I think I have figured out why triode and pentode detectors distort at low volume, unlike a diode detector, which is linear pretty much until it is overloaded.

Turns out that triode and pentode detectors, unless utilizing a grid leak system, are first classified as square law detectors. A weak signal is going to work in the lower knee of the tube's operation, which is non-linear. Give the grid a stronger signal and it becomes a linear plate detector. Same circuit, different strength signal. When stronger signals are fed in, the majority of each wave works in the linear portion of the tube's conduction.

Grid leak detectors, being a tube biased at cut-off, but by a different method, still have the same problem, since a very weak signal is going to work in the non-linear portion of the tube's conduction.

The only solution to this without completely changing the radio's circuitry is to move the volume control to the audio section, and then give the RF section AVC. That will keep the signal much stronger, so the detector will be driven harder. I had a lot of success with this change in my Majestic 180, so I'm going to give it a try in this little Majestic midget set.

T.

:I think I have figured out why triode and pentode detectors distort at low volume, unlike a diode detector, which is linear pretty much until it is overloaded.
:
:Turns out that triode and pentode detectors, unless utilizing a grid leak system, are first classified as square law detectors. A weak signal is going to work in the lower knee of the tube's operation, which is non-linear. Give the grid a stronger signal and it becomes a linear plate detector. Same circuit, different strength signal. When stronger signals are fed in, the majority of each wave works in the linear portion of the tube's conduction.
:
:Grid leak detectors, being a tube biased at cut-off, but by a different method, still have the same problem, since a very weak signal is going to work in the non-linear portion of the tube's conduction.
:
:The only solution to this without completely changing the radio's circuitry is to move the volume control to the audio section, and then give the RF section AVC. That will keep the signal much stronger, so the detector will be driven harder. I had a lot of success with this change in my Majestic 180, so I'm going to give it a try in this little Majestic midget set.
:
:T.
:

I remember a question on the FCC first class radiotelephone license exam was about square law detectors. I memorized the definition as being something like: "A detector that has an output that is the square of the input". I wondered what the devil someone would want something like that for, but we thought the days of TRF were far behind us, the wuperhet was here to stay, so I haven't thought of anything but diode detectors for years and years. You are right, the square law becomes more or less linear at higher inputs, but is very distorted at low levels.
Takes me back to the sixties studing a Q&A book.
Lewis

Another idea I had was to keep the radio's circuitry completely original, and then bias the detector so that it no longer detected. Instead I'd have a diode in the audio circuit. I kind of tried it, though, and it didn't work. Perhaps I have to try a little harder. I'd rather keep the radio's circuitry as original as possible, though AVC is a nice thing to have. Putting the volume control in the audio circuit requires using an ugly modern 1 or 2 meg unit, unless I can find an old 1 or 2 meg unit. I guess I could also gut an old unit that is no longer functioning properly, and hide a 1 or 2 meg unit inside (just the workings...not the whole case).

T.

:Another idea I had was to keep the radio's circuitry completely original, and then bias the detector so that it no longer detected. Instead I'd have a diode in the audio circuit. I kind of tried it, though, and it didn't work. Perhaps I have to try a little harder. I'd rather keep the radio's circuitry as original as possible, though AVC is a nice thing to have. Putting the volume control in the audio circuit requires using an ugly modern 1 or 2 meg unit, unless I can find an old 1 or 2 meg unit. I guess I could also gut an old unit that is no longer functioning properly, and hide a 1 or 2 meg unit inside (just the workings...not the whole case).
:
:T.

Whad wo you think of hiding a Voltage divider somewhere in the audio section so that the volume control has to be turned up to the linear portion of the detector curve at "normal" listeneing levels? Something that can be undone and the radio would be original again?
Lewis

I'm not sure I quite follow.....but with there being only 4 tubes, maximum gain must be utilized. Anything that would reduce the gain would make the radio perform very poorly. In some of the better sets, with more stages of gain, lower value resistors are used in the audio, making the audio section have less gain. That requires turning the volume up higher in the RF section in order to get good volume from the speaker. In those radios I can imagine that it would be more difficult to hear the distortion, since it would occur at very low volume levels. I could hear it in my Majestic 180 at very low volume levels. It was annoying. Fidelity was quite good at all other volume levels, though.

The benefit of AVC is that you can keep the signal higher to the detector. AVC usually maintains the RF circuits at a gain that would be necessary for full volume (since you get full volume when you turn the volume control all the way up). With the volume control now in the audio circuit (in any modern radio), you can still turn the volume down low without reducing the signal strength to the detector. Of course with modern radios they solved that problem completely by using diode detectors.

