Microsoft Questions FCC's 'White Spaces' Decision

narramissic writes "Late last month a wireless prototype submitted by Microsoft and other members of the White Spaces Coalition was rejected by the FCC because it interfered with cable channels. Microsoft, though, claims that the device was malfunctioning when the FCC tested it. From the article: 'In a letter to the FCC Monday, Microsoft said the scanner in one of two prototypes was damaged and "operated at a severely degraded level. The damaged scanner accounted for the entire discrepancy between the Microsoft and the FCC bench test data," said Ed Thomas, a consultant for the White Spaces Coalition and a former chief of the FCC's Office of Engineering and Technology.'"

Why is Microsoft taking on this role?

[Sergeant Pepper]

Unless I've been understanding this wrong, this thing wasn't just a Microsoft prototype. It was submitted by several companies, so why is Microsoft the only one who is questioning it? Are the others backing Microsoft in their complains? Do the others not care enough? Or is there something more nefarious going on - do the others think that the FCC's claims are true?

We Need Wireless Broadband

[grumpygrodyguy]

Well...I'm rooting for MS on this one. Some 3rd world countries have better wireless broadband access than we do.

The telecom and cable monopolies are holding the FCC in their pockets and stifling innovation.

Responsibility

[kaleco]

I would imagine it's the applicant's responsibility to supply a functioning prototype. Otherwise it's like retroactively claiming you were feeling unwell when you sat your finals and didn't get the grade you were hoping for.

FUD -- Microsoft needs to prove it works

[dwarmstr]

The whole point of FCC testing is to confirm the device works to specifications and doesn't violate FCC rules regarding emissions. It failed, and Microsoft needs to submit their design again. To imply the FCC was somehow faulty as is suggested by the "White Spaces" industry wag man (who also is one of those in-and-out regulatory-to-industry guys) is classic FUD. Fix your prototype, MS, and the FCC will certify it.

Re:So this wasn't

[bleh-of-the-huns]

While I agree with you, the fact that a malfunctioning unit could cause a major disruption in service to those around you is still a significant problem. The odds of every device shipping (however many thousands or millions) without a flaw is pretty slim. That being said, more work will need to be done to prevent such an outage should there be a flaw or a malfunction before it passes the FCC.

For the record, I am no fan of either MS or the FCC, but in this case, I would probably side with the FCC.

Forgive my apparent ignorance but...

[Monoliath]

...isn't this exactly what the FCC wants to avoid happening? Failing devices mucking up other channels?

So what's the point of Microsoft saying 'oh, it was screwing up when you were testing it'...

If it mucks up other channels while it is malfunctioning it's not going to be commissioned...that's the whole point of testing it...isn't it?

If it doesn't mess up other channels while it's working fine, then fine...but the whole idea that when it malfunctions it interferes with other transmissions...is the perfect reason not to pass this thing in my mind.

Re:Soemthing smells fishy

[ShieldW0lf]

If a malfunction results in a failure of the local TV signal rather than resulting in a failure of the device, the FCCs decision is the right one.

Devices are expected to fail. Given a long enough timeframe, ALL of them fail.

MSFT needs to understand "fail safe"

[SmoothTom]

With any product that can disrupt other services - in this instance, taking out your neighbor's TV reception or data link if the "scanner" doesn't detect the "channel" is already in use - the product needs to be designed to "fail safe."

In other words, the device should self-test critical functions, and if any do not meet requirements, the device needs to indicate the failure AND NOT TRANSMIT.

Basic rational design.

If the "scanner" fails to detect an "in use" channel properly (self test to ensure it does), the transmitter shouldn't just push ahead and transmit, it should alarm and go to standby.

If the device can just go ahead and transmit, as Microsoft's did, the FCC is absolutely right: The device (and possibly service) should not be allowed.

--
Tomas

Re:hey...

[LunaticTippy]

Imagine if government ran the roads the way they run the spectrum....

You should search "us road privatization" -- we are starting to run our roads the way they run the spectrum.

Right idea but your proposed test has a bug

[Ungrounded Lightning]

In other words, the device should self-test critical functions, and if any do not meet requirements, the device needs to indicate the failure AND NOT TRANSMIT.

Dead on. But:

If the "scanner" fails to detect an "in use" channel properly (self test to ensure it does), the transmitter shouldn't just push ahead and transmit, it should alarm and go to standby.

Which breaks if you bring it up in an environment that doesn't have any "in use" channels to detect. Like in a remote environment (such as my place in a lightly-settled section of Nevada desert) which has zero detectable TV signals and virtually no daytime broadcast radio - exactly the sort of place you'd want to "wire for broadband" with wireless.

IMHO the right algorithm is not an up-front self-test, but a CYA check during turn-up:
  - Check for in-use channel. If not found:
  - Momentarily make a VERY SMALL amount of signal of your own and see if you detect that, to check the detector. If you do:
  - THEN turn on normal transmitter power.

Re:Another public relations misfire

[Control Group]

Right, because of course the other companies involved simply decided that the FCC is all-wise, and have no interest in the decision being appealed. It's utterly beyond the realm of possibility that they collectively decided to use the member with the best PR machine to protest the decision. This is obviously Microsoft branching out on its own hook against the wishes of the other members of the coalition in an evil plot to TAKE OVER THE WORLD, MUAHAHAHA!!

Compliance test is not multiple-choice

[tcgroat]

Yes, that's right. If one sample meets the limit and another does not, how would the test tech know which one was truly defective? In the case of interference testing, a broken product often seems better than the working one (it's likely to have fewer or weaker signals making noise, hence lower interference measurements). The unit was tested as received, and that sample's performance determines "pass" or "fail".

The nature of certification tests is that the test sample represents all products shipped. It's the manufacturers' responsibility to deliver working test samples. If the manufacturer didn't have anybody present to demonstrate the failing unit really was broken, the test technician not only was justified in giving a "test failure" verdict, but as an independent evaluator also was obligated to do so.