bias: cathode current vs plate current

[pompeiisneaks]

Oh also, another reason to not worry about the screen c urrent is it's usually in the 5mA range or so, and thats 'added' current, so if you act like it's not there, you're biasing more 'safely' Say your plate current is 45mA and your screen is 5, you're reading 50mA instead of the real 45. This means you aim for 70% diss, you're really a few points below that and safe. OR if you end up at 72% you're still in the safe zone likely. Most people just don't worry about subtracting it out.

~Phil

[martin manning]

I’ve never understood why amps aren’t built with 1R resistors at the plates to measure plate dissipation directly.

Due to the high voltage, it becomes expensive, you need a high quality 1% resistor. Depending on the tube, let's say it's a 6L6GC at 350VDC. That means at 70% diss you need 60mA so 60mA times 350VDC = 21 Watts. It's best to go up, so 25 watt 1% resistors are EXPENSIVE!

It’s the safety issue. 60 mA through 1 ohm is only 3.6 mW.

[pompeiisneaks]

I’ve never understood why amps aren’t built with 1R resistors at the plates to measure plate dissipation directly.

Due to the high voltage, it becomes expensive, you need a high quality 1% resistor. Depending on the tube, let's say it's a 6L6GC at 350VDC. That means at 70% diss you need 60mA so 60mA times 350VDC = 21 Watts. It's best to go up, so 25 watt 1% resistors are EXPENSIVE!

It’s the safety issue. 60 mA through 1 ohm is only 3.6 mW.

I'm not sure I follow. P = V * I which is 350 * .060 which is how I got 21 watts, how do you get 3.6 mW from the calculations?

What's the safety issue?

~Phil

[sluckey]

I'm not sure I follow. P = V * I which is 350 * .060 which is how I got 21 watts, how do you get 3.6 mW from the calculations?

What's the safety issue?

There's only 60mV ***ACROSS*** the resistor, same as there would be if the resistor was in the cathode circuit. So, .06A x .06A x 1R is .0036W or 3.6mW.

The safety issue is that you have to connect the meter probes across the resistor. There will be 350V on ***BOTH*** probes! I can safely do it. Can you? Can shakey Joe?

[brewdude]

It’s my understanding that it only needs to handle the drop in voltage across the 1R resistor which will be relative to the current drawn by the tube.

However, the high voltage relative to ground seems to be why people don’t choose to use them on the plates—unadvisable to use test points outside chassis.

[martin manning]

Exactly. I would have no problem adding them inside the chassis. Anybody going in there is on their own.

[pompeiisneaks]

I'm not sure I follow. P = V * I which is 350 * .060 which is how I got 21 watts, how do you get 3.6 mW from the calculations?

What's the safety issue?

There's only 60mV ***ACROSS*** the resistor, same as there would be if the resistor was in the cathode circuit. So, .06A x .06A x 1R is .0036W or 3.6mW.

The safety issue is that you have to connect the meter probes across the resistor. There will be 350V on ***BOTH*** probes! I can safely do it. Can you? Can shakey Joe?

D'oh another light bulb moment, I knew this, but had forgotten, that not all of that voltage is dropping across that resistor, gah. I need to remember that, if it was 350 on the top and 0 on the bottom THEN that calculation makes sense.

Gotcha.

and yeah test points outside the chassis at 350VDC is asking for someone to get killed ;D

~Phil

[sluckey]

if it was 350 on the top and 0 on the bottom THEN that calculation makes sense.

No. That calculation is OK for the tube, but not for the resistor. If you have 350V at the top of a 1Ω resistor, and you have 0V at the bottom of a 1Ω resistor, then you have 350A flowing through the resistor! The power dissipated by such a resistor would be 122,500 Watts! For about a nanosecond. Then the resistor would disappear, possibly leaving only two nubs where the leads used to be. Hopefully that would be the only damage.

[brewdude]

I have been using precision 1R/1W resistors on the plates of my last couple builds. I like it and understand the voltage danger... and I build them for myself. I might even put some on the screens and the plates of my next build(?).

[pdf64]

Folks seem to obsess over things they can measure, without considering whether what they are measuring is particularly relevant.
eg tube testers, guys will check their tubes on stupid emission testers, discarding tubes that test bad which in reality would work just fine in an amp, whilst scratching their heads about the tubes that test fine but work like ass in an actual amp.
Same thing for plate dissipation - hey guys here's this thing we can measure and blow out of all proportion.
Even to the point of, when an amp's tech info notes the designer's intended idle plate/cathode current, to decide that the current corresponds to too low a plate dissipation and that they know better than the designer
Plate dissipation is just a limit, not an operating condition.

[martin manning]

Folks seem to obsess over things they can measure, without considering whether what they are measuring is particularly relevant... ...Plate dissipation is just a limit, not an operating condition.

I would say It’s both. It’s a design limit for the tube, but when you are targeting a percentage of that limit in use, it’s an operating point. All of the above is aimed at a more accurate setting of that point, through elimination of a 5-13% error. If you ignore the screen current and set 70% dissipation on an EL34, you are really only at 61 or 62%. If it’s in a cathode biased or Class A amp, and you are targeting 100%, you’ll really only be in the mid-80’s.

[pdf64]

I would say It’s both. It’s a design limit for the tube, but when you are targeting a percentage of that limit in use, it’s an operating point...

The thing is I don't get the rationale of targeting a percentage of the plate dissipation limit? Of course it results in an operating point, but only by default.
It seems a 'tube-o-centric' way of looking at thing, the tube is in an amp, it makes no more sense to target a particular plate dissipation than it does targeting a particular dissipation of some or other resistor in the circuit. (OK I'm exaggerating to make the point )

...If it’s in a cathode biased or Class A amp, and you are targeting 100%, you’ll really only be in the mid-80’s.

Why target 100%, what's the relevance?
If class A operation is the goal then need to be looking at relationship between clipping and cut off; plate dissipation doesn't seem to have any bearing on that?

[martin manning]

Maybe start here: http://www.aikenamps.com/index.php/idle ... 70-percent
As mentioned in the linked article, and to your point, the voltages and load impedance must be chosen appropriately.

[pdf64]

The scope of that page is kinda 'tube-o-centric', ie its purpose is to demonstrate that the 70% thing generally results in the power tubes remaining in their safe area.
Whether the operating point that comes out of that is appropriate for the amp is outside that scope.

I suspect that such an approach may, inadvertently, be leading people into nonsensical beliefs, eg -

1/that tubes have to dissipate a certain % of their limit in order for the amp to operate and sound as intended; I've even seen queries raising concern about preamp 12AX7 idling too cold!
2/ that they have the Aiken seal of approval green light to idle 6550A in their Marshall at 30 watts, 6L6GC in their tweed Fender at 21 watts, ie the tubes will be drawing more at idle than the amp's power supply etc may have been designed to deliver at full power.

Putting the tube's safe operating area before that of the amp they're in seems topsy turvy, akin to putting the cart before the horse.

[martin manning]

Pete, how do you set output tube bias? What are your criteria?