thanks. i'm definitely a bit rusty.
i also just caught up on this 6CB5 craze that seems to have swept the internet in the past few years. if i can get my hands on some, these monoblocks should be a decent platform to try them out down the line. they'll certainly crank out more triode power than the 3d21A's i've designed for. and at 6.6K a-a my OPT's seem like they might just be up to the task. i'll post pics as the build progresses but they're around the same size as 60 watt Edcors. another case of building an amplifier around the tubes backfiring. or in this case it's just lucky i'm overbuilding things so much.
thing is though, i guess after the dust settled after that craze, they're not so easy to find any more? seeing a few isolated singles and pairs on eBay and a few relatively expensive Russian equivalents.
i also just caught up on this 6CB5 craze that seems to have swept the internet in the past few years. if i can get my hands on some, these monoblocks should be a decent platform to try them out down the line. they'll certainly crank out more triode power than the 3d21A's i've designed for. and at 6.6K a-a my OPT's seem like they might just be up to the task. i'll post pics as the build progresses but they're around the same size as 60 watt Edcors. another case of building an amplifier around the tubes backfiring. or in this case it's just lucky i'm overbuilding things so much.
thing is though, i guess after the dust settled after that craze, they're not so easy to find any more? seeing a few isolated singles and pairs on eBay and a few relatively expensive Russian equivalents.
I might have had something to do with that craze. I had remembered the 6CB5 from high school electronics class in 1967 - 1970. They were the biggest and baddest sweep tube available in zero cost suppliers (trash dump, old TV's, etc). Their memories had been stuffed into a dark corner of my brain until AES decided to start exiting the NOS tube business by having a big tube sale. I got some 6CB5's ant "tested" them resulting in their entire stash of 6CB5's and other tubes discussed in this thread vanishing during the sale. This was 9 years ago.
http://www.diyaudio.com/forums/tubes-valves/128533-tube-sale-aes.html
I have seen over 100 watts flow from a pair of 6CB5's in conventional push pull pentode operation. 50+ watts into a 6.6K OPT would be easy in pentode, but I'm not sure they would eat enough voltage in triode without blowing the screen grid to get near 50 watts.
http://www.diyaudio.com/forums/tubes-valves/128533-tube-sale-aes.html
I have seen over 100 watts flow from a pair of 6CB5's in conventional push pull pentode operation. 50+ watts into a 6.6K OPT would be easy in pentode, but I'm not sure they would eat enough voltage in triode without blowing the screen grid to get near 50 watts.
That's a crazy story. Things like the AES sale don't happen all that often and I don't begrudge all the happy builders their tubes.
In my implementation they'll be loafing along at around 18 watts dissipation (370V 50mA) which I hope is a good operating point... the thing is, I don't think my driver stage will give enough swing any more, I had been counting on about 30V out of those 6BZ7's. I might DC couple in a trioded 6J6 to the top of the LTP like I did in my 6L6GC amp.
I'm guesstimating from the partial curves on this page
http://www.bartola.co.uk/valves/tag/6cb5a-triode-mode/
With 10M45S' giving constant current I guess I can always just take a look at what voltage the cathodes are sitting at, but by the time it gets to that it might be a little late to rebuild the driver section
Last edited:
here's what the layout is looking like, using the old 3D21A tubes as stand-ins until the 6P20S' make their way over from the former Eastern bloc.
that's a 12.5uF can capacitor i'll use after the 5AR4. the wee tube in front is a 6J6 i'll use with paralleled plates to direct couple to the 6BZ7 LTP driver.
PT is from the awesome Jan Wüsten at AskJanFirst. Choke is a potted Chicago model, 8H 85 ohm. OPT says "A-405" on it and is supposedly 6.6K a-a.
that's a 12.5uF can capacitor i'll use after the 5AR4. the wee tube in front is a 6J6 i'll use with paralleled plates to direct couple to the 6BZ7 LTP driver.
PT is from the awesome Jan Wüsten at AskJanFirst. Choke is a potted Chicago model, 8H 85 ohm. OPT says "A-405" on it and is supposedly 6.6K a-a.
