[Amps] IM distortion and such

[Tom W8JI]

> I'm not refering to Eimac but the numerous other suppliers 
> and manufacturers.

Richardson's looked at the AL80B manual also. They had no 
problem with it. They own Amperex.

Anyone with a basic understanding of tubes should know a 
thoriated tungsten's life isn't compromised in the least by 
brief excursions over the CCS current rating. IMO all the 
fuss about this is a personal issue. It has nothing to do 
with what the manual says or any other technical problem.

> I should have worded the above differently and was my 
> mis-wording. It should have said deviate from the curve, 
> or be outside the curve. Are you saying Eimacs curves are 
> incorrect and one can operate the tubes outside the 
> recommended curves? This being any tube, not just the 
> 3-500Z. I'm not asking about any products by Ameritron or 
> the 3-500Z question from Rich. This is about all the 
> published information and curves that Eimac has 
> published.>>

Eimac and any other manufacturer can't possibly anticipate 
every operating condition. Their data almost always came 
from a single group of tests in a test fixutre and assumed 
more of a commercial application. Because they don't publish 
a particular set of operating parameters doesn't mean those 
conditions won't work or are unacceptable. Anything I have 
ever done that was outside a clear boundary has always been 
run past engineers at the primary source.

For example I designed a  medical pulse amplifier using two 
MRF150's. The peak pulse power was 1200 watts or more. 
Motorola had no problem with it and out of hundreds of 
generators there was never a failure of FET's.

>>It is impossible to look at one constant current line and
>>predict transfer characteristics or odd-order IM 
>>distortion
>>levels. If the constant current characteristic curves at 
>>one
>>current value set the distortion level why would negative
>>feedback reduce distortion? How could the output be linear
>>when the tube is cut off for more than 140 degrees of the 
>>RF
>>cycle?
>>
>>It is a transfer function problem. The level of IM
>>distortion products is a function of the slope and shape 
>>of
>>the gain change as drive level is varied.
>
> I understand this, but isn't the published curves there to 
> show the maximum operating areas?

Constant current curves primarily allow the designer to 
establish a load line that presents the proper load to the 
tube, and of course can be used to esitmate bias and power 
output. The end result should always be tested. That is the 
only final and accurate way to determine IM.   The 550mA in 
the later tune step has been the focus for many years by one 
person, while the manual's suggested reduction of drive to 
400mA steady current in the next step has been ignored. The 
reason 550mA is chosen is that happens to work out to 
establish a load line where peak clipping won't occur, it 
increases negative feedback by reducing anode load 
impedance, and it optimizes IM performance with drive 
characteristics typical exciters.

As for distortion products, even at 1000 watts PEP out a 
single 3-500Z in GG is significantly cleaner than nearly any 
modern solid state rig. First, Eimac and most commercial 
manufacturers use dB below one tone. Radio manufacturers use 
dB below PEP, making them look 6dB better than they really 
are. Many of the modern radios we use are barely 10db over 
what a class C PA would do, and that's with a steady tone 
test where the lousy ALC systems and bias fluctuations on 
transistor base bias systems don't come into play.

I can't even use a regular modern amateur transceiver to 
measure IM in many tube amps. They are too dirty, and that's 
when the test is a two tone and ALC attack isn't an issue. I 
normally have to either dig out a modified KWM-2 and run it 
though a class A amplifer, or run two RF generators into a 
combiner and amplify up to drive power levels through very 
clean driver amps.

If someone wants to get on a campaign to clean up the 
airways they better do some IM testing on radios and on some 
of the solid state and tetrode amplifers being sold. 
Actually they should test everything before they publically 
complain about it, and not go off attacking based on a wild 
guess.

>>If your read Care and Feeding of Power Grid Tubes 
>>(published
>>by Eimac) all of this information is there. They even say
>>they intentionally design to create an even-order 
>>curvature
>>slope to avoid creating odd-order IM!
>
> I've read it front to back.

Respectfully then, you should re-read this section:

>>I suggest you look at page 96 section 4.3.2 of Care and
>>Feeding. It's all there in black and white and worth
>>reading. It will help sort unfounded opinions from the 
>>true
>>facts. The ideal transfer characteristics for power
>>amplifier conduction angles less than 360 degrees and more
>>than 180 degrees (class AB) aren't even a straight line!

Please point out anything there that disagrees with what I 
have said above. It's the transfer characteristics that 
matter, and a straight line is not always optimum in an HF 
PA.

73 Tom