Dr. Google doesn't give nothing about, but some time ago, while playing with the Broskie software, it comes out(if memory doesn't cheat) that a push pull of 45's, cathode biased, 275V B+ and around 34mA each tube, will give an optimistic 4W on a 8ohm load .Optimum AA load seemed to be 4K . Who knows? Certainly I'll never spend the money they ask for a quartet of 45's.
I have an old budgetary design for PP 45 that makes about 5Wrms per channel with cathode bias. Works well with the same sorts of OPTs commonly found in old 6BQ5 amps. (7K - 10K PP)
Front end half of a 12AX7A per channel, half of a 12AU7A as split load phase inverter - the design originally used a single 7247 per channel, and the filaments of both were heated by the combined cathode current from the output stage - the downside of this is that the tubes had to be well matched as the cathode currents summed, a small additional current from the plate supply was summed with the cathode current in order to provide the 150mA/24V required for the driver tube filaments - there was obviously another series resistor to drop the balance of the bias voltage, and the whole thing was bypassed by a single good cathode bypass cap.. This approach was also used in the Scott 99 and worked well enough to inspire me to purloin the idea..
Fixed bias or independent cathode bias could also be used along with a conventional filament supply..
I can post the design if anyone is curious.. It dates from a time when 45s were a bit cheaper than they are now. It sounded reasonably good in fact..
Front end half of a 12AX7A per channel, half of a 12AU7A as split load phase inverter - the design originally used a single 7247 per channel, and the filaments of both were heated by the combined cathode current from the output stage - the downside of this is that the tubes had to be well matched as the cathode currents summed, a small additional current from the plate supply was summed with the cathode current in order to provide the 150mA/24V required for the driver tube filaments - there was obviously another series resistor to drop the balance of the bias voltage, and the whole thing was bypassed by a single good cathode bypass cap.. This approach was also used in the Scott 99 and worked well enough to inspire me to purloin the idea..
Fixed bias or independent cathode bias could also be used along with a conventional filament supply..
I can post the design if anyone is curious.. It dates from a time when 45s were a bit cheaper than they are now. It sounded reasonably good in fact..
I have a P-P 45 based amp that uses an OPT from a 6V6 based amplifier. It makes about 4.5 watts. I run it self-bias.
The driver is a simple 6SN7 circuit, using a single gain stage direct-coupled into the balanced driver. Very simple, no feedback.
For fun I run it with a balanced input. Pin 2 of the XLR is the input grid, pin 3 is the cathode via the cathode bypass cap, which is tied to pin 3 of the XLR rather than ground. If running single ended I put a jumper in the XLR that ties pin 3 to ground (pin 1). It works beautifully but you need a preamp that does not care about the low impedance of the grounded grid input.
The driver is a simple 6SN7 circuit, using a single gain stage direct-coupled into the balanced driver. Very simple, no feedback.
For fun I run it with a balanced input. Pin 2 of the XLR is the input grid, pin 3 is the cathode via the cathode bypass cap, which is tied to pin 3 of the XLR rather than ground. If running single ended I put a jumper in the XLR that ties pin 3 to ground (pin 1). It works beautifully but you need a preamp that does not care about the low impedance of the grounded grid input.
Why 45's? Because they are the best sounding tube ever made! I have a Tubelab SE with 45's running slightly over spec 320 volts and 30 mA into a 5K Electra Print OPT for a big 2 WPC. Best sounding amp I have, but not very loud on my 87db speakers. Sounds excellent and painfully loud when it went to visit some 106db Lowther based horns.
45's in push pull? No problem. How much power? Oh, only 20 WPC! 20 watts from P-P 45's? Yes and no extreme spec busting is involved, just the same 320 volt B+ used in the Tubelab SE. Idle current is 15 mA per tube, but full crank current is 150 mA for the P-P pair. I used a 3300 ohm load for this test but I think I will use a 3000 ohm pair of UTC's if I ever build the amp. Where did I come uo with this set of operating conditions? They come directly from the RCA data sheet! They claim 18 watts from a 275 volt B+. Yes, it is AB2. And no it doesn't sound like $*#@.
