Optimising transmitter radiated power

[Paul Burridge]

Hi gang,

I've just developed a three stage RF amplifier for 40Mhz and the final
PA stage is shown below. The maximum power transfer occurs when the
load resistor has a value of 140 ohms whereupon it dissipates a
thumping 475mW of RF power! All well and good, but I need to radiate
as much of that power as I can - with only a telescopic whip antenna.
So I snip the load resistor out of circuit and the following questions
arise:

Where do I take the feed point to the antenna from? I realise I don't
have a hope in hell of matching 140 ohms without an "L network" or
whatever, but is that really necessary? Could I get away with taking
the antenna feed direct from a tap off the tank coil?

I only *need* 50mW ERP so am not *too* bothered about losing most of
the useful power.




+--------------|--------+---------------------------+12V
| | |
| | |
| C| | VC1 (5-60pF)
| L1 C| ---
.-. C| --- C3
820R| | | | 1n
| | | | ||
'-' +--------+---------------+||-------+
| | || |
| | |
|| | |/ |
-----||-----+------------| 2N3904 |
|| | |> |
| | |
C1 | +--------+-------+ |
1n .-. | .-. .-.
| | --- | | Rload| |
| | --- | | | |
820R'-' C2 | '-' '-'
| 1n | | 160R |
| | | |
| | | |
| | | |
| | | |
| | | +

+-----------------------+-------+-----------------+----
gnd


created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

 

[Winfield Hill]

Paul Burridge wrote...

I've just developed a three stage RF amplifier for 40Mhz and the
final PA stage is shown below. The maximum power transfer occurs
when the load resistor has a value of 140 ohms whereupon it
dissipates a thumping 475mW of RF power!

A wimp 2n3904 TO-92 audio-frequency transistor? What a choice!
How much power does it dissipate? How much if the load impedance
changes in usual weird ways, becoming inductive, capacitive, etc?

Thanks,
- Win

whill_at_picovolt-dot-com

 

[John S. Dyson]

Hi gang,

I've just developed a three stage RF amplifier for 40Mhz and the final
PA stage is shown below. The maximum power transfer occurs when the
load resistor has a value of 140 ohms whereupon it dissipates a
thumping 475mW of RF power! All well and good, but I need to radiate
as much of that power as I can - with only a telescopic whip antenna.
So I snip the load resistor out of circuit and the following questions
arise:

Where do I take the feed point to the antenna from? I realise I don't
have a hope in hell of matching 140 ohms without an "L network" or
whatever, but is that really necessary? Could I get away with taking
the antenna feed direct from a tap off the tank coil?

I only *need* 50mW ERP so am not *too* bothered about losing most of
the useful power.

As Win implied -- the 2n3904 isn't going to be very tolerant
of varying loads at such a power. I suggest using a tool like
EZNEC (or another of the NEC derived antenna programs), and it
will help you do some CAE on an antenna. It can provide radiation
patterns and feedpoint impedances. You'll be able to somewhat
plan and estimate the required drive (and also simulate some
ugly physical situations.)

I'd suggest using a slightly larger geometry transistor (but
it will present a lower impedance load to the driver.) I havent
done a 40MHz thing (except in simulations), but maybe try
a transistor that could withstand an uglier load and can
use a more effective heatsink -- if you really need the 475mw
RF power. I'd probably try to be ready to handle significantly
higher voltage and current that the steady state (and well
defined) load would cause.

Simulate the circuit with uglier loads also. Take a look at
the peak currents (and voltages) in the transistor. Make
sure that you don't do ugly things to any junctions (not
a good idea to seriously reverse bias the BE junction.)

However, since you only need 50mw ERP, you might do well by
being careful about the antenna design, and then the 2n3904
would likely be adequate.

With the antenna -- how much can you use? Programs like
EZNEC can help to estimate your needs/load/etc.

John

 

[Luhan Monat]

Hi gang,

I've just developed a three stage RF amplifier for 40Mhz and the final
PA stage is shown below. The maximum power transfer occurs when the
load resistor has a value of 140 ohms whereupon it dissipates a
thumping 475mW of RF power! All well and good, but I need to radiate
as much of that power as I can - with only a telescopic whip antenna.
So I snip the load resistor out of circuit and the following questions
arise:

Where do I take the feed point to the antenna from? I realise I don't
have a hope in hell of matching 140 ohms without an "L network" or
whatever, but is that really necessary? Could I get away with taking
the antenna feed direct from a tap off the tank coil?

I only *need* 50mW ERP so am not *too* bothered about losing most of
the useful power.




+--------------|--------+---------------------------+12V
| | |
| | |
| C| | VC1 (5-60pF)
| L1 C| ---
.-. C| --- C3
820R| | | | 1n
| | | | ||
'-' +--------+---------------+||-------+
| | || |
| | |
|| | |/ |
-----||-----+------------| 2N3904 |
|| | |> |
| | |
C1 | +--------+-------+ |
1n .-. | .-. .-.
| | --- | | Rload| |
| | --- | | | |
820R'-' C2 | '-' '-'
| 1n | | 160R |
| | | |
| | | |
| | | |
| | | |
| | | +

+-----------------------+-------+-----------------+----
gnd


created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Remove Rload and C3. Connect a 1/4 wave antenna about 1/3 way down the
coil from the 12 volt side. Experiment with tap location for best results.

