# What Plate Dissipation is - something doesn't seem right

Discussion in 'Amp Tech Center' started by peteb, Jul 28, 2017.

1. ### Old Tele manTele-Holic

May 10, 2017
Tucson, AZ
Sorry, but it *IS* all of them. All those descriptions truly are valid, but they're just individual instances described (and pulled from) a simplified description of what's going on.

We measure voltage and amps, not heat, so ALL descriptions of heat dissipation within a tube are necessarily constrained to only those values and Watts.

As I described earlier, there's DC-power at IDLE, and AC-power when producing output:

IDLE - plate dissipation is DC-power, so: Ppd(dc) = Vp(dc) * Ip(dc), and since there's no AC-power, then Ppd(dc) = DC-power input...ALL the wattage is wasted as heat.

POWER - plate dissipation is INPUT power minus the OUTPUT AC-power being transferred thru the output transformer into the speaker load; so. now, not ALL the power (wattage) is being wasted as heat as it is at idle.

The ratio of how much power the tube produces from the power put into it is its plate efficiency:

P(eff)% = 100%*(P.out)/(P.in)

Last edited: Aug 12, 2017 at 3:42 PM
robrob likes this.

2. ### petebTele-Afflicted

Apr 25, 2003
I agree, plate dissipation means different things at different times.

Old tele man, maybe you always knew that when you measured idle plate dissipation, you were measuring the input power, but how many other TDPRIers know that?

Most in the amp tech communities consider plate dissipation as the wasted heat, and some how they are measuring it when they measure the idle plate dissipation power, but this is incorrect, what is actually being measured is the input power.

3. ### Old Tele manTele-Holic

May 10, 2017
Tucson, AZ
• My father was a retired Chief Electronics Technician, USN, who got me into vacuum tube electronics -- I built my first 40W amp from schematic in back of RCA Tube Manual.
• Spent almost 8 years is USN, operating, servicing and repairing vacuum tube airborne radar systems after attending both intro, intermediate, and advanced electronics schools.
• Taught electronics both in USN and as adjunct college instructor at Arizona Western College, Yuma, AZ.; authored automotive and computer science articles and books.
• After USN, returned home and completed my electronics college degree.
• Then spent 28 years in aerospace/military employment before retiring.

So-o-o-o-o, for ME, I always knew that plate dissipation was a two-component transformation of power (wasted & useful) into heat process. Hence, all my earlier comments about DC-power and AC-power; IDLE power(wasted) and OUTPUT power (useful); and, the approximation equation for their combination (in my post #16):

"Average value of combined zero- and maximum-signals is approximated as:

X(avg) = (Xq/2 + ∆X/4).

NOTE: ∆ represents AC-change in voltage (VAC) or current (IAC) and q denotes quiescent or idle values (VDC or IDC)."

Last edited: Aug 12, 2017 at 9:59 PM
Bendyha likes this.

5. ### RLee77Tele-Afflicted

May 15, 2016
Silicon Valley
You have discovered that the whole equals the sum of the parts. I suspect that is not a revelation to many in this thread. But it is none the less a valuable thing to understand.

Speaking of parts, consider that anything in the chain (only considering the power stage at idle and ignoring all else) between B+ and ground also dissipates power. The OT primary and cathode resistor (if cathode-biased) waste power as well, so plate dissipation does not quite equal input power, but it's close.

Happy to see you are learning useful things from the rca manual, expanding your understanding.

Mar 26, 2014
Northern Germany
I still think post #3 has all the information to answer post #1
Maybe you can understand it now ?

Most of the confusion in defining any topic of discussion is the sloppy use of terminology.
I'm not saying you are the only poster guilty of this, or that I am allways spot-on, but unfortunatly you do tend to jump and twist things quite a a lot, and leave out defining disambiguating words, so in a very breif exchange things can shift from AC to DC, from half wave to full wave, from idle to driven conditions, from the whole circuit to one component, from average to momentary, and the main one; dissipated wattage to output wattage. This means some people are responding to one statement you make, and others to another, and seeing how you often post opposing or differing claims next to each-other, or even intertwined, sometimes combining correct and false statements, then it is easy to see why you think that we are confused. But...
....that you are the one who is confused.

