Hi all,
I am currently pulling together ideas for a two stage Push Pull 6080 amp. I intend to have the transformer in the cathodes so I need a fair bit of gain. I have this perverse desire to build it with two valves per channel. I originally thought of using the ECC91 which is a B7A valve with shared cathode. Unfortunately not enough gain to get the 60V PP I need. I was thinking is there a dual pentode with just a shared cathode connection which could serve as a LTP. I have found plenty of strange valves with common g1 and g2 but with access to the g3, and one or two compactrons, but I was hoping for a B9A valve.
Any ideas.
Shoog
I am currently pulling together ideas for a two stage Push Pull 6080 amp. I intend to have the transformer in the cathodes so I need a fair bit of gain. I have this perverse desire to build it with two valves per channel. I originally thought of using the ECC91 which is a B7A valve with shared cathode. Unfortunately not enough gain to get the 60V PP I need. I was thinking is there a dual pentode with just a shared cathode connection which could serve as a LTP. I have found plenty of strange valves with common g1 and g2 but with access to the g3, and one or two compactrons, but I was hoping for a B9A valve.
Any ideas.
Shoog
Shoog,
I used the 6BN11 compactron dual pentode in my 6as7 amp. It does a pretty good job and is probably the sexiest small tube I've yet to see. The only thing is that you have to expect mis-match between halves. Where-as separated pentodes would allow for you to tube roll one side to get a match. If the cathodes and screens are tied together, all of the mis-match shows up on the plates. Using a trim pot on the screen(s) supply is probably a good idea. I just have to live with mis-match because the screens are being used for feedback/harmonic equalization.
It is a 2 stage, 2 bottle (3 bottle, if you include rectifier tube) PP class AB1 (recently changed from class-A). It puts out about 14Wrms @ 2% thd before the onset of clipping, which is soft. The first 6-7W are class A, so there is about 6dB of headroom class B. Input sensitivity is 2V and it uses 6dB of feedback.
6BN11 isn't the most linear Pentode, but it's not driving the most linear triode either, so the combo works out ok.
The 6BN11's in action
http://img.photobucket.com/albums/v685/JebDePaiva/DSC00327.jpg
I used the 6BN11 compactron dual pentode in my 6as7 amp. It does a pretty good job and is probably the sexiest small tube I've yet to see. The only thing is that you have to expect mis-match between halves. Where-as separated pentodes would allow for you to tube roll one side to get a match. If the cathodes and screens are tied together, all of the mis-match shows up on the plates. Using a trim pot on the screen(s) supply is probably a good idea. I just have to live with mis-match because the screens are being used for feedback/harmonic equalization.
It is a 2 stage, 2 bottle (3 bottle, if you include rectifier tube) PP class AB1 (recently changed from class-A). It puts out about 14Wrms @ 2% thd before the onset of clipping, which is soft. The first 6-7W are class A, so there is about 6dB of headroom class B. Input sensitivity is 2V and it uses 6dB of feedback.
6BN11 isn't the most linear Pentode, but it's not driving the most linear triode either, so the combo works out ok.
The 6BN11's in action
http://img.photobucket.com/albums/v685/JebDePaiva/DSC00327.jpg
I had just done a bit of searching and come up with the 6BN11. Great minds think alike. Theres also the 6AR11.
If you were to force the Pentodes into a LTP with a silicone CCS in the tail then surely the mismatch would sort itself out and the linearity will be improved.
I want really low output impedence and that is why I am putting the transfomer in the cathodes. I want that SS kick in the bass. I found on my previous outing with the 6080 that gain was less than 1 in the finals so having the transformer in the cathodes shouldn't effect the gain unduly.
Fantastic job on the amp. The little Compactrons make a fine partner to the stubby 6080's - though I have some nice Russian coke bottle style.
Shoog
Just looked at the spec on the 6AR11, looks far worse on the curves than the 6BN11. By the way what gain are you getting from the 6BN11?
If you were to force the Pentodes into a LTP with a silicone CCS in the tail then surely the mismatch would sort itself out and the linearity will be improved.
I want really low output impedence and that is why I am putting the transfomer in the cathodes. I want that SS kick in the bass. I found on my previous outing with the 6080 that gain was less than 1 in the finals so having the transformer in the cathodes shouldn't effect the gain unduly.
Fantastic job on the amp. The little Compactrons make a fine partner to the stubby 6080's - though I have some nice Russian coke bottle style.
Shoog
Just looked at the spec on the 6AR11, looks far worse on the curves than the 6BN11. By the way what gain are you getting from the 6BN11?
