|
|
|
mixer noise figure measurement - RF Cafe Forums
|
pigger
Post subject: mixer noise figure measurement
Unread postPosted:
Sat Jul 10, 2004 2:03 am
we are measuring mixer noise figure.
supposedly, it should be a smooth curve, but we saw lots of very high
spurs along that curve. Has anybody here met this situation before?
can anybody help me to analyze what happened?
Thanks in advance
Top
Guest
Post subject:
Unread postPosted:
Sat Jul 10, 2004 12:32 pm
How are you measuring the noise figure?
Are you using an automated noise figure meter that supplies both the
reference signal and the LO, or are you supplying an external LO and
just measuring noise floor difference from input to output, or some
other method?
- Kirt B. :smt024
Top
pigger
Post subject:
Unread postPosted: Sat Jul 10, 2004 3:07 pm
we use the automated noise figure meter (actually, a VNA with noise
figure measurement function) with both the ref noise source and LO
Anonymous wrote:
How are you measuring the noise figure? Are
you using an automated noise figure meter that supplies both the reference
signal and the LO, or are you supplying an external LO and just measuring
noise floor difference from input to output, or some other method?
- Kirt B. :smt024
Top
pigger
Post subject:
Unread postPosted: Sat Jul 10, 2004 3:37 pm
I think I didn't
make it clear.
I use a VNA ( just like an automated noise meter,
I guess, Y-factor method) to measure the mixer noise figure. I swept
the RF and LO frequency and fix the IF. VNA supplies the noise (or signal)
and LO signal.
The phenomenon is at certain freq, there would be
very high spurs instead a smooth curve. And "noise figure overload"
would also be shown on the VNA.
Top
Kirt Blattenberger
Post subject:
Unread postPosted: Sat Jul 10, 2004 8:25 pm
Offline
Site Admin
User avatar
Joined: Sun Aug 03, 2003
2:02 pm
Posts: 308
Location: Erie, PA
Assuming that your tester
is not injecting spurious signals into your DUT, it is posible that
you are generating a mixer spur that falls inside the measurement band.
That can definitely cause an erroneous NF reading.
You might
want to try backing off the mixer input power. If possible, connect
the mixer output to a spectrum analyzer and look to see if there are
discrete spurs stepping across the output band as the input steps.
Kirt B. :smt024
Top
Profile
DW
Post
subject:
Unread postPosted: Sun Jul 11, 2004 6:32 pm
Few
things to check:
1.) are you filtering your output such that
only your IF (assuming a downconversion) is getting through?
2.) Make sure there are no other pieces of equipment in the area that
are on.
3.) Make sure the LO and RF 10 MHz signals synced up.
Not having these synced up can induce some drift which can mess up some
measurements.
4.) If its an active circuit or there are amplifiers
around the mixing circuit, check for oscillations/spurs.
BR,
DW
Top
pigger
Post subject:
Unread postPosted:
Wed Jul 14, 2004 2:32 am
I don't see any obvious spurs when
there are no RF and LO signal coming in. but I will try backing off
the input power.
Kirt Blattenberger wrote:
Assuming that your
tester is not injecting spurious signals into your DUT, it is posible
that you are generating a mixer spur that falls inside the measurement
band. That can definitely cause an erroneous NF reading.
You
might want to try backing off the mixer input power. If possible, connect
the mixer output to a spectrum analyzer and look to see if there are
discrete spurs stepping across the output band as the input steps.
Kirt B. :smt024
Top
pigger
Post subject:
Unread postPosted: Wed Jul 14, 2004 2:35 am
I am not filtering
the mixer IF output, and I also guess maybe it's the LO leakage or RF
leakage who disturbed the NF measurement. But generally is it normal
to use a filter, or it is a must when measuring NF.
DW wrote:
Few things to check:
1.) are you filtering your output such that
only your IF (assuming a downconversion) is getting through?
2.) Make sure there are no other pieces of equipment in the area that
are on.
3.) Make sure the LO and RF 10 MHz signals synced up.
Not having these synced up can induce some drift which can mess up some
measurements.
4.) If its an active circuit or there are amplifiers
around the mixing circuit, check for oscillations/spurs.
BR,
DW
Posted 11/12/2012
|
|
|
|
|
