On Fri, 25 Jul 1997 18:26:59 +0100 Ian White <G3SEK@ifwtech.demon.co.uk>
writes:
>>>Carl, KM1H wrote (about 4CX250Bs and the like):
>>>
>>>
>>>>BUT what is the IMD if one runs say 1700V, proper drive and
>>>stabilized
>>>>screens?
>>>
>>>Worse than at 2000V for the same "attempted" power output; or
>>>alternatively less output for the same general IMD levels - see
>below.
>>
>>That was not what I asked Ian and I am sure that you know that.
>
>It looks like GW4FRX can make some more measurements of IMD to answer
>your question more specifically, Carl. What he proposes to do is:
>
>1. Set up 2*4CX250R with Va 1500V, Vg2 350V and adjust Vg2 to give
>ZSAC
>133mA, all as specified in the EIMAC datasheet.
Why on earth do you want to vary screen voltage Ian? That sounds like
one of Rich's ideas but is not real world practical.
Since we are dealing with just one type of tube, the 4CX250R, why not
follow Eimac spec sheets and adjust Eg for proper idle current? The
screens can be set at various steps as allowed by the popular gas tube
regulators or in say 25V steps if one wishes to use zeners which were
mentioned here also. I doubt if anyone wishing to build a minimal cost
amp would opt for a fully adjustable electronic screen supply. ....great
on paper BUT....
Also at no point did you specify what the test setup consists of.
>2. Adjust drive, tuning and loading to reproduce as closely as
>possible
>the two-tone operating conditions and output power in the EIMAC
>datasheet. The output power will be adjusted to the datasheet value of
>262W PEP per tube, and IMD levels will be reproduced as closely as
>possible (see note below).
Is this with NIB tubes? My experience is that most tubes will load to
their own best point. If you are going to run pure AB1 why not load for
maximum output at whatever voltage you are running? Heavier loading is
always preferred anyway.
Just set the bias for idle and tune for max. You will probably find that
the best output is reached with Is well under the specs. I would also
wager that if you INCREASE loading until output just starts to drop that
will be the point of best IMD. My feeling is that a tetrode should always
be tuned by watching the Bird meter and screen current only. All else
falls into place. The screen meter is also a much more sensitive
indicator of neutralization problems as compared to the plate meter.
I may be getting in a bit deep here but I have a feeling that Eimac specs
were not derived at best case for ANY particular voltage. I have always
found ANY tetrode in good condition to perform better at well under "
typical" screen current specs. For instance the RCA 8122 is typical at
10ma yet they run more efficient at 5ma per pair. Only when they are well
worn do I ever see the need to go higher to obtain peak output.
In VHF+ service we also have the negative screen current to contend with
in almost all ceramic tetrodes....another "X" factor IMO. At 6M this does
not appear to be an issue.
I mention the 8122 a lot since I have over 30 years experience with them
PLUS the Svetlana 4CX400A has almost identical E and I ratings (except
for filament).
>
>3. Increase Va to 2000V - touching nothing else - and measure the new
>output power and change in IMD levels.
>
>4. Re-resonate the output tuned circuit if required - but still
>touching nothing else - and again measure the output power and change
>in
>IMD levels.
>
>5. Adjust both tuning and loading to GW4FRX's own satisfaction, and
>see
>what we get.
>
See my above comments Ian.
>
>Notes:
>
>1. The tests will be done at 144MHz in a carefully neutralized W1SL
>push-pull PA. EIMAC did not specify a test frequency on the datasheet,
>but it might have been as low as 2MHz (see description of their IMD
>test
>setup in 'Care & Feeding'). We'll have to assume that the tank-circuit
>efficincies of the two test rigs are the same.
>
>2. It may be difficult to reproduce the original EIMAC IMD levels
>because they also depend on the power supply characteristics, which
>EIMAC did not specify. GW4FRX has found that stabilized screen
>supplies
>tend to give better IMD than EIMAC found. (This is possibly because
>EIMAC were using simple potential-divider screen supplies with an
>impedance of several thousand ohms, which represented the state of the
>art back in the 1950s when the measurements were made.)
Gee whiz, VR tubes were available too back then. Even I could afford WW2
surplus in the 50's on my paper route money.
>
>3-4. It may not be a fair test to simply jack up the anode volts
>without
>re-resonating the PA. We'll see what difference it makes.
Of course not, if you did that the results would me meaningless. Always
retune for max output and measure. ONLY then should you consider tuning
experiments to see if IMD goes up or down.
>
>5. The tubes will be significantly under-loaded at the higher anode
>voltage, so it's quite possible that the IMD will get worse in tests 3
>or 4. But if the loading is adjusted correctly, staying well on the
>heavy side, there should be a significant increase in power and the
>IMD
>could also improve.
Under loading will always produce inferior results Ian.
>
>
>What do you think, Carl - will those proposed tests answer your
>questions adequately?
With reservations as mentioned above Ian. Since hams...including
myself....tune for maximum output that is how you should run the tests.
Note I said output; I DID NOT say anything about varying driving voltages
to compensate for different bias levels to maintain the idle current
specs. I would guess that drive will also require adjustment to keep at
the same point of "just staying within AB1". In other words, at each
point the drive will require checking to make sure it is just below the
grid current point in key down A0. Only then will you be able to graph a
meaningful picture of the full capabilitiy of the tube. If you want to
vary the screen voltage in steps to see how it affects IMD at any
particular Ep, that would be a bonus but as I said earlier...not always
practical.
Thanks to TOM's suggestions I may be able to set up my own test jig soon;
at least for 250 type tubes. I do not like your idea of using 2 tubes in
PP, for the tests, at all but if that is all you have so be it.
To recap:
Use new or as new tubes
Vary Ep in some predetermined steps.
Set screen voltage as per spec sheet
Set idle current to specs by varying Eg
Adjust loading for maximum output and then continue to increase loading
until power just starts to drop.
Adjust drive to maintain zero grid current when in key down fully loaded
config.
Take measurements with those parameters and only then vary screen voltage
to test IMD. This also requires readjusting bias for idle current.
All in all this is not a simple 10 minute exercise.
I wish I could be there! Sorry to be so long winded also.
73....Carl KM1H
>Thinking further about the underlying question, which I interpret as
>"What happens to the IMD when you increase the anode volts on a
>tetrode,
>leaving the screen volts the same?", the true answer is: it all
>depends.
>Specifically, it depends whether you're allowed to re-adjust the
>loading, because the loading control can make the IMD levels whatever
>you like.
>
>I have to agree with Carl that if 4CX250Rs are operated under the
>2000V
>conditions shown in the EIMAC datasheet, they will sound absolutely
>rotten! That's because IMO the operating conditions do not give a good
>balance between power output and IMD: by loading for 470W PEP output
>per tube, they're being far too greedy. The IMD values quoted
>(3rd-order
>at -23dB below either tone and 5ths at -27dB) are simply not
>acceptable
>by today's amateur-band standards, especially not from a big VHF
>signal.
>
>The reason for our misundertanding is that I had assumed that the
>tubes
>would be operated much more reasonably, and loaded for a power output
>more like 250-300W PEP per tube. Under these conditions, and with good
>power supplies, the IMD can be much better as the data on my web-site
>show.
>
>(Some corrections: the web-site figures are for a pair of 4CX250Rs
>[not
>350As] at 500W PEP output, with 2000V on the anodes, 350-360V on the
>screens and about 4500 hours on the heaters. For easier comparison of
>IMD levels, I will be converting the numbers to dB below either tone
>rather than dB below PEP.)
>
>Another aspect of IMD levels that the common measurements of 3rds and
>5ths fails to point out, is how quickly the higher-order IMD falls
>away.
>In many respects this is even more important than the close-in IMD
>because the higher orders cover far more frequency space. You expect
>some problems if you move in close to a very strong station; but you
>also expect the signal to disappear TOTALLY as you tune away. For that
>reason, there's a lot of advantage in optimizing the loading for lower
>IMD levels far out from the main signal.
>
>
>Moving on to a different aspect, I said:
>>>There was an interesting posting in r.r.a.homebrew several weeks
>ago,
>>>from one of the actual engineers who generate the numbers for the
>>>datasheets of new amplifier devices. The "proper" figures reflect
>his
>>>personal judgement about the best mix of several different orders of
>
>>>IMD (within other device operating constraints of course). He
>happened
>>>to work with transistors, but I'm sure that's how tube datasheets
>are
>>>written too.
>
>Carl replied:
>>I'm not so sure. Most tube design engineers were raised in an era
>where
>>professional integrity still counted.
>
>It has absolutely nothing to do with integrity. It's about engineering
>judgement, nothing else.
>
>The operating conditions for any device can adjusted to give a
>literally
>infinite number of combinations of output power, IMD levels and
>efficiency. The device engineer HAS to use his professional judgement
>to
>decide which particular combinations will be of most interest to the
>user. These will be the operating conditions that get published in the
>datasheets - but they're not the only "correct" ones.
>
>If your judgement is that you want better IMD and are prepared to
>accept
>less output power, you're absolutely entitled to work out a different
>set of operating conditions, and those will be equally "correct".
>
>
>73 from Ian G3SEK Editor, 'The VHF/UHF DX Book'
> 'In Practice' columnist for RadCom (RSGB)
> www.ifwtech.demon.co.uk/g3sek
>
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|