If you make a Visio version, send it to me and I will add the component values. The 300B grids are connected to the output of this board. The uncommited resistors on the right go to the hum pot wipers (mounted off board). They are 10 ohm for measuring tube current.
I will dig the amp out of the closet and trace the power supply. It was built on perf board. At first I had a mosfet regulated supply, but it kept blowing up, so I ripped it out and switched to a choke input simple supply. All the mosfet stuff is still on the board. People have been asking me for this schematic for years.
I have built about a dozen TubelabSE amps using 300B's or 45's. About 40 more are in customers hands though many are still in the construction stage. All have PowerDrive. PowerDrive doesn't gain much additional power output with these tubes, but it sure cleans up the transients and eliminates the blocking distortion that happens with a capacitor coupled output tube. In my analysis the grid only goes positive on strong bass notes which have a low duty cycle ratio unless you try to play disco music at the clipping level. If you do, you are better off with a KT88.
arnoldc said:
It's a good start, but needs some work. First off, I don't think you're gonna get the output swing you need from one LTP, and certainly not the gain for good sensitivity, especially if you include gNFB. This may have to be a two stage front end, and the second stage could be a differential as well to preserve balance.
The input LTP would benefit from an active tail load, as opposed to the passive load shown here. This'll improve balance considerably.
I would also include a grid driver, something like a source follower, a stiff cathode follower, or maybe even a White Cathode Follower. The problem with these DHTs in general is grid current even before the grid actually goes positive. You also need adequate drive to charge up the Cgk + Cmiller + Cstray if you are to avoid the slew limiting problem at the higher frequencies. Where 300Bs are concerned, inadequate grid drive has long been a cause of disappointing results from this type.
I would also include provisions for adjusting the bias of each 300B final so that you can equalize the plate currents. Reducing the DC magnetization will help improve OPT performance, especially considering that PP OPTs aren't designed with gaps to handle DC.
resident said:
MOSFET Source Follower, White Cathode Follower, Kimmel Mu-Stage, Broskie Cathode Follower.
Any of those could work. You could also use an IST and drive it with a low-u, low-rp, triode like a 6BX7, 6AH4, 2A3, 45, 50, trioded-6AQ5/6V6. (This has been done when it's known that gNFB will not be used.)
You could also use a high g(m), small signal pentode like a 12BY7A (seen one of these used to drive a trioded 813, so it'll for sure handle a 300B) wired as a voltage amp with enough current sourcing to drive the 300B.
Lots of possibilities for driving the 300Bs.
That's because it is! I started with Kevins schematic, changed to 9 pin tubes because I had bunches of them. This was about 7 years ago when SRPP was the latest greatest thing, so I threw that in too. I drew up a schematic and laid out the PC board in one pass. It worked the first time so I never looked back. Sometimes you get lucky, most of the time you have to work for success.
I plan to design a "universal" P-P driver board following the same principles but I plan to test several driver stages, SRPP, PowerDrive, augmented CF, and whatever else that I can think of. I would like to have a driver design that can drive anything including big transmitter tubes and screen driven sweep tubes. With this, 300B's are easy. I have far more knowledge and experience than I did 7 years ago, and much better breadboarding, test, and measurement capabilities. I just have far less time to use it all.
hi arnold,
there is a nice discussion here, see post #68: http://www.diyaudio.com/forums/showthread.php?s=&threadid=42111&perpage=25&highlight=&pagenumber=3
you can use a 12B4 to direct couple to the 300b grids, a -150 volt supply would come in handy....biasing the grids of the 300b from the 12b4 grid is possible....
there is a nice discussion here, see post #68: http://www.diyaudio.com/forums/showthread.php?s=&threadid=42111&perpage=25&highlight=&pagenumber=3
you can use a 12B4 to direct couple to the 300b grids, a -150 volt supply would come in handy....biasing the grids of the 300b from the 12b4 grid is possible....
Kevin, your work IS appreciated
My all-diif triode PP amp was inspired by your 300B design, which I saw on the Svetlana site before it got demolished.
The only differences were use of triode-strapped EL34s instead of 300Bs, a pentode CCS for the tail of the 6SL7 splitter instead of a high value resistor, much lower (individual) cathode resistors for the 6SN7, step-network instead of DC coupling between the 6SL7 and 6SN7 and the addition of moderate cross-coupled NFB from OP tube plates to 6SL7 plates (my speakers work better with the lower OP impedance).
My all-diif triode PP amp was inspired by your 300B design, which I saw on the Svetlana site before it got demolished.
