Here's a radical new one that I don't think I've seen yet. They say that you can't take the output from the cathode if you want gain. Here's why this is normally not possible: If you put the load resistor at the cathode, and tap the audio (or whatever) from the cathode, the changing load will also change the grid bias, if the grid is connected to B- in the conventional manner. If the tube starts conducting more, the cathode-grid difference will increase, and the grid will swing more negative (with respect to the cathode). This will cancel out all gain.

HOWEVER! Here's the radical new idea! (Perhaps it's old, but I've never seen it before.) Let's say you're using a 12SQ7. Connect the plate directly to B+ instead of to a 1 meg resistor. Connect the 1 meg resistor to B- and the cathode instead. Tap the audio at the cathode. Now, instead of connecting the grid resistor to B-, connect it to the cathode as well. Normally, in radios, when the cathode is grounded, the grid resistor is grounded as well. Well, if you put a resistor between the cathode and B-, you must put the grid at the same potential as the cathode for this particular tube.

Now, that is kind of exciting, but nothing fantastic. So you're taking the audio at the cathode instead of the plate. Big deal. Now do it to a pentode of any kind. Let's take a 50L6. It requires a 150 ohm cathode bias resistor. Connect this resistor to the cathode. To the other side connect one leg of the primary of the output transformer. Connect the other leg to B-. Where the transformer and the resistor join, connect the grid bias resistor. Connect both the screen grid and the plate to B+. Here's what's so wonderful about this technique: At the plate, some electrons get collected at the screen grid. The plate doesn't realize the full electron flow (this is inefficient amplification). Also, if the tube isn't a beam grid tube, the electrons bounce around. This ruins fidelity at the plate. However, at the cathode, ALL currents that flow through the tube must first flow in here. With the output transformer (or load resistor) at the cathode, full power and fidelity can be realized. There will be no wasted wattage, because the current for both the screen grid and the plate will flow through the output transformer if it is connected at the cathode as previously described.

You would think that with the cathode changing all over the place, the grid would get confused, but its bias changes instantaneously with the cathode, and is always at the correct bias.

I haven't tried this idea with an output tube yet, but I have tried it with a pre-amplifier pentode and triode, and both work quite well with complete gain.

This isn't the idea that I wanted to patent, though I guess I could if noone else has used it yet. I think that someone somewhere has, though.

T.

:Here's a radical new one that I don't think I've seen yet. They say that you can't take the output from the cathode if you want gain. Here's why this is normally not possible: If you put the load resistor at the cathode, and tap the audio (or whatever) from the cathode, the changing load will also change the grid bias, if the grid is connected to B- in the conventional manner. If the tube starts conducting more, the cathode-grid difference will increase, and the grid will swing more negative (with respect to the cathode). This will cancel out all gain.
:
:HOWEVER! Here's the radical new idea! (Perhaps it's old, but I've never seen it before.) Let's say you're using a 12SQ7. Connect the plate directly to B+ instead of to a 1 meg resistor. Connect the 1 meg resistor to B- and the cathode instead. Tap the audio at the cathode. Now, instead of connecting the grid resistor to B-, connect it to the cathode as well. Normally, in radios, when the cathode is grounded, the grid resistor is grounded as well. Well, if you put a resistor between the cathode and B-, you must put the grid at the same potential as the cathode for this particular tube.
:
:Now, that is kind of exciting, but nothing fantastic. So you're taking the audio at the cathode instead of the plate. Big deal. Now do it to a pentode of any kind. Let's take a 50L6. It requires a 150 ohm cathode bias resistor. Connect this resistor to the cathode. To the other side connect one leg of the primary of the output transformer. Connect the other leg to B-. Where the transformer and the resistor join, connect the grid bias resistor. Connect both the screen grid and the plate to B+. Here's what's so wonderful about this technique: At the plate, some electrons get collected at the screen grid. The plate doesn't realize the full electron flow (this is inefficient amplification). Also, if the tube isn't a beam grid tube, the electrons bounce around. This ruins fidelity at the plate. However, at the cathode, ALL currents that flow through the tube must first flow in here. With the output transformer (or load resistor) at the cathode, full power and fidelity can be realized. There will be no wasted wattage, because the current for both the screen grid and the plate will flow through the output transformer if it is connected at the cathode as previously described.
:
:You would think that with the cathode changing all over the place, the grid would get confused, but its bias changes instantaneously with the cathode, and is always at the correct bias.
:
:I haven't tried this idea with an output tube yet, but I have tried it with a pre-amplifier pentode and triode, and both work quite well with complete gain.
:
:This isn't the idea that I wanted to patent, though I guess I could if noone else has used it yet. I think that someone somewhere has, though.
:
:T.

This is why I love posting here! These are great ideas and it gets your mind working!

