PA amplifier mismatch load

[klick]

Hi,
I'm developing a power amplifier whit BFG520 at 915 MHz
these are the characteristics that I have reached:

Vcc=5V
Icq=5mA
Pin=9 dBm
Pout=19 dBm @ 33mA of Ic

these parameters are valid with load of 50 Ohm.
when disconnects the load and i open the output of PA the Ic goes at about 70 mA!!!
how can I do to prevent this?
about what parameters should I work?

Thanks.

 

[vfone]

The reason of this high current when you disconnect the load is because your amplifier starts to oscillate. This is not very unusual for RF amplifiers without proper load.
To fix this problem have to redesign the input/output matching networks.

 

[klick]

thanks for answer

but, when i disconnect the load i don't see even oscillation with spectrum analizer..

to design in/out matching i use network analizer and i optimize S11 and S22 about for less than -10dB...


what is wrong?

thanks

 

[vfone]

Try to disconnect the input power (without load) to see if is oscillating.

If you use input power and the current is going high without load, is possible to be due to high VSWR, which again is normal if the amplifier output match was designed for a specific load.

 

[Abhishekabs]

One of my friend's transistor burned because he left output open!!.
you said you don't see oscillations on SA when you disconnect load, SA it self will act as a load when you will connect to circuit.
Why dont you try this thing, connect attenuators at the output, so your circuit will get 50 Ohm load & you can do input matching.

 

[kspalla]

I would like to know the measurement setup with spectrum analyzer(SA).
If you directly connected to SA and carried the test to find the oscillation condition then the test setup is wrong.

 

[klick]

i know that Spectrum Analizer is 50 Ohm,
but i don't connectd directly the SA, i see the signal whit high impedance probe without touch the amplifier, after i test my PA whit sliding short , and i don't see even oscillation, but the Ic current goes high or low depend of phase of short

I don't understand if this is normal...

although the PA doesn't burn, it worsens the efficency....

 

[kspalla]

Is the IC is DC blocked at the output side?
Some ICs will have bias at the output pin.

 

[ChaoFan]

I think it's normal, the current varies with the impedance. As you said, you can use a sliding short as the load, when you change the load impedance by changing the phase of the sliding short, you can see the current is varying.

Added after 3 minutes:

Actually, sometimes mismatch will casue osillation, but it depends on the stability of the transistor, so sometimes it will not, even vswr is bigger than 30.

 

[Element_115]

How low in frequency do you measure ? It could be a low frequency osc.
Did you look at the devices S-Parameters to see if it is unstable at any
impedance ?

 

[courcirc8]

Your PA can be stable or not, its consumption will almost always vary with the output load impedance!

Take for example a switching PA (worth case). When you disconnect the load, the reflection is total, so depending of the matching network you have, the impedance seen by the chip is somewhere on the edge of the smith chart circle. Now imagine it is close from 0ohm, the PA will consume all what it can!

You can simply check that by adding a transition line (small cable) or better is you have a tuner and plot the consumption vs impedance.

What can you do? I don't know your biasing schematic but if you was using a current-mode biasing, you will be able to limit the current if this is really needed by the application. With a single-transistor PA, a low cost solution is to insert an emitter resistive degeneration (and to bypass if by an RF cap) This simple solution is not perfect, but it may help.

 

[wangyufn]

I think it is normal. If you read some commercial PA data sheet, you can find some specs are defined under certain output VSWR, e.g. 12:1. That is, output is not matched.

Current has two parts, through the transistor and through the load. When load is changing, Icc is changing (not Icq, if you have DC block).

 

[klick]

but, then depends of type of the transistor???


for my application, I use the 2SC5754, this is used for cellular,i think is good
but the mismatch load is NOT defined on data sheet........

If I use BFG21W for my application, is better????
the mismatch load is defined 6:1

 

[ChaoFan]

You are right, If the withstanding mismatch vswr is higher, its ruggedness is better.

 

[courcirc8]

It doesn't really depends of your transistor type, the more important things are the biasing scheme and the operating class. Does it work in class A,B,C,D,E,F,F-1,H...etc...?

This depends on the input drive level, the biasing and the output matching network.
Their eval board is probably a class B PA, but the biasing scheme is so simple that the consumption will vary with:
- temperature
- input level
- output mismatch.

Controlling precisely the maximum current consumption is not simple and it can require a current-limited source for the collector supply. Personally, I would avoid voltage biasing the base as in the application note. If temperature decreases, the current will rise very fast!

Added after 25 minutes:

Sorry..if temperatures rises, the consumption will increase.
Bipolar Vbe temperature dependence is -2mV/deg C
and
Ic% (exp(Vbe/Ut)-1) with Ut=kT/q

 

[klick]

I attach for you my circuit

In origin I used the BFG520 at Vcc=5V, but now I use the 2SC5754 at Vcc=3V

I set Vbias for Icq= 15 mA without RF signal

When I drive the PA whit RF signal at 915 MHz 13 dBm, I get Pout=24 dBm
whit Ic= 120 mA, the efficency is about 70%...This is good for me

but also in this case when I disconnect the load, or I connect the short, the Ic current goes high or low (depend the phase of short)

I attach my circuit if yuo have some advice, where I wrong

Thanks

 

[klick]

This circuit is the final RF amplifier of portable radiocontrol,
the problem is that I mismatch the antenna whit my hand, during the normal operation, the efficency goes low and the battery discharges more fast

 

[vfone]

The circuit is fine. Let like this, and don’t touch anything.
I don’t understand why you need for your amplifier to work without load if the transistor is not damaged at high VSWR.

 

[klick]

I dont' want to work without load,

I tested my circuit on extreme condictions because it is a portable equipment, and the antenna can be mismatched whit hand and the efficency of the system worsens

the current of the system goes high and battery discharges fast

I would like to reduce the consumption in mismatch antenna condictions

 

[vfone]

Sorry, but now you start to mix two different things.
First you mentioned PA without load, and now about PA behavior when antenna is mismatched due to handling.

Working without a load is a TEMPORARY behavior, and if nothing is damaged you don’t have to improve the current consumption of the PA.

Under antenna mismatch due to handling (which is a totally different story), you don’t have to work on the PA side. Have to work on the antenna side, and try to improve the antenna performance when is holding in hand.