Miniature IF Transformers 1

[ohmrun]

Many AM (amplitude modulation) radios utilize 455 kHz IF (intermediate frequency)transformers. My company uses a similar circuit to amplify a single CW (continuous wave) signal in the 100 kHz to 500 kHz band. However, it appears that the IF transformers are becoming harder to acquire. At least the through the hole versions. I would like to design a narrow band circuit (overall Q about 100) to amplify a single CW signal. The circuit voltage gain would be in the neighborhood of 100 dB.

Thanks

 

[nbucska]

The manufacturing of a custom IF will be not cheap. Perhaps
-- if youhave enough space -- you should make an TH to
SMD adapter.

 

[ohmrun]

A very good idea, nbucska. But my fault for leaving an incomplete subject. I would like to design a narrow band circuit with out transformers and still maintain Q > 100 and gain of about 100 dB. I am researching the alternatives. Does anyone have any ideas?

 

[nbucska]

If a data bandwith high, I don't see an easy solution
To replace 2 transformers you need 8 reactive elements (R/C). You would prefer good selectivity and flat top
-- this is difficult without tuning.

if the data BW is low enough
after a single LC stage, mix it down to as low frequency as
possible using a xtal osc. and filter it there.

Else check out xtal or SAW filter.

 

[ohmrun]

Thanks. I have considered down converting. I will look at that and the xtal and saw filter methods.

 

[VEBill]

How about a digital solution (DSP, or FPGA functional equivalent). I know that there are many applications where the signal processing is done all-digitally at frequencies way higher than 0.5 MHz.

 

[logbook]

It seems to me that the first key point is to sort out the spec. You have given a broad range of frequencies and yet a narrow band circuit. I assume this means that you don’t wish to tell us the exact frequency.

You say you need a Q of 100, which defines the bandwidth (in conjunction with the centre frequency). IF stages use this high Q for selectivity. You are using it for noise suppression I suppose.

The question then becomes why you want a transformer at all? Can’t you just use a single LC resonant circuit to get the Q? You would probably want to use a pair of capacitors to give some impedance transformation.

Are you making a transistorised amplifier or an opamp circuit? I would think that this was a trivial job for opamps and that you would have to try quite hard to get good performance (linearity, gain, SFDR) with transistors. (You get a hell of a lot of transistors in opamps for not much money). The bias circuitry is clearly not needed with opamps so the circuit may well be cheaper and simpler using an opamp based design. We are talking about one quad opamp with 25dB per stage. Use an opamp that is not unity gain stable as you need a lot of GBW product.

 

[ohmrun]

Thanks for all of the info VE1BLL and logbook. The amp would operate at a single frequency, hence the narrow band circuit. Currently, the operating frequency is 480 kHz. The high Q is for both selectivity and noise suppression. I have to worry about close proximity of output to input and minimize inductive coupling to prevent oscillation. What kind of op amp would you use in the 100 kHz to 500 kHz range?