The test amp used the driver board developed in the 6L6GC in AB2 thread except for 45 output tubes and a lower B+ voltage.
45's in push pull? No problem. How much power? Oh, only 20 WPC! 20 watts from P-P 45's? Yes and no extreme spec busting is involved, just the same 320 volt B+ used in the Tubelab SE. Idle current is 15 mA per tube, but full crank current is 150 mA for the P-P pair. I used a 3300 ohm load for this test but I think I will use a 3000 ohm pair of UTC's if I ever build the amp. Where did I come uo with this set of operating conditions? They come directly from the RCA data sheet! They claim 18 watts from a 275 volt B+. Yes, it is AB2. And no it doesn't sound like $*#@.
The test amp used the driver board developed in the 6L6GC in AB2 thread except for 45 output tubes and a lower B+ voltage.
I had a breadboard with 7.4 Kohm output transformers and an old UTC interstage 2:1 splitter transformers. 6SN7 cascade -IT-45 PP in cathode bias. Still have a very fond memories about that experiment.It was based on SunAudio 300B PP design.45's in US are still fairly cheap (Just bought a pair of Sylvanias for $25). I'd stay away from globes.
I once came across this one (2a3 tubes, but the design could easily be adapted for the 45), which struck me as interesting:
Index of /2a3_diy
Index of /2a3_diy
The 45 PP is quite an appealing idea,... if only where possible to found the 45's at the price quoted by LIMONO. I never see them for less than 80 U$S a piece.
Heresy by heresy, why not a Magnavox style paraphase front end, made up of 6SL7's?Effectively, I was wrong sticking strictly at the maximum 34mA idle. I understand now, higher B+, higher plate load, and less current, and the five watt are on hand.
I got my collection of used 45's on Ebay over a 10 year period. I have at least 20 of them and never paid more than $10 each. Sadly those days seem to be gone.
I have found that most of them work fine in an SE amp set at 30 mA idle and 320 volts. Some will run out of emission when pushed to 20 watts in P-P. They will display flat topping on the sine wave as the power level is increased.
I have already made a 300B P-P amp so my next DH P-P amp will probably involve 307A / VT225's since I have several NOS tubes. 30 watts is easy without grid current.
Pretty much the same situation here, I bought a lot of used and a few new 45s on eBay and at radio shows, I've never paid more than $40 each new, and generally like George $10 or less. I've loaned out my original 45 SE amp and am thinking of using a nice pair of Twins Audio OPTs in a new 45 SE amp I'll build just for the fun of it. (The GM70s are going very slowly at the moment)
I like to run them on 250V @ 38mA or so into 4K primary Z or greater.. They're pretty hard to beat and not very fussy in use.. IMHO despite designing a PP amp to use them I don't tend to think this is their best use.
Last edited:
I agree that some of the 45 "magic" may be lost in a P-P amp. In my situation the Tubelab SE 45 is the best sounding amp that I have, but 2 watts doesn't go very far with 87 db speakers. I had a 300B P-P amp for several years until an electrolytic cap in the power supply shorted causing a mosfet to explode. I never got around to fixing it, but it sounded very good. The design was based loosely on Kevins 300B P-P amp that was published in VTV about 10 years ago. I have always desired to build a newer verison with good quality parts, and it will happen eventually. That amp convinced me that cascaded LTP's are the way to go if cost and parts count is not a problem. Several driver circuits have come and gone, but is it a coincidence that the driver board that I have been using for these experiments uses a 6SL7 and a 6SN7 is cascaded direct coupled LTPs?
The P-P test amp using 45's was just an experiment. It sounded real nice, but I plan to do a lot more experiments before building anything permanent. I am trying to do a pentode - triode cascaded direct coupled LTP driver design for comparison, but time is scarce right now.
One LTP stage feeding a second one. A schematic can be seen in post 460 in the 6L6GC in AB2 thread.
http://www.diyaudio.com/forums/tubes-valves/133034-6l6gc-ab2-amp.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.