RF signal amplification

[whiz115]

Hello

i'm about to build an RF amplifier for the 88-108MHz range..
i'm thinking to use the 2N3866 as output transistor so i can get about
1W RF power, what i need to know is what common transistors i can use
so i can drive the 2N3866 and also how can i match the impedance between stages? my
source output is ~10mW and it is based on a 2n2222.

also i need a bit of advice on how to minimize overtones since i imagine at such power
output the signal will be very strong across the band.

thanks!

 

[kchriste]

First I'll let you know that it is probably illegal where you live.
2nd, here is a link to an amplifier that may work for you.
There are other examples on his web page too.

 

[whiz115]

First I'll let you know that it is probably illegal where you live.

thanks for letting me know.. i'm aware of that, but i don't intent to open a radio station especially with 1W which is nothing inside a city.

i just want to build it test it a couple of times and then put in a drawer...

your link is helpful in some apects.. probably i need two stages but the BFG193 is not something i can find easily.. is there anything else more common?

also the circuit with the variable capacitors norrows the overtones?!

still i don't know how to match impedance between stages.

 

[Nigel Goodwin]

also the circuit with the variable capacitors norrows the overtones?!

Sorry, that makes no sense?.

still i don't know how to match impedance between stages.

The two pairs of variable capacitors match the impedances - the earlier stages aren't matched.

 

[whiz115]

ok nigel...i was thinking it might be an output filter for the harmonics....

do you think BC817 is suitable to drive the 2N3866?

 

[Nigel Goodwin]

ok nigel...i was thinking it might be an output filter for the harmonics....

do you think BC817 is suitable to drive the 2N3866?

Try one and see, RF is as much an art as a science.

 

[marcbarker]

also how can i match the impedance between stages?

There used to be an online Java Applet thingy that Hewlett Packard put on the net, that demonstrates the principle of RF narrowband impedance matching, usng L, pi and T filters, that are variable. There's an animated Smith Chart with it too. Like a game of golf. Maybe Nigel's seen it before?


edit: somethig like this **broken link removed** but better, it had knobs on it that you could adjust and watch what happend, just like a real RF circuit


edit edit: here it is Agilent Technologies | Test & Measurement

 

[whiz115]

ok marcbarker thanks... i'll check it.

i also want to know how to make a dummy load so the 2N3866don't get baked... can i use a 50ohm/4W resistor in parallel with the output?

 

[marcbarker]

ok marcbarker thanks... i'll check it.

i also want to know how to make a dummy load so the 2N3866don't get baked... can i use a 50ohm/4W resistor in parallel with the output?

light bulb (of the right ohmage that is) is about da best thing u can use for dis job, tune ur circuits 4 max brightness.

if ur using a resistor and it's wire wound, you need 2 stick a cap in series to null out the inductance u work out da value, not me

 

[mneary]

To answer your question in post #3,

also the circuit with the variable capacitors norrows the overtones?!

Yes, it does. It matches the impedance between the output transistor and the load, but since it is a tuned circuit it only does this well only at its resonant frequency. In the process of matching the tuned frequency, the overtones are mostly blocked.

As a dummy load, you can use a 4W 50 ohm resistor, but your typical 4W resistor is wire wound. You may be able to find some old 2W carbon comp 100 ohm resistors that you can wire in parallel.

A light bulb is much more rewarding, since it lights up when things are working well. 1W into 50 ohms is 7 volts RMS at 150 mA, so a pair of 12V 150 mA indicator bulbs in parallel give you a soft glow when it's working.

 

[unclejed613]

NEVER use wirewounds for an RF load. at RF frequencies, they are inductors and are a much higher impedance than their resistance, and for all intents and purposes might as well be an open circuit.

