About the theory that I've made a single ended amp: I measured the output power, it's around 50W. I don't think one 6550 can pull that off alone 😉
That was my initial idea, to use a CCS. Because of some stupid mistake I won't mention here I couldn't get that to work and went for the cathode resistor instead. And then the hum got too annoying.
I'll give it another try soon.
Before going fully regulated B+ I had another go at that CCS. I'm using an LM317 for now, I know it's not perfect but it's an easy start. Applied as in the schematic here.
The trouble here is the same as earlier trials: it works.
A couple of times. When I switch on the amp for the third or fourth time it is toast. This time only 8 mA came out, voltage across the set resistor according. I've done a lot of searching but I can't find much about failure in that application. So I mention it here.
I don't bother with the standby switch and pre heating tubes, just smack on everything. Could that be the culprit?
Lots of stuff on the web about using LM317 as a CCS, and it seems to work troubl;e free there.
The trouble here is the same as earlier trials: it works.
A couple of times. When I switch on the amp for the third or fourth time it is toast. This time only 8 mA came out, voltage across the set resistor according. I've done a lot of searching but I can't find much about failure in that application. So I mention it here.
I don't bother with the standby switch and pre heating tubes, just smack on everything. Could that be the culprit?
Lots of stuff on the web about using LM317 as a CCS, and it seems to work troubl;e free there.
#1 post. The lower valve has a large amount of negative feedback. This will try and keep the plate of the lower valve at a constant voltage. This means that any swing the top valve generates will be inhibited by the lower through the coupling of the OPT. Power will be transferred from the top valve and dumped in the bottom rather than going to the speaker. At the same time any ripple on the HT will appear across the lower winding so will end up in the speaker.
I think you want your lower valve grid straight to ground and a beefy CCS between the cathodes.
I think you want your lower valve grid straight to ground and a beefy CCS between the cathodes.
Last edited:
I got rid of that feedback yesterday as posted. The CCS is now the bottleneck, LM317 fail after a few on/off cycles.
Any garden variety transistor or garden variety IC can take a transient . . . Once.
Get rid of that LM317 or LM317T this way:
(And yes, I have used both the LM317 and LM317T).
Here is a quick and easy one to try . . .
6550 are high transconductance tubes. They can work well with a less than perfect CCS, to create a self inverting output stage.
Let's use one of the simplest "CCS" there is, a common "power supply" choke:
Remove the Schade-Like negative feedback resistor that is on the bottom 6550 (*** See the note below).
Replace the 120uF and 100 Ohm cathode self bias circuit. Use a 5 Henry 200 mA choke (65 Ohms DCR) in series with a 35 Ohm resistor.
Connect the top of the series choke and 35 Ohm resistor to the junction of the two 10 Ohm cathode resistors, and the other end of that less than perfect "CCS" to ground.
On the bottom 6550, bypass the 220k g1 resistor with the 47nF cap (connect across the 220k resistor).
You should be able to use a 5H to 20H choke, as long as it is rated for 200mA or more. And make the sum of the choke DCR plus the series resistor be equal to 100 Ohm. So if the choke DCR is 40 Ohms, use a 60 Ohm resistor in series with the choke.
In case you did not guess, I have used the Hammond 5H 200mA 65 Ohm DCR choke and series resistor in a self inverting Beam Power Tube amplifier.
I call that series circuit, a less than perfect CCS (Constant Current Sink).
*** Note: I have only used that CCS choke for self inverting amplifiers when the output tubes are either triode wired (without Schade negative feedback); Or connecting the screens onto the Ultra Linear Taps (and without Schade negative feedback, the only negative feedback(s) are the UL operation, and the cathode to cathode of the output tubes).
You really do need the Schade negative feedback if you wire the output tubes in either Beam Power mode, or Pentode mode (depending on the tube type, beam power or pentode).
Simple, elegant, works.
Have fun modifying, building, and listening!
