How does automatic antenna tuners work ?

[StoppTidigare]

ad8302 antenna calculation

Hi which components must one have to build an antenna tuner.
Whats the story ?

Kindest regards,
StoppTidigare

 

[flatulent]

They are reactive (L and C) networks that transform impedance. A T configuration with series capacitors and shunt inductor is the usual form.

The transmitter side voltage and current and the phase between them are measured. The ratio of voltage to current magnitude is the impedance seen by the transmitter.

There is a procedure for changing the L and C values to get 50 ohms at zero degrees phase angle.

 

[StoppTidigare]

Hi again,
yes of course but how do I get a value on transmitter side voltage and current and the phase between them are measured, Z_out and angle(Z_out) ? Doesn't one need a vector network analyzer to measure complex impedance ?

Suppose I know how to measure tha above. How should a resonable algoritm look like so it automatically work finds 50 Ohms ?
Kindest regards,
StoppTidigare

The transmitter side voltage and current and the phase between them are measured. The ratio of voltage to current magnitude is the impedance seen by the transmitter.

There is a procedure for changing the L and C values to get 50 ohms at zero degrees phase angle.

 

[flatulent]

I have never designed one of these so do not have details, but here are my ideas.

On manually tuned networks, you set the shunt inductor to a value related to the frequency. The capacitor to the antenna is adjusted for magnitude of impedance presented to the transmitter. The capacitor to the transmitter is adjusted to make the phase angle zero. If you cannot get the 50 =j0 you want, you try the next inductor tap and retune.

You sense voltage by a resistive divider on the line and current with a torroid with the line through the center and many turns of fine wire for the output. I think that there is a 90 degree phase shift, but am not certain. At HF you can use an ADC and do the calculations in the microprocessor. You can also use XOR gates to get the phase after using comparators to get logic levels. You use analog multipliers/dividers to get the magnitude part. Since phase is not used in the magnitude calculation, you convert the voltage and current signals to DC by rectification and filtering.

Very few companies make theses and I suspect their firmware is kept a secret. The SGC brand is a copy of a military system.

Here is a suggested start. Set the shunt iductor and output capacitor to the center of the range. Step the input capacitor in a successive approximation manner to get zero phase or close to it. Step the output capacitor up and down by one increment and measure the magnitude. Select the step that puts the magnitude closer to 50 ohms. Cycle through the above to get as good as you can. Then step the inductor up and down by one increment and do the complete above.

There is probably a way of trying the 64 combinations of 4 sets of values of each arm at the start to get the closest measurement. Then try the further 64 sets of combinations with smaller variations and select the best combination. This may converge faster.

 

[mr_ghz]

Hi StoppTidigare

For measuring phase and amplitude of two signals 'vector analyzer' there exists a one-chip solution. Look for AD8302 at analog's website (www.analog.com). I have already used this chip to create a simple network analyzer, it works fine.
The chip can't measure phase differences of more than +/-90°. But with two chips and a little bit of logic also this problem can be omitted.

Bye

 

[pool_77]

You can tune it by measuring SWR.

 

[twinsen]

You may be able to learn more here. This is a project in the 2003 ARRL Handbook

h**p://www.miami.muohio.edu/president/personal/w8zr/eztuner/ezcontents.htm

 

[mgf]

HI mr_ghz,

would be nice to see how you deal with 2 IC for measuring more than 90 deg of phase shift.

I have allready 2 ad8302 for that purpose, but I was thinking to strip out all other values of more than 90deg (plus or minus) in the mcu.

Also - supose you find the phase shift and amplitude:
- how do you know from where to start checking L and C values ?;
- or you just make some complex math, maybe Q factor involved here and find exactly what the C and L values are ?.

Is anybody out there using complex math with PIC's ?

regards,
mgf

 

[mr_ghz]

Hi mgf

The absolute phase measurement works like this:
The problem of the AD8302 is, that a phase difference of +20° gives the same output signal as -20° (and also +40° is the same as -40° and so on). I work with two AD8302 in parallel. One input has the same (microstrip-line) length to both chips. The other input has different line length. Because I always measure the frequency, I can calculate this phase difference. So the second AD8302 gives me an idea if I am in the upper (>0°) or lower (<0°) part of the diagram.

Bye

 

[mgf]

mr_ghz,

well, how to implement which channel leads or lags the other in phase ? -
I think the situation can change for different DUT.
I'm not so sure that I understand how u do it, but I believe in your practice, and therefore u can enlighten me just a bit please , 'cause I'm loosing my nerves here, or I'm so stupid I can't see it.. :? ...
with your VNA can u determine the complex impedance of a DUT ?

brgds,
mgf

 

[mr_ghz]

Hi mgf

Yes, I can measure the complex impedance of a DUT. As an example I would explain measuring the complex S11:
At a defined moment my output frequency of my LO is 5GHz. This signal is driven over a directional coupler (known characteristics) to my DUT. My coupler has a forward and backward output. The forward output is fed over a power-divider and two identical lines to both port A of the two AD8302. The backward output feeds a divider, and from this divider two lines with different lengths go to the two port B inputs. Because I know the frequency (5GHz) and the length-difference, I can calculate the phase difference between the both port B inputs, say it is -20°.
Now I measure (with the AD8302) the phase difference from port A to port B of the first chip. As an example i measure 43°. But now I don't know if it's +43° or -43° (see datasheet of AD8302). The phase output of the 2nd chip gives the solution: If I have +43° this output will have +23°, if I have -43° it will have -63°. So the 2nd chip is only used for the decision which of the 2 possible points is the correct one...
This works similar for S21 measurements.

