[Amps] RF2K+ LDMOS linear

[Jim Thomson]

From: "Jeff AC0C" <keepwalking188 at ac0c.com>
To: "Ken K6MR" <k6mr at outlook.com>, <amps at contesting.com>
Subject: Re: [Amps] RF2K+ LDMOS linear

<For me, especially with RTTY contesting, the capability of running this amp 
<on a So2r basis should not be assumed to be workable until it's actually 
proved.  This gets back to my main point.  The gee-whiz it will be perfect 
buzz on a lot of tech really needs to come with a list of disclaimers to 
cover the exceptions and restrictions.  But they never do; meaning the only 
way to really get a handle on them is for someone to get one and give it a 
hard workout.  That would at least be one solid data point.

And you can't really make a compare to the Elecraft stuff - the same issue 
confronts the KPA1500 (it's an unknown new entity - until it's proven in 
combat).  However they have a proven track record of making a decent product 
and improving it when there are issues.  These other various guys who want 
to get into the amp biz are unknown entities by comparison and while they 
may build perfect products and have perfect support, to my thinking until 
that's actually proven it needs to seen with a skeptical eye until proven 
otherwise.

73/jeff/ac0c

###  The flex power genius  video depicts  ONE FLEX xcvr  and ONE FLEX
power genius amp, running 1.5 kw on CW...and toggling rapidly back and forth between
40 + 20m.   So no, you dont require two amps.  If a 120mm fan is good, a 150mm is even
better, esp when 2 are used in push-pull.   Typ 150mm fan is  234 or 243 cfm..like the papst,
with ball bearings.   Stick it in another room, or in a sound proof box, end of noise issue...and end
of cooling issues.   At work, we would put noisy printers inside these sound proof boxes, with flip up
plexiglass covers.  They were lined on 5 x sides with  honey comb foam.   Dead quiet with that lid down.

##  the only fly in the ointment, when toggling between any two bands is if the auto tuner is used on both the  2 
bands in  question.  But with the use of relays, and no motor driven caps, their simple relay switched L network
would still switch bands fast like.  IF the tuner was only required on just one band, then the tuner would be bypassed when
toggling to the 2nd band. 

##  the acid test is simple, stuff  1.5 kw  CCS CXR through the auto tuner into an outboard   manually operated tuner....which is 
mal-adjusted so the swr =  3:1....then into the big dummy load.    Then let is sit for at least 1 hr...and see what happens.  Do
this on 160, 20, 10m .   If it doesnt blow its brains out, you have a winner on your hands.  Ok, then let some rtty contester beat on it
for 48 hrs. For rtty contest, leave the fans on full bore CCS for the entire 48 hrs, ditto with any contest.   Dont let stuff heat up 1st..then
try and play..catch up. 

##  13 db down for the 3rd harmonic means 75 watts reflected back to the combiner.   How many db down is the 5th and 7th and 9th harmonics ? 
That 75 watts is not absorbed by any LP filter, nor is it routed to ground.   The combiner is not a tuned circuit like  a PI net, so the 75 watts has nothing
to bounce off of.   So where does it go ?    Probably back into the PA.  This would be a good application for a Diplexer setup... which is just a LP and HP filter
with their inputs in parallel.  Output of LP filter goes to TR relay to ant.   Output of  HP filter  goes into a 50 ohm dummy load.  Flex used  7 x diplexers
on its latest..power genius..with an on board 50 ohm dummy load.    The 7 x HP filters only have to handle low power, like the 75 watts from the 3rd
harmonic, plus  what ever level harmonics on the 5-7-9.    Simplest method would be to  either use an onboard 50 ohm dummy load, or route some 58 or 213
through a SO-239  then into an external dummy load.   There goes another huge load off of the heatsink.   2nd harmonics is typ –40..as more on the 4-6-8th
harmonics, typ of any push- pull arrangement. 

##  The serrated heatsink concept is a good one, easy method to increase surface area.   Eimac used  serrated fins on some of its tubes... depicted
in some of their qst ads a long time ago.   Notice that Eimac uses very very  thin fins and lots of em, on their  3CX-3000A7  and also 3CX-6000A7. 
On both those tubes, what you think are thick fins at the top and bottom are actually struts.....used to support the large diam cylindrical piece
that surrounds the internal thin fins.    With either tube sitting  vertical on a white sheet of paper, with loads of light on the paper, and viewed  from above,
you can clearly see the myriad of paper thin, silver plated copper fins, going out like spokes of a wheel.  At their outer edges, they are fold back a bit, like a hairpin. 

##  yaeus got it right with their 100%  copper heatsink they used on their  400 watt version of the FT-9000MP.   Aluminum only conducts heat 57%  as good
as copper.   Copper will extract heat a lot faster than aluminum. I have tested this.   But you then still have to transfer that heat ...to the air. 
A copper spreader plus  serrated aluminum heatsink + high cfm fans, esp in a push pull arrangement should suffice for all demanding needs.  

## pre-distortion output on the german or any other amp, ss or tube, should be standard.  A pair of gigavac  GH-1 relays, sped up for qsk, one on the input,
and one on the output would ensure 100%  reliability for qsk cw...or vox  cw  + ssb.  With a vac relay on both the input + output side, then the amp will
handle 2 kw on bypass mode.   That comes in handy when calibrating the internal wattmeter, using a 2nd amp..with 1st amp on bypass mode. Also
comes in handy when wiring 2 amps nose to tail in series, like for a redundant setup.  A simple spdt relay, or switch ensures only 1 amp can be keyed 
at any one time.  Any gigavac vac relay can be bought with either a 12 vdc or  26.5 vdc coil.   They are field replaceable too...so you can easily swap to a 
different voltage coil if you like.  If you are voltage limited, the 12 vdc relays can easily be sped up with just 18+  vdc.    A dc to dc converter could also be
used to obtain the required higher vdc, if the amp only has a 12 vdc lower aux source. Or the 50 vdc could be used,  with correct drop resistor.  To speed it
up even more,  the correct value cap can be wired in parallel with the drop resistor.  Now you are down to  1.6 msecs..which is ample for qsk.     

## If the LDMOS devices  become unobtainium  down the road,  I would buy spares, NOW,at the same time the amp is purchased.... and store them away. 
If the LDMOS  devices blew up, buy some new ones, if still available, and not touch your stashed away spares.   Then at least if  and when they become
unobtainium, you have a working set in the amp..and spares, which means the amp is still usable for a long time...and also has a high resale value.  
Or just replace the entire pallet with the latest and greatest devices.  

Jim  VE7RF