[AMPS] Roller Insuctors and Core Materials

[Rich Measures]

>
>
>Hi,
>
>A bit of a thread has appeared recently on the Towertalk
>reflector about "QRO" antenna tuners;  the discussion
>has touched on the subject topic,rollers,  Delrin and ceramic
>cores.  A few quotes from there:
>
>Jim,  KH7M wrote,
>
>Inside an AT4K tuner by Palstar:
>"Roller Inductor: This is the surprise within my unit.  The
>inductor is a tape wound obviously heavily silver-plated unit,
>using tape which measure 3/8" wide, edgewound, with a tape of
>0.062" thickness, The inductor, per the manual, is rated at 28
>uHy, 5 kV and 30 amps. 

€  E.F. Johnson and Multronics  rate 1/2" wide ribbon at 15 RF 
amperes-rms.  

>A large, maybe 0.75" or so pinch "wheel"
>rolls around inside the tape or the air core unit as shown in one
>of a neat series of photos of the unit, facilities, and of Paul
>Hrevnek himself holding an AT4k, at
>
>http://www.palstarinc.com/contact.htm 
>
>(takes a couple of minutes to load pictures, then they appear
>sequentially one at a time).
>
>The inductor in my unit appears visually to be similar in
>construction and color to the unit Surplus Sales of Nebraska
>illustrates on the back cover of their catalog, 8; that is heavy,
>red plastic end plates, etc.  However, surplus sales rates their
>inductor as 40 uHy, 12 1/2" long, but probably the same
>manufacturer, I would guess.  Inductor in my unit is about 9
>inches long.
>
>As my unit was rcv'd, the bottom 14 turns of the inductor are
>shorted to ground; this can be removed if more L is needed down
>on 160 meters, but should be replaced, per Paul, when on 15
>meters, as the full 28 uHy coil has an internal resonance on 15
>meters.  If, for some reason use of 15 meters is not anticipated,
>the jumper may be removed permanently.
>
>Appears completely unlike the apparently bare copper inductor in
>the AT4K photo appearing at the Palstar web site."
>
>Tom,  W8JI,  comments in response:
>
>"It looks like there is little difference in these tuners, based on 
>your post. [KH7M: that is between the Palstar AT4K and the
>Ameritron ATR-30 units] That assumes there is nothing wrong 
>with the components in the 4K, and the ATR 30 is built to spec.
>
>"I'm more comfortable talking facts than subjective opinions. 

€    During the grate-debate, Mr. Rauch rejected standard AC Circuit 
Analysis of a vhf suppressor of his own design, when the result proved to 
be somewhat unfavourable.  .  

>Let's go 
>through the facts...
>
>> Capacitors: Input and Output 6kVm 350 pf, with an added 350 pf
>> 7.5 kV fixed doorknob C which can be switched in,  in parallel with
>> the network output capacitor. This is said to be necessary when
>> on 160 or 80 meters, and the antenna impedance is below 25 or 30
>> ohms; switching in the added 350 pf will reduce the current
>> within the inductor, and thus losses and heating therein.
>
>"Untrue, as you will see. The change in heating is insignificant. As a 
>matter of fact the capacitor most likely reduces power capability. 
>
>"Let's look at the important components in three tuners, the 
>components that routinely fail. Antenna switches, nuts, bolts and 
>wires are the least of our worries. Consider a tuner with 350 pF 
>max C, one with 500pF, and one with 350pF variable input and 350 
>pF shunted across the output variable on low Z loads at 4000 watts. 
>Capacitor voltage is peak (since that is the failure determinant) on 
>the highest voltage capacitor and inductor current is RMS (since 
>that is the failure determinant). All at 1.8 MHz since that is worse 
>case.
>
>For 50 ohms load 1.8 MHz:
>
>350 pF /350 pF Vc 3300v  / 11.6 uH  17.6 A 

