[TriLUG] solved: can't connect to wap when move it to alternate location

[Neil L. Little]

Great trouble shooting Joe.
Glad you successfully found the root cause of the problem.

73,

Neil, WA4AZL

Joseph Mack NA3T wrote:
> On Tue, 13 Apr 2010, Joseph Mack NA3T wrote:
>
>> On Mon, 12 Apr 2010, Neil L. Little wrote:
>>
>>> Crap, they went out of business! Ok there was a place over on 
>>> Yonkers road in Raleigh but I see now they went out of business last 
>>> year.
>>
>> I did find a place that would rent a spectrum analyzer. Only 
>> $1901/mo. At that price, I should buy one.
>
> Story so far ...
>
> Linksys wap works fine at home, take it to the friend's store, I can't 
> connect to, or find the wap with netstumbler or iwconfig. Take the wap 
> home it works fine again. Same thing with two other Linksys waps and 
> another laptop (admittedly the same type of wifi card). There are two 
> other waps in the vicinity I can pick up, which I assume are operating 
> fine. Since my laptop can pick up other waps and not my waps, I assume 
> the problem is with my wap(s) and not my laptop(s). If I disconnect 
> the antennas to the wap, I can connect fine till I'm about 10' away. 
> If I put paperclips in the antenna jacks, I can connect till I'm about 
> 20' away. If I put the antennas back on, I can't connect at all.
>
> The store had a 5.8GHz phone system, which I eliminated as the cause, 
> by turning it off. My 900MHz phones at home had no effect on the wap. 
> Pete Soper told me that some 5.8GHz phones use 2.45GHz for one of the 
> directions, which would be in-band.
>
> The waps were non-operational during the week, but when I returned to 
> the store in the weekend, when I could safely mess with the network, 
> the wap worked fine. Presumably the interference isn't constant.
>
> hypothesis: wap is going CDMA deaf. Is the interferring signal in-band 
> or out-of-band? A spectrum analyser ($$$$) would be just the thing 
> here. Was there a bank of microwave ovens on the other side of the wall?
>
> John Mitchell kindly lent me his wi-spy v1, an in-band 2.45GHz band 
> scanner. I expected it would take a little while to figure this thing 
> out, but it was very easy to use. I could identify waps easily. It 
> didn't take long to find that 4 different microwave ovens were 
> radiating from ch7-11. I was surprised that they were all on the same 
> freq and relatively narrow - I thought they were 100's of MHz wide. 
> Looking with google, I find they're quite tightly regulated freq wise. 
> The microwave ovens weren't detectable more than 20' away.
>
> For John's device there is a Linux driver, by the kismet people, which 
> I got going in short time, and which gives a nicer display than the 
> windows wi-spy display. John's device used to be about $99, but isn't 
> made anymore. It's been replaced by a $199 device with similar specs 
> and there is a new $99 device which spec-wise is about half of John's 
> v1 device. I could have gone for the device I had at $99, but not the 
> replacement at $199.
>
> After a null trip to the store in the weekend when I found the wap 
> working fine, I returned during the week when the wap wasn't working. 
> Then the wi-spy to show no inband interferrence. All I saw was the 
> waps I could already see with iwconfig.
>
> Thus there was no in-band interference. It had to be out of band 
> interferrence. I did the following
>
> o put a 4db (and 8db) attenuator in the antenna line - the wap started 
> working. Presumably the 3rd order intercept point had been dropped 
> enough for the device to start working again.
>
> o put a 4 pole 60MHz bandpass filter centered on ch 6 in the antenna 
> line - the wap started working.
>
> The interference was almost certainly out-of-band.
>
> I didn't know the brand of the other waps in the vicinity which 
> appeared to be working just fine. They weren't bothered by the 
> putative out-of-band signal and I assumed they were getting the same 
> amount of signal as were my waps. Presumably I could have walked 
> around and found out what they were. I assumed they had a different RF 
> frontend than the Linksys. Looking in the digi-key catalog, I find 
> ceramic filters for wifi front ends at $0.80 each. Presumably the 
> LinkSys box doesn't have a filter, and I wasn't up for surgery on it. 
> Presumably I needed any other front end. I assumed all wap 
> manufacturers had their own front ends, presumably many with filters. 
> I bought a TP-Link wap from Intrex and it worked first time.
>
> Thanks to John for his wi-spy.
>
> Joe
>