[NLRS] More on 902 junk

[w0zq at aol.com]

Hi John and all.

At my QTH it appears the biggest problem is the big signal at about 
890.8 MHz. There is also stuff below that but its farther from 902 and 
not quite as strong. There is also some stuff at 930 MHz that is 
strong. On this spectrum analyzer the signal at 890.9 measured -44 dBm, 
the signal at 930 MHz is about -55 dBm while the noise floor between 
these two is flat at about -78 dBm. Spectrum analyzers typically don't 
have a very sensitive front end, but these signals are strong enough to 
be easily detected. I took these measurements using the instruments 
power average tool averaging 100 samples (makes the trace nice and 
clean looking).

My first attempt to reduce the crude was to use a 4 pole cavity filter 
that is available from HyperLink Technologies (and others), Model 
BPF900, that is designed for the 902 - 928 MHz ISM band. The initial 
spec's looked interesting as it can handle 50 watts, is physically 
small (good for roving), not to pricey ($100), and has only about 0.5 
dB of insertion loss. The 40 dB out of band rejection is 815 and 1015 
MHz. Unfortunately it appears that its -3dB points are around 890 and 
940 MHz which means that it won't help with my close in 890.8 and 930 
MHz stuff. Testing with the spectrum analyzer confirmed this. It did 
knock down some crude at around 870 MHz by about 10 dB, but the 890 and 
930 MHz stuff was not phased. I did not try to retune it .... I simply 
used it as it was delivered. The tuning screws are accessible, so it 
MAY be possible to tighten the filter up, perhaps at the expense of 
increased insertion loss (more on this below).

The big cavity was provided to me by a local. I believe its from cell 
phone equipment and it did need to be retuned to use it for the 902 
band .... it was centered on 902.5 to cover both 902.1 and 903.1. I 
believe its a five cavity and physically its much bigger than the 
HyperLink filter used above. When it is in-line the 890.8 MHz crude 
drops from -44 dBm to near the noise floor at about -74 dBm, or down 
-30 dBm. The crude at 930 MHz disappears into the noise floor at - 77 
dBm, or at least -22 dBm rejection at that freq.

Perhaps my QTH is sort of a worst case as the problem is that big 
signal close in at 890.8 MHz. In any case, these numbers may provide 
some guidance as to how much rejection may be needed in a "good" 902 
filter .... like -30 dB at 891 MHz and 930 MHz. Based on RF power 
measurements it appears that the insertion loss in this big cavity 
filter is about 1.5 dB which is unfortunate, but being only somewhat 
deaf is much much better than being cold stone deaf.

I do wonder if the HyperLink filter can be tuned to provide -30 dB 
rejection at 890.8 and 930 MHz, and if so, what the insertion loss at 
902 MHz would be. If this is doable, it does provide one route towards 
finding a filter at about the $100 level. I may offer my sample up to 
experimentation .... seven screws to turn.

73, Jon
W0ZQ



Jon, are you aware of other sources for cavity filters? Some of the 
companies that make repeater duplexers have cavity filters that would 
cover the 902 band, but I suspect that the price new is pretty high. If 
you have two, there must be SOME other source out there! 
 
73 de WØJT 



________________________________________________________________________
Email and AIM finally together. You've gotta check out free AOL Mail! - 
http://mail.aol.com