how well such a set will work in practice. Maybe you could comment on this.

One thing I have been wondering about is the purpose of that second tuning condenser-the one with less plates. It appears to be an attempt to tune or peak the secondary of the 1st untuned RF transformer. This appears unusual and fruitless to me and goes against normal practice of the time. Does adjusting this control help to tune in stations? Can you peak up a station to maximum volume or strength with it? If it doesn't do much, then I suggest you disconnect it if only temporarily to see if the radio would perform better without it. As I see it right now, that condenser is acting as a short to ground for radio signals appearing at the secondary of that transformer. These signals (stations) go on to the grid of the middle tube to be amplified and if they are not making it to the grid of that tube then that tuning condenser is not helping the set. 

If disconnecting this tuning condenser does help improve performance then consider the following: It may not be very satisfying to have an unused control so prominent on the front of this receiver but keep in mind that the set may not have worked for the original builder back in 1922 or 1923 and had been set aside, but fortunately for us, preserved down to this day.

I made the changes you suggested Rod, now the volume goes up and down with the rheostat adjustment like it should, but the volume is very weak.

I am looking over the circuit now. The first 201A tube is part of the tuner circuit, it is a tuned Grid RF amplifier, its grid being connected to the vario-coupler and tuning condenser. The plate circuit of this tube is untuned hence no need for neutralization. I also see that the middle 201A tube is an untuned RF amplifier as well as a reflexed AF amplifier. The untuned RF amplifier section works into the crystal Detector stage. The critical connection from the plate of the middle tube (as the First Audio Stage) to "P" of the second audio frequency transformer by way of the headphone jack is correct. However, the "B" connection on the second audio transformer is not correct. That can be easily changed in my estimation.

The first step here to take to correct the problem you have described is to disconnect the "B" connection of the second audio transformer from the common rheostat circuit connection after the on-off switch. Next, connect "B" of the second audio transformer to B+. You can make this connection anywhere it is convenient for you. For example, to the B+ connection at the two phone jacks. This should correct the problem and restore good performance. Let me know how it turns out just in case a further change is needed.

So far so good. That does sound odd what you described as you turned up the amplifier filament rheostat. I will take a look at the rest of the schematic-maybe I'll notice something that will explain it. 
Rod

Made some progress Rod, thanks to you. Made both changes and now getting some sound out of the amp section. When I turn the rheostat for the amp tube up the sound increases until I get to about the half way mark. I can hear it coming out of the speaker, but not very loud. Then as I continue to turn up the power, the volume starts dropping again. Never had that happen before with an amp.

Ok. I just took another look at the picture of the audio amplifier section. I see now what I referred to as the "F-" connection (grid return) is actually marked as "A" on the ERLA audio transformers. "A" is the grid return for the secondary windings of the audio transformers.  The two "A" connections are tied together in your schematic diagram as they should be. One of the audios also has a mica condenser from "G" to the  common "A" connection as well. Likewise, one of the RF transformers has a mica condenser from its "G" to the common "A" connection.

Here is what I suggest: make a connection, that is, run an insulated wire (use black spaghetti on buss bar wire if you have it) from the common "A" connection of the audio transformers and the mica condensers to the A- terminal on the terminal strip at the back of the radio. Solder the connection or firmly screw it together at each end.

After you have done this and should you find that the audio stages are now working, i.e. you are getting output from the speaker jack, then celebrate! Then you can think about moving the B- connection at the back terminal strip of the set from A- to A+ as I suggested before. This small change should reduce B+ current consumption as well as reduce overloading and distortion on strong signals.

What connections should I add or change Rod?

Well, I am just looking at the audio stages for now. I take it the headphones are wired to the Detector stage? If not, then maybe the following won't help. This is what I see or don't see: The audio transformer secondary returns (F-) are connected together as they should be but I don't see a connection from there back to the A- terminal. The amplifier sections won't work properly if that is the case. Also, a minor point is that the B- should be connected to the A+. That way the amplifier stages (RF or AF) will get some helpful negative grid bias from the 6VDC filament A battery connections.

I bought this radio quite a while ago in non-working condition. I pull it off the shelf once in a while and tinker with it to try to get it running. I've gotten the detector section of the radio to work fairly decently ( being a reflex, it will never be great), but am stumped on getting the amp section working. I've had some smart experienced radiomen try to help me on this but so far no luck. I'm sure its something simple since I can listen to the radio through the headphone jack loud and clear, but when I plug a speaker into the speaker jack, nothing. I've taken both jacks apart to clean and adjust them so they are working properly. I've check all the connections and resoldered several of them just to be sure. I've also pulled the tube base to clean it and test the connections between the binding posts to the pin tabs and finally to the pins, all check good. There is no schematic available for this ERLA radio so I have drawn one up. I have not changed any connections and am wondering if it has a wrong connection somewhere. Both audio transformers check good, no open windings. The RF transformers are mounted in tube sockets as you can see from the photos, they check good also. Any ideas?