GK-71 - economical build.

[Paul Barker]

Some many years ago our very own Paul Barker reckoned they would be ace for driving GM-70.

.

If I did I was wrong.


Agree with Mark.

[pre65]

I'm not sure how you'd use this one, but I'm sure Paul , Nick or James could comment better . It seems to be aimed at Class B usage, and the data sheet shows quite low current for Class A operating points . I'm concerned about grid current on the input for the higher current class B type operating points . It'll need to run at maybe 25mA to drive a GK71 properly , although I still have no idea what your B+ is going to be, or why you can't answer that question - it will tell me what sort of current and voltage swing you are aiming to acheive into the grid of the GK-71 .

HT will be about 600v, not quite sure till the PSU is tested.

The CV1285 is run at 6 ma and cathode about -5v with both halves in parallel for class A driver.(as per the data sheet)

[JamesD]

6mA is really not enough to drive the 6N7 - I examined in depth using 6AM4 at 10mA and that is really not enough but just enough as a demo with tiny valve big valve. 6N7 at 6mA isn't going to do it.

However 300V at 12.5mA and -2.5 volts on the grid works - yields +/- 87Volts of voltage swing and should just work to drive the GK-71 at say 525Va and Vg=-75 volts and Ia=90mA.. 5K load yields about 9Watts.

[IslandPink]

Thanks for the info on HT Phil .
I'd still favour either a beefy pentode , or an Aikido, myself.
In each case you can get plenty of gain with only one coupling cap, and no low-end time-constants ( eg. from chokes or cathode by-pass caps ) of any significance, in the driver.

Pentode could be EL86 or 4P1L running 25 to 30mA .
HT for those could be 450V or so .
For the Aikido , 5751 or ECC83 driving 5687 .
I would love to hear a 6SL7/5687 Aikido, though .
Aikido would need maybe 350V .

[IslandPink]

Interesting....yes...
Maybe a bit lower on current, 30ma, and G2 at 150V . I know the data isn't on the sheets, but Thorsten reckoned as a driver ( pentode ) you should typically keep the G2 at or below the lowest swing of the anode from it's nominal position.
Plenty of headroom on dissipation.
I assume you have some ?

[IslandPink]

Hang on a sec.. what's going on here
You just posted about 6V6 .... and I'm replying to that ; but your post disappeared
Help !

[pre65]

Hang on a sec.. what's going on here
You just posted about 6V6 .... and I'm replying to that ; but your post disappeared
Help !

I realised that it was most probably not enough gain, unless you know better ?

[IslandPink]

Ah, moving target !
In pentode mode, no bypass cap, 35ma, but screen shunt current ( OA2 ~15ma) going through the cathode , you should get about 20 gain ; with cathode by-pass cap, about 30 gain .

[pre65]

Moving target indeed !

I'm still wandering around (like a tit in a trance ) trying to decide on a driver valve.

I may be leaning towards the 12GN7a (12HG7) as a pentode, but without a cathode bypass cap.

Gas tube (150v) on G2 as per previous suggestions.

Can one use an anode CCS with a pentode ?

[IslandPink]

Sounds good apart from the CCS idea. Why do you want to use a CCS ?
Normally not a good idea, only application I've seen for that is JC Morrison using a CCS and a high-value load resistor in parallel to achieve insane amounts of gain in a phono stage .
The anode resistor sets the gain in a pentode stage : by-passed gain is approx Rl x S ( transconductance ) . You will get lots of gain with a 12GN7 , although it'll reduce to about 2/3 of that value without the cathode by-pass cap, assuming the OA2 shunt current also goes through the cathode resistor .

Have a look at Steve's circuit for the 6B4G monoblocks, that used 12HG7 pentode driver . Anode load may be about 4K and I'd guess the cathode resistor somewhere around 60 to 100R . You'll probably need about 400V for the supply on this, should be OK though , you'll need to drop down from the main HT with a smoothing stage anyway , to keep the noise into the pentode minimal .

[pre65]

Looking at Pete Millett's site I think I will stick with the 12GN7a but as a triode with CCS load.

Voltage gain (with unbypassed cathode) is 40 ish.

http://www.pmillett.com/pentodes.htm

[IslandPink]

OK, sounds like a plan . Current is good, gain is good ( thanks to the CCS ) . You've avoided a by-pass cap, but there's a CCS in the circuit. I haven't had much luck with them sounding natural, but try it out and see how it goes .
Watch out for dissipation in the CCS if you're running 30mA . Choose the heatsinking and supply voltage with some consideration ; dissipation for the unit will be current x voltage-drop .

[little eddy]

Perhaps a couple of silly questions.

Watching this thread with interest but personally would not like to exceed a B+ of 500V, perhaps 550. Many circuits with lowish B+ have a grid bias of 45 to 50V, but all but 1 schematic I found fall into two categories:
1.Self bias using big resistor
2.Fixed bias using IST

If I had a bias voltage available, why couldn't I use a de-coupling cap and fixed bias via a grid resistor to the bias supply (maybe 10R resistors from cathodes to 0V)?

If feasible, this way I could get increased VA and output power.

Also regarding the gain stage/driver, a 12gn7 often cited might run at 28mA according to the data sheet. A good driver should have low output resistance, large voltage swing capability but also high current if A2 were to be considered. Does anyone run the GM70 into A2? Many designs using ISTs often run at lower currents in the 15 to 20mA range (I guess due to IST current ratings). Would this work in A2, or are they limited to designs where there is sufficient B+/Va so as not to need to consider A2 operation from an output power perspective?

[pre65]

I've got the basic layout sorted now, it's a 3 piece amp.

two monoblocks and the HT supply.

HT is a single mains transformer, hybrid bridge with 6D22s damper diodes, then split into two with 1.5KV choke and Mikes fetreg (e-choke) per channel.

The psu for GK-71 filament and 12HG7 DC supply are on each monoblock.

Here is the pre-production picture.

[Paul Barker]

IST