Hi Shoog,
the QQE03/12 is a suitable double tetrode with common cathode an screens. This tube is obtainable for a reasonable price.
Data sheet:
http://tubedata.tigahost.com/tubedata/sheets/030/q/QQE03-12.pdf
Original intended for mobile VHF radios, it is also a good powerful tube for audio applications.
regards from Hamburg
Wolfgang
the QQE03/12 is a suitable double tetrode with common cathode an screens. This tube is obtainable for a reasonable price.
Data sheet:
http://tubedata.tigahost.com/tubedata/sheets/030/q/QQE03-12.pdf
Original intended for mobile VHF radios, it is also a good powerful tube for audio applications.
regards from Hamburg
Wolfgang
Apart from the fact thats its an output tetrode it looks really sweet and quite linear. Unfortunately running it as a driver would probably take it out of its happy place
Still it would effortlessly push the 6080's into class B. However my prefered choice of CCS's in the tails of the output prevents Class B as an option.
Tubelab - can't find a datasheet on the 5656 which probably makes it to obscure to be easily obtainable. If you remember the discussion on cathode follower output stages - you never did say how much voltage the top MOSFET needed to stay happy.
Shoog
Hi Shoog,
I have specs of the 5656 but no curves.
Heater 6,3V, 0,4A
Plate 150V
screen 120V
Ug1 -2V
Plate current 15mA
Screen current2,7mA
Transconductance 5,8 mA/V
plate resistance 60 kOhms
plate dissipation 2,5W
Pin 1 common screens
Pin 2 grid 1
Pin 3 grid 2
Pin 4,5 heater
Pin 6 common cathode
Pin 7 Plate 1
Pin 8 Plate 2
Pin 9 common cathode
thats all I´ve got.
The ELL 80 is a fine small double Pentode , they are wide spread used in European Radios when stereo FM was introduced in the sixties . The ELL 80 are 2 x EL95 in one bulb .
Today an ELL80 is a absolute hard to find item. I own one brand new tube in my stock . As far as I remember I keep this tube since more than 20 years.
regards from Hamburg
Wolfgang
I have specs of the 5656 but no curves.
Heater 6,3V, 0,4A
Plate 150V
screen 120V
Ug1 -2V
Plate current 15mA
Screen current2,7mA
Transconductance 5,8 mA/V
plate resistance 60 kOhms
plate dissipation 2,5W
Pin 1 common screens
Pin 2 grid 1
Pin 3 grid 2
Pin 4,5 heater
Pin 6 common cathode
Pin 7 Plate 1
Pin 8 Plate 2
Pin 9 common cathode
thats all I´ve got.
The ELL 80 is a fine small double Pentode , they are wide spread used in European Radios when stereo FM was introduced in the sixties . The ELL 80 are 2 x EL95 in one bulb .
Today an ELL80 is a absolute hard to find item. I own one brand new tube in my stock . As far as I remember I keep this tube since more than 20 years.
regards from Hamburg
Wolfgang
The Raytheon data sheet is on Franks with curves. I have no clue how common the tubes are. I found some while going through my collection. I looked up the data sheet and decided that they were worth saving. Never got time to try them though.
I thought I did somewhere, but I can't find it now. The mosfet needs enough voltage to work and stay linear. The purpuse of the mosfet is to keep the voltage across the cathode follower output tube constant. This means the B+ on top of the mosfet needs to be 30 volts or so above the most positive excursion of the output voltage plus the desired voltage across the output tube.
I have had good luck using regulator triodes (6080, 6AS7, 6336A) or big sweep tubes (6LW6) in triode mode for the output tube. These work well with 75 to 100 volts across them. So, figure out the desired P-P output voltage swing, take half of that (the positive going peak) + the 75 to 100 volts for the output tube + 30 volts or so for the mosfet, to arive at the minimum B+ voltage. The cathode voltage will go below zero due to the inductive nature of the OPT so the mosfet will see 2 X the supply voltage - the 75 to 100 volt drop across the output tube. The idle voltage across the mosfet is fairly high. It is the B+ voltage - the output tube voltage (75 to 100 volts).
You don't need to use a mosfet for the top device either. I have an amp running that uses tubes. Now the 30 volts across the mosfet becomes 100 volts or so for the tube.
It seems reasonable to use 30V for the extra linearity. i certainly can live with that.
I was thinking of trying to do some cross coupled feedback to the screens of the input pentodes. Would I be right in thinking that this could lead to more AC inbalance because the feedback wouldn't be correcting the same sides ?
Shoog
I was thinking of trying to do some cross coupled feedback to the screens of the input pentodes. Would I be right in thinking that this could lead to more AC inbalance because the feedback wouldn't be correcting the same sides ?