. BTW, to get 1 watt out of the 2N3866, you need 100mW input, as the transistor's power gain is only 10db. 1 watt is +30dbm (db referenced to 1 milliwatt). subtract the power gain from the desired level, and you get +20dbm, which is 100mw. so to get to 1 watt out from 10mw, you need another stage with a power gain of 10db.

below is the 400Mhz test circuit for the 2N3866. if you DOUBLE both the cap values AND the inductor values, it will work at 100Mhz. if you build an amp, even for 100Mhz, you will need to keep all component leads VERY short. "ugly construction" on a ground plane would probably be the best construction method. with 1W of RF out, the transistor dissipates 5V of heat, so the transistor may have to be heat sinked, which may not be easy, because the collector is the output element, and heat sinking will introduce a lot of capacitance to ground. as for output filtering, use a good 2 or 3 stage PI filter to eliminate harmonics. harmonics will not show up in the FM band, but at around 200, 300, 400, and 500Mhz (assuming a 100Mhz output). 200 and 300Mhz harmonics will be the strongest, so your PI filter should have a cutoff frequency of 120-150Mhz.

 

[mneary]

@unclejed613, cannot find your 400 MHz example.

Have you looked at the amplifier referenced in post #2? Do you have any comments or suggestions? Is the amplifier in post #2 an example of what you had in mind when you wrote "additional stage" in your post #11? Any thoughts on the OP's question in post #5 about using BC817 instead of BFG193?

Many low power transmitters use one-stage pi networks. -When you only have a +30 dbm carrier you need 30 db less attenuation than you need on a +60 dbm (1kW) transmitter.

 

[unclejed613]

sorry 'bout that, must have missed a step in the upload...

the amp below is from the data sheet for the 2N3866

BFG193 has an Ft of 4Ghz, the BC817 has an Ft of 100Mhz. at 100Mhz, the BC817 has no gain, but the power gain of the BFG193 is 15db.

i'd prefer the dual PI filter, since the harmonics of the FM broadcast band tend to be in parts of the spectrum used for public services, and safety of life becomes an issue.

 

[whiz115]

@unclejed613 thanks for the information you gave me.. still i don't know what transistor to use so i can drive the 2N3866.. i'm thinking if i might try the BF199 which is more into RF instead of the BC817 that i previously said.

concerning the heatsink of the 2N3866 it will be like a small ring, check this photo **broken link removed** i don't know if it
adds too much capacitance or if it is enough so i can keep the transistor cool

now about the dummy load...either i'll use a bulb as marcbarker suggested or the following circuit...which is also useful to measure the rectified output.

 

[unclejed613]

at 100Mhz, the bulb might also be too inductive. go with the circuit you have.

 

[whiz115]

at 100Mhz, the bulb might also be too inductive. go with the circuit you have.

marcbarker also said i could use the resistor in series with a capacitor
isn't it an alternative too?!

what about the BF199 do you find it suitable so i can i try it or is it too small?

 

[Nigel Goodwin]

marcbarker also said i could use the resistor in series with a capacitor
isn't it an alternative too?!

No, that's a really bad idea - it's essential that you use only non-inductive resistors. Place a number of smaller carbon resistors in parallel, to give you the require 50 ohms.

To adjust the transmitter into it, use an SWR meter, a dummy load gives a perfect SWR when the transmitter is correctly adjusted to 50 ohms.

 

[whiz115]

No, that's a really bad idea - it's essential that you use only non-inductive resistors. Place a number of smaller carbon resistors in parallel, to give you the require 50 ohms.

ok i'll use the circuit i uploaded...

 

[Hero999]

What load is the transmitter going to drive?

If it's a mono/dipole it should be tuned to 75Ω, if it's a Yagi array, it should be 50Ω and if it's a folded dipole you need 300Ω. Non-inductive resistors for use as dummy loads are available quite cheaply.

 

[unclejed613]

marcbarker also said i could use the resistor in series with a capacitor
isn't it an alternative too?!

what about the BF199 do you find it suitable so i can i try it or is it too small?

it has a gain of 10 at 100Mhz (the data sheet listed the GBW as 1100Mhz (instead of Ft, some data sheets are now listing it as GBW)). while the device might be a little on the lightweight side, i think it might get you enough drive for the 2N3866