Forum

In no way, shape or form is this type of beast:

http://cp.literature.agilent.com/litweb/pdf/5968-8156E.pdf

the equivalent of this:

http://www.oscium.com/products/wipry-spectrum-spectrum-analyzer

Again, we cannot simply say "one is better than the other". We have to look at what we are trying to accomplish ( noise spectral density, carrier plus noise spectral density/noise spectral density, phase noise measurements, intermod determination etc ). Each has it?s niche, but one is definitely not the equivalent of another.

Dave

Oscium

Going back to the previous question, are you guys saying that the unit will only accept levels above -40 dBm. Please check if that is a misprint on your webpage, or if your technical people say that is correct. Based on the answer to that , we can then move on to some other questions

Dave

Hi, dave.

While you are at it, you might want to ask them what on earth this info on their website means:

1. Wireless Internet Problem?
It is now easier than ever to find a clear channel on the 2.4-2.5GHz band. If you're accessing the internet through a busy channel, there can be interference. Internet packets collide and have to be re-routed, slowing down your internet access.

So, that's how WiFi works.....

Hi Enesem

Yes, that's a classic. Some may think I'm being "too hard" on them. That is not the case. Manufacturers have a responsibility, morally and otherwise to provide the most accurate information to their potential customers. I have spent much of my career trying to sort out horrific networks where the poor customer told me ?But the salesman told me that wouldn?t happen/but he said we could do it that way etc etc ). Can you imagine the uproar if a manufacturer called Ciscow systems appeared and said ?With our APs using 802.11n, you will be able to achieve 10 Gbit/s? ? People would instantly jump on that, as the vast majority of people in the Wi-Fi industry ?just know? that there is no way that could happen.

Spectrum analysis is a bit of ?magic and science and an art form?. It takes time and dedication to trully get a ?good feel? for what the specifications mean. It?s my bread and butter. I remember the first time I saw a spectrum analyzer. Hadn?t a clue what it was.

Older engineers not only wouldn?t help me learn about it, they deliberately fed me misinformation ( job security and all that ). I had to dig through manuals, then buy electronics books to understand the specs etc etc. I well remember my first time seeing that unit. It may as well have been from the planet Mars. I hadn?t a clue about it. We are all newbies at some time.

This is true for many people in the Wi-Fi industry, who are just starting. Like myself, when I first started, the specifications were gobbledegook to me. The CWNP program is helping to change that.
Spectrum analysis is difficult enough without ( and I?ll be kind here ) misprints in the documentation.
A newbie or even an experienced Wi-Fi engineer who has not had much exposure to spectrum analysis could be easily misled. ?Hey Bob come look at this display !!.....cool !! Rockin music on the video?Air guitar???.Let?s buy a dozen !!

Metageek has a little freebie called Inssider. I use it often for a quick of things. It?s not meant to be anything fancy. It?s a simple little thingy that gives you a quick snapshot of a few bits and pieces for free. You know exactly what it can do and what it cannot. If it could talk, it might say as Popeye would say ?I am what I am, and that?s what I am?.

The next step is getting a device that you have to pay for. Measurements in Wi-Fi can vary from fairly straightforward to quite complex. For most users, a few simple things like received signal level and noise floor measurements are fine. If you don?t have that, then you have the equivalent of a device that shows pictures ( still extremely useful ).

The website advertises power measurement capabilities at an extra cost ( I?m fine with that ). However, if we can?t get proper responses to questions about the basic calibration indications/specs of the basic unit, how much faith will we have in a power measurement option ?

Oscium, please respond to the question about minium power level. I write on a number of different forums ( microwave, satellite, Wi-Max, cellular etc ) under different names. I haven?t posted anything there yet, as you have been responsive to my concerns, and made some changes. That is a step in the right direction. My main goal is to try and ensure at the end of the day that your website is as accurate as can be. That is not only for the sake of potential customers ( and again, I think the whole I-pad/I-phone thing is a great idea, and you have some nice features in the unit ), but also for your bottom line.

Please respond. This is not a ?trap?. This is simply a way to find out if you guys have misprinted something or whatever. I don?t have one of your units with me, but I would find it very surprising if ? 40 dBm was the minimum level that your device can actually ?see?. You might be short changing yourselves here.

Once we get that out of the way ( after a discussion of why I?m so ?hooked up? about that particular value of -40 dBm), we can move onto the other parts of Ben?s posting.

If you?re prepared to ?work with me?, I?ll do my best to give suggestions that could help everyone out. The ball is in your court.

Again, this is not some sort of a thing to ?catch you out?.

Dave

We appreciate the feedback. It's always welcomed. We've got a few of our guys out of the office and I'd love to get their insight before responding, if that's cool. Should have something solid by the end of next week. In the interim, we've nixed the customer's quote that had the word equivalent in it, as suggested. In truth, we don't know which SA he was using to compare to our product. But regardless, if you were mistaken, someone else could be as well. And since that wasn't our intent, we've nixed it. In addition to dave1234, thanks to enesem and Wlanman for jumping in. You're helping us and for that we are grateful. Please stand by.

That's fine. No problems. The only spectrum analyzers that weigh that much are professional desktop units which have a very large number of features and are very expensive. Depending on the manufacturer, the prices vary from ( very roughly ) $50,000 upwards. The data sheet below is representative of a unit that I have used for many years, along with many others:

http://www.testequipmentconnection.com/tecspecs/HP_8560E.PDF

I can pretty say with certainty that I've worked on spectrum analzyer models from just about all the manufacturers at one time or another. These units are used in labs and for interference analysis etc. In other words they are specialized devices. It would be wrong to compare the two as each is designed for a different market.

