Topband: W8ji ATR-10 design 160M?

[Tom W8JI]

Besides all of the focus on loss, one has to consider component values, 
expense, dissipation at different inductance values, connection losses, and 
matching range of the tuner.

The L in this example needs over 5000 pF of capacitance, and all of the heat 
in any of the high capacitance systems would be concentrated in few turns of 
an inductor. So while we can have less total loss and heat, the heat is more 
concentrated. A power rating curve or impedance range curve is far more 
complex than paper efficiency or inductor Q.

Not only that, we have significant wiring and connection losses that do not 
show up in the paper tuner.

The advice I would give anyone is pretty simple. Don't design or use an 
antenna that has extremely low or reactive feed impedances, unless you have 
no other choice. If you must use one, plan on low efficiency and other 
problems.

It really would require a book of text and drawings to cover this properly, 
just like it would with many one page or one line "always do this" or "it 
works this way" answers.



----- Original Message ----- 
From: "Paul Christensen" <w9ac at arrl.net>
To: <topband at contesting.com>
Sent: Thursday, October 17, 2013 8:31 AM
Subject: Topband: W8ji ATR-10 design 160M?


> >*L network tuners (like the Ten-Tec, Nye Viking, etc) handle more*
>>*power into impedances near 50 ohms, but often do a poor job *
>>*matching reactive or very low impedance loads on low frequencies.
>
> That's a true statement only if we severely limit the C value in an L 
> tuner. But if we use a very high value of C in an L type -- or high C 
> values in a T, losses are significantly minimized when either tuner is 
> terminated into low Z loads on 160m.
>
> I just computed tuner losses between a high-pass L, low-pass L,and 
> high-pass T.  For the comparison, I kept coil Q at 200 although a roller 
> inductor's Q is greatly affected by its mechanical design.   See this link 
> that shows typical Q variance in a common roller inductor: 
> http://www.karinya.net/g3txq/temp/rollercoaster_q.png
>
> I also assumed a resistive 5-ohm load at the output terminals of the tuner 
> models.  Q for the C values was held at 1000.  Frequency is 1810 kHz.
>
> Example #1 - High Pass L
>
> L =1.47 uH
> C = 5800 pF
> Loss = 0.08 dB (1.8%)
>
> Example #2 - Low Pass L
>
> L = 1.33 uH
> C = 5200 pF
> Loss = 0.08 dB (1.8%)
>
> Example #3 - High Pass T (500 pF Output C)
>
> L= 11.4 uH
> Cin = 180 pF
> Loss = 1.09 dB (22%)
>
> Example #4 - High Pass T (1000 pF Output C)
>
> L =  5.8 uH
> Cin = 343 pF
> Loss = 0.57 dB (12%)
>
> Example #5 = High Pass T (5000 pF Output C)
>
> L = 1.44 uH
> Cin = 3100 pF
> Loss = 0.11 dB (2.4%)
>
> See a pattern here?  To get minimum loss in a high-pass T with low-Z 
> terminations, it takes C values approaching the high values required in 
> either L type.  This should be of no surprise.  There's no clear winner 
> here except the T does offer an attribute not yet mentioned:  We can 
> easily control the phase shift through the T for use in various phasing 
> projects like directional antenna systems.  We can't easily do that with 
> just an L tuner.  In that case, controlled phase shift needs to be 
> attained by another method, like changing line length.   With adequate C 
> size and reasonable coil Q we can get low loss on low bands -- and don't 
> need the XMatch to get it .  However, based on my limited knowledge of 
> that device, it employs a lot of switched C on the output and should work 
> very well into low Z loads.
>
> So even with the high-pass T, we need a ton of output C (and nearly 
> commensurate input C) to get low loss into low Z terminations -- way more 
> than what you get when you buy a T tuner off the shelf.
>
> For the examples, I used a very low load Z value of 5 ohms.  Apart from a 
> mobile installation, these are not antennas I want to use.  Even on 160m. 
> If the input end of a line is anywhere near that value, most antenna 
> systems will be very short.  Thanks, but no thanks.  I will do just about 
> anything to ensure an antenna length that's long enough such that the Z 
> seen at the input end of the line, no matter the line length and without 
> any other external components -- is at least 50 ohms and don't care if it 
> rises well into the K-ohm area.  Almost any simple T or L tuner will 
> perform the matching function in this case.  When using multiband wire 
> antennas where the lowest operating frequency is a half-wave radiator 
> length -- or base-fed verticals that are not unreasonably short, then no 
> matter the line length, the Z at the input end stays well into the 
> double-digits and tuner loss is reasonably low.
>
> On my QRZ.com page, you will see a motorized balanced L tuner that uses 
> Eimac vacuum relays to switch a Jennings vacuum variable cap either in 
> front of, or behind the balanced coil pair.  If I was to build the tuner 
> today, I would eliminate that expensive piece.  That part of the circuit 
> was designed when the input end of the line is less than about 50 ohms. 
> Again, unless it's a mobile installation, I really don't want to operate 
> with short antennas that result in low Z at the line input.
>
> Moral of my story: If your 160m tuner will terminate into some really low 
> Z values: (1) try and keep coil Q high; and (2), make damn sure you can 
> switch-in some big-value caps for the job.
>
> Paul, W9AC
> _________________
> Topband Reflector
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2014.0.4158 / Virus Database: 3614/6755 - Release Date: 10/16/13
>