Proper way to adjust ATU...

AM Radio discussion. Directional arrays are FUN!
Post Reply
FirebirdTN
Posts: 51
Joined: Wed Jun 08, 2011 11:24 am

Proper way to adjust ATU...

Post by FirebirdTN » Tue Jul 26, 2016 11:12 am

Don't know if this is beyond the scope of the forums, but in this situation, I can't imagine its too darn hard...

Got a station I do work for, and during thunderstorms it always seems to blow up the same cap in the ATU. There was some question about what value it should have been (when I started working there it got the snot blown out of it last time!), but think we nailed the value okay.

But I'm curious what the PROPER way to adjust the coils are?

Its a very SIMPLE setup: 1kw feeding a single ATU comprised of a "T" network feeding a folded unipole.

I have an impedance analyzer and have NO problem getting it set at a 50 ohm resistive load, but my OCD tells me just cause its 50 ohms doesn't mean its right.

Anyone want to take a stab at schooling me?

Thanks!

-Alan

User avatar
RGORJANCE
Posts: 1388
Joined: Mon Dec 31, 2007 7:48 am
Location: RACINE, WI

Re: Proper way to adjust ATU...

Post by RGORJANCE » Tue Jul 26, 2016 12:11 pm

I won't address the tuning on this post, rather I want to point you to something else. The fact that you continue to lose the caps might be traced to another area.

1) Check to see if you have the ball gaps set properly
2) Check to see if there is a good static drain path to ground
3) Check the ground connections of the ATU to verify solid connection
4( Check capacitor voltage rating/type---G1, G2, etc, or the 290 series
5) If tower is lighted, check to see if the three wire lighting choke is properly wired up

Fossil

User avatar
Deep Thought
Posts: 3217
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: Proper way to adjust ATU...

Post by Deep Thought » Tue Jul 26, 2016 4:16 pm

Assuming you don't know what the design phase shift is, sometimes it is just easier to start with a simple -90 degree network. You'll need the tower impedance, and remember that the tower X is part of the tee network output leg.

For example, assuming 30 +j70 for your tower Z and a -90°network matching that to 50 j0 ohms at the input and you're on 1000 KHz to make the math easy:

Input leg: j((50 * cos(-90)) -sqr(50 * 30))/sin(-90) = ((50*0)-38.7)/-1 = +j38.7 ohms reactance. The coil would be 38.7 / (2 * pi * FmHz) uH. At 1 MHz that would be 6.16 uH.

Shunt leg: j(sqr(50 * 30))/sin(-90) = 38.7/-1 = -j38.7 ohms reactance. The capacitor would be 1/(38.7 * (2 * pi * FmHz)) uF. At 1 MHz that would be 0.0041 uF. A 0.002 uF capacitor (-j79.6 ohms) is a good choice here at 1 MHz to give you some adjustment room. You'll need to offset 79.6 - 38.7 ohms of reactance with a series coil, or 40.9 ohms. That is 6.51 uH at 1 MHz.

Output leg: j((30 * cos(-90)) -sqr(50 * 30))/sin(-90) = ((30*0)-38.7)/-1 = +j38.7 ohms reactance.
Now, we have to compensate for the +j70 the tower adds to the output leg, so that needs a capacitor in series of 38.7-70 = -j31.3 ohms reactance. We also want some adjustment range so let's double that to end up with a reasonable series coil. That means -62.6 ohms or 0.00254 uF at 1 MHz. Let's use an 0.002 here (-j79.58 at 1 MHz) with a coil in series to offset the -j48.28 ohms of excess capacitive reactance. That coil will need to be set at 7.68 uH.

So...your settings are a 6.16 uH input leg with no series capacitor, a 6.51 uH coil in series with a 0.002 uF capacitor for the shunt leg and a 7.68 uH coil in series with a 0.002 uF capacitor for the output leg.

Easy peasy. :mrgreen:

90 degrees is a good place to start on a tee network because as you can see above it bypasses a lot of calculations since sin(90°) =1 and cos(90°) = 0. Once you have the shunt leg calculated (which is just the square root of the product of the input and output R) all you have to do is reverse the sign for the input and output values, and then correct for the tower X. You can calculate for other values of phase shift using the formulas above, which can be handy for balancing the sidebands for best modulation performance with some transmitters.

You can also see from the formulas why you don't want to approach 0° or 180° phase shift in a tee network...the trig tries to divide by zero. :lol:
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

User avatar
Deep Thought
Posts: 3217
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: Proper way to adjust ATU...

Post by Deep Thought » Tue Jul 26, 2016 8:57 pm

Adding to what Bob said...
Its a very SIMPLE setup: 1kw feeding a single ATU comprised of a "T" network feeding a folded unipole.
Even though you have a unipole skirt it still needs to have a static drain to ground on the output side of the ATU. The tower may be grounded but that skirt isn't and the high reactance aggravates any lightning discharges because the skirt already has a high voltage at the feedpoint even at 1 KW. If there is no drain I would bet that is why you keep blowing things up.
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

COMMENG
Posts: 95
Joined: Wed Dec 18, 2013 4:30 pm
Location: Midwest

Re: Proper way to adjust ATU...

