Hey Jeff,et al, whaddya thunk a this?

Just noticed it requires usb 2.0; I only have 1.0 so this would be better?

uknowwho wrote in news:gpetbo$tj7$ snipped-for-privacy@tioat.net:

Twice as good, thats why it wants 2 instead of 1..... 1 +1 = 2 ! :)

Yo. Dat's me.

Dunno. What problem are ya trying to solve? Whatcha got to work with?

What desktop top dilemma? I'm not going to search through all the past articles in order to find your dilemma. I'm lazy. Make it easy for me and others by posting the article number or supplying a description.

Dunno. RadioLabs has been around for quite a while. I don't think they'll turn you in for buying their products. If it puts out too much power, I might turn you in to the EPA as a gross RF polluter. In general, the FCC runs on complaints. Ordinary intereference complaints don't count. It has to be something with ties to terrorism or national security, or the FCC won't spend the money or allocate the resources. Welcome to the post 911 era.

200mw for 802.11b and 100mw for 802.11g is high power? I don't think so. Most client radios belch about 35mw. That's a 7.6dB boost in transmit power, which is good for about a 2.3 times increase in transmit range. However, you won't get any improvement in receive range as the receiver is the usual (unspecified) chipset. No way are you going to get the claimed 10x the range of a typical notebook computah. The only real benefit over a notebook wireless derrangement is that the USB radio allows you to be creative about the type and location of the antenna, which has a big effect on range.

You might be ok with USB 1.0 as you'll be limited to 11Mbits/sec thruput. If you're trying for the DX (distance) record, you'll be running at the slower wireless speeds anyway, so the faster USB port isn't going to buy you much. I don't know what OS you're running so I can't tell if there will be driver issues.

Also, the $130 price tag seems too expensive when there are similar admittedly lower tx power products available (e.g. Hawking HWUN3) for half the price.

Caveat Emptor.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

So if I know the chipset where do I find performance info on that? Seems I remember there was a gov. link for that but probably can't find it now, for the tests? But look at their advertised receive sensitivity ratings, that sounds good on the surface -91dbM at 11mbs

Have to check their return policy but the specs says usb 2.0 "only". The first unit is 2.0 only the second one (200mw) is 1.0 or 2.0 The first radio is 500 mw output via an amplifier, but I'm more concerned about the receive end of matters. Also any amplifier will also increase the noise level on the receive side right?

Yeah the price is a ripoff, but might be worth it for convenience of usb unit. Have to find out the chipset as I also want to know what software it is compatible with. The OS is not an issue, they state they have drivers for 98SE, but I've heard that story before. Thanks for your usual good answer, this group would be toast without you.

yeah, no guarantee, no buyee

"Peter Pan" wrote in news:i7ednfvC7JJnkSbUnZ2dnUVZ snipped-for-privacy@earthlink.com:

say, did your Mother have any children that lived?

You don't. Most chipset spec sheets require an NDA from the vendor in order to obtain a copy. However, by the time the chips get into the final product, the manufactories tend to just claim their sensitivity is the same a that of the chipset. Never mind all the circuitry and hardware in between the antenna and the chips. In a few cases, the manufacturers actually measured the sensitivity and used real numbers for the data sheets. I've collected a few:

That number was probably borrowed directly from the chipset and not measured.

Yep. If the noise figure of the amp is better than the noise figure of the chipset, and there's at least 6-10dB of gain, then you have a net improvement. Anything less isn't going to help very much. If you have a long piece of coax between the antenna and the receiver, then an RX amplifier at the antenna will take the coax losses out of the equations. Otherwise, it takes a really good low noise figure amp to do anything useful.

Also, remember that the amplifier amplifies both the signal and the noise equally. The general idea is to improve the final SNR (signal to noise ratio). If you amplify both the noise and the signal by exactly the same amount, there's no benefit. Worse, the amplifier contributes a small amount of noise, which is also amplified, potentially resulting in a worse SNR. Unless you're trying to compensate for a lossy coax cable run, I don't recommend adding an RX amp.

Interesting that the 624M (MIMO) is said to have a 3dB improvement at every point over the basic 624.

Incidentally, the same data in spreadsheet form:

+/- 2dB or worse was approximately my measurement error. Even the same unit measured a few months apart will result in such variations (due to digital noise pickup across the board). It may be for real. It appears that D-Link is one of the few manufacturers that actually has measured the receiver sensitivity using a BER/PER reference. However, even those numbers look cooked. When I had my pile of borrowed test equipment, I was never able to reproduce the idealized curves shown. My data points were all over the place. It was not unusual to have a few of the higher speeds show miserable sensitivity. Maybe if I had some better test equipment:

Since MIMO has to be done in a test box, using simulated independent spatial diversity paths, there will be substantial variations depending on assumptions and methodology. My guess(tm) is the DI-624M rx was probably tested without MIMO. It's also possible that the numbers were generated in a simulator. See:

Replace the range axis on the graphs with sensitivity or path loss and you have something suitable.

A few points:

1. The numbers showing sensitivities that are equal to the chipset values are probably contrived.
2. Every lab and manufactory has a slightly different way of measuring rx sensitivity. To do a proper test, a PER/BER generator and analyzer is required, which takes hours. Most labs just look for the error threshold and use that number. Probably good enough but
3. Everyone lies, but that's ok because nobody really understands what the rx sensitivity numbers really mean.

How about this piece of marketing hype?

The article explains a few things but doesn't mention anything on how their -100dBm rx sensitivity was actually measured. I don't think so.