Now that it has been a few months, how have people found the WI-SPY to be?
I emailed the WIFI folks regarding the external antenna version - here is their reply:
"We are hoping to have the external antenna version available in 3-4 months; it will have an SMA (or reverse SMA) connector.
We will most likely ship with a small SMA antenna so that Wi-Spy can be used out of the box, but users are welcome to attach any antenna to the Wi-Spy (for directionality, gain, etc).
Due to changes in the plastics and higher manufacturing cost this version will likely sell for [for somewhat more money -- they aren't
100% sure of the price, so I'm not posting their proposed price, as per their request].
If you need the directionality of the external antenna I would suggest waiting, otherwise the current version is fine for troubleshooting wifi."
I went to the hardware store and got the fittings per Jeff's photo, but I found an old beat-up photoflood reflector at a flea market that looked closer to a parabola (if you ignore the dents), and that seems to work okay. The fittings are nice in that they can go into a lot of different reflector types.
I don't know if Wi-Spy people will read this, but I did notice a couple of things they could do better with the software. You can put the cursor on any of the three traces, but the dB readout on that cursor still bounces around with the instantaneous trace - - it would be better if it followed the actual trace the cursor was assigned to. The auto-scaling of the average trace is nice, but why can't there be something that tells what the new scaling is? - - dB per division, maybe. If the peak trace goes to flat top limiting, is the electrical peak truly limited, or is it just the display that limits. If the electrical peak isn't limited, then perhaps the average trace is still valid through what looks like overload signals. For judging antenna performance differences, the average trace would seem to be more meaningful. I set a D-Link router to a beacon interval of 3 milliseconds to increase the average energy readout in an antenna study, and the average trace became quite useable. But when I went to beaconing at 1 millisecond intervals, I think the transmitter overheated, judging by the way it lost symmetry in the channel it was assigned to.
I find the device to be quite useful, and with further enhancements it will really be a valuable item for antenna, router and bridge optimization and troubleshooting - - looking forward to whatever they can come up with.
Well, it's a stainless steel salad bowl I picked up at the hardware store for about $8. Unfortunately, I drilled through the bar code label. I recall that it was made in India. It's NOT a parabola and could easily be improved. Meanwhile, search Google for "stainless mixing bowl" and you'll find a collection of possible bowls. Try to get one with a flat bottom as they're easier to drill and punch.
See: |
formatting link
for some discussion on the design. Bashing the 1 3/8" D hole should be done with a chassis punch and NOT a circular hole saw, which makes a mess.
If you don't wanna deal with the mixing bowl, then a flat plate or corner reflector will work. Make the flat part 1/2 wave diameter, and bend the edges up to form the corner reflector. Sorta looks like a parabola but it's really a corner reflector. Kinda like a solar cooker reflector.
You shove the dongle into the 1" PVC pipe with a USB extension cable. It takes some effort to find the optimum focal point. The nice part is that it requires no modifications to the dongle.
One change I wanna try is to through drill the two PVC electrical joints so that the 1" pipe can be shoved all the way through them in one piece. I'm digging through a large pile of ancient end mills as we speak but managed to cut my fingers. Oops.
Balance? It hardly weighs anything. Not a problem. Use your imagination for the back piece of 1" PVC. You could install a PVC "T" connection, which would form a pistol grip handle. Drill it for a
1/4-20 bolt and mount it to a tripod. Be creative.
One of my non-geek friends saw the salad bowl antenna in my truck and commented that it was nice that I was building a bird bath or fountain. Sigh.
Hmmm, chassis punches seem to be about $50... A bit steep for one time use (especially considering Wi-Spy will be updated shortly). I'll ask some handy neighbors, but I doubt they'll have one either.
The Greenlee punch isn't that rare - - any electrician would have a set of conduit punches including the one for the 1" fittings that Jeff showed in the picture. I don't think I paid more than $40 for my set of 4 Greenlee punches at the flea market. The 1" punch is also stamped 34.6mm and I measured mine at 34.5mm, so that's pretty close to the actual hole size. Ebay looks like a good source for the 7235BB 1/2" to 1 1/4" set.
Another way to make this hole would be to use a nibbler. I think Radio Shack sells a nibbler of sorts, and there is also the venerable Adel Nibbler that has been around for ages, but the price has ballooned with all the inflation. I also found an excellent compound tin snips that cuts left, made by Midwest in USA (Wiss is not as good), and this would do the job, too. The final method, and the least desirable due to all the work involved is to drill a bunch of holes with very little metal between them in a circle, and then snip through the separating metal with diagonal cutters, and file the rough edges smooth.
