[TowerTalk]Tuner Losses & Bird ratings

[K7GCO@aol.com]

In a message dated 11.05.00 13:20:19 Pacific Daylight Time, 
w8ji@contesting.com writes:
<<  << 
 >  The most accurate way to measure loss in a tuner with different 
 >  source and load impedances is to measure the temperature rise in 
 >  the tuner, not the power levels. >>
 > 
 >  . 73, Tom W8JI
 >  w8ji@contesting.com
 > 
 > Right on Tom.  Basically, if a tuner can match the Z at the end of the
 > feedline (1:1 SWR is obtained at the input of a tuner), 100% transfer of
 > power occurs to the antenna less the losses in the feedline which are
 > predictable and the losses in the tuner as measured by the "heat test." 
 > There are exceptions.  I had a load that the Johnson MtatchBox matched 1:1
 > but wasn't transferring all the power without getting warm--with 100 W as
 > the inductors were heavy.  With 1 KW I had arcing which tipped me off of a
 > problem but no apparent heating with short tests. 
 
 For a low power test, if you placed the tuner in a styrofoam 
 container and measured temperature rise in the box over a period 
 of time, you could measure the power loss.
 
 Of course high power, as you found out, shows the problem right 
 away.
 
 Most tuners can only dissipate 50-100 watts without severely 
 overheating or failing, because that loss is concentrated in one or 
 two components. So if you are running several hundred watts or 
 more and nothing is melting, you can be sure efficiency is high.
 
 Arcing can be a different issue, arcing is not related to loss. It is a 
 voltage breakdown problem. Arcing does not necessarily mean 
 losses are high.
 
 73, Tom W8JI
 w8ji@contesting.com  >

Right on again Tom.  In the arcing across the variable Xc plates case, I 
still had 1:1 SWR but comparison with another known reference antenna on 
receive weren't what I though they should be.  Reversing the connections of 
the differential Xc eliminated arcing, the RF ammeters read higher and I will 
duplicate the test for an actual dB difference.  As I pointed out, it then 
required less of the tank coil for resonance and 1:1 SWR which lowered the 
peak voltages therefore no arcing with 1 KW.  Now the combinations produced 
1:1 SWR and more output.--the desired goal.  This was a Johnson MatchBox that 
I constructed with back to back rotary coils with a differential mechanism 
between them so they turn in the right directions.  When I changed inductor 
values I tuned a shaft and didn't have finger contact with the coil itself so 
as to detect heat.  I will also do the heat test.  The conductors were fairly 
large and transmissions were short so heating, whatever it was would be 
minimal.  

Switch contacts are usually smaller in surface area than the conductors they 
are connected to and can heat even though the conductor doesn't.  Feel them 
also.  They only make pressure contact and are another source of resistance 
and heat.  With repeated heating there is evidence that the spring pressure 
could lighten and that will cause even more heating.  The standard back to 
back finger switches are adequate for receive signals as far as I'm 
concerned.  I like those switches with big silver pads.  I have paralleled 4 
decks of the light switches in tuner circuits to get the necessary switch 
contact current carrying capacity based on heat tests.  Switches are 
typically the RF current carrying weak-link in tuners.  I sometimes skip 
contacts to increase the spacing between conductors connected to it.  This 
reduces capacity coupling between them and increases the voltage needed for 
arcing.  I wish I could show pictures of some of my modified switches.  
Eventually I will have a site where I can do that.  

The home made plug in 20M tank coil that got warm at one end due to a 50 year 
old bad solder joint was detected only because I removed it right away.  I 
did notice unbalanced currents in the RF ammeters in series with each lead of 
the open wire and was a tip off something was wrong.  They balanced up when I 
soldered the joint properly.  Note!  RF ammeter calibration can be off 
between identical meters so reverse them and observe their readings.  In 
commercial tank coils, there are pressure contacts created by the banana plug 
not screwed tight. Sometimes they get loose and heat can be generated at the 
pressure contact also.

Another source of heat is the 4 flexible fingers in the tip of the banana 
plug are press fitted into the body.  Sometimes they loosen with use and it 
gets warm with a KW.  I developed a process and practice of soldering the 
banana plug fingers even if there is a tight compression fit.  I think they 
are nickel plated and solder doesn't stick easy.  I spun them in a lathe and 
filed the nickel off at the joint, heated it with a torch, coated with solder 
paste and the solder flowed evenly over the joint.  Only a small amount is 
needed and made a great solder joint.  I replaced the retaining screws with 
brass ones that I had silver plated.  In the next silver plating session I 
have In will silver plate the whole banana plug.  I plan to operate QRP 
someday.  I have a copper coil 2' in diameter and 3' long with 3/4" tubing 
I'm just dieing to use on 40M.

The heat test is a great test of tuner components and arcing suggests the 
spacing is not wide enough. It also suggests the configuration used to obtain 
1:1 SWR may just create excessive voltages (and currents)--try another one 
and compare output and heat generation.  What I'd like is a small diameter 
instrument that will measure temperature on contact.  I could get in there in 
the tight spots like on switches and check for heat generation.  I use a lot 
of tuners with big components and they can cover up an inefficient matching 
configuration.  Use a low power tuner of the same configuration for initial 
tests of heating and arcing.  With "Parallel and Series Tuners" using plug in 
coils for a particular band, have about 3 with different inductance 
values--even the next higher and lower bands The inductance of typical tank 
coils may not be optimum for maximum output.  Flea markets are still loaded 
with plug in coils at very good prices.  

I have a balanced L tuner that I really like with 2 rotary coils using 1/8" 
wire that don't even get warm with 1 KW on 10M feeding a 100 ohm balanced 
coax load.  Balanced L tuner currents seem to be lower than resonant tank 
circuits.   

There are temperature patches machinists use that can be placed on components 
for heat tests.  I'd start with 100 degree patches.

K7GCO










--
FAQ on WWW:               http://www.contesting.com/towertalkfaq.html
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm