getting good range .. - ditching USB wireless adaptor. What PCI one?

snipped-for-privacy@90.usenet.us.com hath wroth:

Yeah, I just hate it when reality ruins a good theory. Many of the antenna models have various shortcuts to reduce calculations. One of them is for symmetrical antennas to simply calculate the half hemisphere and assume that the other half is identical (mirror imaged). I do this all the time. The result is usually something like a hemisphere.

You're correct that it's not a full sphere that's symmetrical in 3 axes. While the shrinking antenna more closesly resembles a point source radiator, the relative size of all the scrap metal hanging around the antenna causes all kinds of goofy distortions in what would normally be a perfect sphere. Reality sucks.

Well, again is a matter of what direction the antenna models are created. By convention, the PCB is oriented horizontally, which is what I called the "normal" orientation.

Well, I could try it here if I could find what I did with the DWL-122. I think it's on my neighbors roof.

Right. I can see that Hawking mini dish as being best to recommend to somebody as a guaranteed solution. Personally, I really like the idea of being able to change antennas. But maybe it's best to already have a high gain antenna.

Maybe not, I've not used it. It's the claim they make.

Have you compared them? Just curiosity more than anything. I believe you, because a high-gain antenna can do so much more than just a 200 mw amp..(I'm going on memory on that), just wondering if you've got any experience with that the Senao USB...

Cheers, Steve

Seriously James, consider the other solutions I suggested before getting a PCI card. It does not look like it will help you unless you get antenna cables to extend it.

The fact that you have not yet even tried placing the usb adapter in a better place tells me that you are jumping the gun.

Placement can be critical. Antennas make a world of difference. The bus that your radio connects on is not the issue.

Get that USB extension cable FIRST ! You will need it somewhere else someday anyway.

Steve

On 6 May 2007 09:32:39 -0700, seaweedsteve wrote in :

I have not directly compared them. I just based what I wrote on the Senao USB needing real magic to beat it. ;)

The more I look at the Hawking the more it does look like the one to recommend. The fact that it has signal LEDs on it means that anybody can point it. And it is showing slightly better transmit power on it's specs as the Senao. -23~24dbm @?speed Hawking vs 22dbm @6-24Mbps -Senao ) for whatever that's worth. In other words, equivalent.

Here's the data for the claimed increased sensitivity on the Senao

Senao- Receive Sensitivity (Typical)

l2.412~2.472G(IEEE802.11g)

6Mbps@ -91dBm; 54Mbps@ -76dBm

l2.412~2.472G(IEEE802.11b)

11Mbps@ -91dBm;

1Mbps@ -96dBm

On 7 May 2007 11:19:51 -0700, seaweedsteve wrote in :

The problems are that there is no rigid standard, and Everybody Lies(sm) to a greater or lessor degree, so specs alone aren't really comparable or dependable.

I'll add it (later) to my slowly growing list of lies and fabrications for RX sensitivity specs. See:

Not too shabby but fairly close to the middle of the lies and fabrications. Most manufacturers specify exactly what the chip vendor suggests for sensitivities. Never mind all the connectors, matching, coax cables, and diversity switches in between the chip and the antenna connector. When I had the equipment and time to bench test the product, my results were rarely as good or as consistent as those of the manufacturers data sheets. There was even one manufacturer (name forgotten) who had identical numbers for every one of their products. Yeah, sure.

Anyway, treat receiver sensitivity numbers with suspicion unless you can verify that the manufacturer actually made the tests, and under what test conditions (i.e. at what bit or packet error rate).

Everybody lies, but that's ok because nobody listens.

On Thu, 03 May 2007 22:44:53 -0700, Jeff Liebermann wrote in :

Macro shots of the Hitachi mini 2.4 GHz antenna I used:

Picture of my ThinkPad T41 with the keyboard lifted up (metal plate on bottom) to show installation of modem+Bluetooth card and antenna.

The antenna comes with sticky tape applied, which I used to stick it to the edge of the plastic battery compartment, just behind a memory socket. The short black cable is routed straight to the modem+Bluetooth card in the same indentation used for the much longer gray Wi-Fi antenna cables. (The Wi-Fi card is out of view at the front [lower part of image].)

Still working fine.

Ahem. There most certainly are 802.11 measurement test proceedures and standards. They're all in the IEEE 802.11a/b/g specifications.

See:

for an overview of the really cool but overpriced test equipment required to perform the measurements. At the bottom of the page, are the applicable IEEE document names and section numbers. Of course, it's not really intended to understood by mortals. For example, from the 802.11g rx sensitivity spec:

19.5.1 Receiver minimum input level sensitivity The packet error rate (PER) of the ERP-OFDM modes shall be less than 10% at a PSDU length of 1000 bytes for the input levels of Table 91 of 17.3.10. Input levels are specific for each data rate and are measured at the antenna connector. A noise figure (NF) of 10 dB and an implementation loss of 5 dB are assumed. The PER of the ERP-DSSS modes shall be as specified in 18.4.8.1.

Ummm... right, whatever, ok, say what, huh? Like I previously said, reading the 802.11 specifications will turn one's brains to mush.

Anyways, I don't lie. I prefer to guess(tm), which is sufficiently close to lying to be equally effective.

On Tue, 08 May 2007 02:59:19 GMT, Jeff Liebermann wrote in :

My reference was to the lack of agreed standards for claims in marketing materials, as in the case of audio amplifier output.

On Tue, 08 May 2007 02:31:13 GMT, Jeff Liebermann wrote in :

Thanks to deregulation of advertising. Thank Republicans for that.

He was listening. Lots of other people do too.

Any reason to think that radio (vs antenna) sensitivity varies much in these various USB devices? Is there simply no way for "mortals" to judge?

