The Tube Beast at ETF2010

[Dyna Saur]

A totally different approach to all-triode power. This interesting little triode PPP amp was presented at European Triode Festival 2010 (ETF2010)This is afectionately known as The Tube Beast, it has one hundred 6N3Ps (equivalent to 2C51 / 5670 / WE396A) on four PC boards, each containing 25 tubes, with all sections in parallel.  Just think of each of these four boards as being q really BIG power triode.  The differential driver board contains eight triode strapped EL84s as drivers and cathode followers, directly coupled to the bower boards' grids.  Add in six large toroid transformers and mount it all on a one meter square wooden base (39 3/8 inches on a side).  It is reportedly capable of 100 Watts per channel.

Howzabout a new  Bottlehead Kit for this one ;')   8^)

As Will Smith said in "Independence Day", I gotta get me one of these...

/ed B in NC

 

[JC]

Is there a picture or a link somewhere?  I'm not having much luck with the search engines on this one...

 

[paulw]

link to ETF website

http://www.triodefestival.net/

and is this the 'Beast'?

 

[Grainger49]

I like the "Warning !  Lightning Bolts" sign on the table.

Thanks for the link!

 

[JC]

I appreciate it!

I wonder if the WAF on this has a "-" in front of it?

 

[HF9]

I hope he's using a matched hundredtet of tubes

 

[RayP]

As I understand it, the tubes don't need to be matched. The idea is that there are so many imperfections, it all averages out. I would imagine each tube has its own cathode resistor however.

A while back I built an amp that had 5 12B4A's per channel and I really liked it. I should drag it out again and have a listen.

ray


ray

 

[Paul Joppa]

As I understand it, the tubes don't need to be matched. The idea is that there are so many imperfections, it all averages out. ...

Some years ago, I made a mathematical model of real-world triode nonlinearity in order to assess the importance of it. The model I chose was to average a large number of triodes in parallel, with the mu of each triode chosen from a normal statistical distribution. It worked quite well, and even allowed me to extract a nonlinearity factor from published curves. The nonlinearity factor I chose was the standard deviation of the distribution of mu values.