22w, 55mA...ok for el34 cathode biased?

[iknowjohnny]

Or do you think it will eat tubes like this? I have 450R one each cathode which should be the same as 1/2 that (225) for a shared R. What do you say? Bias even lower or can i get reasonable life out of a set of winged C's like this?

[Andy Le Blanc]

EL34\6CA7 have a max plate dissipation of 25w..... plus 8w dissipation in the screen grid....
that said...... its good to err on the safe side.... and the cathode method of
bias is the safest and recomended method for the longevity of the tubes life....
I have rig with jj's thats cathode biased and very happy at 27w plate dissipation
but I wouldn't push it more than that..... its based on fender tranny's (50w)with a 200ohm bias resistor
make sure your method of measurement is correct and that the math is good......

[iknowjohnny]

Well, my math is horrible. It's my Achilles heel. But i had several people do the math for me with my measurements and it seem to be at 22w, so i guess i'm fine if you're at 27. Thanks

[FunkyE9th]

Isn't it usually good to not exceed 70% of max dissipation for class AB? I suppose if you like how it sounds where it's at, then that's what matters.

[Jana]

he's biasing class A, 100%. You can't bias into class AB with cathode bias. Well, you can, but it doesn't work very well.

[FunkyE9th]

The way I understand it, you would have to bias the tube at a lower power in order to ensure that a push-pull class A amp stays in class A. Otherwise, one tube will go into cutoff while the other swings to maximum. Are you sure one tube is not going into cutoff while the other tube is at maximum swing? Not trying to argue...just trying to understand things better. I know there is a lot of debate on what amp is considered A or AB. So amps that people consider "class A" is not really class A, but just a hotly biased AB (e.g. AC30). I think a lot of it stems from the assumption that cathode bias equals class A.

[Andy Le Blanc]

the 70% thing is an insurance against crossing the static plate dissipation
boundry....... ever........ on paper..... its not a bad thing when dealing with a
fixed bias...... where reliability meets the expectation of loudness.....
but jeez...... the design maximum rateings where in place long before......

now class definition is the fun stuff..... one way is to look at the conduction angle
class B is strickly 180 degrees and class A is 360......

the conduction angle is 2 x cos-1(-IPq/IPpk)...... this may scare you but its not that bad

IPq....... is the static idle current the one you measure for bias

IPpk......is the maximum power current or peak current.....
so if your doing output testing this is the current (measured the same way as bias) you get
that coincides with the largest voltage you can measure across the dummy load at the output....

now...... its a ratio......and if it equals 1.... its class A...... if its 0..... its B
and it gives a good way to judge the AB region......

pg. 387 in RCA technical series RC-28 gives the class A1 ratings for the 6L6

so the zero signal plate current is .054 A and the max is .066A
which is .8181 ....where their class AB1 has a static or zero-signal of .116A
and a max of .210A...... or .5523 ...... which is near the middle of AB
there are recommendations for this...... with the lowest distortion being found
at ratios between 1/5 and 1/10 or .1 and .2 or more toward class B

please notice that the class A is not strictly class A by the definition....

[FunkyE9th]

Don't mean to hijack the thread... Why is the 70% thing only applicable to fixed bias?

[Phil_S]

A not so technical answer...

The 70% thing applies to fixed bias because the tube is regulated by the negative grid voltage, which is constant (fixed). As you push the amp, circuit demands will change, and that 30% is needed to accommodate those demands and output watts will vary some. Without some "unused" watts, the amp would sound like poo as soon as you inject a signal of any significance and your tubes would last a very short amount of time.

In a cathode biased amp, the tube is regulated by the cathode resistor. As demands are made, the tube is said to be self-biased because the cathode resistor will allow the tube to balance the plate voltage-mA draw so the wattage is relatively constant. For this reason, the idle static plate dissipation in a cathode biased amp can be close to 100%.

Andy will you please correct for any seriously misleading statements? I can't do the technical stuff the way you can.