€  How was this calculated?
>
>500 pF/ 500pF Vc 2325v /  8.4 uH 17.5 A
>
>For 20 ohms load 1.8 MHz:
>
>318 pF/ 500 pF Vc 3500v / 22.78 A 
>
>350 pF / 548 pF Vc 3190v / 22.75 A
>
>"You can plainly see, even at 4000 watts,  capacitor voltage is the 
>least of our worries in either tuner. Current is the problem.  The 
>extra 350 pF doesn't do much, because the INPUT capacitor sets the 
>system Q.
>
>"The doorknob capacitor actually becomes the component that limits 
>power. At 20 ohms, it must handle just under 2/3 of the current in 
>the output capacitor.  That current is 14.12 amperes, so the doorknob 
>must carry 9 amperes. The current rating of an x50 series doorknob is 
>about 3.2 amperes at 1.8 MHz. 

€  Depends on the value.  The 100pF unit is rated by ITT Jennings at  
3.4A-rms for 1 MHz and 7.2A-rms for 10Mhz.   My guess is that the current 
carrying ability at 1.8MHz is c. 4.5A.  

>An x57 series is 4.2 amperes, and an 
>x59 series 7 amperes. Even the largest standard doorknob won't handle 
>that current without heating.

€  The highest-capacitance 50-series is 100pF.  For padding a 350pF 
variable, (3) 100pF units in parallel would be needed, yielding a total 
capability of >13A.  13A into a 50-ohm load is  8.45kW.  

>> As my unit was rcv'd, the bottom 14 turns of the inductor are
>> shorted to ground; this can be removed if more L is needed down
>> on 160 meters, but should be replaced, per Paul, when on 15
>> meters, as the full 28 uHy coil has an internal resonance on 15
>> meters.
>
>> If, for some reason use of 15 meters is not anticipated,
>> the jumper may be removed permanently.
>
>"Actually that resonance moves around. It starts out in lower VHF 
>(by six meters) at mid-inductance setting  and crosses ten meters and 
>eventually moves down to 20 MHz as the roller is cranked to the ends. 
>It never stays on one spot, and the band it "kills" really depends on 
>the load impedance you are trying to match and the amount of C 
>being used. A more accurate statement would be "it causes 
>major problems somewhere above 18 MHz, depending on load and 
>adjustments".

€  agreed
>
>"More inductance will help match wider impedance ranges on 160, 

€  Higher impedances, yes.  Lower impedances, no.  

>but the power handling will be severely compromised because voltage 
>ratings of the capacitors becomes an issue with more inductance.
>
>"Actually there are two limits to power handling. The ARRL Antenna 
>book treats this in more detail, but misses one very important point. 
>It considers the dissipation limit of the inductor as a constant, 
>but that isn't true at all.   
>
>"Our Handbooks, and many people, assume that as more C is added 
>power handling goes up. That isn't necessarily true, since as shown 
>above current in the inductor barely changes.
>
>"While reactance is reduced 28% in a maximum C change from 350 to 
>500 pF (driving 50 ohms), current is reduced under 1%!!!! At some 
>point more capacitance may cause roller failure, because heat is 
>concentrated in a smaller area. 
>
>"Rollers safely dissipate less power as they are cranked down to less 
>and less inductance. That's because the heat, even though slightly 
>less is generated, is now concentrated in a smaller and smaller area 
>of the roller! Too much capacitance, or too little, can cause more 
>heating. There is a "sweet spot" where power handling is maximized 
>for a given set of components, and it isn't the spot software 
>predicts. 
>
>"Looking at the ratings of a roller, or even the size, can be very 
>misleading. As a matter of fact, most ratings I've seen are simply 
>pulled from a certain place near the back pocket of the person giving 
>the ratings. A 30 ampere roller (at 5 MHz) is a BIG component. Try 
>5/8 inch copper tubing and perhaps a foot in diameter, that'll get 
>you there. 

>"Even calculations based on data like Q and impedances often fails to 
>tell the story. The real test is to run power and measure the heat. 
>
€  amen.  

