Chinese Toy Recall:
Possible Lessons for the
Spectrum Community

The papers are now full of stories about the recall of more than 10 million toys from China that are dangerous due to lead paint and other issues.Most articles blame the Federal Government for insufficient vigilance on the safety of imports. In the wireless technology area there are also many foreign imports. These are required by law and FCC regulation to meet certain electromagnetic compatibility regulations in order to prevent interference to other spectrum users.

But like Lord Nelson at the Battle of Copenhagen who put his telescope to his blind eye and failed to see the signal ordering him to disengage, for more than a decade FCC has been using its blind eyes to watch over EMC compliance of radio products subject to its regulations. There has been criticism that the Consumer Product Safety Commission (CPSC) has only 15 people watching imports of products subject to its jurisdiction. I doubt if FCC has one full time equivalent (FTE). Let me emphasize that this is not the fault of the current commissioners and current top managers, but rather a situation that developed over more than a decade.

The spectrum products involved here include unlicensed Part 15 transmitters, but also include licensed transmitters and incidental emitters like personal computers and microwave ovens and radio receivers (yes, receivers have been known to cause interference in some cases)*. All get little or no market surveillance from an FCC whose limited enforcement efforts are almost exclusively complaint driven. (Generally only complaints from prominent lawyers and trade associations get any attention.)

What is the bottom line problem here? What if 10,000-100,000 or more transmitters/emitters are sold in the US with too strong out-of-band emissions in a public safety band or a valuable commercial band like a cellular band or even a broadcast band? Once they are out in the marketplace they are a lot harder to get back than lead-painted toys, especially if they perform their intended function well.

What little enforcement of equipment authorization rules done at FCC today is almost exclusively in reaction to complaints from well heeled competitors with prominent attorneys. (I recall several years ago when a former FCC chairman in private practice came in to meet with the head of OET with a cardboard box containing an AC-powered battery charger that was underselling his client. An amazing number of people in OET and EB in the next few weeks were obsessed with battery chargers.) There is virtually no market surveillance or spot checking of products or the test data that allegedly shows the products are compliant. A look at recent enforcement actions shows very few in the equipment authorization area. By the time a competitor notices a product is hurting them competitively there may be 10,000 or 100,000 units in use capable of causing interference. In general, they will stay in use until they fail of old age.

FCC's lab used to test all transmitter models before they could be sold in the US. This cause major backups in the summer when new Christmas products were being readied for shipment. Approval moved to a system where equipment was tested by private labs and then results approved by FCC and later Telecommunications Certification Bodies (TCBs) that could do all the approval themselves. At each step the Commission was promised that the resources saved in not reviewing every application would be diverted to sampling and market surveillance. But under the constant budget pressure of the last decade and the basic lack of any constituency for equipment marketing enforcement, there is little if any market surveillance and sampling. I recall in 2001 as part of the UWB policy deliberations I asked the Lab for recent test reports of personal computer emission checks and was shocked to hear that no one had ever requested such test reports from manufacturers and no one was sure how to write a letter requesting such.

The realistic sampling of equipment in stores by FCC is minimal due to budget limitations. FCC has the right to demand equipment from manufacturers/importers at no cost for testing under Section 2.945 . However, this is generally ineffective due to the "lab queen" issue. (This does not refer to FCC employees who are crossdressers, rather to the practice of sending to FCC units of equipment that are sure to pass the test.) Lab queens are produced by either modifying units to meet regulations, e.g. reducing their power, or sampling a large number of highly variable units and selecting only the units that pass.

I recall a case about 10 years ago in which there was a complaint against a 2nd tier Japanese manufacturer who was accused by a competitor of selling over powered cordless phones. The samples supplied by the competitor and samples bought by FCC were overpower, but the one sent in by the manufacturer met the regulation. Why? Opening all the units showed that the "lab queens" differed from the other units in 3 different resistors that controlled output power. Not only the values were different, but the cutting of the component leads and the solder holding them in were visibly different from all the other components. The 3 resistors had been changed after production, probably at corporate headquarters before sending them to FCC. The manufacturer was made to change new units and let off with a minor fine.

The status quo will only change if their is a constituency for enforcement. the manufacturer trade associations such as Consumer Electronics Association, Telecommunications Industry Association, and SDR Forum all seem to believe that all their members, and indeed all manufacturers, are nice people who would never make harmful products so there is no need to have more effective enforcement. Wake up! There are sleazy Far East manufacturers who make and market 10W cordless phones that work in the aviation bands and drive the FAA and aviation industry crazy. There is no limit to how depraved some marginal producers can be to maximize profit. The policy challenge is to balance costs and risks and I think the current balance is all wrong and just waiting for an import disaster like in the toy industry recently.

