ad: SteppIR-1

The T-network

Discussion in 'General Technical Questions and Answers' started by KL7AJ, Dec 2, 2009.

Thread Status:
Not open for further replies.
ad: L-HROutlet
ad: l-assoc
ad: l-gcopper
ad: Subscribe
ad: L-rfparts
ad: l-innov
ad: l-Waters
ad: l-rl
  1. KL7AJ

    KL7AJ Ham Member QRZ Page

    Probably the majority of low cost antenna tuners uses a C-L-C T configuration. This configuration gives quite a bit of versatility, doesn't require a roller inductor for a perfect match, and works equally well for low-to-hi-Z and hi-to-low-Z matching. However, this circuit can easily degenerate into a high pass filter under certain conditions. Fortunately, the conditions where this is least likely to occur is also where the greatest efficiency is going to be...maximum C and minimum L.

    One of the disadvantages of the T-network is the fact that you can have multiple settings that can give you a conjugate match. Only one of these settings will result in the greatest efficiency.

    However, there's another subtlety about the T network that can compensate for this shortcoming. The T network allows you to compensate for PHASING independently (well, not qute independently) of impedance matching.

    Why would you care about this? One obvious case would be if you're using a phased array. If you have two T network tuners, each feeding a different antenna, you can steer the antenna array using your tuners.

    Now, since there IS some interaction between the phasing adjustment and the impedance matching of any T network, it can be quite tricky to get the phasing AND impedance matching right. It may be easier to add yet another T network tuner to handle just the overall matching function, but yet has no effect on the DIFFERENTIAL phasing between your antenna elements.

    In fact, this is precisely the arrangement you have in most directional A.M. broadcast stations: one T network at the output of the transmitter, and one Tee feeding each antenna's transmission line. These are typically all enclosed in a single "Phasor cabinet".

  2. AG3Y

    AG3Y Ham Member QRZ Page

    Yes, but what was the QUESTION ? ? ? :eek: :rolleyes: :p
  3. K9ZMD

    K9ZMD Subscriber QRZ Page

    There's been a good psychic path twixt SoCal and SoCold lately. Eric just knew that USPS is slowly bringing me a T-network tuner and I've been wondering, "What good is it?" and "Did I make a mistake?" Also, "What if it won't do what I hope it will do?" So thanks, Eric. Now that I know more about T-networks, tell me . . . when will it arrive?
  4. KU3X

    KU3X Ham Member QRZ Page

    A roller inductor has nothing to do with a tuner being Pi, T, L/C or what ever tuner you use or build. An inductor is an inductor when it comes to using one in a tuner. All a roller inductor does is let you have less taps on a waffer switch. I've seen tuners with 24 tap positions and that's a lot.

    When using a T or Pi network, you always want to use the setting the uses the most about of capacitance for coupling the antenna.

    Do you want a good tuner? Then build the Ulitamate Antenna Tuner that is in the ARRL hand book. They modified the Ulitamate Antenna Tuner into, if I remember these letters correctly, the SPC tuner.

    If you are using parallel feeders, then you want a link coupling tuner.

  5. KL7AJ

    KL7AJ Ham Member QRZ Page

    My first antenna tuner was the Ultimate Transmatch I bought in 1972, and I still have it. Never did the SPC modification on it, though.

    You reminded my of something when you mentioned all the taps. Standard practice is to SHORT OUT the unused part of the inductor...this is usually the case whether you're using a roller inductor OR a tapped inductor. This practice is supposedly so you don't have self resonances happening with the unused turns "flapping in the breeze". Actually, this is very seldom a problem, and if carefully designed, any self resonances will be out of any meaningful frequency range.

    I do NOT short out the unused turns in my homebrew transmatches. Shorted inductors can generate a LOT of heat under certain conditions. My motto is, "When in doubt, DON'T short it out!

    I'm sure I'll get some flack on this one. Flack away!

  6. AI3V

    AI3V Ham Member QRZ Page

    24?, that's just getting started!

    I own a Ten Tec #228


    46 taps, and don't you forget it!:p

    Coil is wound on a toroidal form, the tap switch moves a wiper arm like on a AC Variac. You get nice "detents" for each tap.

    EDIT: The unused coils are not shorted.

    Last edited: Dec 3, 2009
  7. AI3V

    AI3V Ham Member QRZ Page

    If you are taking flak, that just means you are over the target!

  8. KD0CAC

    KD0CAC Ham Member QRZ Page

    That's good .
    I seem to be near the target too often .
  9. WB2UAQ

    WB2UAQ Ham Member QRZ Page

    Hi Eric,

    I have to comment on the min L for setting you suggested at the start.

    First, I subscribe to the transmatch loss measuring method that Frank Witt, AI1H introduced many years ago now. This method was first knocked down but I learned not long ago from a follower that the ARRL changed their mind about that and published something about it but can't remember where:(

    With this method I can quickly determine the loss in the transmatch so I did some tests where I had a good match with the highest inductance and one with the lowest inductance. The maximum inductance resulted in min loss in the transmatch and the widest SWR bandwidth. The difference in loss was considerable and some where in the serveral dB range.