[jjman]

I possibly less technical guess:

The presence of a cathode resistor provides “curtailment” to the AC (output signal swing) as well as to the DC (bias current.) However, the absence of a cathode resistor, (with negative DC-grid bias, aka “fixed” bias) presents no “curtailment” to the AC signal output. The second has nothing providing resistance to the AC component of the cathode’s current path.

The cathode resistor is “biased against" the AC-signal as well as the DC-idle.

What I don’t understand is when you add a cathode bypass cap. Under this approach why does the 70% rule not apply as with fixed bias?

........because in AB cathode bias, a stronger output increases the (average) cathode current which increases the cathode voltage which "curtails" the output. I think that's it. Like a govnah!

[FunkyE9th]

If I understand correctly, without the cathode bypass cap, you get a type of negative feedback (i.e degeneration, the bias voltage changes as input voltage at the grid varies), so the gain goes down. But just because the gain goes down, it does not mean the maximum swing of the PA changes. It just takes a bigger input to get to max swing. But when you add the cathode bypass, you get higher gain, so less input is required to get to max swing, but max swing is still the same. If I understand correctly.....

[iknowjohnny]

That explains why when i set my switch (i added a switch to cut out the bypass caps) to no caps, the output volume is reduced considerably.

[Andy Le Blanc]

the 70% rule comes from the graphical examination of operational characteristics
when plotted over your average plate characteristics for the given tube type
the first thing you do is observe and plot the recommended limits on plate dissipation....
so that when you next plot your intended load line..... you can see if, or how
significantly they intersect..... by plotting the dissipation limit at 70% you
insure that the dynamic operation of the tube ( which graphically forms an
elipse around the linear load line because the load is reactive as well as resistive)
does not impinge upon the limits of static dissipation...... this means less distortion in practice
and longer operational life of your tube....
this is more significant with fixed bias because it magnifies minor differences
in tubes used for push-pull and does not have the built in degeneration that a
cathode bias has..... so when a tube goes in to a runaway condition its very destructive...

its a good arguement.... but a guitar amp is not ment for HI-FI....
distortion is the tone color your seeking..... and with a simple rig your better off going balls to the wall
but if your investing in a quad of tubes and looking for a hundred watts
with a clean responce with a master volume it would pay to conservative
its the applcation that counts

[Andy Le Blanc]

to futher describe the amp mentioned early in the thread.....
its a cathode biased push pull amp with jj el34 and fender replacement tranny
50w rated NSC 018343 opt..... NSC 022798 power...... 5u4 rectifier

B+ today was running at 398v..... the winds been blowing and theres been
an outage so the power is a bit off today.... it usually runs hot at the wall
but not today....

the amp was tested with an 1k HZ 200 ma. input

so...at idle plate voltage was 373v (measured from plate to cathode)
with static dissipation of 132 ma. for two tubes (measured across a 1 ohm between the cathode resistor and ground)

373 V x .132 A = 49.2 W (for two tubes) 24.6 W static dissipation
a bit off today.... last test yielded 27W

with an 8 ohm resistive load the signal was applied and the amp produced
an absolute maximum voltage of 18V at the output across the load

(18 ) 2/8 = 324/8 = 40.5 W (RMS)

at which time a max signal plate current of 199 ma. was measured...

with a CA (conduction angle) of 263 degrees

2*cos-1(-IPq/IPpk) = 2*cos-1(-132/199) = 2*cos-1(-.6633)

2*131.5520 = 263 degrees

this of course is not strictly class A....... but it is toward the "A" side of "AB"
the good thing about this is that it gives the home builder a yard stick to
judge where the amp is opperating...... the results will be different of course
if I were to measure for undistorted output.... but the amp in question uses
a paraphase inverter exactly for the early onset and color of the distortion
that it produces

I like the number that the math gives you..... its more valid to have a reproduce-able test with a result..... that tells you where in the AB region
it is working.... instead of jumping up and down about whether or not its truly "A" "AB" "AB1" "AB2"..... just need to get a consenus of where these
conditions reside between "1" and "0".....