:Here's a radical new one that I don't think I've seen yet.

Cathode follower/common cathode amplifier.

http://www.tubecad.com/2005/August/blog0053.htm

Z-

I've seen cathode follower before. My idea isn't cathode follower.....I guess it's similar, though, since the load resistor would be kind of the same as the cathode follower tube.

T.

Yeah....I'm looking at these schematics, and they're not what I'm talking about. They still show the grid being biased at conventional bias points and/or the audio output being derived at places on the cathode circuit other than what I said, which is why those circuits don't work for gain.

T.

:Yeah....I'm looking at these schematics, and they're not what I'm talking about. They still show the grid being biased at conventional bias points and/or the audio output being derived at places on the cathode circuit other than what I said, which is why those circuits don't work for gain.
:
:T.

Please provide a schematic. It's easier to understand with a schematic than with words.

Z-

I will. Also, if you read through it, you can draw it as you read it.

T.

Hi Thomas

Isn't this a standard cathode follower circuit? In this type of circuit there isn't any voltage gain. Instead you will have current gain. In addition there is no signal inversion like the more common plate output stage.

Norm

:Here's a radical new one that I don't think I've seen yet. They say that you can't take the output from the cathode if you want gain. Here's why this is normally not possible: If you put the load resistor at the cathode, and tap the audio (or whatever) from the cathode, the changing load will also change the grid bias, if the grid is connected to B- in the conventional manner. If the tube starts conducting more, the cathode-grid difference will increase, and the grid will swing more negative (with respect to the cathode). This will cancel out all gain.
:
:HOWEVER! Here's the radical new idea! (Perhaps it's old, but I've never seen it before.) Let's say you're using a 12SQ7. Connect the plate directly to B+ instead of to a 1 meg resistor. Connect the 1 meg resistor to B- and the cathode instead. Tap the audio at the cathode. Now, instead of connecting the grid resistor to B-, connect it to the cathode as well. Normally, in radios, when the cathode is grounded, the grid resistor is grounded as well. Well, if you put a resistor between the cathode and B-, you must put the grid at the same potential as the cathode for this particular tube.
:
:Now, that is kind of exciting, but nothing fantastic. So you're taking the audio at the cathode instead of the plate. Big deal. Now do it to a pentode of any kind. Let's take a 50L6. It requires a 150 ohm cathode bias resistor. Connect this resistor to the cathode. To the other side connect one leg of the primary of the output transformer. Connect the other leg to B-. Where the transformer and the resistor join, connect the grid bias resistor. Connect both the screen grid and the plate to B+. Here's what's so wonderful about this technique: At the plate, some electrons get collected at the screen grid. The plate doesn't realize the full electron flow (this is inefficient amplification). Also, if the tube isn't a beam grid tube, the electrons bounce around. This ruins fidelity at the plate. However, at the cathode, ALL currents that flow through the tube must first flow in here. With the output transformer (or load resistor) at the cathode, full power and fidelity can be realized. There will be no wasted wattage, because the current for both the screen grid and the plate will flow through the output transformer if it is connected at the cathode as previously described.
:
:You would think that with the cathode changing all over the place, the grid would get confused, but its bias changes instantaneously with the cathode, and is always at the correct bias.
:
:I haven't tried this idea with an output tube yet, but I have tried it with a pre-amplifier pentode and triode, and both work quite well with complete gain.
:
:This isn't the idea that I wanted to patent, though I guess I could if noone else has used it yet. I think that someone somewhere has, though.
:
:T.

Hmmm....well, in my circuit, each stage is isolated by the typical capacitor. I guess what I've done, though, is take a cathode follower system and over complicate it with capacitors. With the idea I came up with, though, tubes are not cascaded. The audio signal is linked in the conventional way. Also, I do get voltage gain with my set-up.

T.

:Hmmm....well, in my circuit, each stage is isolated by the typical capacitor. I guess what I've done, though, is take a cathode follower system and over complicate it with capacitors. With the idea I came up with, though, tubes are not cascaded. The audio signal is linked in the conventional way. Also, I do get voltage gain with my set-up.
:
:T.

Thomas, I agree with Norm in that this is a cathode follower also known as "common plate" amplifer in which the voltage gain cannot exceed mu over mu + 1. However, the power gain can be substantial AND you can artificially achieve a voltage gain by using a transformer in the cathode circuit as a load, then you can get any kind of phase reversal you please by simply changing your transformer arrangement. This would also allow you to cascade stages should you so desire.

Well, I was going to post my schematic, but I give up. I've tried two sites so far, and both of them are confusing. Whenever I post a picture, I either cannot locate it later, or it shows up in a small unchangable size.

T.