Post by COMMENG » Wed Jul 27, 2016 4:13 pm

In a similar vein, one cold use some MatLab code as below that takes into account any degree of lag. ("%" are comments and not code)

% This is a simple three-element T-Match Low Pass with Forth Element +j
% for Matching 50 ohm coax to a Low Impedance Antenna with XL.
% Note: This is for a -N degrees phase shift Low-Pass T circuit
% This is a more general Formula than for -90 degrees only where XL1 = XC2 = XL3
% Can be used in Phase Shift networks to also match impedances
clc
% "T" type ATU calculations
Fc = 1.0e6; % ENTER Assigned Frequency
pi = 355/113;
twopie = 2*pi;
Omega = twopie*Fc;
Rin = 50 % ENTER Impedance of Coax Feed Line
RANT = 30 % ENTER Resisitive part of Antenna Impedance
PhaseShift = 90; % ENTER PhaseShift (Note: 75 degrees seems to yield most sensible values)
Nphaseshift = (-PhaseShift/57.3); % Neg sign for Lagging; divide by 57 degrees/radian
SINEN = sin(Nphaseshift);
TANIEN = tan(Nphaseshift);
% Calculate Reactances
Xcshunt = sqrt(Rin*RANT)/SINEN; % (Second Element)
XL1series = (Rin/TANIEN) - Xcshunt; % (First Element)
XL2series = (RANT/TANIEN) - Xcshunt; % (Third Element)
% Calculate Component Values
L1firstseries = XL1series/Omega
C1Shunt = -1/(Omega*Xcshunt)
L2secondseries = XL2series/Omega
%
% Calculate any Series Capacitance needed to Cancel Antenna Inductance
XLant = 70 % ENTER measured j in ohms; Here Antenna j = 70;
Ccseries = 1/(Omega*XLant) % (Fourth Element)

The Results for a 30 ohm antenna with an Inductive Reactance of 70 Ohms at -90 degrees Phase Shift were:

L1firstseries = 6.1631e-06, or 6.16 uH or 6.2 uH (First Element),

C1Shunt = 4.1094e-09, or 4,109 pF or 0.004 uF (Second Element)

L2secondseries = 6.1635e-06, or 6.164 uH or 6.2 uH (Third Element)

Ccseries = 2.2736e-09, or 2,273 pF or 0.0023 uF (Fourth Element)

FirebirdTN
Posts: 51
Joined: Wed Jun 08, 2011 11:24 am

Re: Proper way to adjust ATU...

Post by FirebirdTN » Thu Jul 28, 2016 5:47 am

Thank you everyone for chiming in. Kind of in a rush this morning, so I will go over all the replies a little later. I just wanted to address a couple comments that stood out:

There is NO SDC. I didn't think one was required in a folded unipole antenna? My main employer has a (now defunct) folded unipole, so I will have to take a peek in the ATU and see if it is present in ours.

Thanks!

-Alan

User avatar
Deep Thought
Posts: 3217
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: Proper way to adjust ATU...

Post by Deep Thought » Thu Jul 28, 2016 8:08 pm

That's a common mistake...the tower is grounded but that skirt is not and it charges up real nice in a storm. Installing one will almost certainly stop the capacitor carnage.
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

User avatar
kkiddkkidd
Posts: 559
Joined: Mon Dec 03, 2007 11:13 am
Location: Lawrenceburg, TN

Re: Proper way to adjust ATU...

Post by kkiddkkidd » Sun Jul 31, 2016 5:55 am

Another thing to check is the ground system and general grounding. IF the builder subscribed to the "we don't need no stinking ground system with a unipole" theory... You might not have one. Or minimal gs or poorly connected or...

All the SDC's in the world won't do much if there isn't a path to ground.
--
Kevin C. Kidd CSRE/AMD
WD4RAT
AM Ground Systems Company
http://www.amgroundsystems.com
KK Broadcast Engineering
http://www.kkbc.com

Tom Osenkowsky (RIP)
Posts: 11
Joined: Wed Mar 16, 2016 4:38 pm

Re: Proper way to adjust ATU...

Post by Tom Osenkowsky (RIP) » Thu Aug 04, 2016 6:31 pm

The phase shift of the ATU also affects bandwidth. The rotation of the sidebands with respect to the carrier is critical only at the RF generation point. This would be at the plate(s) of the final RF amplifier in a tube transmitter or at the PA combining point in a solid state transmitter. The manufacturer of your transmitter can advise you what the optimum rotation should be at the antenna terminal since this is easily accessible. Remember the transmitter has an RF output matching network and the phase shift of this network must be considered in the overall picture. The transmission line phase shift msut also be accounted for in the overall picture.
To make things less complicated I developed software long ago to calculate the the required leg reactances which will produce the required phase shift to optimize impedance bandwidth. The sideband rotation may appear unacceptable at the ATU input terminals or transmitter antenna terminal but remember these are just physically easily accessible locations to make measurements.

