SE 245 driverstage

[Diy martin]

I will build a powerfull driverstage for my new SE 245 and have the following tubes in my collection

71A-VT25/10Y-56-46-30-31-6SN7GTB-6J5G-27-5842-STC4019A

I will not use a preamp so the gain must be high enough for a cd player input.
What will be the best choice ?
The output transformers are Tribute Amorpe core 5K-16 ohm.
I have also a pair of 5k:5k 50ma interstage transformers laying around.
The speakers I use are the Oris 150 horns with AER drivers

[kmaier]

The 45 triode is not overly difficult to drive. It's quite easy to obtain a little over 2-watts output without stressing the tube. The driver stage can be RC-coupled, but ensure you have a decent driver current. I use ~2.8ma on the driver and get a flat response to about 45KHz within 1dB (with a good OPT). You also want to ensure it can provide a minimum of 150 volts peak-to-peak to ensure you stay in a more linear operating range.

I've done several designs using the 45 in single-ended mode. One of my designs has fairly high gain, ~250mv input for 1-watt output and can easily be used with any decent line-level source. After working with multiple twin triodes, I finally settled on a 5814A (commercial version of a 12AU7). The reasons for this were:

- lowest output noise from the input/driver
- best overall linearity
- lowest distortion
- wide bandwidth
- AC filament operation
- filament to cathode voltage rating exceeds driver stage requirements
- excellent NOS versions are easily obtained

The stages are direct-coupled, first stage biased at 1ma and the second stage biased at 2.8ma. Overall gain for the input/driver is ~150 volts per volt or ~43dB. It will vary a bit based on the particular tube (zero feedback) and can swing 150 volts pk-pk output without any clipping.

For the output stage, I prefer self-bias and AC filaments for the 45. Don't use a grid resistor larger than 250K. Preferred bias current is 34ma and a plate load of 5K. I used a bit higher voltage which resulted in 60 volts of self-bias and about 295 volts cathode to plate. With a good 45 you can get a solid 2.25-watts output.

With an AC filament (and a good 45) S/N is more than 85dB below 1-watt output. Note that not all 45s will do this well. I've not found any balloon 45s (aka 245/345) to be this good. I developed a unique balance arrangement (as I've never seen any other design like it to date) where I use a fixed DC balance and an adjustable AC balance. I have found this to provide much better hum balance and eliminate any need for a DC filament (again, assumes good quality 45s).

So, that's my input... sorry it doesn't match one of your driver tubes... but a pair of 6J5s would probably work well in the same topology. Of course, just my $0.02.

Regards, KM

[kevinkr]

A single 5842 per channel will make an extremely linear driver, particularly when CCS or choke loaded, rp is a scant 2K or so, and there is sufficient gain for a cd player or indeed most lower level line sources.

Depending on the primary inductance of your IT it could be that this is a match made in heaven for the 5842. (Watch out for ultrasonic resonances.)