I hope this would help

 

[mgf]

HI mr_ghz,

well, the picture is clear now,but with one exception:
and now I quote you : "The phase output of the 2nd chip gives the solution: If I have +43° this output will have +23°, if I have -43° it will have -63°." .... According to the ad8302 datasheet you cannot determine the sign of phase diff, wright ? So, how can you read "-63°" on the output of the 2nd chip. In the first example I understand that you substitute
the computed value of -20° (+43°-20°=+23° ok) ,but on -43°?.

Or maybe you have done some Vphs center point changes on the 2nd chip.
Also nice to hear that it works on 5Ghz.

regards//
mgf

 

[IICCEE]

This is some reference:

[1] , “Design of an electronically tunable microwave impedance transformer,” in IEEE MTT-S Int. Microwave Symp Dig., 1997, pp. 647–650.
[2] Y. Sun and J. K. Fidler, “High-speed automatic antenna tuning units,” in Proc. 9th Int. Conf. Antennas and Propagation, vol. 1, Apr. 4–7, 1995,
(Conf. Publ. 407), pp. 218–222.
[3] J. de Mingo, A. Crespo, and A.Valdovinos, “Input impedance antenna automatic matching system,” in Proc. IEEE Int. Symp. Personal, Indoor and Mobile Radio Communications (PIMRC 2002), Sept. 2002,
pp. 1872–1876.

 

[mgf]

FINE IICCEE,

but I guess one must be a member of IEEE to read those documents.
no other ideea on reading those docs.
anyway, noted.

regards,
mgf

 

[mr_ghz]

Hi mgf

When the phase difference is -63°, the chip gives the same signal as with +63°. So I have to check two possibilities: Either I have +23°, so the input phase is +43°, or I have +63° (in reality -63°), so the input phase is -43°.
Look at the AD8302 datasheet at page 10, in the upper left you see how the chip trades negative phases.
There exists (theoretically) one worst case: if I have to decide of the same phase, but with different signs. This can be omitted with a good calculated delay-line.
The 5GHz signal in my example was not the best choosen frequency: the chip only works up to 2.7GHz, perhaps with 3GHz. Higher frequencies in my VNA are handled with input mixers, so even frequencies above 3GHz can be mesured.

Bye

 

[vfone]

https://www.sgcworld.com/
Here you can find a lot of schematics and theory about auto-antenna tuner.

regards

 

[mgf]

HI mr_ghz,

maybe I'm using a different datasheet for ad8302 (I have Rev.A) and on page 10 upper left corner there is a diagram TPC25 re VPHS vs Phase input difference and in my oppinion with no relevance whatsoever on how the negative phase difference is delt with (this chip can not make this difference).
I'll give it a try with the known delay line, but I hope it must be some more easy way (like I will still use one ad8302 only for the value of the phase difference, and with some simple DBM I will get the sign) . Your posts appreciated.

brgds//
mgf

 

[mr_ghz]

Hi mgf

When you found a more simple way using only one AD8302, let me know!

Bye

 

[plasma]

i think you look for Programmable Coaxial Tuner,look in http://www.stepelectronics.com.tw/Focus/focus.htm
model CCMT-1808 (0.8 to 18 GHz)
you can find there Manual Microwave Tuners use them to get
the value of "L" and "C"
pl

 

[on5di]

Hi mgf

The absolute phase measurement works like this:
The problem of the AD8302 is, that a phase difference of +20° gives the same output signal as -20° (and also +40° is the same as -40° and so on). I work with two AD8302 in parallel. One input has the same (microstrip-line) length to both chips. The other input has different line length. Because I always measure the frequency, I can calculate this phase difference. So the second AD8302 gives me an idea if I am in the upper (>0°) or lower (<0°) part of the diagram.

Bye

Hi
That is a good trick ! i want to use the ad8302 too for an auto tuner, but i was going to use an rc network to create a 90 degree shift (close to 90 in fact ). the ad8302 looks nice and i suspect you can use it over a larger power range ( comparing to the schematics of an sg230 tuner where they use a minicircuits mixer to get the phase. )
What how much delay do you add (in degrees ) , i plan to use it from 1.8 to 30 Mcs . From your username i suspect you use it on UHF/SHF .
Regards Johan ON5DI (HAM callsign)