    I will repeat this test again. It has been >10 years ago.

    I reasoned at the time that with minimum inductance I had max current flowing to ground thru the shunt L. Max circulating current increased the losses in the wiring and the inductor. The bandwidth was the smallest as well due to the higher Q of the circuit.

    I would appreciate your comments!

    73, Pete WB2UAQ
  10. N0SYA

    N0SYA Ham Member QRZ Page

    What Pete said above is exactly what I have seen too.
  11. W5DXP

    W5DXP Ham Member QRZ Page

    There is a T-tuner simulator at:

    With a 10 ohm load at 10 MHz, it yielded the following results:

    Input C pf      Coil uH       Output C pf         loss dB
       124.1           0.7            250              0.4     (Autotune)
        68.3           1.4            126.7            0.8     (Manual)
    Doubling the inductance doubled the losses.
  12. G3TXQ

    G3TXQ Ham Member QRZ Page


    It looks like you didn't adjust the inductor Q between the two settings. When you do, the losses are 0.4dB and 0.6dB - the higher inductance setting still shows the greater loss, but not by quite as much as your figures ;)

    Steve G3TXQ
  13. KL7AJ

    KL7AJ Ham Member QRZ Page

    Hi Pete:

    You have observed the difference between loaded Q and unloaded Q, which at times seems to create a paradox. It's not too complicated though. Let's look at it closely:

    A transmatch (or any tank circuit for that matter) has an unloaded Q which is limited only by the internal resistance, generally the resistive losses in the coil (and to a much lesser extent, the capacitor. This "resistor" appears "in the loop" in series with the inductor and capacitor. (Remember that without an outside reference you can't really define a coil or a capacitor as being in seris or in parallel! That is, until you apply an external sourse!)

    Now....for any single layer solenoid inductor, all things being equal, the inductance goes up as a square of the number of turns, but the resistance only goes up linearly with the number of turns. (Resistance directly proportional to the length of the wire). We can ignore parasitic capacitance between windings and such, since they're second order effects.
    What this means is, a higher inductance coil will have higher Q than a lower inductance coil, of the same construction. We know that, for series circuit, Q=X/R. So, considering ONLY series resistance, a higher inductance, higher Q coil should be more efficient.

    Now here's the rub. Your load appears as a PARALLEL resistance across the tank circuit. If you recall, the Q of a PARALLEL circuit is not X/R, but rather R/X! This is why the Q of a pi network goes UP as the capacitance is increased, since the reactance (Xc) goes down.

    In both cases, however, series alone, or parallel alone, efficiency is proportional to Q. However, when you have BOTH the series and the parallel aspects to contend with, we have some conflicting criteria, don't we?

    When we put everything together we come up with this equation for the efficiency of any tank circuit:

    Efficiency= 1/(1-(Qloaded/Qunloaded))

    Hope this helps a bit!

  14. G3TXQ

    G3TXQ Ham Member QRZ Page


    If you have the ARRL TLW software, Eric's point is well illustrated. You can force different values for L in their T-tuner simulator and observe the current through the coil - for a fixed load it varies very little for a wide range of inductances.

    If you assume the coil resistance rises as SQRT(inductance) it's clear that the higher inductance values must mean more coil dissipation - in fact the power dissipated in the coil rises directly in proportion to SQRT(L).

    Steve G3TXQ
    Last edited: Dec 3, 2009
  15. W5DXP

    W5DXP Ham Member QRZ Page

    Steve, educate me and others on that subject. How do you determine the Q of the coil for the different settings?
  16. G3TXQ

    G3TXQ Ham Member QRZ Page


    As an approximation, I assume the number of turns will increase as SQRT(L) and that loss resistance will be proportional to the number of turns. So in your example, if the Q is set for 100 when L=0.7uH I would reset the Q to 141 when L=1.4uH.

    It'd be a lot easier if these simulators had an option to do that automatically!

    Steve G3TXQ
  17. W5DXP

    W5DXP Ham Member QRZ Page

    Seems that assumes a separate non-tapped coil for each setting. What about the fact that tapped coils have a lower Q than non-tapped coils? Could that be the source of some of the discrepancies?
  18. AI3V

    AI3V Ham Member QRZ Page

    Uh, the definition of Q is the ratio of reactance to resistance.


    EDIT: You measure Q with a Q Meter!:)

    Last edited: Dec 3, 2009
  19. KL7AJ

    KL7AJ Ham Member QRZ Page

    To really see what's happening, you need to measure the coil current. A simple toroidal current probe at one end of the coil should make this an easy task. (I actually do this in my A.C. electronics class to prove that KCL works even for A.C.!)

  20. AI3V

    AI3V Ham Member QRZ Page

    What does Potassium Chloride have to do with Air Conditioning? :D

Thread Status:
Not open for further replies.

Share This Page