 

[Paul Burridge]

A wimp 2n3904 TO-92 audio-frequency transistor? What a choice!
How much power does it dissipate? How much if the load impedance
changes in usual weird ways, becoming inductive, capacitive, etc

You must have exceptionally good hearing, Win.
Fairchild rate this device as an effective RF amplifier up to 100Mhz!
It's capable of about two-thirds of a Watt, as well - comfortably more
than I need. The biggest problem I face is knowing what the hell the
radiation resistance of a telescopic whip with a small GP is. If I
knew that, I'd have some useful clue about matching the stage's 140
ohm output impedence to it.

 

[James Meyer]

Where do I take the feed point to the antenna from? I realise I don't
have a hope in hell of matching 140 ohms without an "L network" or
whatever, but is that really necessary? Could I get away with taking
the antenna feed direct from a tap off the tank coil?

Yes, it (the extra network) is necessary. The short wire is a very high
impedance load. Taping the tank coil only results in a match to a lower
impedance, not a higher one.

The simplest matching network would be a second coil with many more
turns than the tank that was coupled to the tank by being placed close to it.
The antenna is attached to one end of the extra coil and the other end is
grounded.

Make the coupling variable by allowing for the coils to be moved closer
together and farther apart. Unfortunately, the reactance of the short antenna,
and anything close to the antenna, will be coupled back into the tank circuit
and will have a tendency to de-tune it. Vary the spacing between the coils to
vary the power delivered to the antenna being sure to re-tune the tank after
each change.

Jim

 

[John Smith]

You must have exceptionally good hearing, Win.
Fairchild rate this device as an effective RF amplifier up to 100Mhz!
It's capable of about two-thirds of a Watt, as well - comfortably more
than I need. The biggest problem I face is knowing what the hell the
radiation resistance of a telescopic whip with a small GP is. If I
knew that, I'd have some useful clue about matching the stage's 140
ohm output impedence to it.

Hi, Paul -

The radiation resistance of a 1/4 wave (approximately 6 feet at 40 MHz)
vertical ground plane antenna is approximately 35 ohms (half the value of a
half-wave dipole) in free space. You need a 6 ft vertical radiator and 4
radials of about 6 feet or slightly longer. Your standing wave ratio would
be about 140/35 or about 4:1.

As a vertical gets shorter than 1/4 wave, it becomes reactive capacitively
and its radiation resistance goes down.

You can quadruple the radiation resistance of the vertical monopole by
folding it over. That is, make a 6 foot hairpin out of the vertical. To
visualize this, imagine a vertical rod going up from the feed point to about
6 feet, making a 180, and heading back down to the feed point where it
attaches to the ground radials. You should wind up with about 4 times 35 or
140 Ohms of radiation resistance.

If you are going to put the antenna some distance away, then see if you can
find some 140 ohm coaxial cable. If not, you can use 75 ohm coax and suffer
the ~2:1 SWR. But it won't hurt much. In fact, if you make the coax physical
length equal to the velocity factor times the free space half wavelength or
about .66 times 12 feet or about 8 feet (or an integer multiple of a half
wave), then your transmitter will never know the coax doesn't match the
antenna.

If you are going to have the antenna attached directly to your circuit, just
try out the folded monopole I described above. You really should have some
kind of ground plane, but, what he heck, try it anyway.

If you absolutely need to use a shortened whip (less than about 6 feet), you
may need to compensate for its capacitive reactance and low radiation
resistance. Or just put it in there and try it. You may get enough radiation
to do what you want to do anyway.

Good luck.

John

 

[dyson]

You must have exceptionally good hearing, Win.
Fairchild rate this device as an effective RF amplifier up to 100Mhz!

It is pretty obvious that it is useful much higher than audio,
when considering the Ccb and Cbe type characteristics along with
the Ft.

It's capable of about two-thirds of a Watt, as well - comfortably more
than I need.

I strongly suggest being careful about the junction currents and
the voltages. If all you need (in the system) is 50mw, then
the 2n3904 type device will work fine. If you really need the
power amp output to be more than a few 100mw, and you have
a poorly controlled load, then the longevity of the component
might be at risk.

An anecdote: when I was a kid back in the late 1960's/early
1970's, I used a 2n5089 as an output stage for a small FM (100MHz)
transmitter. The 2n5089 is likely less well suited than the
3904, but it performed remarkably well.

One significant disadvantage of using Audio/Switching transistors
in low level RF/IF applications is probably more related to
the Ccb type feedback (and associated limited gain or instability.)
Even though such transistors can be used in those applications,
it is better (for the low power RF/IF stuff) to use the components
that are designed to have the lower feedback. (Some RF/IF compoenets
also have an interesting reverse AGC capability.)

For your application, the audio components should work just fine
(modulo the safe operating region type issues.) Nowadays, audio
transistors can be quite wide band (not like the old Germanium
components where fT would be fairly unimpressively low.)

John

John

 

[Spajky]

Remove Rload and C3. Connect a 1/4 wave antenna about 1/3 way down the
coil from the 12 volt side. Experiment with tap location for best results.

IMHO also a 1µH coil after that at the bottom of antenna would help to
match impedance of it (even more than that value if it is less than
1/4 wave type one ...

-- Regards, SPAJKY
& visit site - http://www.spajky.vze.com
Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
E-mail AntiSpam: remove ##

 

[Ken Smith]

A wimp 2n3904 TO-92 audio-frequency transistor? What a choice!
How much power does it dissipate? How much if the load impedance
changes in usual weird ways, becoming inductive, capacitive, etc?

I've used 2N3904s at frequencies this high. The cut-off frequency
(250MHz IIRC) is high enough that they are ok for smallish signals. Their
large signal characteristics aren't so good though.