T.

:I'm not sure I quite follow.......

If you had a volume control in the grid of the AF tube, then you could run the antenna control to a point where the detector would be in the linear portion of the square law detector. In other words, turn the volume up with the antenna control until you reach the linear, then turn the audio down until it is at a comfortable (and hopefully undistorted) level. Not something you have tear the radio up to do, anyhow.
Lewis

Ah. Yes, that's what I'm going to do, though I'm not going to keep the antenna strength at full. Instead I"m also going to install an AVC circuit (very simple--one fast acting signal diode, a .0001 MFD cap off the detector plate, a 3.3Meg series resistor off of the diode, and a .05 MFD cap for filtering at the antenna end. The AVC circuit will automatically regulate the antenna circuit at a gain appropriate for linear detection.

T.

:Ah. Yes, that's what I'm going to do, though I'm not going to keep the antenna strength at full. Instead I"m also going to install an AVC circuit (very simple--one fast acting signal diode, a .0001 MFD cap off the detector plate, a 3.3Meg series resistor off of the diode, and a .05 MFD cap for filtering at the antenna end. The AVC circuit will automatically regulate the antenna circuit at a gain appropriate for linear detection.
:
:T.

Thumb to kink of it, I haven't looked at the schematic of this radio. Maybe I orta do that before giving advice. What sort of critter is this, anyhoo?
Lewis

Well, it's kind of on par to this:

http://www.nostalgiaair.org/PagesByModel/119/M0016119.pdf

I have homework to do, so I will draw a schematic and post it later, as well as a schematic of what modifications I will be making.

T.

:Well, it's kind of on par to this:
:
:http://www.nostalgiaair.org/PagesByModel/119/M0016119.pdf
:
:I have homework to do, so I will draw a schematic and post it later, as well as a schematic of what modifications I will be making.
:
:T.

Yeah, that'll be cool. I had forgotten that in the days before AVC, they used RF gain for audio output, exactly where you don't want to be on the curve of a square law detector.
Lewis

http://www.nitehawk.com/rasmit/ras_appl6.pdf

Nice. I guess solid state diodes become square law detectors at low voltages, too. I forgot about the .7 volts that needs to be overcome before a diode conducts.

One thing I did try with my radio was to have a variable detector bias. At really low volume levels I found that the detector would not distort if I made the bias less negative. I tried to come up with a method of automatic bias compensation, but it didn't work very well. It was somewhat like AVC. Audio was taken from the grid of the output tube through a diode with the anode facing away (negative supply), and then through a high resistance. It was then filtered with a .05 MFD capacitor. The grid coil of the detector was lifted from B- and then connected to the compensation circuit. The cathode bias of the detector was reduced until the circuit would self compensate at all volume levels. Various resistor values were tried from the detector grid to B- as well to see which one reduced the automatic bias to an appropriate level. At low volume levels not much negative voltage developed, so the grid was less negative. As the volume was turned up, more of a negative voltage developed, which biased the detector more negative. It was unreliable, though. A better method would be with a dual section potentiometer.........or to just eliminate the problem once and for all with AVC, and a volume control in the audio circuit. The reason why I haven't simply converted the radio yet is because I don't have any switched 1 meg pots at hand. Once I have one, it's getting converted. I want a nice listenable radio that performs well at all volumes. All new wire will be cloth, and the new potentiometer will be hidden inside of an old one, so it'll look nice. I can't find a schematic for this radio, so probably noone will know the difference anyway.

It is interesting reading about the problem, though, and then experimenting to find other ways to correct it (none of which have been successful for me so far).

T.

