John,

My comments are below your text snippets:

I have for the time being disonnected the first mixer, and am able to feed
in a 45MHz signal straight into the firts IF amp if I wish.
Now if I look at the output of the second IF with a spectrum analyser,
with
the power off to the first IF, but with the 2nd LO running, I can wind up
the second IF's gain, and the noise floor goes up, which I guess is
expected, but there are no peaks at 455KHz.

This would be expected, as there is nothing to bandlimit the output of the
high-gain 455 KHz IF, so you're seeing its broadband input noise amplified
by its gain.

If I connect the power to the first IF, and then wind the gain of the
second IF up, I get a noisy peak at 455KHz, that varies in aplitude with
the gain. The input to the first IF has a dummy load.

This is also to be expected, as the input noise of the first IF is amplified
by its gain, then bandlimited before amplifying in the second IF.


The second LO is a crystal oscillator, see my previous questions regarding
this.

The question is, should I expect to see a peak, I dont think so.
What could be causing this? The first IF is in a double sided PCB
enclosure.

I hope someone can help as this is driving me nuts....

Don't go nuts - it sounds normal. It would be a lot easier to say if
everything truly is normal if you could supply some numeric data -
fortunately you have a signal generator and a spectrum analyzer. I suggest
that you put a signal into the first mixer and observe the output of the
last IF on your spectrum analyzer. What input signal level does it take to
get 10 dB output signal-to-noise ratio. What level does it take to get the
same result feeding a signal into the first IF chain, then the second IF
chain, etc ??. Don't worry so much about gain during these tests, just the
signal-to-noise ratios. BTW, make sure the analyzer's resolution bandwidth
is approximately the same as the receiver's IF bandwidth for these tests,
otherwise you'll have to apply a noise bandwidth correction).

Good luck

Joe
W3JDR

Hi,
The receiver that I am building, has a 455KHz second IF. This consists of a
common gate amp followed by a collins mech filter 6KHz BW, and then two
AD631 gain controlled amps, with a max gain of 100dB.

The first IF is at 45MHz, and consists of a 20KHz BW monolithic crystal
filter with dual gate mosfet amps either side.

The second LO is at 44.545MHz, and the mixers are SBL-1's.

I have for the time being disonnected the first mixer, and am able to feed
in a 45MHz signal straight into the firts IF amp if I wish.

Now if I look at the output of the second IF with a spectrum analyser, with
the power off to the first IF, but with the 2nd LO running, I can wind up
the second IF's gain, and the noise floor goes up, which I guess is
expected, but there are no peaks at 455KHz.

If I connect the power to the first IF, and then wind the gain of the
second IF up, I get a noisy peak at 455KHz, that varies in aplitude with
the gain. The input to the first IF has a dummy load.

The second LO is a crystal oscillator, see my previous questions regarding
this.

The question is, should I expect to see a peak, I dont think so.
What could be causing this? The first IF is in a double sided PCB
enclosure.

I hope someone can help as this is driving me nuts....

Thanks,
John

On Sat, 17 Jun 2006 11:32:36 GMT, W3JDR wrote:

John,

My comments are below your text snippets:

I have for the time being disonnected the first mixer, and am able to feed
in a 45MHz signal straight into the firts IF amp if I wish.
Now if I look at the output of the second IF with a spectrum analyser,
with
the power off to the first IF, but with the 2nd LO running, I can wind up
the second IF's gain, and the noise floor goes up, which I guess is
expected, but there are no peaks at 455KHz.

This would be expected, as there is nothing to bandlimit the output of the
high-gain 455 KHz IF, so you're seeing its broadband input noise amplified
by its gain.

If I connect the power to the first IF, and then wind the gain of the
second IF up, I get a noisy peak at 455KHz, that varies in aplitude with
the gain. The input to the first IF has a dummy load.

This is also to be expected, as the input noise of the first IF is amplified
by its gain, then bandlimited before amplifying in the second IF.


The second LO is a crystal oscillator, see my previous questions regarding
this.

The question is, should I expect to see a peak, I dont think so.
What could be causing this? The first IF is in a double sided PCB
enclosure.

I hope someone can help as this is driving me nuts....

Don't go nuts - it sounds normal. It would be a lot easier to say if
everything truly is normal if you could supply some numeric data -
fortunately you have a signal generator and a spectrum analyzer. I suggest
that you put a signal into the first mixer and observe the output of the
last IF on your spectrum analyzer. What input signal level does it take to
get 10 dB output signal-to-noise ratio. What level does it take to get the
same result feeding a signal into the first IF chain, then the second IF
chain, etc ??. Don't worry so much about gain during these tests, just the
signal-to-noise ratios. BTW, make sure the analyzer's resolution bandwidth
is approximately the same as the receiver's IF bandwidth for these tests,
otherwise you'll have to apply a noise bandwidth correction).

Good luck

Joe
W3JDR

Thanks Joe,
I have just managed to get some more time on my project, after some delay,
I did some measurements as you suggested.

For the 2nd IF with gain at min:-
10dB increase in output with -80dBm input
20dB increase in output with -67dBm input
30dB increase in output with -59dBm input

gain at max:-
10dB increase in output with -119dBm input (very noisy output)
20dB increase in output with -109dBm input
30dB increase in output with -97dBm input

I then connected the 2nd to the first IF:_
gain at max
10dB increase in output with -100dBm input

When I disconnect the power to the 1st IF the noise floor is at -20dBm at
max gain.

With power to the first IF the noise floor is +10dBm with max gain.

As I drive in a higher amplitude signal the noise floor drops around the
signal, but the peak is still at +10dBm.

Is all this OK?

Regards,
John.