Take a look at this site:

http.//www.tubecad.com/2005/June/blog0048.htm

Have a great day

MRO

Sock it to me...via E-mail attachment and I'll do it fer ye.

But just ye remember.... a cathode follower is a cathode follower....IS A CATHODE FOLLOWER.

The tube amps builders" reinventing" it, resplendently
complete with its "audiofool" "transparency" vague descriptives.
http://www.valvediy.com/pnppg2.html

BIGGGGGG BUCKS fruitless patenting:
http://www.patentstorm.us/patents/5859565.html
73's de Edd

I'm not looking to patent it. It's just an idea I had that seems to work out well.

T.

Ok. I'm going to send you my schematic. It is different than the others, though. Here's my difference: I don't send the grid bias to B-. Sending the grid to B- with a big load resistor on the cathode is obviously going to cancel out all gain.

T.

Oh yeah! Maybe this might work. It's a bit small, so the text is hard to read, but you can still see some stuff.

http://www.metroflog.com/bigbadpiratetom

This seems to be to be like the non-inverting side of a phase inverter, only without the plate resistor in the inverting part. Cathode, grid resistor, grid, cathode resistor, ground. Output taken from the cathode to feed a grid of a push-pull output circuit.

Lewis L.

:Hi Thomas
:
: Isn't this a standard cathode follower circuit? In this type of circuit there isn't any voltage gain. Instead you will have current gain. In addition there is no signal inversion like the more common plate output stage.
:
:Norm
:
::Here's a radical new one that I don't think I've seen yet. They say that you can't take the output from the cathode if you want gain. Here's why this is normally not possible: If you put the load resistor at the cathode, and tap the audio (or whatever) from the cathode, the changing load will also change the grid bias, if the grid is connected to B- in the conventional manner. If the tube starts conducting more, the cathode-grid difference will increase, and the grid will swing more negative (with respect to the cathode). This will cancel out all gain.
::
::HOWEVER! Here's the radical new idea! (Perhaps it's old, but I've never seen it before.) Let's say you're using a 12SQ7. Connect the plate directly to B+ instead of to a 1 meg resistor. Connect the 1 meg resistor to B- and the cathode instead. Tap the audio at the cathode. Now, instead of connecting the grid resistor to B-, connect it to the cathode as well. Normally, in radios, when the cathode is grounded, the grid resistor is grounded as well. Well, if you put a resistor between the cathode and B-, you must put the grid at the same potential as the cathode for this particular tube.
::
::Now, that is kind of exciting, but nothing fantastic. So you're taking the audio at the cathode instead of the plate. Big deal. Now do it to a pentode of any kind. Let's take a 50L6. It requires a 150 ohm cathode bias resistor. Connect this resistor to the cathode. To the other side connect one leg of the primary of the output transformer. Connect the other leg to B-. Where the transformer and the resistor join, connect the grid bias resistor. Connect both the screen grid and the plate to B+. Here's what's so wonderful about this technique: At the plate, some electrons get collected at the screen grid. The plate doesn't realize the full electron flow (this is inefficient amplification). Also, if the tube isn't a beam grid tube, the electrons bounce around. This ruins fidelity at the plate. However, at the cathode, ALL currents that flow through the tube must first flow in here. With the output transformer (or load resistor) at the cathode, full power and fidelity can be realized. There will be no wasted wattage, because the current for both the screen grid and the plate will flow through the output transformer if it is connected at the cathode as previously described.
::
::You would think that with the cathode changing all over the place, the grid would get confused, but its bias changes instantaneously with the cathode, and is always at the correct bias.
::
::I haven't tried this idea with an output tube yet, but I have tried it with a pre-amplifier pentode and triode, and both work quite well with complete gain.
::
::This isn't the idea that I wanted to patent, though I guess I could if noone else has used it yet. I think that someone somewhere has, though.
::
::T.

This seems to be to be like the non-inverting side of a phase inverter, only without the plate resistor in the inverting part. Cathode, grid resistor, grid, cathode resistor, ground. Output taken from the cathode to feed a grid of a push-pull output circuit.

Lewis L.