What should be done?

1. The manufacturer trade associations should reconsider their opposition to effective enforcement and view enforcement as in their best interest, particularly because it can limit opportunities for low cost non compliant competitors and level the playing field. Use trade associations like NAB and CTIA should also advocate equipment enforcement. This FCC will not be more aggressive about enforcement without industry support. forget what you learned in high school civics.

2. FCC should modify 2.945 to decrease the likelihood of lab queens. This could be done by requiring a notarized statement by a person subject to US jurisdiction that the units submitted were taken out of normal inventory without a priori testing and without modification subject to penalties of perjury along with a clear FCC intention of pressing DOJ for criminal prosecution of any violators.

3. Increase FCC budget for random store purchases of commercial equipment for testing.

4. Amending Part 2, Subpart J to permit the FCC to request from any grantee, e.g. company approved to manufacture, market, or import equipment, a free "coupon" permitting an FCC agent to redeem the coupon for a unit of equipment on sale at a retail store with the grantee compensating the store for reasonable costs. This would allow ready access to real samples of equipment. tested equipment would be returned to manufacturers. If there is reasonable doubt about whether FCC has authority to do this under present legislation, it should request such authority at its next oversight hearing.

Also amend Part 2, Subpart J to permit the FCC agents to select from normal warehouse inventory units to be sampled. This has been done at times in the past when inventory can be identified - usually only for the case of large importers and manufacturers. But there are lingering doubts about whether it is legal and a non cooperative manufacturer/importer might stonewall FCC legally.

5. Routinely call in for examination test data for a few per cent of all the models of electronics sold in the US for at least a spot check. The fact that this data is rarely requested may well be encouraging people to not pay for the required EMC tests. (The industry knows what is being requested and what isn't it is no secret in industry circles how little FCC enforcement is going on.) In the previously mentioned request for PC test data in 2001, it took one major manufacturer over a month to send in the requested data - I suspect he had never measured it for that model.

A little bit of enforcement goes a long way deferring antisocial behavior by manufacturers and importers. I hope the spectrum community can work together to urge FCC to step up its enforcement in this area so we never have Pandora's box opened.

Broadcast interests may see this as an argument against the Docket 04-186 white space proposals, but it really is an independent issue and needs attention whether or not the 04-186 proposals are ever adopted.

---------------

* For over a decade the FCC turned a blind eye to repeated complaints from both VSAT interests and the Canadian Government that "fuzzbuster"-like police radar detectors, illegal in many states, had local oscillator emissions that interfered with VSATs . They were at the time not subject to regulation because 15.111 did not apply to receivers above 960 MHz. They are now subject to emission limitations , per 15.109(h), although other consumer grade receivers above 960 MHz, e.g. PCS, AWS and DBS receivers, are subject to no emission limits and testing at all! Presumably we have to wait for a tsunami of interference complaints from>960 MHz consumer receivers before FCC considers any emission regulations for them. (>960 MHz receivers were probably exempted from 15.111 decades ago because all equipment at the time at such frequencies was limited production commercial equipment used only in commercial locations. The "fuzzbuster" problem shows how wrong that assumption was. Now consumer electronics reaches as high as 77 GHz, it's just that the FCC Rules live in a time warp until the powerful interests complain.)

Washington Post Supports White Spaces

For those "outside the Beltway", here is an editorial from today's Post

The White Open Spaces:
The FCC should allow unlicensed use of unused TV band spectrum,
when and if the technology is ready.

Thursday, August 16, 2007; A14

COVETED BITS of the radio spectrum called "white spaces" -- unused areas of spectrum wedged between licensed TV channels -- may soon be freed up by the Federal Communications Commission. Right now no broadband devices are allowed to use these parts of the spectrum, but the FCC is considering whether to let companies sell FCC-certified wireless devices that would be used without an exclusive broadcast license in these slivers of bandwidth. Such white-space devices (WSDs) would be low-power and so would emit signals over very small geographic areas.

White space within the TV band is unlicensed, like WiFi, but is physically better suited than WiFi for broadband transmission. Given the innovation that WiFi access has spurred, as well as the potential for broader coverage both in rural areas and in urban community wireless networks (such as the free WiFi network in Dupont Circle), the FCC has already decided to allow WSDs that are fixed in one location starting after TV's digital transition in 2009. The more controversial issue the commission is considering is whether to also allow portable WSDs, which could be used in products such as laptops or personal digital assistants. Portable WSDs are more difficult to design because they'd need to instantaneously identify which channels are being used in different regions.

