how to measure the output impedance of a transmitter ?

Thread Starter

shengwuei

Joined Aug 22, 2008
33
Hi Sir,

Hope I am posting at the right place.

My transmitter system block diagram is as below. I need to transmit a bipolar pulse at maximum 200Vpp/50MHz, but for now I cannot achieve acceptable output waveform at the load side(Vout), see below :



output waveform :
  • cyan / orange : time and frequency domain at Vtx
  • green / pink : time and frequency domain at Vout



As you can see, waveform at Vout is highly distorted, frequency peak at Vout is only about 15MHz, guess I need to do impedance matching on the load side.

However, for a transmitter I believe it is meaningless to use a network analyzer to measure its output impedance while it is not active. Also, it is not possible to measure its output impedance when transmitting pulses due to network analyzer is not able to sustain 200Vpp high voltage. So I need some other way to find out its output impedance.

In RF field as I know, a "load-pull" test is performed to find out the output impedance of a PA (for example : https://www.microwaves101.com/encyclopedias/load-pull-for-power-devices), though the test equipment is expensive and I do not have access to this kind of equipment.

I tried the simplest matching topologies like shunting an inductor or capacitor to ground, adding a series inductor or capacitor with the load, but none of them gave me a better waveform on Vout.

So, is there a feasible way to measure the output impedance of my transmitter ? Or is there a feasible way to do impedance on the output side ?

Thanks in advance !
 

Attachments

BR-549

Joined Sep 22, 2013
4,928
What does that load symbol represent? Is it some kind of transducer?

What is this circuit for? And what are you doing?
 

Thread Starter

shengwuei

Joined Aug 22, 2008
33
the load is an ultrasound transducer, typically a R-jX(capacitance) impedance,
I had used network analyzer to measure the impedance of transducer, but need to find a way to measure the impedance of transmitter(+ cable), thanks.
 

BR-549

Joined Sep 22, 2013
4,928
What is the impedance of the transducer....at 50 MHz? Let's start there.

Have you ever worked with(or matched) coaxial feed lines before? The easiest adjustment is line length. If that is not possible.....then a stub can be used.

Assuming the load can dissipate 50 MHz energy.
 

Thread Starter

shengwuei

Joined Aug 22, 2008
33
impedance of the transducer at 50MHz is shown above, 8.531 - 1.074J Ohm.

I have to know the impedance of the transmitter (the purpose of this post), then I may be able to match load impedance to be equal to transmitter impedance(no reflection) or its complex conjugate(maximum power delivery), thanks.
 

BR-549

Joined Sep 22, 2013
4,928
Well, usually the transmitter is designed to have a common impedance output........usually the expected feed line characteristic impedance.......such as 50, 75, 300 ohms.

The power amplifier of the transmitter usually has a matching output circuit.
 

RichardO

Joined May 4, 2013
2,270
Is the mechanical resonance of the transducer 50MHz? What is the "mechanical" load on the transducer? Both of these can effect how the transmitter reacts with the transducer.
 

KL7AJ

Joined Nov 4, 2008
2,229
Hi Sir,

Hope I am posting at the right place.

My transmitter system block diagram is as below. I need to transmit a bipolar pulse at maximum 200Vpp/50MHz, but for now I cannot achieve acceptable output waveform at the load side(Vout), see below :



output waveform :
  • cyan / orange : time and frequency domain at Vtx
  • green / pink : time and frequency domain at Vout



As you can see, waveform at Vout is highly distorted, frequency peak at Vout is only about 15MHz, guess I need to do impedance matching on the load side.

However, for a transmitter I believe it is meaningless to use a network analyzer to measure its output impedance while it is not active. Also, it is not possible to measure its output impedance when transmitting pulses due to network analyzer is not able to sustain 200Vpp high voltage. So I need some other way to find out its output impedance.

In RF field as I know, a "load-pull" test is performed to find out the output impedance of a PA (for example : https://www.microwaves101.com/encyclopedias/load-pull-for-power-devices), though the test equipment is expensive and I do not have access to this kind of equipment.

I tried the simplest matching topologies like shunting an inductor or capacitor to ground, adding a series inductor or capacitor with the load, but none of them gave me a better waveform on Vout.

So, is there a feasible way to measure the output impedance of my transmitter ? Or is there a feasible way to do impedance on the output side ?

Thanks in advance !
The best way to measure transmitter output impedance is with a RESISTIVE load of known value (dummy load). The piezo transducer is highly reactive. Get a good wide bandwidth resistor, and then measure the difference in voltage between the resistor connected and then disconnected. The voltage drop will be proportional to the load resistance. If the voltage drops to 50% of the unloaded voltage, the output impedance is 50 ohms, etc.
 

danadak

Joined Mar 10, 2018
4,057
and I do not really know what do you mean by "thru V measurement for various tuner settings one could extrapolate back to unknown Ztxout", could you explain more on this ? Thanks !
If there is a black box, terminated in 50 ohms R, and that box is connected to transmitter
one would think that for various settings in tuner, L, C, and a measurement made that one can
infer the Txzout.

Example, we match tuner to max Pout,, so its conjugate matched, then m,esure the L and C
in tuner, that of course had to be conjugate of Txzout.

Regards, Dana.
 

KL7AJ

Joined Nov 4, 2008
2,229
If there is a black box, terminated in 50 ohms R, and that box is connected to transmitter
one would think that for various settings in tuner, L, C, and a measurement made that one can
infer the Txzout.

Example, we match tuner to max Pout,, so its conjugate matched, then m,esure the L and C
in tuner, that of course had to be conjugate of Txzout.

Regards, Dana.
Yes....another variation on the non-reactive resistor trick.
 

KL7AJ

Joined Nov 4, 2008
2,229
Yes....another variation on the non-reactive resistor trick.
It might also be useful, if feasible, to temporarily remove the transmission line, and connect the load directly across the transmitter. This takes out the impedance transformation of the transmission line....which isn't hard to calculate with a Smith Chart...but one more item to complicate matters. This is especially true since the Piezo load impedance is far removed from 50 ohms.
 

Thread Starter

shengwuei

Joined Aug 22, 2008
33
Hi KL7AJ and all,

I tried to remove the transmission line and put a 50 Ohm load on transmitter output, then measuring output waveform/voltage level from 1MHz up to 50MHz.
When frequency is low(1~30MHz), output waveform is clearly showing a square wave, meaning that transmitter output impedance is purely resistive soI can calculate the "resistance" simply by RL*(Vi/Vo-1) (the voltage divider rule). But when frequency is high(40~50MHz), output waveform is highly distorted, showing at high frequencies transmitter output impedance is away from purely resistive, in this case I cannot find the resistance and high frequencies are exactly where I need to do impedance matching comparing to low frequencies.

  • 45MHz, 5 cycle pulse
    45MHz_5cycle.jpg

  • 50MHz, 5 cycle pulse
    50MHz_5cycle.jpg

Appreciate anyway, thank you all and further suggestions would be highly appreciated.
 
Top