Using tube load lines with plate chokes and other load line questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,716
Location
Helsinki, Finland
I have recently learned the basic usage of tube load lines when setting up a common cathode triode stage. I bumped into the fact that I must select plate resistor about 2-5 times the value of internal plate resistance of tube for desired gain and impedance.

http://www.aikenamps.com/CommonCathode.htm

With some tubes it's difficult to get low THD and low output impedance without ridiculously high B+, so one (expensive) choice is to use a plate choke to replace the plate resistor. As far as I understood plate now sits at B+ (or near) since there's no voltage drop from Rp.

How does this affect the usage of tube load lines, since they assume a plate resistor? Where do I fit the choke into the calculations in the link above, and how to bias the tube with it?

Thanks,
Mike
 
Same as transformer-coupled (see which).

The choke DC resistance plots as a near vertical line. Run that up to your desired current, put a dot.

The ideal choke AC impedance is infinite. This plots as a horizontal line through your dot.

In any practical case, you have a load. Maybe 1Meg and 30pFd stray C. Maybe something else. This plots as "in parallel with" your choke impedance. If your choke is any good, its impedance is higher than your load. So don't bother plotting the choke, plot the load through the dot.


> difficult to get low THD and low output impedance without ridiculously high B+, so one (expensive) choice is to use a plate choke to replace the plate resistor.

Low THD goes with low signal level, particularly small ratio of signal current to idle current.

Low output Z on a triode happens near zero grid bias and plate voltage raised almost to the dissipation limit. Fairly low tube voltage and HIGH current. This also helps keep THD down. Using a resistor to a higher voltage lowers both gain and output Z. Choke-coupling will generally give higher gain and output Z.

THE way to low output Z is a hot fat lazy triode. Or NFB, but that only reduces small-signal Z, not large-signal Z.
 
PRR said:
Same as transformer-coupled (see which).

The choke DC resistance plots as a near vertical line. Run that up to your desired current, put a dot.

The ideal choke AC impedance is infinite. This plots as a horizontal line through your dot.

Thanks for the info. I suspected this but had to make sure.

PRR said:
THE way to low output Z is a hot fat lazy triode.

Cheaper as well.

I acquired some cheap Russian 6N8S "6SN7" and the root of my above questions was to try to mold it into a graceful driver for 10k/600 output transformer (that's the only one I have available). Not gonna happen. And plate chokes are too expensive.

But I also have some 6922 and ECC99, and either of these will drive it with ease.

Pre/line amp preliminary plan:
1:7 input transformer to two stages of 6SN7 to one half of of the above driving 4:1 output. very little or no feedback, three bottles for stereo, not insane heater requirements (but quite high).

related question, what's a ECC99 or ECC88 dual triode equivalent in a single triode package (for just one channel of the above)? Socket type unimportant.
 
PRR said:
Same as transformer-coupled (see which).

What did you mean by this?

I guess single ended where transformer primary replaces, or at least augments plate resistor.

Or maybe parafeed, which I understood is a transformer with plate choke-like winding which is cap coupled to a separate impedance conversion winding that doesn't take DC.
 
Heya  :)

JJ make a new-production ECC88 - called the E88CC.  Substitutes are 6922, or 6DJ8.  Get this little piece of marvellous software, it'll help a lot!
http://tdsl.duncanamps.com/
"Tube Data Sheet locator" - also gives you subs for tubes.

Remember that chokes and transformers have a max amount of DC current, if you're putting them directly into a high-current-for-low-THD situation the output transformer is going to get pretty big!  You get around it by cap-coupling;  if you're cap-coupling your output transformer anyway, why not feed it from a cathode follower? Easiest way to low output Z I know of.

HTH  =)
 
lucidtone said:
JJ make a new-production ECC88 - called the E88CC.   Substitutes are 6922, or 6DJ8.  Get this little piece of marvellous software, it'll help a lot!
http://tdsl.duncanamps.com/
"Tube Data Sheet locator" - also gives you subs for tubes.

Very familiar with those already, but as long as the Russians and Ukrainians provide dirt cheap cold war military NOS tubes extremely cheap, I won't bother with anything new production.

I decided this time I won't go with ECC88/6922 or 6N23P (direct ancestor, probably sounds better), or 6N1P (higher plate resistance, not so great for low impedance). 6N6P seems like the right tube, which is somewhat beefier than JJ-ECC99 for example (even lower plate resistance, 8W max dissipation!). I think this is the "hot fat lazy triode" to use. Could even drive a 2:1 output transformer with ease. So linear it makes plate choke pointless.

lucidtone said:
if you're cap-coupling your output transformer anyway, why not feed it from a cathode follower? Easiest way to low output Z I know of.

Because I don't want to waste gain in this particular design and I could instead boil some eggs and make coffee with 6N6P.
 
Another related question.

In this picture I have biased a tube to a good usable range. I know what happens when tube runs out of voltage: distortion, and sometimes a good sound.

But what happens at the question mark, when tube runs out of current? What happens to the sound? Some kind of sharp clipping?

index.php
 

Attachments

  • out-of-current.png
    out-of-current.png
    28 KB · Views: 203
Is that not where blocking distortion occurs?

Things might be getting quite crunchy before then, depending on how you set up this stage.
 
I'll file this for a future experiment then. In the above case I would have a good answer with about 30k plate resistor, then try 600ohm, 1.7k and 3k cathode resistors. will overdrive and hear what happens.

Interesting: http://www.aikenamps.com/BlockingDistortion.html

seems like it would make pretty sharp clipping with a recovery time constant. I think I recognize the sound from those descriptions.

further, I find it odd none of the solutions proposed by the above link mention changing bias point, or going for higher B+...
 
> what happens at the question mark, when tube runs out of current?

The tube does NOT "run out of current".

If Grid is negative of cathode it draws "no" current". But if Grid is positive of cathode it draws "large" current.

The transition from "no" current to "large" current puts a severe kink in whatever is driving the grid.

Imagine a child's swing near a mud cliff. If the swing does not touch the cliff, it swings freely. If the swing hits the cliff, it can "swing" into the mud, but it is hard work. The graceful swing motion becomes a slamming.

For most audio we don't want to do this.

The 6J6 is specified for positive grid use; see attached image.

For negative or zero grid, there is no grid current, and we can swing to say 80V 10mA.

If we force the grid positive to say +6V we can do 80V at 50mA or 20V at 30mA. BUT we must force 15mA-25mA into the grid to do this. You can not do this with a 12AX7! And you can't do it with R-C coupling to the grid: it will de-bias, "block".

Because there is such a lot of increased current possible, some older and some hi-power amps do force the grid positive. This usually means another power tube and a transformer to do the forcing. The 300 Watt Fender used a 6V6 to drive the output tube grids. The early SVT used beefy cathode followers. If you need 25,000 Watts of audio to modulate an AM radio transmitter, class B2 triodes are a handy method. In almost any "sane" audio situation, you do not want to go there.
 

Attachments

  • 6J6-6V.gif
    6J6-6V.gif
    14.7 KB · Views: 33
Thanks,

the above link seems to call it class A2. 2 for marking the part where grid current flows. Or AB2.

I will test the "slamming motion" effect with the next p2p preamp project I build. Won't even try any A2 tricks as it seems to be a power amp thing and tubes I use won't be specified for positive grid.
 
Back
Top