Last edited:
plagiarizing from my previous designs, and by extension, all the great people who have basically told me how to do everything on these forums ,)
i have a couple questions. first one is: is it kosher to run cathode feedback here even though i'm using a CCS? is it weird to have the OPT only grounded through that 1M resistor? if not I'll give up on the CFB idea but since i generally avoid global feedback this seemed like a nice little feature.
second main question: i completely forget how to figure out plate resistors in an LTP with a CCS in the tail. can anyone tell me how that goes again? my desired operating point is as follows:
or do i just calculate the 130V voltage drop at 3mA to get to that 240V operating point? so 43K?
i have a couple questions. first one is: is it kosher to run cathode feedback here even though i'm using a CCS? is it weird to have the OPT only grounded through that 1M resistor? if not I'll give up on the CFB idea but since i generally avoid global feedback this seemed like a nice little feature.
second main question: i completely forget how to figure out plate resistors in an LTP with a CCS in the tail. can anyone tell me how that goes again? my desired operating point is as follows:
or do i just calculate the 130V voltage drop at 3mA to get to that 240V operating point? so 43K?
Last edited:
i'm pretty sure, yeah. that's why i designed a new frontend
i'm just not 100% sure it's a viable one, hence all the questions...Hint and Caution:
DO Ground the secondary 4 Ohm tap. You must provide a return for the DC current of both cathodes. A 1 Meg Ohm resistor will not work.
Now, you do have made it more difficult to use a scope to test the amp output.
But it is OK, this is how to test the amplifier output:
Connect a 4 Ohm power resistor (or two 8 Ohm power resistors in parallel) to the grounded secondary 4 Ohm tap.
Connect the other end of the 4 Ohm power resistor (or the other end of the two 8 Ohm power resistors in parallel) to the secondary 16 Ohm tap.
(it is actually 4 Ohms of 'secondary turns' from the secondary 4 Ohm tap to the secondary 16 Ohm tap). That is why cathode feedback is from 'Common' to '16', with the 4 Ohm tap grounded, that tap is in the middle of the total secondary turns.
Connect the scope probe ground to the grounded secondary 4 Ohm tap,
and the scope probe tip to the secondary 16 Ohm tap.
The test will be for a 4 Ohm load.
The voltage across 4 Ohms worth of secondary winding will be 0.7 times the voltage across 8 Ohms worth of secondary winding.
Just remember that P = Squared/R, and R is now 4 Ohms.
When you connect an 8 Ohm load across 8 Ohms worth of secondary windings,
it will have 1.4 times the voltage you got across the 4 Ohm resistor across 4 Ohms worth of secondary windings. But that will be the same wattage for both.
Ready for listening?
Connect the 8 Ohm speaker (-) to the secondary Com tap.
Connect the 8 Ohm speaker (+) to the secondary 8 Ohm tap.
Listen to the amp, and let us know how it sounds.
DO Ground the secondary 4 Ohm tap. You must provide a return for the DC current of both cathodes. A 1 Meg Ohm resistor will not work.
Now, you do have made it more difficult to use a scope to test the amp output.
But it is OK, this is how to test the amplifier output:
Connect a 4 Ohm power resistor (or two 8 Ohm power resistors in parallel) to the grounded secondary 4 Ohm tap.
Connect the other end of the 4 Ohm power resistor (or the other end of the two 8 Ohm power resistors in parallel) to the secondary 16 Ohm tap.
(it is actually 4 Ohms of 'secondary turns' from the secondary 4 Ohm tap to the secondary 16 Ohm tap). That is why cathode feedback is from 'Common' to '16', with the 4 Ohm tap grounded, that tap is in the middle of the total secondary turns.
Connect the scope probe ground to the grounded secondary 4 Ohm tap,
and the scope probe tip to the secondary 16 Ohm tap.
The test will be for a 4 Ohm load.
The voltage across 4 Ohms worth of secondary winding will be 0.7 times the voltage across 8 Ohms worth of secondary winding.
Just remember that P = Squared/R, and R is now 4 Ohms.
When you connect an 8 Ohm load across 8 Ohms worth of secondary windings,
it will have 1.4 times the voltage you got across the 4 Ohm resistor across 4 Ohms worth of secondary windings. But that will be the same wattage for both.
Ready for listening?
Connect the 8 Ohm speaker (-) to the secondary Com tap.
Connect the 8 Ohm speaker (+) to the secondary 8 Ohm tap.
Listen to the amp, and let us know how it sounds.
What would the maximum voltage for triode connection of the 6CM5?
I repaired a commercial amplifier using using 807's running at 600 volts as triodes 36 ma for a claimed 50 watts and built an triode 807 amp using 560 volts the tubes never needed replacing; I don't think 6CM5s would handle anything like that on the Screen grid.
Phil
I repaired a commercial amplifier using using 807's running at 600 volts as triodes 36 ma for a claimed 50 watts and built an triode 807 amp using 560 volts the tubes never needed replacing; I don't think 6CM5s would handle anything like that on the Screen grid.