I just took a look at the webpage, and have no problem with the customer reviews:

http://www.oscium.com/products/wipry-spectrum-spectrum-analyzer

Those are personal impressions of using the device in the arenas for which it is intended. That's fair enough. We just want to make sure that apples are compared with apples and not oranges.

Just as a recommendation, it might be advisable to re-word the phrase that enesem pointed out. The whole "internet packets" thing is just not good at all.

I would suggest you guys "beef up" some of the descriptions of the good points ( easy to use, short manual, cursors etc ), but get rid of that whole description of internet packets etc.

Now that we've got all that other stuff out of the way, hopefully next week we can get down to the nitty gritty about the various power levels etc.

Dave

Already working on the re-word. It's just not done yet. Looking forward to continuing the convo next week...

Oscium

This will be my last post on this forum, as I?m moving onto other things after receiving a very interesting phone call re one of my posts today, and I need to start writing and doing some research tomorrow for upcoming work.

Many desktop spectrum analyzers are fitted with a calibration output. A calibrated cable is used to connect the output of the calibration port to the input of the analyzer. All cables and connectors being of the appropriate characteristic impedance. In a proper environment, the accuracy of the calibration point of the spectrum analyzer should be traced back to some standard, preferably a national one. In this manner ( along with some other tests, perhaps using an external calibration source ), we can be reasonably sure that the analyzer will measure power correctly. Older units had a resolution bandwidth correction factor, units following on from that, generally used digital filtering. By integrating the ?area under the curve?, accurate measurements could be made. In order to determine whether a particular wireless station ( not necessarily Wi-Fi ) was transmitting at too high a power, many times in the past, I have had to make accurate measurements using a spectrum analyzer.
The two companies mentioned are correct in their measurement values. I know this, because I have used benchtop spectrum analyzers and compared what is being received with those units to the units mentioned. Over the weekend I had a friend in the UK do another set of tests and the values recorded by the benchtop unit were comparable to the Cisco and Airmagnet units.

From Ben?s post:

?If you look at the second picture I posted above, you'll see that the signal from the access point in the office I write at showed a 17.74 dBm signal strength. I compared this with Metageek Chanalyzer (using the WiSpy DBx USB adapter) and I saw a signal strength of around -45 dBm at the same place from the same AP.

When I emailed back and form with my contact at Oscium, his response was essentially everybody's lost but me. His position is that their spectrum analyzer reads the signal strength correctly, and that Metageek (and Cisco, and AirMagnet, and Apple, and Broadcom, and Atheros...) reads the signal strength incorrectly?

How do we know the benchtop analyzers are accurate ? Multiple reasons. One of the main ones is this: satellites have a fixed amount of power. All the major satellite organizations use a network of interlinked benchtop spectrum analyzers to record how much power is being utilized on the satellite. If they were wrong ( and they?ve doing this for over thirty years ), they would literally lose billions in revenue. Satellite bandwidth is expensive. Those analyzers are accurate. NASA uses them for it?s Deep Space Network. When I was working on communications support for NASA for the Shuttle Program, we would have lost communications with the orbiter if the analyzers were not accurate. Lives were literally at stake.

I would suggest that someone has a good look at the whole calibration setup.
From a purely Wi-Fi point of view, having a lower limit of ? 40 dBm would mean that a huge proportion of Wi-Fi signals would not be observable.
What level is a Wi-Fi signal at ? That?s like asking how long is a piece of string. As some point of reference, however, I would suggest looking at mimumum values, as per a client adapter spec sheet, or the minimum values recommended for voice use ( sometimes seen as the lower end of -60 to -70 dBm ), or whatever.
Otherwise you guys will doing yourselves an injustice.

http://www.cisco.com/en/US/prod/collateral/wireless/ps5678/ps6973/ps8382/product_data_sheet0900aecd806b7c5c.pdf

I won?t be continuing this discussion here, but hope someone in your engineering group takes a good look at this.

I will be available on PM for about a week if you wish to get in touch with me. After that, I won?t be checking my messages.

Good luck to you and everybody else on this forum.

Dave

We'll miss you Dave. Your knowledge is unsurpassed. I wish you the best of luck.

But, back to power levels - I routinely test to -84 dBm and better (conducted) to measure Sensitivity of b/g radios. With the proper setup I can measure that low radiated, ie with an antenna, in a medium sized RF enclosure.

I consider -40 to be a really strong signal and if your radio needs that strong of a signal to operate, you're probably only getting about 30 feet of usable range.

My question to the Wipri people is this: Is this the same -40dB reading I would get from a device performing a measurement using the correct IEEE measurement, taking into account rise and fall slopes in the signals' power graph, ignoring the peak in the first moments of overshoot, and then averaging the signal in between the gated measurements? I don't see how it could.

If not, you may be short-changing your selves with your -40 dB claims.

b/g signals are "bursty" signals afterall, and if you were going to compare your readings against those of a WLAN radio manufacturer (using the correct IEEE methodology), you have to use the same technique.

Compliance engineers often forget this, but then their tests are measuring to FCC standards not IEEE, and for a completely different purpose.

This is not to say your device doesn't have a place in the Wi-Fi space, but I think your sensitivity levels needs further explanation. What exactly was your goal when you first decided to produce this device?

Dave - all the best in your newest opportunity. We'll still be providing what was promised by the end of the week.