[stixx]

Hope I'm not starting to bore people by repeating that a very similar amp is playing music for several weeks now...

 
It's my second attempt with the C3G triode wired, and after putting my new SE output transformers from Reinhoefer to use I must clearly say it's the best amp I built so far. Even beats my parafeed Aikido in terms of speed and authentic sonic reproduction.
With a single Salas shunt reg (set at 80mA) to supply around 32mA to the C3G's it easily fulfills Salas' recommendation to run the reg hot.
Absolutely NO stability issues at startup or else plus it's deadquiet.
 
C3G reqire loctal bases which is a bit of a drawback, they need considerable break-in and warm up time, but deliver they do.
Get a pair and listen to any well recorded music involving piano, drums and female voices... goosebumps guaranteed...

 

 

[adamus]

lovely!
 
i am beginning to think DHT with rod colemans supplies, electraprint or sowter SE 10k : 300r could be on the cards....

 

[FrankCooter]

Excellent work  stixx! Really appreciate you sharing your schematic. Definitely intend to try something like this myself at some point.
 
I've seen the C3G around,but never tried it. Judging by the D3a, you Germans seem to win the prize for quality control for the difficult to manufacture high gm types. Time to go to Ebay!

 

[FrankCooter]

Quote:

lovely!
 
i am beginning to think DHT with rod colemans supplies, electraprint or sowter SE 10k : 300r could be on the cards....

Bet you'll like it if you try it! I have a pair of 5k:300 @60ma Electra-Print transformers originally purchased for  preamp outputs. My first attempt at a headphone amp  was a breadboard "spud" using these with a pair of GM70's stolen from my speaker amp. Used 4 car batteries for the filaments and ran the plates at about 600v. A total absurdity, but sounded great with  HD600's. 

 

[stixx]

 
Quote:

I've seen the C3G around,but never tried it. Judging by the D3a, you Germans seem to win the prize for quality control for the difficult to manufacture high gm types. Time to go to Ebay!

C3G still is a bit of a sleeper altough prices have gone up over the last three years or so. When I built my first C3G amp (OTL) in 2004 I purchased them for about 4€ each, brandnew in the box!
Now this is about three times that...;-(
I once read in a documentation about the C3G that since they were intended for long distance broadcasting they had meticulous quality control resulting in going price of about 250 marks (125€ or about 170$ a pop!!). That was in the 60's...
 

 

 
Doesn't this look tasty...

 

[adamus]

any movement on this regal?
 
i bought a couple of cg3's to go head to head with some d3a. will report back.

 

[regal]

I finished the chassis work for the PS and mounted everything inside.  Still a long way to go and I have to work this weekend.  Couting on a week off after the Further concert over thanksgiving.
 
I studied both the 5842 and the 417A datasheet and no mention of the maximum grid leak resistor value like mandated for the 6C45pi,   so I think I am going to raise the value of the grid leak to 500k and the pot to 47k,  as a tube DAC could one day be in my future.
 
Does the cg3 work with a 47k volume pot, or is it like the 6c45pi ?

 

[dsavitsk]

Quote:

or is it like the 6c45pi ?

Where are you getting this?  You keep talking about the 6c45 as if increasing the grid leak will bring about the end of the world.  The input capacitance for the 6C45 is 11pF, mu is 52.  The input capacitance for the 5842 is 9pf and mu is 43.  The difference here isn't that big. But, more importantly, the 6C45 does just fine with a 50K, or even 100K, pot ... unless, of course, you think you need frequency response out to 100K -- but if that's the case you are using the wrong topology.  I'd be more concerned that the 5842's Gm is a little low for this application.

 

[stixx]

 
Quote:

the 6C45 does just fine with a 50K, or even 100K pot...

 
And so does the C3G... both my C3G amps have 100K pots and definitely do not lack any treble extension.
 
 

 

[adamus]

Quote:

Where are you getting this?  You keep talking about the 6c45 as if increasing the grid leak will bring about the end of the world.  The input capacitance for the 6C45 is 11pF, mu is 52.  The input capacitance for the 5842 is 9pf and mu is 43.  The difference here isn't that big. But, more importantly, the 6C45 does just fine with a 50K, or even 100K, pot ... unless, of course, you think you need frequency response out to 100K -- but if that's the case you are using the wrong topology.  I'd be more concerned that the 5842's Gm is a little low for this application.

I too have read on quite a few occasions people saying it doesnt tolerate a gridleak of over 50k.... maybe a forum myth?
 
i havent fired the c3g up yet. should i take any special measures with it like i had to with the d3a (grid, anode + cathode stoppers). So far i have just stuck 1k right on the grid pin. CSS is a pnp cascode with led ref's.
 

 

[regal]

Quote:

Where are you getting this?  You keep talking about the 6c45 as if increasing the grid leak will bring about the end of the world.  The input capacitance for the 6C45 is 11pF, mu is 52.  The input capacitance for the 5842 is 9pf and mu is 43.  The difference here isn't that big. But, more importantly, the 6C45 does just fine with a 50K, or even 100K, pot ... unless, of course, you think you need frequency response out to 100K -- but if that's the case you are using the wrong topology.  I'd be more concerned that the 5842's Gm is a little low for this application.

 
 
Straight from the datasheet max grid leak is 150k  http://www.tubes.ru/techinfo/AmateurRadio/6c45p-e.html   ,  if you are using using a 50k pot and a 10x50k grid leak you basically have a grid leak thru your potentiometer from the 6c45,  not good for pot life or SQ if the datasheet is accurate.

 

[dsavitsk]

I see your confusion: unless you put a cap between them, the pot and the grid leak are in parallel meaning that the pot is effectively the grid leak.

 

[nikongod]

What does "mKA" mean for grid leak? 

 

[adamus]

Quote:

I see your confusion: unless you put a cap between them, the pot and the grid leak are in parallel meaning that the pot is effectively the grid leak.

whats considered best pratice? i typically just parallel a a largish resistor with the pot and dont bother protecting the pot (or the entire input stage) with a cap.

 

[regal]

Quote:

whats considered best pratice? i typically just parallel a a largish resistor with the pot and dont bother protecting the pot (or the entire input stage) with a cap.

Its only an issue I've seen with the 6c45,  its an unusual tube witch evidently leaks alot of current off the grid.  So you are forced to use a 10k pot feeding a 100k grid leak (following the 10x rule).   With most tubes you use a 50k pot and a 10x50k=500k grid leak.  But if you do this with the 6c45p considereable current will follow the pot shunt to ground.  That is if the datasheet is correct. 
 
Could be a non issue as I have yet to find a pair of 6c45p's that matched close enough to make a stereo amp so I don't speak from direct experience just what I saw on the datasheet.  I know there are lot of complaints with 6c45pi sounding hard due to oscillations but this could be part of the issue as well IMO.