34+ mm????? I sure hope you read that wrong. 1" = 25.4mm. I can understand a little bit of difference, especially since US/Imperial measures tend to be "nominal", but a 30% error seems extreme.
One inch would be the internal diameter of the PVC. This is fairly hefty stuff. 1" sched 40 PVC for water is 1-1/4" outer diameter. That's the hole you would be drilling.
If you don't have a chassis punch, you should be able to drill through the bottom by nailing it to a piece of scrap wood, and drill it with a hole saw. Two nails or drywall screws at 1" spacing, drill bit centered between them... I think that would be okay, but Jeff's the one who has experience in mangling a bowl. I've drilled through some sheet metal successfully, drilling into backing wood. The bowl makes it a little more interesting.
The alternative would be a flange instead of the feedthrough coupling, and a much smaller hole, just for the cable. The bolts to secure the flange could bolt to something on the back side that provided a handle.
I found that 1 1/4 was a bit tight. It would work, but the PVC parts ended up a bit tilted as the stainless shredded the plastic thread. I also wanted a little bit of slop to compensate for centering inaccuracy. Methinks 1 3/8 is the correct punch size.
Just a reminder... use the electrical gray PVC fittings and NOT the white plumbing PVC fittings. The electrical type are straight thread, while the plumbing are tapered. The electrical goes together and comes apart quite easily. The white plumbing type are very difficult to manage.
You proceedure works just fine with a drill press. However, the bowl was far too large to fit in my cheapo junk drill press. So, I tried it using a hand drill and eended up with a mess. It might work with a hole saw and a block of wood in a drill press, but I haver wast able to try it.
The cable will fit through a smaller hole but the typical USB connector will not. The USB connectors are permanately attached to the cable and cannot easily be removed without cutting the cable and then splicing it back together.
I found the UPC bar code on the salad bowl "Item #108" 21614-00071 Made in India I haven't tracked down the manufacturer yet.
If I look around local stores -- what are some things to look out for?
I guess as close to parabolic as possible? Size between X and Y? I see on that other website, alot of mesh or metal strip like bowls -- which makes it easier to create something without drilling... how useful are those?
Avoid salad bowls with almost vertical sides. Vertical sides don't contribute any gain. It should resemble a parabola.
There's 2 ways to use a wok (shallow parabola). One is as a parabolic reflector which will have a focus quite far from the dish. The other is just as a flat plate reflector, where the USB dongle antenna is about 1/4 wavelength from the reflector. At that point, you might as well use something flat, like a pizza dish. The gain is about the same.
Yep.
Nope. The bigger the dish, the higher the gain. Try not to get rediculously big. Also, materials are an issue. Copper, stainless, and aluminum are good. Steel, galvanized, and anodized are not so good.
They work. Look at it this way. ANY antenna is better than the circuit board antenna found in the typical USB dongle. If you don't wanna drill, a pie plate will do. The results won't be optimum and a proper design will yield more gain, but that requires drilling, calculating, measuring, and some effort.
I noticed that there is now an upgrade for the software that runs the Wi-Spy dongle - - latest version is 2006.3.17.0. There is no obvious mention of the reason for the upgrade in the User's Guide, but maybe something was changed through the uninstall-install process. I haven't checked the Metageek FAQ lately.
I've been checking the dB calibration of the display with an attenuator set that's good to about 4 GHz, and find, as I suspected, the numbers are almost meaningless. For display levels above -47 dB, there is a consistent 1.5 dB change for every 1 dB difference in the Wi-Spy display. Below -47 dB, the actual change runs about 3 dB for every 1 db difference in the Wi-Spy display. This information might come in handy when comparing antennas on an Access Point with a Wi-Spy receiver. Using a 1.2" whip on an N connector as a reference, I found a Biquad antenna gives about 9 dB gain, while a "Cantenna" made with 2 large "Bush's Beans" cans and a 4" to 6" heating duct adapter gives about 12 - 15 dB gain. We also built some Helix antennas on
42mm and 33mm forms, both about 16" long, and found them only about 3 dB better than the 1/4-wave whip! The polarization loss doesn't help, I guess. So your recommendation, Jeff, to put up a Biquad is very good advice. The Can-with-horn performance was a surprise, and coupled with its greater directivity and lower sensitivity to off-axis sources for lower interference and noise, it might be good for critical situations, but it's larger and more wind-prone, so it's not so easy to apply. I made a cradle out of some plywood and a couple of strips of angle aluminum, all held to the "Cantenna" with stainless hose clamps. The plywood then can be U-bolted to a mast.
Has anyone attempted to install an antenna connector on his Wi-Spy? You know these units without a connector won't be worth much after the "connectored" product arrives, so maybe it's worth the effort to tear into the ones we have and put on-line a few photographs and a mod procedure.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.