I have just finished restoration of a 1930's Canadian Marconi 162. They took an interesting approach to the linearity issue by using a "button" style 1 1/2 volt cell in a weird looking holder in the circuitry.I don't know what the life of the cell was expected to be,it was very dead when I did the work. I replaced the actual cell with a modern one and it works fine. They current is in the micro amp range, so even with the battery in circuit, life should be years. The radio has great fidelity with push pull 6F6's driving an unusually compliant 10 inch:Nice. I guess solid state diodes become square law detectors at low voltages, too. I forgot about the .7 volts that needs to be overcome before a diode conducts.
:
:One thing I did try with my radio was to have a variable detector bias. At really low volume levels I found that the detector would not distort if I made the bias less negative. I tried to come up with a method of automatic bias compensation, but it didn't work very well. It was somewhat like AVC. Audio was taken from the grid of the output tube through a diode with the anode facing away (negative supply), and then through a high resistance. It was then filtered with a .05 MFD capacitor. The grid coil of the detector was lifted from B- and then connected to the compensation circuit. The cathode bias of the detector was reduced until the circuit would self compensate at all volume levels. Various resistor values were tried from the detector grid to B- as well to see which one reduced the automatic bias to an appropriate level. At low volume levels not much negative voltage developed, so the grid was less negative. As the volume was turned up, more of a negative voltage developed, which biased the detector more negative. It was unreliable, though. A better method would be with a dual section potentiometer.........or to just eliminate the problem once and for all with AVC, and a volume control in the audio circuit. The reason why I haven't simply converted the radio yet is because I don't have any switched 1 meg pots at hand. Once I have one, it's getting converted. I want a nice listenable radio that performs well at all volumes. All new wire will be cloth, and the new potentiometer will be hidden inside of an old one, so it'll look nice. I can't find a schematic for this radio, so probably noone will know the difference anyway.
:
:It is interesting reading about the problem, though, and then experimenting to find other ways to correct it (none of which have been successful for me so far).
:
:T.

:I have just finished restoration of a 1930's Canadian Marconi 162. They took an interesting approach to the linearity issue by using a "button" style 1 1/2 volt cell in a weird looking holder in the circuitry.I don't know what the life of the cell was expected to be,it was very dead when I did the work. I replaced the actual cell with a modern one and it works fine. They current is in the micro amp range, so even with the battery in circuit, life should be years. The radio has great fidelity with push pull 6F6's driving an unusually compliant 10 inch:Nice. I guess solid state diodes become square law detectors at low voltages, too. I forgot about the .7 volts that needs to be overcome before a diode conducts.
::
::One thing I did try with my radio was to have a variable detector bias. At really low volume levels I found that the detector would not distort if I made the bias less negative. I tried to come up with a method of automatic bias compensation, but it didn't work very well. It was somewhat like AVC. Audio was taken from the grid of the output tube through a diode with the anode facing away (negative supply), and then through a high resistance. It was then filtered with a .05 MFD capacitor. The grid coil of the detector was lifted from B- and then connected to the compensation circuit. The cathode bias of the detector was reduced until the circuit would self compensate at all volume levels. Various resistor values were tried from the detector grid to B- as well to see which one reduced the automatic bias to an appropriate level. At low volume levels not much negative voltage developed, so the grid was less negative. As the volume was turned up, more of a negative voltage developed, which biased the detector more negative. It was unreliable, though. A better method would be with a dual section potentiometer.........or to just eliminate the problem once and for all with AVC, and a volume control in the audio circuit. The reason why I haven't simply converted the radio yet is because I don't have any switched 1 meg pots at hand. Once I have one, it's getting converted. I want a nice listenable radio that performs well at all volumes. All new wire will be cloth, and the new potentiometer will be hidden inside of an old one, so it'll look nice. I can't find a schematic for this radio, so probably noone will know the difference anyway.
::
::It is interesting reading about the problem, though, and then experimenting to find other ways to correct it (none of which have been successful for me so far).
::
::T.
Sorry about putting my post in the wrong place in the dialogue box! Rooky mistake...

Hi Thomas:

The first radio I bought was a 1928 Philco 531 (511 chassis) a few months back and it has the same setup you mentioned. The 'volume control' does nothing more than attenuate the signal on the antenna lead. Maybe this has been thought of alredy (or wouldn't work). Remove the 'volume control' pot wiring from the antenna circuit and replace with a fixed resistance value that gives good performance without overload on strong stations. Take the pot and re-use it to attenuate the audio reaching the AF stage. In the 511 schematic for example, what if a pot was put in series with the grounded lead on the first AF transformer's secondary? My thought is that raising that leg from groud would result in less signal passing to the first AF.

It is interesting reading about the problem, though, and then experimenting to find other ways to correct it (none of which have been successful for me so far)

Yes. I've done that, but I don't leave the antenna circuit at a fixed gain. I give it AVC. More later....

T.

:Yes. I've done that, but I don't leave the antenna circuit at a fixed gain. I give it AVC. More later....
:
:
:T.

Thomas:
A thought. Wouldn't you need a remote cut-off tube to apply AVC bias on the grid for gain control? It seems to me that the only reason to have remote cut off would be for AVC action. Could you substitute a latter day remote cut off tube that would match the appearence of your RF tube?
Just thinking.
Lewis

6D6 is a remote cut-off tube, which is exactly how they achieve volume control with the potentiometer in the cathode circuit. Making the cathode more positive with respect to the grid is the same as making the grid more negative with respect to the cathode.

AVC works wonderfully with the 6D6.

T.