And, 4V of ripple is far too much for B+ (just my opinion).
Connect a B+ with that much ripple to a Schade feedback resistor, and see what happens.
My self inverting amplifiers have less than 100uV ripple at the output transformer secondary.
Get rid of that LM317 or LM317T this way:
(And yes, I have used both the LM317 and LM317T).
Here is a quick and easy one to try . . .
6550 are high transconductance tubes. They can work well with a less than perfect CCS, to create a self inverting output stage.
Let's use one of the simplest "CCS" there is, a common "power supply" choke:
Remove the Schade-Like negative feedback resistor that is on the bottom 6550 (*** See the note below).
Replace the 120uF and 100 Ohm cathode self bias circuit. Use a 5 Henry 200 mA choke (65 Ohms DCR) in series with a 35 Ohm resistor.
Connect the top of the series choke and 35 Ohm resistor to the junction of the two 10 Ohm cathode resistors, and the other end of that less than perfect "CCS" to ground.
On the bottom 6550, bypass the 220k g1 resistor with the 47nF cap (connect across the 220k resistor).
You should be able to use a 5H to 20H choke, as long as it is rated for 200mA or more. And make the sum of the choke DCR plus the series resistor be equal to 100 Ohm. So if the choke DCR is 40 Ohms, use a 60 Ohm resistor in series with the choke.
In case you did not guess, I have used the Hammond 5H 200mA 65 Ohm DCR choke and series resistor in a self inverting Beam Power Tube amplifier.
I call that series circuit, a less than perfect CCS (Constant Current Sink).
*** Note: I have only used that CCS choke for self inverting amplifiers when the output tubes are either triode wired (without Schade negative feedback); Or connecting the screens onto the Ultra Linear Taps (and without Schade negative feedback, the only negative feedback(s) are the UL operation, and the cathode to cathode of the output tubes).
You really do need the Schade negative feedback if you wire the output tubes in either Beam Power mode, or Pentode mode (depending on the tube type, beam power or pentode).
Simple, elegant, works.
Have fun modifying, building, and listening!
And, 4V of ripple is far too much for B+ (just my opinion).
Connect a B+ with that much ripple to a Schade feedback resistor, and see what happens.
My self inverting amplifiers have less than 100uV ripple at the output transformer secondary.
Last edited:
Did you try to short circuit the grid of the lower tube to ground?
As mentioned the lower tube does not need any feedback as it only serves as a current balance to be able to use a non gapped transformer.
As mentioned the lower tube does not need any feedback as it only serves as a current balance to be able to use a non gapped transformer.
I've got a few of those, I'll look into that. I'm real new to designing current sinks...
Time for an update.
I fed everything from the regulated supply. Ripple on that is a few mV.
I tried 6A3sUMMER's suggestion to use a choke on the cathodes. Worked alright, but still too much hum to my liking.
I'm going back to a LTP phase splitter and separate cathode resistors and caps on the output stage.
Thanks for all the help guys.
I fed everything from the regulated supply. Ripple on that is a few mV.
I tried 6A3sUMMER's suggestion to use a choke on the cathodes. Worked alright, but still too much hum to my liking.
I'm going back to a LTP phase splitter and separate cathode resistors and caps on the output stage.
Thanks for all the help guys.
1. For the schematic in Post # 1:
The input pentode has an Un-bypassed 250 Ohm cathode self bias resistor.
A pentode has lots of gain.
If the 6.3VAC that powers the pentode filament is not properly grounded, then any leakage of the filament to cathode will cause hum in the amplifier that will be amplified by the input tube, and the output tubes, all the way to the output transformer secondary.
Properly ground the 6.3VAC:
Either use a Center Tapped 6.3VAC secondary, Or connect each end of the 6.3VAC secondary through a 100 Ohm resistor to ground (2 secondary ends, 2 resistors, one for each winding end; this forms a Pseudo center tap).