Shoog
It's a pretty tricky concept that I've toyed with quite a bit. If you have a habit of thinking of a push pull transformers as two separate transformers that only share a secondary (as do I). It helps to think of the primary as a single winding and ignore the fact that it's center is tied to B+ (AC ground).
The hotter side 1 output tube runs, the more it will reduce the gain of side 2, which in turn will also increase the gain of side 1 (since side 2 is providing less feedback to side 1 driver). This defiantly leads one to think it is causing imbalance. However, when the phases reverse, it's doing the same thing but in the opposite direction. Since the output tube that is swinging towards saturation will always be the one running hotter.
So, bottom line is that the transformer will see improved AC balance, because current through the transformer will be matched more closely between + and - cycle. However, it does promote class AB operation (even though it can operate in class A), so I don't think a CCS on the tail of the output stage would be a good match. However, you could do a current mirror with AC bypass caps across them. That would give you the balancing advantage of the CCS, but allow for class AB (which is good for speakers with impedance dips).
Cross coupling is probably the most effective feedback mechanism I've tried. It's predictable, does a good job at lowering odd-order harmonics along with even and extracting that last bit of power. It also lowers Zout by a predicable amount and really smooths out the transition from class A to B. I'd like to experiment more, but it's my only amp running at the moment, so I can't afford much down time. Though, I'm almost done with a new amp.
The one thing you have to keep an eye on, is that for the drivers, if you let the screen voltage swing too low, it will saturate pre-mature (same principal that applies to ultra-linear output stages).
The arrangement I usually implement for the output stage is to have a current source in each cathode and then bypass between the cathodes with back to back caps. The node is referenced to ground with a 1meg resistor. This behaves as differential. Is this similar to what you are describing?
In this particular implementation this is not possible because the difference in biasing points would create a DC current across the output transformer (which is cathode coupled) and lead to saturation. So the plan is to have each primary side seperated and run down through a CCS to ground. This is a move away from differental but eliminates DC in the transformers.
I would think about going DC coupled and using a voltage divider down from the plate to the screens so that I was only using a bit of feedback. The whole beauty this scheme is that the CCS do all the adjustment for you.within reason. I have used it on my current DC coupled partial feedback amp. Its a bit iffy to get stable (those CCS love to blow) but when it works it really works.
If all this works it should be the best amp I have built yet.
Shoog
In this particular implementation this is not possible because the difference in biasing points would create a DC current across the output transformer (which is cathode coupled) and lead to saturation. So the plan is to have each primary side seperated and run down through a CCS to ground. This is a move away from differental but eliminates DC in the transformers.
I would think about going DC coupled and using a voltage divider down from the plate to the screens so that I was only using a bit of feedback. The whole beauty this scheme is that the CCS do all the adjustment for you.within reason. I have used it on my current DC coupled partial feedback amp. Its a bit iffy to get stable (those CCS love to blow) but when it works it really works.
If all this works it should be the best amp I have built yet.
Shoog
I'm not completely clear what your plan is. I'd be curious to see a conceptual schematic.
This is what I was describing:
It only regulates I2 to match I1, but the total current isn't regulated. So it provides good balance without limiting the amp to class-A.
If you are planning to run the output stage as cathode follower. Cross-coupling won't work. You will want feedback from the cathode of the output stage triode to the screen of the driver on the same side. You always want the signal of the screen to be in phase with the signal on the plate to get negative feedback.
A trick I use to limit the amount of feedback to the screens, is to place a resistor between the two screens (this assumes you have some dropping resistance going to each screen). Since they are out of phase, they will see the center of this resistor as AC ground. Shrinking the value of this resistor decreases feedback, increasing will do the opposite. Similar to the way it's done on the Baby huey amp. Except feedback to the screens rather than plates of the driver.
This is what I was describing:
It only regulates I2 to match I1, but the total current isn't regulated. So it provides good balance without limiting the amp to class-A.
If you are planning to run the output stage as cathode follower. Cross-coupling won't work. You will want feedback from the cathode of the output stage triode to the screen of the driver on the same side. You always want the signal of the screen to be in phase with the signal on the plate to get negative feedback.
A trick I use to limit the amount of feedback to the screens, is to place a resistor between the two screens (this assumes you have some dropping resistance going to each screen). Since they are out of phase, they will see the center of this resistor as AC ground. Shrinking the value of this resistor decreases feedback, increasing will do the opposite. Similar to the way it's done on the Baby huey amp. Except feedback to the screens rather than plates of the driver.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.