Phil
Yes, you could keep DC off the secondary, like was suggested above.
One thing to remember is what tube you use. Some of those only have a u = 3.8.
Triode mode, plus cathode feedback is going to require a lot of drive at the tube grid; and there will not be much swing at the plate/screen connection, unless the bias is quite negative. I see 6CB5, 6CM5, etc. mentioned.
u = 3.8, Hmm, anybody for 300B (u = 3.85)
Most of these pentodes were made for very large swings on the plates, and no swing on the screen grids (pentode mode).
The plate connection at the top allowed for very high voltage (but at zero current).
These drove flyback transformers.
That is actually similar to some RF tubes in class C mode (i.e. the 807).
The plate tank was resonant, and acted like a flywheel.
That is unlike a class A single ended; or a class AB or class B push pull amp.
Resonant tank in an audio amp and music do not go together very well.
Using them in triode mode with very high voltages could be tried, but caution is the word.
Using a 100 Ohm resistor from plate to screen would allow easy measurement of screen current with a DMM. The screen current times the screen voltage = screen watts should never be exceeded.
For plates at least, there can be an effect of current crowding that happens at high voltages. You can sometimes see a small spot of the plate glowing, where the crowding occurs. Of course these tubes are not made for that.
I wonder if the screen can suffer from a similar current crowding. But you won't see it glow, it may just melt and short.
One thing to remember is what tube you use. Some of those only have a u = 3.8.
Triode mode, plus cathode feedback is going to require a lot of drive at the tube grid; and there will not be much swing at the plate/screen connection, unless the bias is quite negative. I see 6CB5, 6CM5, etc. mentioned.
u = 3.8, Hmm, anybody for 300B (u = 3.85)
Most of these pentodes were made for very large swings on the plates, and no swing on the screen grids (pentode mode).
The plate connection at the top allowed for very high voltage (but at zero current).
These drove flyback transformers.
That is actually similar to some RF tubes in class C mode (i.e. the 807).
The plate tank was resonant, and acted like a flywheel.
That is unlike a class A single ended; or a class AB or class B push pull amp.
Resonant tank in an audio amp and music do not go together very well.
Using them in triode mode with very high voltages could be tried, but caution is the word.
Using a 100 Ohm resistor from plate to screen would allow easy measurement of screen current with a DMM. The screen current times the screen voltage = screen watts should never be exceeded.
For plates at least, there can be an effect of current crowding that happens at high voltages. You can sometimes see a small spot of the plate glowing, where the crowding occurs. Of course these tubes are not made for that.
I wonder if the screen can suffer from a similar current crowding. But you won't see it glow, it may just melt and short.
A watchout for using the secondary for cathode feedback...
I recently tried this with some old Partridge transformers. They have dual 4 ohm secondaries that I connected in series with the center tap to ground, and I thought they would be reasonably symmetrical. However when I tested, there was a low level oscillation at about 50 kHz (not caused by incorrect feedback polarity) and I needed zobel networks from each end of the secondary to ground to get rid of it. With global feedback applied I had problems with stability that I could mostly compensate, but square wave response was still a bit funny looking. I ended up not using the secondaries for cathode feedback.
I think Tubelab mentioned something similar in some old threads, success was a bit hit and miss. Bear in mind that this was with high gm 7591's in pentode mode, with fixed bias. Triode mode should be more forgiving. Worth giving it a go as there are definitely gains to be made if it works.
I recently tried this with some old Partridge transformers. They have dual 4 ohm secondaries that I connected in series with the center tap to ground, and I thought they would be reasonably symmetrical. However when I tested, there was a low level oscillation at about 50 kHz (not caused by incorrect feedback polarity) and I needed zobel networks from each end of the secondary to ground to get rid of it. With global feedback applied I had problems with stability that I could mostly compensate, but square wave response was still a bit funny looking. I ended up not using the secondaries for cathode feedback.
I think Tubelab mentioned something similar in some old threads, success was a bit hit and miss. Bear in mind that this was with high gm 7591's in pentode mode, with fixed bias. Triode mode should be more forgiving. Worth giving it a go as there are definitely gains to be made if it works.
that's really interesting. what kind of value would work here? i guess it has to be large if we're talking electrolytics.
as for not providing a return to ground for the cathodes: from what i've been told, since these are CCS' in pure class A, the net current going between the junction of both cathodes and ground should be 0. everything just slides back and forth between one cathode capacitor and the other, from what i understand.
tikiroo,
What may have happened on the Partridge case you site (and other cases with a similar problem) is the leakage reactance of the two halves of the secondaries. That does cause a phase shift that increases with frequency. The transformer frequency response is also starting to drop off there too.
