Higher Education

Have a pretty technical question here.  I now know that the AP-225 has better receiver filtering for DAS frequencies.  Does anyone know what kind of receiver filtering is on the radio (or any AP for that matter) for the 902-928 ISM band?  We have some scientific research groups looking at placing 915 MHz wireless monitoring for varius freezers and other critical lab equipment.  Some of the repeaters have an output of 1/2 of a watt, which is pretty strong at 915mhz.  Was just wondering if I could get a response from an Aruba RF engineer.  Many thanks!

 

 

Our RF engineer says that there is ~50 dB of filtering rejection.

---

@cjoseph wrote:

Our RF engineer says that there is ~50 dB of filtering rejection.

---

Excellent!  Thanks for the information.  Not too bad.  Is that for all radios or certain AP models?

 

 

Specific to the 220 series.

Great. Thanks again!

Matthew L. Bonadies
Campus Network Operations
Wireless Networks
Indiana University - Bloomington
mbonadie@iu.edu

I’d like to preface this post by saying I know just enough about RF to be dangerous.  I’ve never been sure what to make of Aruba’s filtering.  I assume it has value, especially since they’re spending money to put it in APs when it’s not the type of thing that will drive many purchasing decisions.  But it seems to me that we have to accept this on faith.

 

I'm coming into this making the following assumptions, any number of which may be incorrect:

 

1. Power from a 900MHz transmitter at or near 2.4GHz is very low.

2. Receive sensitivity of the AP’s radios at 900Mhz is very high (insensitive).

3. If 1 & 2 are correct, it follows that the filter rejection at 900MHz really doesn’t matter.

4. The cellular filtering is one or more band rejection filters, designed to have low rejection in the 2.4Ghz spectrum.

5. The relevant interference is not in the intended band of the interfering transmitter(s), but the 2nd order IMPs which will fall around 1800MHz.

6. To know if there will be enough interference to worry about, or if there will be enough filtering to deal with it, you need to know the characteristics of both the transmitting and receiving radios and the path loss between them in addition to the filter rejection.

 

Sadly, as Wi-Fi designers we rarely know much about the radios we have to deal with (out of band characteristics, 3rd order intercept, etc.), so concerning ourselves with the with filter rejection values seems pointless.  All we can really do is test.  Am I wrong?

 

Chuck Enfield

Manager, Wireless Systems & Engineering

Telecommunications & Networking Services

The Pennsylvania State University 

Hi Chuck-

Good to hear from you. We met at a CIC wireless meeting a year ago or so in Chicago? Here are my ramblings. I think you are right, except the system in the 902 band can put out 500 mw. That's pretty strong for that frequency, especially close to a very small filter. There still is a chance for front-end overload receiver interference. Anyhow, here are my ramblings:


1. Power from a 900MHz transmitter at or near 2.4GHz is very low.

a. Normally yes, except for a repeater

2. Receive sensitivity of the AP's radios at 900Mhz is very high (insensitive).

a. Not very high. It should be very low due to a band-pass filter that effectively notches the 902 band. This means that the AP should be very "selective" in the 2.4 and 5.8 bands only, and deaf at 902(plus now the DAS system frequencies)

3. If 1 & 2 are correct, it follows that the filter rejection at 900MHz really doesn't matter.

a. agreed

4. The cellular filtering is one or more band rejection filters, designed to have low rejection in the 2.4Ghz spectrum.

a. Agreed. Maybe that same filter works in the 902 band to some extent?

5. The relevant interference is not in the intended band of the interfering transmitter(s), but the 2nd order IMPs which will fall around 1800MHz.

a. Agreed.

6. To know if there will be enough interference to worry about, or if there will be enough filtering to deal with it, you need to know the characteristics of both the transmitting and receiving radios and the path loss between them in addition to the filter rejection.

a. Agreed. That's why I am an RF dork.

I don't think you are wrong about anything, and I think we are talking the same stuff here. To you it may seem pointless, but I need to know everything about radios because I am a radio/RF dork. It makes me happy. Now, I can test what the engineer claims and get real world results. It would be nice to see an image rejection sweep from when they confirmed the 902 band for the AP-225. But that would only satisfy my dorkness........... See you down the road!

Matthew L. Bonadies
Campus Network Operations
Wireless Networks
Indiana University - Bloomington
mbonadie@iu.edu

I thought that was you, but I wasn't sure.  I've met a few Matt's from IU and I wasn't positive about your last name.

 

Regarding #2 (receive sensitivity), I was never completely sure how to word this.  If the sensitivity is low (AKA, poor receiver) then the number value is high (less negative, maybe even positive).  I'm not sure what's customary when referring to receive sensitivity.  Anyway, we were on the same page there. 

 

I definitely didn't mean to suggest it's pointless to worry about interference such as this, but I don't generally have access to the radio specs required to do the calculations before the fact.  I’ve seen some wireless mic systems that publish the OIP3, but otherwise where do you get that info?

Right. Not sure. It may be a competitive edge for Aruba to keep that information somewhat internal. RX sensitivity is in dbm, just like we see it when looking at the noise floor levels. -95dbm is awesome, -50dbm not so much. In radio, I look for how many microvolts it takes to do the same thing. Very little mV= very sensitive, more mV=slightly more deaf. If a receiver can successfully de-modulate and -95 signal and has high selectivity and image rejection from adjacent frequencies, all is honkey-dorey.

This is how I understand RX filtering. See example pic from google (to benefit other readers):

[http://users.innercite.com/kj6ko/images/902fil.JPG]

See the big null at Mark #2? That would be representative of the notch filtering presented to the receiver. Our pass-band of interest would start a trace at 902 and stop at 928. For the AP-225, that notch would "probably" be less effective at the band edges and probably have a max null at around -50dbm. At least, that's what experience tells me. It just depends on the design. Also, you need big (physically) filters to have high selectivity. -50 might be what the filter is limited to because if it's physical size. I hope this helps others understand RF filtering.

Matthew L. Bonadies
Campus Network Operations
Wireless Networks
Indiana University - Bloomington
mbonadie@iu.edu