Friday, November 30, 2012

LANforge Attenuator v1.0 is complete.

It took a bit longer than originally hoped, but today we release the LANforge RF Attenuator, version 1.0.  This system has 3 ATS0760-95 attenuator modules, with 0 to 95.5dB attenuation programmable in 0.5dB steps.  These modules support 0.7Ghz to 6Ghz frequency, so it handles all current WiFi frequencies.  It can be tightly integrated with the LANforge WiFIRE testing tools, and can be used standalone as well.

Full source code, including Arduino shield PCB files are released under GPL and CC-BY-SA licenses.

We are still waiting for the semi-rigid SMA cables (and hoping they will give us a bit better isolation), but even with just the standard double-shielded cable we feel the results are acceptable.  Here is a link to some test results using the LANforge WiFIRE test systems (based on ath9k NICs) and the attenuator.

If you have any questions, please send email to:

Monday, November 19, 2012

WiFi cable differences

In the previous post, I described how I tried to wrap some cables with copper tape to increase shielding.  That did not work as well as I had planned, and so new cables were ordered.  By now, I've taken to always ordering a few extra cables and I then cut one up to inspect it in more detail.  My first attempt at SMA to SMA patch cables used RG-316 cable, which is just a single-shielded cable.  It turns out this thing leaks RF like a sieve.

In a bit of paranoia, I also cut open one of the LMR-195 SMA to SMA patch cables, which I am using to connect the AP and Station machines to the attenuator.  Per spec, these are >90 dB shielding, and they are shielded with both aluminum foil and a wire braid.  But, in some testing, I found that just grabbing the 3 cables and gently lifting them up above the table they were hanging from took the maximum attenuation from around -86 dB to -65 dB.  This indicates to me that these cables must be leaking more than expected.

To improve the SMA to SMA patch cable performance, I ordered some new assemblies based on KSR-100 cable to connect the attenuator modules to the faceplate.  This cable has an inner aluminum foil shield and standard woven shield on the outside.  It claims >90db shielding, and it did seem better.  (The attenuator project can now go down to about -86 dB signal, and before it was tricky to go below -82).  I now have some supposedly even better SMA to SMA cables on order, for both inside the attenuator and to connect the attenuator to AP and Station machines....

My u.fl to SMA pigtails were also in bad shape.  I was also using pigtails made of RG-178, which is similar to RG-316 but smaller in diameter (and likely even leakier than the RG-316).

After looking and asking around, I found some vendors that would make pigtails wit double-shielded 1.32mm coax cable.  I have not been able to find any details on how well this shielding works, but on other stats, such as attenuation, it seems to perform better than most options that can handle u.fl connectors.  From what I can tell, it is likely that all '1.32mm coax cable' is double shielded, but I am not 100% certain this is the case.  While it seems you can always find a better SMA to SMA cable assembly, I have not found anything that looks better than the 1.32mm cable for u.fl to SMA pigtails.

For comparison, here is a smaller diameter single-shielded pigtail that came standard with a pci-e to mini-pci-e adapter card.

So, for now, I'm using 1.32mm coax for the u.fl to SMA pigtails, and have some 670-141 semi-flexible cable, that claims  > 100dB shielding....