Page 1 of 3

Clementine 6L6 Build

PostPosted: Wed Aug 17, 2011 6:39 pm
by TomMcNally
I put this together this afternoon after work ... it kicks ass !
Clementine 6L6 build, Edcor XPWR-005 (PA-774 clone) and
Edcor GXSE15-8-3.5K, Hammond 156R choke, Russian
6SL7 and 6L6 equivalents.

Better pix coming ...


Image

PostPosted: Wed Aug 17, 2011 7:43 pm
by badassbob
Nice work Tom. How do you like it in comparison to the GSG? I know theyre both way different, but how would you describe the differences between the two?

PostPosted: Wed Aug 17, 2011 8:05 pm
by TomMcNally
It's only a few hours old, but I'd say it has more power than a
get*set*go, but probably a little more distortion when cranked up.
It's quiet, no hum with my ear up to the speaker.
Certainly a cheaper build, mine has $13 worth of tubes!

The circuit is very similar to the JE Labs EL-34SE, and the
new Clementine board makes it simple to build and wire.

PostPosted: Wed Aug 17, 2011 8:20 pm
by Ty_Bower
The Clementine board would certainly make for a more compact build. On the other hand, the GSG board naturally leads you towards vacuum tube rectification. The GSG also has the 6 volt DC filament supplies on board. You could easily build a GSG out for 6L6, but it would be a bit of work to turn a Clementine for 6B4G.

Build them both, I suppose. :))

PostPosted: Wed Aug 17, 2011 9:43 pm
by TerrySmith
SWEET! Is it triode, pentode or UL connected?

PostPosted: Wed Aug 17, 2011 9:50 pm
by TomMcNally
Triode strapped with 100 ohm resistor plate<>screen.
The schematic is posted. The OPTs have a tap,
so UL might work well too.

PostPosted: Thu Aug 18, 2011 5:50 am
by Shannon Parks
Unfortunately, ultralinear connection doesn't seem to have benefits for the hifi listener. I initially wanted to do a 10W 807 SE UL amp, but the distortion curve tracks along with the triode-strapped. That is, you get 5% distortion at 5W using either method. But then my UL tests exhibited a very inferior damping, or amplifier impedance. Damping factor on a UL amp was like 0.5, whereas a triode-strapped amp was back in the 3-4 range. All these numbers are off the top of my head, but it is easy for anyone to duplicate.

Shannon

Re: Clementine 6L6 Build

PostPosted: Thu Aug 18, 2011 6:17 am
by Shannon Parks
TomMcNally wrote:Clementine 6L6 build, Edcor XPWR-005 (PA-774 clone) and Edcor GXSE15-8-3.5K, Hammond 156R choke, Russian
6SL7 and 6L6 equivalents.


Awesome, Tom! If you get the chance, I'm curious what the peak B+ at C7 is on a cold amp at startup. Won't get my XPWR-005 for several more weeks! :'(

Since these outputs are really for a GSG, I have a mod you can do. You can use a 750 ohm resistor at R27 & R28. Hookup J11 and J10 to the output + (pins 1) and - (pins 2). This will be 6dB of global feedback, or you could use 2K for 3dB. Heresy to some, but I think this mod will lower the impedances enough to really hit a sweet spot and sound as good as any high-falutin' triode. In particular, it will help match those 3.5K outputs better. I think the standard transformer set I'll recommend will be the Edcor XPWR-005 (PA-774 clone) and Edcor GXSE15-8-5K (same as Darling). Just need to convince Brian to start stocking them.

Shannon

PostPosted: Thu Aug 18, 2011 7:51 am
by TomMcNally
I'll experiment later today, check some voltages and let you know.

The Chubby Clementine 6L6 amp was on all night, the Edcor XPWR-005
was still comfortable to the touch this morning.

By the way, I know a few guys would go to more complex methods,
but a quick SPARK test is the easiest way to make sure the 6.3V
filament windings are in phase. Two leads will easily fit in the
connector if they aren't tinned.

... tom

PostPosted: Thu Aug 18, 2011 8:21 am
by Shannon Parks
TomMcNally wrote:By the way, I know a few guys would go to more complex methods, but a quick SPARK test is the easiest way to make sure the 6.3V filament windings are in phase.


That was a fun thread! I just connect two ends together and test for 0VAC across the other two. If I get 12.6VAC, I know I have them miswired.

Shannon

PostPosted: Thu Aug 18, 2011 4:38 pm
by TomMcNally
More pix, taken before adding feedback ...

ImageImageImageImage

Image

Re: Clementine 6L6 Build

PostPosted: Thu Aug 18, 2011 4:41 pm
by TomMcNally
separks wrote:Awesome, Tom! If you get the chance, I'm curious what the peak B+ at C7 is on a cold amp at startup. Won't get my XPWR-005 for several more weeks! :'(


I got about 465 cold, dropped to 424 after the tubes warmed up.

separks wrote:Since these outputs are really for a GSG, I have a mod you can do. You can use a 750 ohm resistor at R27 & R28. Hookup J11 and J10 to the output + (pins 1) and - (pins 2). This will be 6dB of global feedback, or you could use 2K for 3dB. Heresy to some, but I think this mod will lower the impedances enough to really hit a sweet spot and sound as good as any high-falutin' triode. In particular, it will help match those 3.5K outputs better.


I used 1K ... definitely more kick in the bass sounds really good.

... tom

PostPosted: Fri Aug 19, 2011 10:25 am
by dougm216
Very nice Tom. I'm looking for next amp project. What speakers do you use? What efficiency? Most of my speakers are Klipsch My 35 is more power than needed. Any suggestions? I do like the looks of this one.

Doug

PostPosted: Fri Aug 19, 2011 10:48 am
by TomMcNally
Doug -

My test (aka bedroom) speakers are Klipsch KSB 3.1 ...
they are 94 db, work great with anything!

The Clementine in Darling or 6L6 (Chubby Clementine) mode
is a nice easy to build amp.

never say never...

PostPosted: Fri Aug 19, 2011 2:19 pm
by EWBrown
One possible way to use 6B4Gs on the Clementine, would be to run both of their filaments in parallel (along with the 6SL7 driver) , off a common 6VDC, 3A supply, and then add fixed grid bias for the 6B4Gs. This would also necessitate lowering the B+ voltage suitably, in order to avoid exceeding the max safe P-F voltage.

Note that the 6B4Gs' filaments would have to have a low-impedance (or low resistance) reference toDC ground, in order for this to work properly (if at all). 6.3VAC on their AC filaments might tend to be a tad too "hummy" , DC would be better.

Still a bit complex, but it should not require hacking up the PC board. Any additional support circuitry could be built up on a couple of terminal strips.

Just another krazzy idear.... ;) (lol) :))

/ed B