The only differences were use of triode-strapped EL34s instead of 300Bs, a pentode CCS for the tail of the 6SL7 splitter instead of a high value resistor, much lower (individual) cathode resistors for the 6SN7, step-network instead of DC coupling between the 6SL7 and 6SN7 and the addition of moderate cross-coupled NFB from OP tube plates to 6SL7 plates (my speakers work better with the lower OP impedance).
What's cookin'
Caveat: I'm not a designer, but a builder so please let me know some obvious which I have missed...
The input will be using a dual triode WE396A in a differential configuration, with a CCS on the cathode inspired by Gary Pimm. I'm trying to figure out what voltage swing will give me the least distortion and I'm not sure if the adjustable resistor will help or even work. It will be cap coupled to the next stage, though direct coupling will be more appreciated.
The driver will be using a pair of 56/76 transformer coupled (sorry Miles) to the 300B using my existing LL1660S in push pull to push pull configuration. I wanted good voltage swing which I think can be achieved in this configuration.
The finally, the 300B will be fixed bias, but since it's IT coupled, I can't have individual bias or can I?
I'd really appreciate your comments guys...
An externally hosted image should be here but it was not working when we last tested it.
Caveat: I'm not a designer, but a builder so please let me know some obvious which I have missed...
The input will be using a dual triode WE396A in a differential configuration, with a CCS on the cathode inspired by Gary Pimm. I'm trying to figure out what voltage swing will give me the least distortion and I'm not sure if the adjustable resistor will help or even work. It will be cap coupled to the next stage, though direct coupling will be more appreciated.
The driver will be using a pair of 56/76 transformer coupled (sorry Miles) to the 300B using my existing LL1660S in push pull to push pull configuration. I wanted good voltage swing which I think can be achieved in this configuration.
The finally, the 300B will be fixed bias, but since it's IT coupled, I can't have individual bias or can I?
I'd really appreciate your comments guys...
Arnold,
Gary Pimm's CCS designs are terrific. However, for the "tail" of a difference amplifier stage, GP's stuff is overkill. Cascoded DN2540N5s are more than sufficient for this job. You can't ground the CCS. The CCS must connect to a negative supply that provides voltage to operate it and compliance for the I/P signal. More coupling caps. are needed, if a bipolar PSU is not employed. TANSTAAFL!
You need a grid leak resistor at the I/P of the inverting triode. Since the non-inverting triode is undriven, connect it directly to ground. Thanks to the "tail" CCS, a trim pot. in the plate circuit is probably unnecessary.
The use of individual bias adjustments for each 300B, which I think is very important, forces you to cap. couple the interstage trafo to the grids and add grid leak resistors. The grid leak resistors connect to the wipers of the bias trim potentiometers. Don't forget to "stand" the 300B cathodes on unbypassed resistors. 330 Ohms is about right, when the tubes are setup to operate in Class "A".
Gary Pimm's CCS designs are terrific. However, for the "tail" of a difference amplifier stage, GP's stuff is overkill. Cascoded DN2540N5s are more than sufficient for this job. You can't ground the CCS. The CCS must connect to a negative supply that provides voltage to operate it and compliance for the I/P signal. More coupling caps. are needed, if a bipolar PSU is not employed. TANSTAAFL!
You need a grid leak resistor at the I/P of the inverting triode. Since the non-inverting triode is undriven, connect it directly to ground. Thanks to the "tail" CCS, a trim pot. in the plate circuit is probably unnecessary.
The use of individual bias adjustments for each 300B, which I think is very important, forces you to cap. couple the interstage trafo to the grids and add grid leak resistors. The grid leak resistors connect to the wipers of the bias trim potentiometers. Don't forget to "stand" the 300B cathodes on unbypassed resistors. 330 Ohms is about right, when the tubes are setup to operate in Class "A".
AFIK, the Lundhall 1660 can be connected as separate secondaries so that individual biasing of output tubes can be done. I believe this is better than using grid leak resistors, because the secondaries will have much lower dc resistance than a grid leak resistor.
if class A only operation is contemplated, i believe that a fixed secondary load needs to be reflected to the LL1660 primary, 100k is a start.
assuming a B+ of 350volts, the 300B is biased -60v so my guess is you need 240volts peak of grid to grid drive...
A look at the Lundahl LL1660S data sheet indicates that the circuitry for the PP 300Bs can be kept separate. That is good, as DC blocking caps. are not needed and the DC resistance between grid and ground is low.
Use the resistance Lundahl recommends across each secondary winding to achieve correct loading.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.