The freeware program "Toolbox" has various options which allow the user to design or analyze a Tee network. You can download Toolbox from The BDR at:
http://www.thebdr.net/resources/software/index.html Look under RF Software. Choose option #17. Toolbox is freeware. This is NOT an advertisement since there is no other version of this program, paid or otherwise. It is a 16 bit program. No fancy graphics, etc.
You may enter measured impedance data and observe the results of the sideband rotation. If entering a transmission line be sure to use the electrical (not physical) length to account for velocity factor.

User avatar
Deep Thought
Posts: 3217
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: Proper way to adjust ATU...

Post by Deep Thought » Thu Aug 04, 2016 11:30 pm

If you want to run Tom's programs on 64 bit Windows, DOSBox will do it nicely.
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

FirebirdTN
Posts: 51
Joined: Wed Jun 08, 2011 11:24 am

Re: Proper way to adjust ATU...

Post by FirebirdTN » Sat Sep 03, 2016 7:16 am

I know its been a minute since I've posted. Just wanted to say THANK you for all who contributed.

If I am honest though-I have LONG forgot much of the math presented here. I could *probably* figure it all out, but am wondering if that is even necessary?

What I was wondering was more along the lines of the procedural method of getting the ATU tuned correctly.

As I mentioned above I have an impedance analyser, and have no problems fiddling with the coils to get 50j0, but somehow I am convincing myself that just because it presents a happy load to the transmitter, it doesn't make it right.

Just some example questions:

I know all the adjustments are somewhat interactive, but do you adjust one specific coil for primarily the resistance, another for the reactance, and the final for bandwidth? Or something else that popped in my wee brain-do you adjust the shunt leg of the "T" for resonance, and THEN adjust the other two coils for proper match?

Sorry guys...this is just stuff I don't fiddle with on a day-to-day basis. Any direction would be appreciated!

Oh, and for the software...Computer's are pretty much my niche', so 16bit software isn't a problem for me. Hell, I have done most automation tasks at work using good old batch files (since I am not a programmer). But on a personal level, have moved permanently away from Windows to Linux.

Thanks,

-Alan

W2XJ
Posts: 89
Joined: Mon Dec 03, 2007 12:15 pm

Re: Proper way to adjust ATU...

Post by W2XJ » Sat Sep 03, 2016 10:35 pm

I guess we start with there not being any original documentation. You really need to know first what the transmitter wants to see. As Tom O already said, sideband symmetry is the goal and every transmitter type has it's own requirements. In general, though, it often boils down to having a load at the TX that is a horsehoe on the Smith chart pointing either up or down. The phase shift through the T network functions in conjunction with the length of the transmission line. If those factors are known it is easy to design a T network that has the appropriate transformation. Then put that into a simulator and tweak in software. After that take the vaules for each leg of the network and systematically work towards the antenna. Terminate the input (coax) side of the network with a 50 ohm non reactive resistor and sequently bridge each element to match the simulator values. When complete, the entire network should show the conjegate impedane of the tower on the tower side of the network.

I do also agree with those who strongly suggest a static drain choke on the skirt wires.

User avatar
Deep Thought
Posts: 3217
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: Proper way to adjust ATU...

Post by Deep Thought » Sun Sep 04, 2016 9:45 am

FirebirdTN wrote:I know all the adjustments are somewhat interactive, but do you adjust one specific coil for primarily the resistance, another for the reactance, and the final for bandwidth? Or something else that popped in my wee brain-do you adjust the shunt leg of the "T" for resonance, and THEN adjust the other two coils for proper match?
It's not a good idea to just "wing it" with this especially with a solid state transmitter but since you asked...

The shunt leg of a tee network will affect the resistance (R) more than the reactance (X), while the input leg is all about the reactance you see at the input (since it is in series with that). The output leg will affect both.

Generally speaking, you set the output leg to the correct (design) value taking into account the load reactance, tune the shunt leg for proper R and then the input leg for j0 X. You still need to do the math (it isn't that difficult with any calculator) to get the proper output leg setting that produces the required phase shift.
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

Tom Osenkowsky (RIP)
Posts: 11
Joined: Wed Mar 16, 2016 4:38 pm

Re: Proper way to adjust ATU...

Post by Tom Osenkowsky (RIP) » Sun Sep 04, 2016 8:05 pm

If you use the TOOLBOX program you don't need to know the math, the equations are already in the program. Your first step is to contact the manufacturer of your transmitter. Ask them what the optimal sideband impedances at +/- 10 kHz from carrier are at the transmitter antenna terminals. This is an easily accessible location to measure the sweep.
Now measure the impedances at that location. Choose option 17, then start with option "B". You can enter the electrical length of your transmission line (taking velocity factor into account) and input the known values of shunt and output leg capacitors if the ATU is a common lagging Tee network. Now the program will iterate phase shifts using the appropriate keys. Select the phase shift and component values that produce the impedance rotation specified by the transmitter manufacturer. It is not as difficult as it sounds. All the calculations are performed for you. Keep in mind the component ratings to be sure you do not exceed the capacitor current maximum. One way I do this is to specify an operating power which takes modulation into account i.e. 1500 Watts for a 1 kW AM station.

Post Reply