I knew what you guys thought about claims etc, but still tend to go for the product that is marketed as "more powerful/more sensitive" over the one that says, um, "easy to use".

I expect claims to be exaggerated but perhaps having some foundation....

Probably just superstition.

Steve

John Navas hath wroth:

Oh, that wasn't clear. There actually are advertising standards for products. The basic rule is "cover thy ass" which means that any claims and numbers need to be substantiated if challenged in court. For example, I could purchase a random router, submit it to an FCC test lab to verify that it meets both FCC specifications and published specifications, and sue for false advertising. This has actually happened in some industries. The usual result is something dumb like the court deciding that RF issues are not within their jurisdiction or a settlement that allows the manufacturer to tweak that specs with something like "up to 108Mbits/sec". The dividing line for actionability is very clear. If damage is done, there's cause for legal action. If no damage can be demonstrated, don't bother.

I had kinda hoped that my personal campaign for "secure by default", where router manufacturers would make at least a half hearted effort to ship their products with security fully enabled, would have an effect on either market advertising guidelines or litigatory precedent. It didn't happen for the simple reason that damage could not easily be demonstrated. There just aren't enough malicious hackers out there to demonstrate real damage caused by inadvertently operating in an insecure manner. I had some hope that endangerment might be a suitable issue, but that didn't work either. So, the manufacturers continue to claim that their products are "secure" but ship them with no security enabled. Lovely.

As for the audio amplifier example, have patience. Once wireless becomes a real commodity, the hype will follow shortly. We're close with the Buffalo "MIMO Performance" baloney. Same with ridiculously high speed claims that don't specify test conditions. The 3X range and 2X speed attempts to simplify product comparisons are also close. It will only be a matter of time before we have "virtual hyper compressed cached peak accelerated and turbo-charged" un-reproducible data rate tests along with the "Mega Thruster IsoTropical" antenna. With this magic converter, you can turn your house wiring into a giant Wi-Fi antenna. I can't wait.

John Navas hath wroth:

Now, why can't I take pictures that look that good?

The Hitachi antenna seems to be made of polysulfone PCB which is way better for low loss than the typical G10/FR4. I couldn't tell if it's silver plated but I wouldn't be suprised. Nicely done coax transition and elevated mounting.

Like I proclaimed, that can't work. You've got the antenna sandwitched between to large pieces of sheet metal. It's also directly on top of the RAM memory, which is good for at least a little RFI/EMI noise.

I'll admit I'm stumped and somewhat amazed. According to conventional wisdom, it shouldn't work. However, nicely done. RF really is magic.

On Tue, 08 May 2007 09:43:11 -0700, Jeff Liebermann wrote in :

What camera do you have? What lighting? The best way to ensure good macro shots is a good macro lens (and matching body) and good lighting (e.g., ring light).

This was taken with my Panasonic DMC-FZ5.

I'm not sure, but I think the bottom is actually high-grade plastic, not metal. And there are significant openings to the back and sides.

On Tue, 08 May 2007 02:48:53 GMT, John Navas wrote in :

Permanent image repository:

Oh yes. Chip specifications are usually specified as "Maximum", "Nominal", and "typical". All of these are ways to hide the fact that such specifications vary substantially with manufacturing processes, materials, tolerances, measurement method, test equipment calibration, packaging, circuit board losses, and position of the moon.

How mortal do you want to be? It's easy enough to setup a bench test for sensitivity. You can use a wireless access point as a signal generator. Spend some time with a good spectrum analyzer or RF wattmeter measuring the power output on channels 1, 6, and 11. Also, spend some time with a precision step attenuator getting it calibrated at 2.4GHz.

Now comes the hard part. Shield the transmitter with whatever is available to prevent it from radiating around the attenuator. That's a potentially complex mechanical project. Everything will need to be shielded and filtered including power leads. Something like this:

Now, do it again for the unit under test. Instead of leakage, the problem here is picking up garbage from other WLAN's and 2.4GHz sources. So, you need *TWO* shielded boxes. I can say from experience that unless you build or buy the shielded boxes, you're wasteing your time.

Put everything on the bench start sending packets. Your reference level is a 10% PER (packet error rate) for 802.11g and 8% PER for

802.11b. If you have a computah, you can use netstat -e or something similar to measure errors. When your PER is 10%, read the attenuation off the attenuator and calculate your sensitivity.

Having worked in the advertising industry, I can instantly recognize a victim of "sloganism". That's where the marketing catch phrases don't quite mean as much as you think they do. For example "more powerful and more sensitive". More powerful and more sensitive THAN WHAT? Without a reference for comparison, such slogans are worthless. More powerful than a spark gap transmitter? Sure. Similarly, 3x faster makes no sense for the same reason, but with the added muddle of not bothering to specify the conditions where it's 3x faster. Is it 3x faster in the presence of interference? 3x faster at what range?

As for "easy to use", that's usually a catch phrase for "crippled".

They all have some basis or foundation. Under some conditions, some product really is 3x faster than another product. Whether you actually will ever see those conditions is rather improbable.

Also, don't forget the speed claims. Failure to meet published range and speed specifications (or outrightly misleading consumers) is legally actionable. Here's Belkin at its best:

Yep. RF is magic.

.....sigh. I guess I'll go with antedotal evidence...

seaweedsteve hath wroth:

It's not that horrible. I've done the shielded box thing using cardboard and aluminium foil. It wasn't great, but it worked well enough to make measurements. The attenuator was a piece of junk that wasn't really calibrated at 2.4GHz. I didn't need DC to X-band so I just calibrated to 0.1dB at just 2.4GHz.

Doing it the right way:

Real test equipment: