[Amps] A tale of two IMs What happens?

[Jim Thomson]

Date: Sun, 11 Apr 2010 14:45:24 -0400
From: Roger <sub1 at rogerhalstead.com>
Subject: [Amps] A  tale of two IMs What happens?

 "No one has touched the question on how reducing power without returning 
affects IM I now have another one.

Given that most of todays transceivers have an IM or roughly -35db "so 
I've been told",

###  You can look em ALL up on qst's extended lab reports. Any given
xcvr, will have imd figures, that vary all over the map..depending on
band... with the worse case usually  being 12m  for some reason. The
QST results only depict the worse case imd.  The extended lab  results
will  show all  9 x bands.. one at a time.  In the case of the kenwood 870,
they also did it 3 x times.. just for 20m.  IMD  was -30db pep [-24db below one tone]
at 100 w pep.. which improved a whole bunch, when power dropped  down to 85W.
Then improved a whole bunch more, when tested at 50 w.  -30db pep is what you 
get from a drake T4XC  with sweep tubes, from 1977 era.  And in order to pull that
off, idle current was 70 ma @ 700 vdc = 49 watts.   Max 2x tubes diss is only 52 watts.
So they had to idle em at 99% of their diss ratings.. just to get the lousy imd resuts they do. 




 and we put a amp behind it that ALSO has an IM of 
-35db, what is the resultant IM?

## I researched this out yrs ago... and the consensus was , at best it would be the
same... and at the very worse, it could degrade imd  by 6db max.  In most cases,
it appears to degrade by 1-2 db.   




  What if the amp has an IM of -55db?
Do they add, subtract, or go with the lowest number? 

##  they go to the lowest number.  A  -30 db pep imd  xcvr, driving a -57db pep
3CX-3000A7  will result in  -30db imd  at the very best...and  -28db  at the very
worse.   That's  dealt with ad nauseum in ..." SSB systems and circuits" .  The same
classic book also explains why the 2 x tone imd test is flawed. The IMD3.. and IMD5
will  appear to be  lower than what it really is.  The  IMD7-9-11-13  products  will tend to
cancel out the IMD3  product, leaving bogus results in any 2 x tone tests.    The test 
is flawed anyway, since everything is in a steady state condx.  Neither the B+.. nor the
bias vdc, screen vdc, etc, etc, is varying at all.   It's not a dynamic test.  Also, by carefully
varying the SPACING  of the 2 x tones,  you can hit a ..'sweet spot' ... and get better
imd results.  A steady state 2 x tone imd test.. won't factor in ALC overshoots, overdriving
input audio circuits,  overdriving everything else down stream from the mic jack etc. 

##  If you use the ALC cable  from the xcvr to the amp.. and the  xcvr's ALC  has a different time
constant than the linear's  ALC,  you just lost it right there.   



 IOW it is the amps 
job to "faithfully" reproduce the input signal, but that really only 
happens when running class A, if the user is lucky.

## Correct.  So if you really want -42 db pep imd  from the linear... the xcvr
had better be -42db pep imd as well.  Most of the yaesu class A  xcvr's  will
far exceed -42db pep.. by a whole bunch.... so then you end up with the
amplifier being the limiting factor.  What they need to do with these Class A 
xcvr's is to incorporate either EBS in em [ with some type of TRI-state bias] 
so they don't burn up... OR use some kind of 'sliding bias' scheme... like 
Krell does in their Class A  stereo amps.   My yaesu MK-V   sucks 10A @ 30 vdc
when in class A =300w diss !   That's  10A.. whether you apply audio drive or not! 
When driven with a  1 khz sine wave, the mk-v puts out 75w.  300-75= 225 w diss.
So the xcvr runs HOTTER, when keyed, but no audio applied. 

## They need adjustable bias too.  No point in having -63db pep imd, in Class A...
when the amp is only good for -38db.   They need to have their bias tweaked, so the
xcvr imd is the same.. or slightly better than the amp..... and incorporate sliding bias,
or tri/quad state EBS  bias schemes. 





Intuitively, "I would think" that the two figures would add, but if that 
were the case the amp with -35db and the exciter with -35db would have a 
pretty ratty signal. OTOH in the case of the -55db amp behind the -35 db 
exciter does the amp "clean up" the exciter signal? Doesn't seem likely.

##  see above. The amp cleans up nothing.  In SSB systems and circuits,
they hate GG triodes  driven by a high imd xcvr. ['ham stuff']  The writer's of the book
[collins engineers]  prefer grid driven tetrodes.  Then the  drive power requirements
are so low, that it's easy to  run the driver in Class A.  

## a buddy across town built a hb  4CX-1500B  3 yrs ago.. grid driven.  He can get an
easy 1500w  out of it.. with just 10 watts of drive... and a 50 ohm globar resistor across
 the input of the amp.  If a 200 ohm resistor was used.. and a 1:4 un-un... the drive power
would be even lower still.  As is... 10 watts is obtainable from Class A.. from yaesu 
xcvr's. 


It'd be interesting to see the IM figures for today's exciters, (and 
amplifiers) rather than than adds that just say "Provides amazingly low IM".

## Both the amps and all the xcvr's that have been tested in the arrl,
have been written up in both the qst.. and esp the extended lab reports.

##  I know several ESSB ops who drive 8877 Ameritron amps with Class A,
that's as  clean as it gets.  My self,  I use noise gates, levelor's, compressor's
and split band peak limiter's, /distortion cancelled clippers,  in the audio chain..
then drive the analog BM in the mk-v,  in my ESSB set up.   I had to add an extra
RCA jack + coupling cap, to drive the BM directly.  The  analog TX  noise floor on the 
MK-V is 20 db QUIETER, than when using the DSP  100-3100 hz TX mode. You can
see that right away, when toggling between analog + DSP mode, via menu.... while
listening on a 2nd RX..3' away.  [keyed, but No audio drive applied]   Driving the 
analog BM  directly  via new jack means the noisy stock yaesu audio pre-amp is
completely  bypassed.. a huge diff. The outboard class A audio pre-amp eats the
stock inboard yaesu pre-amp, hands down. 




## Then it's easy to obtain 1.5 kw  from an amp... and have NO ALC showing on the
 xcvr at all!  Then the alc overshoot problem is solved once and for all.  Method #2 
uses a  fixed vdc  fed to the xcvr's  ALC jack.  This also works on older radio's of the
tube xcvr era.  A few volts of -vdc  is fed to alc jack on xcvr.. and vdc tweaked
for  say.. 85 watts pep.  Scream all you want, and the xcvr will never put out more
than 85w pep. No alc overshoot. The alc voltage doesn't have to be derived.. after
the fact.. since you have injected it into the xcvr's alc buss.. and on a continuous
basis.   The trick here is... if you whisper into the mic... your power will drop off..
since you no longer have RF compression  [normal alc] .  You get around this
problem easily... by either using a boom mic, that's always 1" or less  from your
mouth.. and/or... use outboard  noise gate /compressor/peak limiter in the audio.

##  after all this song and dance.. it's  fun to be able to dial up 1386 watts.. and never
see the AS wattmeter  go higher than 1386 watts.... and never see any alc action on the
alc meter either. 

##  yaesu isn't there yet  with the Class A  concept.  They need to incorporate
the various bias scheme I suggested earlier..AND bigger/more effective heat sinks.
Copper heat sinks  conducts heat almost twice as good as Aluminum.  If you look
at any Eimac big metal tube, the fins INSIDE are razor blade thick.. and lots of em.
They are all silver plated copper.  Not to be confused with the 'struts', that also
resemble fins.  The 'struts' at the top and bottom of the tube  are there to hold the
anode cooler  cylinder assy in place.  Shine a light down from the top... with a white
sheet of paper below the tube.  All the heat has to be ultimately transferred to the air
and lots of thin fins works better than fewer thick fins.  Try  sticking steel, AL, then CU
into a grinder, and you will soon see [and feel] the super fast heat transfer of CU.

later........ Jim   VE7RF


I also see I need a refresher on working with logs.<:-))

73

Roger (K8RI)