>"The ATR-30's "sweet spot" is 150 ohms. At 150 ohms on 1.8 MHz, it 
>will handle 7.5 kW carrier with a 50% duty cycle. The only 
>concern is the roller shaft heating, something soon to be moved up 
>higher by a material change. The goal is a tuner that handles 3 KW 
>continuous carrier on any mode on any band into the widest load 
>range possible.
>
>"The penalty of a high power rating is a more restricted matching 
>range, but proper choices of feedline lengths usually eliminate that 
>problem.
>
€     For a multiband antenna fed with ladder-line, the tuner must be 
capable of dealing with any eventuality. 

>"If you look at the currents above, and consider the current drops at 
>half the rate of RF power reduction, you'll see why tuners that use 
>thin wire (#12 to #16) inductors fail on 160 meters (and 80 meters). 
>Measures (and perhaps others) who reach empirical conclusions on a 
>regular basis blame the Delrin, but the dissipation factor of the 
>Delrin is not the real issue at all. 

€  Please prove that D-factor doesn't matter.  

>The electric field  is spread 
>out over a wide area of the inductor. It isn't a capacitor, it's an 
>inductor and the bulk of the electric field is outside the Delrin  
>making dissipation factor virtually unimportant in this application. 

€  Nevertheless, even though you profess that it shouldn't Delrin 
assuredly melts in antenna tuner applications. When I stick a neon bulb 
into a tuner, it lights up plenty next to the roller-coil.  If the 
roller-coil is wound on a Delrin form, the Delrin is undoubtedly exposed 
to plenty of RF voltage.   When a Delrin coil form melts down, the 
bubbles form throughout the Delrin, not just where the Delrin touches the 
wire.   I have seen the same effect with antenna end insulators that 
melt.  The melt-zone is not where the plastic touches the antenna wire, 
rather it is a few cm from the contact point.  .   

>The roller operates at well under 3kV, not 30,000 volts like an 
>antenna end insulator. 

€   In my 160m tuner, which feeds an end-fed 1/2-wave Hertz antenna, when 
driving it with100V-peak to the  50-ohm input (100w), I measured 
1100V-peak at the end of the 55uH coil.  Using 1500w, the coil potential 
is over 4000V. 

>
>"The real problem is the small wire size, the very high currents, and 
>the fact Delrin (and many other plastics) melt at low temperatures. 
>The solid core also restricts airflow and reduces the ability of the 
>component to dissipate heat. If the Delrin was replaced with 
>Polyethylene, the power rating would remain nearly the same. 

€  There is a substantive difference in the D-factor between 
Polyethelyene and Delrin.  

>Only Ceramic would stand the heat and increase power ratings.
>
>"The XMatch uses a small gauge wire roller, but it has a ceramic 
>form. The Q difference between that roller and the Delrin rollers is 
>minor, but the ability of the ceramic to handle heat is a major 
>factor in the XMatch's ability to handle higher power.   
>  
>"Truthfully, most things are sold on pure "fluff".  Like antennas and 
>SWR, the equipment that works best is the stuff you are happiest 
>with. Examples of "fluff" abound....
>
€  Obviously, the early MFJ "3kW" tuner was never tested at anything like 
3kW. 

>> provides large tables of exactly what antenna R and +/- XL
>> can be tuned, and keep the losses less than 20 % of the power;
>
>"I would hope losses would ALWAYS be less than that!  At 4000 watts 
>20% loss would be 800 watts. I did some rough calculations, and the 
>inductor would be nearly incandescent if loss were that high. 

€  indeed

>At 10%, 
>it would be a dull red glow. Power loss is the least of our worries, 
>component failure is the real worriment. 
>
>"The Delrin roller fails at 70 watts dissipation. You don't have to 
>lose very much power to make a component fail. Only 30 watts of heat 
>causes the #8 wire inductor in my 4 square ATU to get hot."
>
€  With #8, that would take a bunch of amps of RF.  

>Thought the above might add some info to the conversation on
>going here about RI-40.50,  Multronics, inductors and cores.
>
>73,  Jim,  KH7M
>
>
>
>
>--
>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
>



Rich...

R. L. Measures, 805-386-3734, AG6K, www.vcnet.com/measures  


--
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