SAW bandpass filter does'nt filter subharmonics of signal in passband ?

Thread Starter

Pifanatic

Joined Sep 14, 2018
9
Hello, i am trying to filter a CW of 2.4 Ghz of an Hackrf one with a SAW Banpassfilter with bandwidth of 120MHz at 2450 MHz. But the subharmonic of 1.2Ghz and others are not filtered although a CW of 1.2 GHz from the same hackrf get filtered. So i am asking if this is particular to SAW filter ? Thanks for yours responses !
 

Thread Starter

Pifanatic

Joined Sep 14, 2018
9
I have made the measurements several times with the same set up , also with an other (chinese) SAW filter and find the same results. So i am asking if this has to do with the mechanicals waves or something else.
 

Irving

Joined Jan 30, 2016
3,843
So what you're complaining about is the fact its only 30dB down at 1200MHz. Well that's SAW filters for you. They have very steep sides to the bandpass but relatively low stop-band attenuation. You need to precede the SAW filter with a broader but deeper bandpass to deal with sub-harmonics. Normally when you use such a device as, say, an IF filter for a scope, the input frequencies to the front-end mixers are higher than the IF which is itself higher than the maximum frequency of the scope.

eg. a 1GHz scope would have an input filter 3dB down at 1GHz and >30dB down at 2GHz. You then mix that input with a 6GHz signal giving 6 - 7GHz and then with a VCO from say 3.6GHz to 4.6GHz giving rise to a 2.4GHz IF. Other mixing products fall out of band and there are no sub-harmonics to deal with.
 

Thread Starter

Pifanatic

Joined Sep 14, 2018
9
30 dB down at 1200 MHz is ok, but it is only a few dB down at 1200 MHz when a stronger signal is present at 2400MHz (itself also a few dB down due to insertion loss)
 

Irving

Joined Jan 30, 2016
3,843
OK, I see what you're saying, red is with filter, green without - so in diagram 1 without filter, 2.4G is -9dB, 1.2G -23dB but with filter installed 2.4G is -15dB but 1.2G is -29dB rather than -53dB you expected?

Where does the 1200MHz sub-harmonic originate?

Is the BPF response measured in same setup?
 
Last edited:

Thread Starter

Pifanatic

Joined Sep 14, 2018
9
Exactly,
The 1200MHz subharmonic originate from the hackrf one itself which has a lot of différents kinds of signals.
The BPF response is mesured with the tracking generator of the same SNA.
 

Irving

Joined Jan 30, 2016
3,843
I have to say I'm perplexed. Its been a long while since I last used a SAW filter in anger, but I don't ever recall seeing anything like this. The fact that two different manufacturers give the same result suggests the 1200MHz signal is bypassing the filter somehow, but thats not borne out by the 1200MHz test...

What's your intended goal with this SAW filter?
 

Thread Starter

Pifanatic

Joined Sep 14, 2018
9
I want to transmitt to the QO100 satellite at 2.4 GHz with clean signal, this principally for experience and learning.

I will try a LC Bandpass filter which i am awaiting to compare. One could make the experience with any other SAW BP filter and two added RF générator, one in the passband and the other in the half frequency, but i haven't yet the material to get further. Although it would have confirm that this is a real phenomenon.
I figure perhaps the mechanical wave from half frequency could take advantage of the strong enough mechanical wave of frequency in passband to go through. But haven't yet read something about this.
 
Top