By putting zebecs across them, the phase and amplitude response is changed. It seemed to terminate the 50kHz resonance or at least significantly change the phase enough that it no longer was a problem for the cathode feedback mode.
What may have happened on the Partridge case you site (and other cases with a similar problem) is the leakage reactance of the two halves of the secondaries. That does cause a phase shift that increases with frequency. The transformer frequency response is also starting to drop off there too.
By putting zebecs across them, the phase and amplitude response is changed. It seemed to terminate the 50kHz resonance or at least significantly change the phase enough that it no longer was a problem for the cathode feedback mode.
sorenj07,
Suppose you have 2 35mA current sources. That is 70mA.
Now in push pull with AC cathode coupling, suppose the current in one tube goes to 45mA, while the other tube current goes to 25mA. That is still 70mA, and is what we expect, they just traded current.
But ... the current has to come from somewhere. B+ is referred to ground. So it's return IS ground. But the cathodes and current sources have no return to ground in the schematic.
I hope that explains it. You need to ground either the 4 Ohm tap, or you need to ground the current sources.
By the way, if the primary and secondary ever short to each other, and the secondary is not grounded, you become the return path for B+ when you grab the speaker wires!
I have never had a primary to secondary short. But I did have a Brand New Unpowered output transformer that had a primary to Laminations short. I returned it to the vendor.
Just in case you ever have a primary to LAMs short, it is always a good idea to be sure that you penetrate the paint on the end bells and return to ground. The easy way to do this is be sure to use Lockwashers that bite into the end bell tab where you attach them to the chassis.
I use nuts that have the lock washer captivated.
(screw, chassis, end bell tab, lock washer/nut).
Suppose you have 2 35mA current sources. That is 70mA.
Now in push pull with AC cathode coupling, suppose the current in one tube goes to 45mA, while the other tube current goes to 25mA. That is still 70mA, and is what we expect, they just traded current.
But ... the current has to come from somewhere. B+ is referred to ground. So it's return IS ground. But the cathodes and current sources have no return to ground in the schematic.
I hope that explains it. You need to ground either the 4 Ohm tap, or you need to ground the current sources.
By the way, if the primary and secondary ever short to each other, and the secondary is not grounded, you become the return path for B+ when you grab the speaker wires!
I have never had a primary to secondary short. But I did have a Brand New Unpowered output transformer that had a primary to Laminations short. I returned it to the vendor.
Just in case you ever have a primary to LAMs short, it is always a good idea to be sure that you penetrate the paint on the end bells and return to ground. The easy way to do this is be sure to use Lockwashers that bite into the end bell tab where you attach them to the chassis.
I use nuts that have the lock washer captivated.
(screw, chassis, end bell tab, lock washer/nut).
Yes the electrolytics should be large, several hundred uF or more at a guess depending on the impedance at the cathode. I'd start with 220uF or thereabouts.
I am with PRR on the need to ground the secondary for safety, and I have unfortunately had a primary to secondary short in an output transformer and it would, umm, have been an electrifying experience were the transformer secondary not grounded.. (Sorry )
I have also found that cathode feedback is not universally applicable even with quite good transformers and depends heavily on the quality and design of the output transformer.
I am with PRR on the need to ground the secondary for safety, and I have unfortunately had a primary to secondary short in an output transformer and it would, umm, have been an electrifying experience were the transformer secondary not grounded.. (Sorry
)I have also found that cathode feedback is not universally applicable even with quite good transformers and depends heavily on the quality and design of the output transformer.
I have seen CFB clean up some cheap OPT's and squeeze a bit more power and bandwidth out of them, and I have seen it make things worse.
I have a bunch of push pull OPT's that were made for guitar amps with cost as the only factor. Zero interleaving, just one half primary, then the entire 0-4-8-16 secondary, then the other half primary. They work far better than the spec sheets indicate (80VA from 80 Hz to 4 KHz) but I have tried CFB in several amp designs, always with the same results. The distortion and DF improve in the 300 Hz to 5 KHz range and gets worse everywhere else. It seems that these transformers are not symmetrical enough at the frequency extremes.
I have found that these same transformers work quite well and sound excellent when driven by a low impedance source like 300B's or sweep tubes with plate to grid feedback applied. They also sound pretty bland when fed by wimpy tubes like EL84's without feedback.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.