Correctly worded, many of you very simple questions can be very simply answered.

The main problem for most of us, seems to be, that you are taking a subject that seems so blatantly obvious, and is bound up in fixed constants and the unchangable nature of the components, so as to be just a fact, and the redefining of it is unrelated to the development of an understanding of the design or building of amplifiers.
Basicaly, all we are dealing with here is Ohm Law, which we work with, and the Laws of thermodynamics, which we live with.

You say;

Well maybe ... and that is a totally acceptable practical analysis, thinking that will not set your foot too wrong.

But I don't think the concept of "wasted heat" is what anyone thinks, have you ever heard " I have set the tube to waste 180°C ", who measures heat ?

As to a technically correct definition, compared to the way that it is used in praxis, well everybody up to now has seemed to agree on the standard use of the term "plate deffinition" and it has caused no problems, so why try and change it to some text-book definition. And then when and where does one stop this praxis ?
It's like when someone saying "look at that rainbow", and then trying to explain to them that it is impossible to look at the same rainbow, because everybody sees there own unique rainbow......we have great words for people like that over here in Germany.

Most, if not all of us, consider plate dissipation as product of Amperes and Volts that can be measured across the tube at idle conditions, and use this as an indicating level to show us how the tube is biased, the dissipation itself is the part of the measuremenrt that can be used as an indicator of where within the recomended maximum dissipation limits the chosen bias setting has been placed.
If and when we are measuring this product of Amperes and Volts across the tube at idle conditions, then this IS the input power, there is no signal input, so that is what we are measuring, and it is what we think we are measuring. Having measured Volts and Amperes, we have also got the Power measurement....and what is happening with this power when no signal is being amplified.... it is at idle... all it is really doing is giving off heat.

You have, I believe, tried to draw a correlation between the input signal swing, or the output power, and idle dissipation power.
That the input signal changes the momentary dissipation is true, but the limiting figure for the maximum wattage heat dissipation is taken as an average measurement, because variations in heat dissipating is much slower and more constant than the signal variation.

That the signal can change this average...in Class A...not really all that much(excluding due to power supply sag caused by the whole of the amps current demand draw changing) the maximum average dissipation is set at idle, it will useually drop under signal conditions. (there are exceptions)
In Class AB this is not the case, or at least not the same, with low power signals the tube is in Class A and the dissipation usually remains fairly constant, but this will change when the gradual slide into the Class B happens, with the dissipation rising, but more input signal may push the tube further and the plate dissipation may drop again, but screen dissipation will keep rising. Without seeing the knowing the exact loading,bias setting, voltages and current draws of the whole circuit, a more exacting prediction of the dissipation levels or output capabilities, or efficiency, can not be made.
......A change in filament voltage is a change in input power that changes dissipation in several ways, and output power.
......Screen-grid current limiting resistors will change dissipation, more in the dynamic state than the static.
......So many variables, many of them interrelated, some cumulative, some opposing...

So is there a direct fixed correlation between signal input and idle plate dissipation ?....No.
So is there a direct fixed correlation between output power and idle plate dissipation ?....No.

So is there any advantage to exclusivly defining plate dissipation as a measurement of input power, rather than seeing idle bias plate dissipation as wasted heat ? Are there any conclusions you can draw from your new isight ?

I'll let you answer that one peteb.....

Last edited: Aug 12, 2017 at 10:02 AM

7. ### robrobPoster ExtraordinaireAd Free Member

Dec 29, 2012
United States
It seems petep may also be confusing input power (idle, no signal plate voltage & current) with AC signal voltage on the grid.

8. ### Old Tele manTele-Holic

May 10, 2017
Tucson, AZ
RobRob -- Can you (1) overlay, onto that 12AX7 Eb-Ib plot that you posted above, the Ip.q*Vp.q "rectangle" that represents the DC-power input at idle? That might help peteb visualize that at idle there's power going into the tube, but nothing USEFUL coming out (yet).

