John Swenson's BDT preamp - finally!!

[jkeny]

After a very long hiatus, I'm finally finishing this BDT preamp build.

Tests just now are producing it's first sounds but I have to check out how I have the pot wired. I'm also just using a pair of headphones for testing so the sound is pretty low because I'm using the output transformer designed for driving into an amplifier.
Schematic:

It's a really interesting & unique topology using Beam Deflection Tubes which were originally designed for colour TV as colour subcarrier demodulators. here's what John S says about these tubes: "they just happen to be one of the most linear amplifiers ever made and behave like differential amplifiers with balanced in and out (they make great phase splitters too), with one really neat feature: the gain can be adjusted with a DC voltage on the control grid." So what we have is a balanced tube version of the lightspeed with gain! Main thread on this is here

I'm looking at various possibilities for using this pre-amplifier:
- A headphone amplifier if a suitable output transformer is used - I'll be trying a Toroidal PS transformer as John has suggested dual primaries/dual secondaries - secondaries of 24V will give about 100ohm output impedance - depending on how low the headphones' impedance is secondaries of 12V or 10V can be tried. This should make a really nice H/P amp and can easily be run in balanced configuration.

- I'm also thinking of being able to switch the output transformers so as to use this as the front end for a Susan Parker Zeus amplifier. This will be an interesting configuration I reckon.

- I'll also be using this as a separate pre-amplifier!

Has anybody else built one of these?

[DF96]

Do you find any problem with partition noise? A BDT has two sources of partition noise. One is the normal g2 (called accelerator in BDT) like a pentode, the other is partition between the two anodes. The second would be the worst. This noise is not a problem when a BDT is used for its intended purposes, but I wonder if it is noticeable in this audio application?

[jkeny]

Originally Posted by DF96

Do you find any problem with partition noise? A BDT has two sources of partition noise. One is the normal g2 (called accelerator in BDT) like a pentode, the other is partition between the two anodes. The second would be the worst. This noise is not a problem when a BDT is used for its intended purposes, but I wonder if it is noticeable in this audio application?

I'm still in preliminary testing so too soon to say - I'm using headphones to test at the moment as my main amp is out with a reviewer. I'll be trying to optimise the sound tomorrow & will post my findings but the sound I heard on my headphones had no signs of noise. JS had two issues with these tubes
- one, they are sensitive to magnetic fields so need to be located away from the PS trafo.
- two, I think he said that they are a bit sensitive to vibration.

I have to test both of these aspects as I have not taken any special precautions in these areas, yet!

This is the first time I've heard about partition noise - have you used these tubes before in the TV field?

[mashaffer]

Interesting tube. My first question of course was how linear is it without all the sand on top? I was curious to find out what might be achieved by more normal methods so I did some searching. From Mr. Swenson's write up here...

Tube DIY Asylum

it sounds like it might not be all that linear. It would be interesting to see a comparison. He also mentions that it is susceptible to magnetic fields and to heater hum (thus DC heaters). I like the idea of being able to control the gain via grid voltage (remote control via. SS pot) but it looks like you have to beat the tube into submission with some pretty big hammers.

Makes me wonder what other odd ball tubes might also be repurposed.

[jkeny]

Well I've briefly listened to it without the mosfet cascodes & I don't hear any non-linearities - that's not to say there are none -again this was through headphones (32ohm) so the vol was down. There's a max gain of 4 without the mosfet cascodes. Not sure if it has to be beaten into submission to linearise it though. The cascodes only require a little CCS & a couple of parts - not difficult. The main build-up work is in the PS but with some boards these are not too difficult. I built mine on perf-board before I got pcbs

I think he also mentions using 6DJ8 tubes which he used in a low power amp along with the 6AR8 front end - here's what he said "For the fun of it I also built a power amp just using BDTs, you can get a couple watts out of it. I used a 6AR8 driving a 6DJ8 (with the cascode on the 6AR8 to get enough gain), it sounds fantastic on very sensitive speakers. It would probably make a really nice headphone amp but I've never actually built one in that configuration. The headphone amp could be built with just a 6AR8 per channel. "

[smoking-amp]

The BDT has signal on only one of its deflectors. It might perform better with a symmetrical input. If only one input is available, I would try tapping R17 to get the same signal amplitude (maybe 1/2 since the output will double with both deflectors working) as on the active input, and connect that to the grounded deflector (cap coupled with a high value R to ground for 0 V DC). You could also try that with both deflectors, as long as these positive feedbacks have less than unity gain (adjust tappings for a little less than unity gain).

[jkeny]

I'm using both inputs (balanced), smoking-amp but thanks for the advice for SE use as I will want to connect it SE at times - I don't really follow what you are saying though, - could you diagram it?

[smoking-amp]

You just split R17 into two series resistors. Their sum approx. the same resistance as the present R17, but the smaller resistor/larger resistor should = (input signal V)/(2X present output signal V) at the BDT plates. Smaller resistor goes to ground as before. Tap point gets a cap to the presently grounded deflector. A high R resistor from that deflector to ground maintains 0 DC level on the deflector still.

[jkeny]

Originally Posted by smoking-amp

You just split R17 into two series resistors. Their sum approx. the same resistance as the present R17, but the smaller resistor/larger resistor should = (input signal V)/(2X present output signal V) at the BDT plates. Smaller resistor goes to ground as before. Tap point gets a cap to the presently grounded deflector. A high R resistor from that deflector to ground maintains 0 DC level on the deflector still.

Ok, I got it but the Vout on the plate changes based on the vol control, I think - so does this negate your idea?

I seem to remember JS stating that it will work better in balanced mode

[smoking-amp]

"but the Vout on the plate changes based on the vol control, I think - so does this negate your idea?"

No problemo, input and output levels track.