2. For the schematic in Post # 23:
The input stage's top triode cathode can pick up any hum from the filament to cathode leakage. Some tubes have this insulation broken down, and leaky (like the 100k DC resistance on the 7199 tubes I used to use, over time that insulation went bad).
The solution was to use a resistive divider. Two resistors in series from B+, and divide B+ to about +90V; the bottom resistor to ground has to be bypassed by a capacitor.
Then, the junction of the divider resistors and bypass cap, is connected to the 6.3VAC filament secondary (Either to the Center Tap, or the Pseudo Center Tap of two 100 Ohm resistors, each connected to an end of the 6.3VAc, and the junction of the 100 Ohm resistors to the elevation voltage from the B+ divider circuit.
3. Be careful when using a Choke as a CCS. It will pick up any magnetic fields. Examples: the the B+ filter choke, and the power transformer.
Good luck Beating the hum out!
The input pentode has an Un-bypassed 250 Ohm cathode self bias resistor.
A pentode has lots of gain.
If the 6.3VAC that powers the pentode filament is not properly grounded, then any leakage of the filament to cathode will cause hum in the amplifier that will be amplified by the input tube, and the output tubes, all the way to the output transformer secondary.
Properly ground the 6.3VAC:
Either use a Center Tapped 6.3VAC secondary, Or connect each end of the 6.3VAC secondary through a 100 Ohm resistor to ground (2 secondary ends, 2 resistors, one for each winding end; this forms a Pseudo center tap).
2. For the schematic in Post # 23:
The input stage's top triode cathode can pick up any hum from the filament to cathode leakage. Some tubes have this insulation broken down, and leaky (like the 100k DC resistance on the 7199 tubes I used to use, over time that insulation went bad).
The solution was to use a resistive divider. Two resistors in series from B+, and divide B+ to about +90V; the bottom resistor to ground has to be bypassed by a capacitor.
Then, the junction of the divider resistors and bypass cap, is connected to the 6.3VAC filament secondary (Either to the Center Tap, or the Pseudo Center Tap of two 100 Ohm resistors, each connected to an end of the 6.3VAc, and the junction of the 100 Ohm resistors to the elevation voltage from the B+ divider circuit.
3. Be careful when using a Choke as a CCS. It will pick up any magnetic fields. Examples: the the B+ filter choke, and the power transformer.
Good luck Beating the hum out!
Last edited:
@Miniwatt
You may already have set on a compromise that works for you, but thought I would add this article from TubeCad regarding cathode biasing, there Broskie is offering interesting ideas on both CCS and CCS-hybrids.
https://www.tubecad.com/2011/08/blog0210.htm
Also should point out that using a true CCS only works with LTP's running in class A only, so if the LTP is running in class AB it will not work as soon as one of the halves enters class B region it will demand for more current than the CCS can provide, the article above covers shortly on it too.
ps. I ought to be humble and backpedal a bit on the separate cathode resistors, was in a different thinking mode when thinking through the self-splitter, it's still an LTP and no matter whether tubes or silicon devices there are no perfect pairs, good luck!
You may already have set on a compromise that works for you, but thought I would add this article from TubeCad regarding cathode biasing, there Broskie is offering interesting ideas on both CCS and CCS-hybrids.
https://www.tubecad.com/2011/08/blog0210.htm
Also should point out that using a true CCS only works with LTP's running in class A only, so if the LTP is running in class AB it will not work as soon as one of the halves enters class B region it will demand for more current than the CCS can provide, the article above covers shortly on it too.
ps. I ought to be humble and backpedal a bit on the separate cathode resistors, was in a different thinking mode when thinking through the self-splitter, it's still an LTP and no matter whether tubes or silicon devices there are no perfect pairs, good luck!
A "phase splitter" is about the easiest and most linear stage possible. All of this Sturm und Drang seems very misplaced in context. Beware internet Lore, eyes on the prize.
All good fortune,
Chris