Next, (2) throw a 100kΩ plate-load line onto that graph (omit output loading for now) showing how that load line intercepts/defines the X- and Y-coordinate 'corners' of that input (idle) power rectangle and mark the idle spot with a dot or circle.

Finally, (3) overlay a thicker line 'on' that load line showing the Vp(ac) movement on the load-line up & down from the idle point.

Then, maybe, peteb can "see" for himself how the much smaller Ip(ac)*Vp(ac) "output power" dissipation 'rectangle' area is within that much larger idle-dissipation 'rectangle' area...and...possibly grasp the concept of wasted (DC) and useful (AC) power dissipations occurring within the tube at the two different operating conditions of idle and full power?

GOAL? (1) to show that DC-power is when NONE of the plate dissipation is producing a useful output (ie: no AC-power); (2) to show how little the output power is in Class-A operation; and, (3) to illustrate how AC-output is simply the diversion of wasted heat from the plate into some OTHER load, be that a plate resistor (small-signal) or output transformer (power).

A constant glowing lightbulb conveys no signal, but a lightbulb varying in light/heat output does, especially if the variation contains information such as speech or music! And, a tube is little more than a voltage-controlled hi-tech lightbulb.

Last edited: Aug 12, 2017 at 4:34 PM
RLee77 and Modman68 like this.

9. ### RLee77Tele-Afflicted

May 15, 2016
Silicon Valley
And perhaps dispel the misconception that average plate dissipation remains constant as the stage is driven to full signal power.

10. ### Old Tele manTele-Holic

May 10, 2017
Tucson, AZ
The last Operation Characteristics plot, "Peak Input Signal Grid-to-Grid," found on page 9 of 9 of the GE 6L6-GC data sheet, dated 3-59, definitely dispels that misconception:

http://tubedata.milbert.com/sheets/093/6/6L6GC.pdf

It shows both plate-dissipation and screen-dissipation (per tube!) for the Class-AB1 55W example as input grid-to-grid drive voltage is increased from 0Vg (idle) to 85Vg(pk-pk)...clearly showing individual plate dissipation dropping and screen dissipation climbing.

Last edited: Aug 12, 2017 at 10:04 PM

Mar 26, 2014
Northern Germany

RDH p.26.

When a resistance-loaded triode is operated under steady conditions, the power dissipation is indicated by Fig. 2.20.

The area of the rectangle OCDB represents the total power (Ebb Ib0) drawn from the plate supply.

The area of the rectangle OCQK represents the plate dissipation of the valve (Ebo lbo) and the area of the rectangle KQDB represents the dissipation in the load resistor (Ebb — Ebo) Ib0 = Ibo2RL

Under dynamic conditions the plate dissipation decreases by the amount of power

output, and the load resistor dissipation increases by the same amount, provided that there is no a.c. shunt load and that there is no distortion.

RDH p.559

The plate circuit efficiency is defined as the ratio of the maximum signal frequency power output to the d.c. power input under these conditions. (Owing to rectification, the d.c. power input under maximum signal conditions is usually greater than with zero signal input).

The plate circuit efficiency cannot exceed 50% in any Class A amplifier, and usually does not exceed 25% with Class A triodes. The plate circuit efficiency in an ideal Class A triode is given by n = RL/(2RL + 4rp)

The plate dissipation in a Class A triode is greatest with zero input signal. If there are no rectification effects, the power input remains constant and the power output is simply power transferred from heating the plate to useful output. With normal rise of plate current for 5% second harmonic distortion, the plate dissipation is more nearly constant, but it is still sufficient to base the design on the zero-signal condition.

RLee77 and Old Tele man like this.

12. ### Old Tele manTele-Holic

May 10, 2017
Tucson, AZ
...and, if the shaded area is increased to the A-to-B(Ebb) rectangle area, which is the input power (Pi) from the power supply, one can see that plate dissipation and load resistor dissipation (Po) are truly LESS than HALF of the input power (Pi) in accordance with the plate efficiency equation:

Class-A efficiency(%) ≤ 100%*(Po / Pi)

...which, in the RDH example above, is somewhat less than 25%, reinforcing Bendyha's statement.

IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.