ECF80 driving EL84 PP

[Tesla88]

Hi , i'm building an EL84 PP Amp and i want to try a pentode in the front-end , since i've only 1 socket for input stage and phase splitter i would try ECF80 and i need gain to accomodate GNFB.

Found these schematics on the web :
http://www.ankaudiokits.com/_images/ANK_EL84_REV_E_add_LED_(2).GIF

http://i49.photobucket.com/albums/f268/nibb100/el84.jpg

Now i'm using the first link (audio note) but i was wondering if it's correctly designed or if it can be improoved.
I lowered R4 from 560 to 330K , seems to sound a little better

Any advice ? Anyone used a ECF80 as gain phase splitter ?

Thank you guys!

 

[Tom Bavis]

R4 value will affect operating point of both stages since they're direct coupled. Adjust value for about 1/3 to 1/4 of B+ at cathode of triode phase inverter. Ideally it could be set for minimum distortion.

 

[artosalo]

Direct coupled voltage amplifier and cathodyne is always a compromise, but can anyhow work quite well when sufficient amount of GNFB is used since it masks the distortion generated at the front end.
Better to separate the stages and bias both to optimum.
R4 should be adjusted, as already told in previous post, with distortion meter and at high power level, not by ear engineering.
Do you really need R8 / C7 ?
How much GNFB you have now ?

 

[gingertube]

R8/C7 is going to put the HF -3dB voltage point at 11kHz. That does seem a bit agressive to me.
It will depend upon the Output Tranny quality.

For a B+ of +330V then you will want the grid of the ECF80 Triode at between +70 and +80V. That is probably not too low for a pentode anode. The pentode screen (G2) would then want to be around +50V. artosalo's advice to separate the stages and bias each optimaly is GOOD. If you use auto (cathode) bias for the triode stage you can return its grid leak to the top of the 33K which will effectively bootstrap the grid leak resistor and allow use of a small good quality coupling cap whilst maintaing a very low frequency roll off (a zero) so that low frequency stability should not become an issue.
Cheers,
Ian

 

[Tesla88]

R4 value will affect operating point of both stages since they're direct coupled. Adjust value for about 1/3 to 1/4 of B+ at cathode of triode phase inverter. Ideally it could be set for minimum distortion.

R4 should be adjusted, as already told in previous post, with distortion meter and at high power level, not by ear engineering.

Ok , i will check voltage and also thd !

Better to separate the stages and bias both to optimum.

Well , and gingertube explained it well 🙂

Do you really need R8 / C7 ?

R8/C7 is going to put the HF -3dB voltage point at 11kHz. That does seem a bit agressive to me.

I left R8/C7 away , don't need them , but a grid stopper of 2k for triode section o ECF80

Anyway i think i've to explain the whole story..

 

[Tesla88]

Whole story

Months ago i've purchased an amplifier from electro kicker jukebox
Here you can see the original amp :
Jupiter Jukebox

Not having 7189 i went for EL84 , maybe i'll go for EL84M one day...
Output xformers comes with UL taps

I'll started with a classic CC triode gain stage DC coupled to a concertina splitter using a ECC81 as in the Red Light District amp ...
Test shows good response without GNFB down to 35Hz to 20KHz , bad things happen at 20Hz , so i introduced GNFB , good improovement but the overall gain is too low to drive the amp directly from an ipod or smartphone ...

So i tried the Baby Huey (power tubes were cathode biased not CCS , matter ?) , nice approach and sound more detailed than previous design , still have problems at 20-25Hz but in a "softer" way ... but still need 0.7-0.8Vrms at input to drive the amp at full output.

Finally i tried the ECF80 and done some test last evening , ok now i can introduce lot of GNFB (note : i don't like too much of it) and drive the amp with 250/300mVrms ...

Here i discovered a probably mistake ,output xformers are unknown and so the output taps ,most of tests were at 8 ohm load applied to that i assumed to be the 8 ohm taps, with that load the reflected load is more than 10K plate to plate and i measured 380-390V AC plate to plate , is a bit high , isn't it?

So i tried another winding an give worse results at 8 ohm but putting a 16 ohm load lead to good performance down to 20Hz also with reduced GNFB (10dB or so) , reflected load is 7.8 K and Plate to Plate voltage is 320V AC.

If i'll post transformer measures and scope traces , would you be so kind to help me with transfomer impedance matching and to solve distortion at 20Hz?

Also, if you were me, will you go for BabyHuey or ECF80 ?

many thanks for all support!

Apologies for my english 🙁