hi,

I have chosen the IF frequency to be 8mhz (since I have crystals for
that freq) in my superhet project. So, VFO frequencies should be:

band (mhz) VFO (mhz)
---------------- ---------------
1.8 9.8
3.5 11.5
7.0 15.0
14.0 6.0
21.0 13.0
28.0 20.0

Well..now here is a question: can I use single chebyshev low-pass
filter for all those frequencies? Or maybe it's better to use different
filters with different cut-off frequencies?

thanks!

band (mhz) VFO (mhz)
---------------- ---------------
1.8 9.8
3.5 11.5
7.0 15.0
14.0 6.0
21.0 13.0
28.0 20.0


Hmm, did you check those frequencies for spurious responses? Low-side
injection can be a real problem. Remember even if you have a very clean
LO system the mixer itself has to be switching at the LO frequency and
will generate some pretty significant harmonics.

With a 20 MHz VFO on 28 MHz, the second harmonic would be 40 and minus
8 would put a spur response at 32.

At 14.35 you'd have a third harmonic of 6.35*3=19.05 19.05-8=11.05

I guess you might be OK, but you should look at all of that and the
bandwidth of the RF stages between the antenna and mixer.

does anybody have a link to any project/schematic of heterodyne
frequency generator?

On 16 May 2006 02:43:28 -0700, wrote:

I have chosen the IF frequency to be 8mhz (since I have crystals for
that freq) in my superhet project. So, VFO frequencies should be:

band (mhz) VFO (mhz)
---------------- ---------------
1.8 9.8
3.5 11.5
7.0 15.0
14.0 6.0
21.0 13.0
28.0 20.0

Well..now here is a question: can I use single chebyshev low-pass
filter for all those frequencies? Or maybe it's better to use different
filters with different cut-off frequencies?

Unless you are using a true multiplicative (four quadrant multiplier)
as a mixer, i do not see why you would want to put a low pass filter
between the VFO and the mixer.

However, if you are using some switching type double balanced mixer,
such as diode ring or hard driven Gilbert cell (such as MC1496, NE602
etc.) mixer, these should be driven with a clean symmetric square wave
to ensure accurate switching.

A symmetric square wave contains a rich series of odd harmonics that
could mix with the signals in the antenna. With the VFO at 6 MHz and 8
MHz IF, the third LO harmonic would be at 18 MHz, forming spurious
responses at 18+8=26 MHz and 18-8=10 MHz, the 5th harmonic would
produce other spurious responses.

Since you are apparently not designing a general coverage receiver,
but a ham band only receiver, IMHO, it is a much better idea to put
the effort to build filters into the front end (e.g. band pass filters
for each ham band or a constantly tunable filter) ahead of the mixer.
In the example above, the 14,0-14,35 MHz front end band pass filter
would strongly attenuate the spurious responses at 10 and 26 MHz.
While there would still be mixing products with the harmonics of the
local oscillator, with proper input filtering, the RF port of the
mixer would only contain some thermal noise from the input stage(s),
which at HF would not be an issue.

As others have pointed out, your VFO frequencies are quite high and
would ruin the stability. Also running the VFO below the received
frequency can easily cause harmonics to fall on the RF passband or IF
frequency.

For local oscillator frequency stability you might have to use the
mixing method (band crystal +/- VFO) or the offset-PLL method
(VCO-VFO) / divide-by-N and phase comparator. In both cases you should
check that none of the auxiliary frequencies (band crystal, VFO, VCO,
VCO-VFO difference, PLL reference crystal or PLL reference frequency)
or their harmonics fall on the input frequency band or the IF
frequency.

With a limited coverage receiver this should not be too hard to do.
Put all the generated frequencies and their harmonics into a
spreadsheet and check for several points on that band that all
spurious signals are outside the ham band under investigation and that
these frequencies do not come too close to the IF bandwidth.

Even if you do not currently plan to implement the WARC bands, it
would be a good idea to include them in this spreadsheet analysis, in
case you sometimes would want to implement them too.

Paul OH3LWR