:Hi Thomas
:
: Isn't this a standard cathode follower circuit? In this type of circuit there isn't any voltage gain. Instead you will have current gain. In addition there is no signal inversion like the more common plate output stage.
:
:Norm
:
::Here's a radical new one that I don't think I've seen yet. They say that you can't take the output from the cathode if you want gain. Here's why this is normally not possible: If you put the load resistor at the cathode, and tap the audio (or whatever) from the cathode, the changing load will also change the grid bias, if the grid is connected to B- in the conventional manner. If the tube starts conducting more, the cathode-grid difference will increase, and the grid will swing more negative (with respect to the cathode). This will cancel out all gain.
::
::HOWEVER! Here's the radical new idea! (Perhaps it's old, but I've never seen it before.) Let's say you're using a 12SQ7. Connect the plate directly to B+ instead of to a 1 meg resistor. Connect the 1 meg resistor to B- and the cathode instead. Tap the audio at the cathode. Now, instead of connecting the grid resistor to B-, connect it to the cathode as well. Normally, in radios, when the cathode is grounded, the grid resistor is grounded as well. Well, if you put a resistor between the cathode and B-, you must put the grid at the same potential as the cathode for this particular tube.
::
::Now, that is kind of exciting, but nothing fantastic. So you're taking the audio at the cathode instead of the plate. Big deal. Now do it to a pentode of any kind. Let's take a 50L6. It requires a 150 ohm cathode bias resistor. Connect this resistor to the cathode. To the other side connect one leg of the primary of the output transformer. Connect the other leg to B-. Where the transformer and the resistor join, connect the grid bias resistor. Connect both the screen grid and the plate to B+. Here's what's so wonderful about this technique: At the plate, some electrons get collected at the screen grid. The plate doesn't realize the full electron flow (this is inefficient amplification). Also, if the tube isn't a beam grid tube, the electrons bounce around. This ruins fidelity at the plate. However, at the cathode, ALL currents that flow through the tube must first flow in here. With the output transformer (or load resistor) at the cathode, full power and fidelity can be realized. There will be no wasted wattage, because the current for both the screen grid and the plate will flow through the output transformer if it is connected at the cathode as previously described.
::
::You would think that with the cathode changing all over the place, the grid would get confused, but its bias changes instantaneously with the cathode, and is always at the correct bias.
::
::I haven't tried this idea with an output tube yet, but I have tried it with a pre-amplifier pentode and triode, and both work quite well with complete gain.
::
::This isn't the idea that I wanted to patent, though I guess I could if noone else has used it yet. I think that someone somewhere has, though.
::
::T.

Enlarged Schematic referincing:
http://img441.imageshack.us/img441/893/tdscathodeoutputjpegmp3.png

Aside observations:

With the first amp stage you are configuring as a cathode follower ...e.g. current amplifier stage , however, the abnormally high cathode resistor value sugggests of convential voltage amplifier logic.

The final output circuit stages effective power level is not aided by that parasitic 150 ohm resistor . Since the power is then proportionatively divided between it and the O/P xformer.

Screen current consumption is typically minumal, in its trade off benefits of its static accelerating potential of and amassing the electron beam stream .

73's de Edd

The 150 ohm resistor is necessary for biasing the grid with respect for the cathode. If you connect the grid to B-, it'll be too negative, since the output transformer's primary is a lot more than 150 ohms. This is the problem with most circuits that derive output from the cathode. They still put the grid at B- potential, and forget that the huge load resistor in the cathode circuit is going to severely affect the bias, and will add severe negative feedback. One must remember to bias the grid with respect to the cathode, no matter where the cathode may be.

T.

Since you have the circuit built up already, whereas others would have to dig up parts ,
tear down or recreate that circuit up from scratch how about your evaluation using :

Hook up your initial circuit and use an audio source to confirm that circuits definitive potential max output level.

Now, agreed on the bottleneck 150 ohm series resistor, but why not rethink your manner of the acquisition of that biasing voltage.
With your present selection of 470 k for that grid resistor, the max required currrent draw that might ever be permitted through that path would be ~ 450 ua worst case.
Try eliminating the 150 ohm and connecting the O/P direct to the cathode and ...considering that you want to keep within the same low Z that was initially presented by that leg of the circuitry... install a 150 ohm (or upwards in value) pot across the O/P xfmr and take your then acquired, adjustable voltage off its CT to feed to the grid resistors low end.

There possibly might be some unknown transitional threshold within its adjustment range that
might initiate a shift from mild negative feed back towards positive feedback, so be prepared
for additive mild capacitive filtering of that derived bias voltage as referenced to either / both cathode and / or ground.

Re-evaluation for power outtut and performance ?

73's de Edd

You could do that, too, or have a transformer with a 150 ohm tap. The 150 ohm resistor as I have it wired, though, is in a conventional 50L6 circuit anyway, so I don't see it as parasitic. True, though, that you can gain a bit more by omitting it, and placing a tap or pot. in in the output primary.

Thomas

:You could do that, too, or have a transformer with a 150 ohm tap. The 150 ohm resistor as I have it wired, though, is in a conventional 50L6 circuit anyway, so I don't see it as parasitic. True, though, that you can gain a bit more by omitting it, and placing a tap or pot. in in the output primary.
:
:Thomas

You'd need 150 Ohms DC for the grid bias, and that would be how many Ohms reactance to the audio?

Lewis L.