6EM7 project: maybe use more sand?

[mashaffer]

As I looked at this project and the desire to include tube crossover and the desire to get it in a small cabinet I started to look at other options.

To summarize from other posts...

The original plan was 6EM7 as treble amp (over 200Hz) using the natural rolloff of the driver and 1st order HP via the input cap with active 3rd order PB for woofer line out and 3rd order LP for sub line out. Various things cause me to have to compromise on the tubes used for the filter CFs and the size of the case.

Talk of power drive and other source follower usage got me to thinking that using a high voltage MOSFET instead of the tube for the sub out might be a good way to save on heat, heater wiring, and space. So I simulated in LT spice changing nothing in the circuit besides the active device and it appears to work fine. Well actually the source resistor is a lot smaller than the cathode resistor was.

So first pass it seems that this should be fine for subwoofer duty. The next thought is that some guys I really respect are using these SFs in full range amps and the reports are that the sound is anything but sandy. So then I thought that I might even try using the same approach for the woofer BP.

Is this heresy? What do you think of this approach?

[Michael Koster]

Here's a suggestion I plan to try: Use CCS-loaded MOSFET source followers for the active filter buffer stages (assuming the Sallen-Key or other unity-gain topology). There is probably no sonic benefit to using a tube CF over a MOSFET SF in this duty.

Cheers,

Michael

PS, on your original question, in my mind there are real benefits to a low output impedance driving a woofer with a high damping factor. Especially with a high-Q woofer or tuned enclosure, etc.

[mashaffer]

This one is I suppose best answered by George since he has gobs of experience with these things but I will gladly accept any insight. I was looking at using 2SK2700 which is used in the power drive circuits but LTspice had no model for it so I used SPA11N60C3 to model instead.

In the sims I used voltage divider to bias the gate in the center of the 300V supply and used 3.1k source resistor to set to about 45mA and the results seemed quite good.

Now here is the deal. My PT is rated at 150mA of which I use about 100mA for the EM7s leaving 50mA for the filters. I will need 6 FETs total (2 input for BP, 2 output for BP, 2 output for Sub LP). Obviously I can't run 45mA a piece. With the CF models 8mA through a 6N1P looked quite good so the questions are...

Can the 2SK2700 do a good job at say 8mA for the output SFs and 2 or 3mA for the input SFs? And since I don't need 100V+ of swing should I make the voltage divide asymetrical to bring the voltage across the source resistor down to allow the same (or similar) small value to be used? And while I am at it is voltage divider even the best bias method to use?

Thanks for your patience.

[mashaffer]

OK, here is a first stab at what I have in mind. The target was 10 to 60 or 70Hz sub and woofer from that 60 or 70Hz to about 200Hz. I had to diddle with the individual filter specifications somewhat to compensate for the interaction of LP and HP parts of the woofer filter. Additional massaging is likely.

The impedances may be less than optimal as I am pretty much guessing at that based on the idea that the inputs to the filters will be in parallel to the grid of the VAS triode of the 6EM7.

The choice of MOSFET is nothing magic as it was just one of the few high voltage devices in the model list of LTspice.

You opinions are highly coveted.

[mashaffer]

The overall response of each section and the sum. 10hz - 100kHz. The ultrasonic behavior looks OK to my untrained eye.

[mashaffer]

And a close up of the results in the crossover region. The crossover is closer to 70 than 60Hz but that is not entirely a bad thing. Should still be low enough for the main woofers to have a significant contribution for tympani.

[mashaffer]

PS. I did a couple of sims with CCS and there was a slight increase in maximum attenuation outside of the pass band but it was not huge. I don't know about what would happen to distortion and dynamics of course.