To test the feasibility of such devices, last year the FCC started soliciting designs for devices that can identify unoccupied channels and then transmit wireless signals that don't interfere with licensed broadcasts. Two prototypes, submitted by Microsoft and Philips, recently failed to meet the proposed sensing and non-interference requirements, the FCC says. Microsoft is disputing the test results for its prototype.

Broadcast companies (including those owned by The Washington Post Co.), afraid of potential interference from WSDs, are lobbying against use of white space. Groups such as the Association for Maximum Service Television have argued that the FCC should never allow portable WSDs because the failure of these recent prototypes proves no portable WSD technology could work without destroying TV as we know it.

Certainly the FCC shouldn't approve WSDs that will obliterate TV. But just because these prototypes fell short doesn't mean the technology can never work. The limited success of these devices and another designed at the University of Kansas certainly gives hope that someday a non-interfering product could exist. After all, low-power wireless microphone operators often already use white spaces for similar short-distance broadcasts without a license -- although they're supposed to get licenses -- and they coexist peacefully with TV stations. (Wireless microphone operators also oppose sharing white spaces with unlicensed Internet service providers.)

Given the good that could come out of using this unoccupied bandwidth, the FCC should continue to encourage WSD research and development.

Why the

Secrecy

About

SES Bonuses

at FCC?

Although you would never know it from reading the FCC website and FCC public announcements, the 40 or so FCC career employees* who are members of the Senior Executive Service (SES) are eligible for annual bonuses based on their performance under the provisions of 5 USC 5384.

In the 1980s when Mark Fowler was FCC Chairman, these were awarded at a major public ceremony, sometimes complete with a military honor guard. Indeed, I recall that Bob Foosaner, Chief of the Private Radio Bureau (WTB's predecessor) used to brag publicly that he got a big bonus every year! More recently, Broadcasting & Cable magazine somehow got tipped off when the bonuses were given and pressured FCC for a list that it then printed annually. But for several years even this source of information has dried up.

The Office of Personnel Management (OPM) publishes an annual report about the SES bonuses throughout the government, but FCC is not large enough to get itemized treatment.

SES bonuses come from the Civil Service Reform Act of 1978 that, among other things, created the SES. Also in the same legislation was the present 5 USC 4507 that created special awards for the very top of the SES: Presidential Rank Awards:

Distinguished rank recipients receive a lump-sum payment of 35 percent of their basic pay; Meritorious rank recipients receive 20 percent of basic pay. All recipients receive a framed certificate signed by the President.

These are usually awarded at a White House ceremony by the President. Thus the granting of such awards to FCC can not be slid under the carpet. Oddly, in the almost 30 years since such awards have been given, only 2 such awards have been given to FCC employees: James McKinney, former Chief of the Broadcast Bureau, and Jerry Vaughan, former Deputy Chief, WTB and a key figure in organizing the first spectrum auctions. By contrast, OPM records show that last year the smaller National Transportation Safety Board gave out one distinguished and one meritorious award, the National Science Foundation**gave out 2 distinguished awards and 3 meritorious awards, and the Nuclear Regulatory Commission gave out 2 distinguished awards and 8 meritorious awards.

So why does FCC have such an odd track record in recognizing the accomplishments of its senior managers? 8th floor readers - I hope you ask this question.

------
* SES employees come in two flavors: career and non-career. Career employees need not be lifetime civil servants before getting the job, but must be vetted by both the agency and OPM for qualifying professional and management experience. Non-career appointees are basically political appointees. Only career SESers are eligible for bonuses, but FCC in the past decade or so has appointed some people as career SES without any obvious management experience. Apparently the OPM review has become rather loose despite the statutory requirement.

** NSF is smaller than FCC, but due to the nature of its work has more SESers. NRC is somewhat larger than FCC and has many more SESers.

Applying Market Principles to Spectrum Management

[Reprinted with permission from Policy Tracker, a London-based European spectrum policy newsletter. Hence, the spelling used below.] A market-based approach is the way of the future according to a new book by three of the architects of the UK approach to liberalisation. Here Martin Cave, Chris Doyle, William Webb argue that obstacles like interference can be overcome:

The key purpose of spectrum management is to maximise the value that society gains from the radio spectrum. This can be achieved in principle by allowing as many of the highest value users as possible to access the spectrum while ensuring that the interference between different users remains manageable. However, achieving this in practice is difficult. Identifying the highest value users in advance is prone to error and the balance between allowing additional users while ensuring that interference remains within appropriate bounds is highly complex and changes over time. Most spectrum is currently managed under a “command and control” approach whereby the regulator decides on most, or all, aspects of the use of the spectrum, including the application and technologies allowed and the organisation of bands of spectrum. Such an approach is unlikely to achieve the overall objective set out above because the regulator is not well placed to understand which use of the spectrum will lead to the highest value nor which point of balance between interference and additional usage is economically optimal. Further, a command and control approach is becoming more difficult to manage as an ever expanding range of applications appears. An alternative is for regulators to make use of economic management methods to achieve their duties. This allows the market to make decisions as to optimal usage and interference rather than the regulator. The market often has more information and can make better decisions, and if established appropriately can rapidly modify its decisions in the light of new information. Our new book sets out in detail the way that economic management of radio spectrum can be implemented; the remainder of this paper provides a summary of its findings. Technology The advance of technology is having some impact on spectrum management. Multi-modal radios are gradually reducing the advantages of international harmonisation, making it easier for regulators to allow the use of market forces. Technologies that provide “overlays” (cognitive radio) may be best enabled simply by providing spectrum owners with sub-leasing capabilities while “underlays” (UWB) can be accommodated as an increased noise floor for existing owners. The key tools that a regulator needs to deploy in order to allow market forces to manage spectrum are auctions, trading and property rights. In addition, some powers to address anti-competitive behaviour may also be required. Spectrum auctions have been used extensively around the world for assigning many thousands of licenses covering many different uses and types of user. The design of auctions has evolved to accommodate a wide range of situations and new concepts continue to be developed. Trading Trading is likely to work well where no single player has market power, property rights are well defined, full information about previous trades and the current trade is available and there are no unforeseen externalities. Possible issues with trading include a lack of harmonisation and windfall profits, but we do not believe these are serious. Where the costs of regulation would outweigh the benefits, there may be more effective ways of managing the radio spectrum. For example, for some uses and frequency ranges it might be more effective to make radio spectrum a public good (i.e. licensed exempt spectrum). Property rights A set of property rights is critical to facilitate trading and enable change of use. The key technical components of property rights comprise limits on in-band, out-of-band and geographical emissions. From these limits, neighbours will be able to calculate their likely interference levels. These rights and obligations will need to be carefully defined since the interference experienced by one user depends on the rights granted to other users. Within these rights easements should not usually be allowed, but rights should be allocated in ways which take account of the economic value, and interference potential, of new technologies such as UWB. Licences should be perpetual but spectrum licensees should not pay any charge which discourages efficient trading. Competition issues Concentration of spectrum holdings in the hands of one or a small numbers of operators can provide a means to monopolise service markets. The extent of the risk depends on policies adopted at the time of a spectrum award and the rules governing spectrum holdings thereafter. There are grounds for designing award rules in ways which will prevent any operator from immediately gaining market power in a services market. Also, intervention by the spectrum regulator can reduce the risk of market failure associated with abuse of power in spectrum markets. For these reasons, concerns about competition are not a valid basis for rejecting spectrum markets. Sufficient instruments exist to combat monopolisation of spectrum in a market context, and the alternative command-and-control methods can have equally deleterious effects on end users. In addition to these basic tools, there are other areas of spectrum that might require a different management approach. These include spectrum commons and spectrum owned by public or Governmental bodies. Also, a somewhat different approach should be adopted to the management of spectrum in developing countries. Commons Spectrum should be unlicensed where there is little probability of congestion. Despite arguments about the ability of “spectrum commons” to alleviate congestion, congestion across key parts of the spectrum is likely for the foreseeable future. Congestion is unlikely where short range communications are used and can be made less likely by regulatory insistence on, e.g. politeness protocols. Hence, there should be a mix of licensed and unlicensed spectrum with the unlicensed approach restricted to bands and applications where congestion is unlikely. Market mechanisms cannot be used in determining how much unlicensed spectrum would be needed so instead the regulator needs to make a judgement. This could be based on a top-down approach which estimated the total amount of unlicensed spectrum that might be needed. Alternatively, a bottom-up approach could be adopted which considered spectrum on a band-by-band basis, using a structured process to determine whether licensed or unlicensed use would be most appropriate, generally based on the approach predicted to lead to the greatest economic value. Public spectrum There is no reason why public sector spectrum should be treated as ‘exceptional’ and be immune from application of the market processes. In practice, however, reforms in this area may have to take a more gradual course, relying on intermediate stages such as valuation and administrative prices. There remains the pervasive problem, under almost any spectrum management regime, of generating incentives towards efficient use of spectrum by the public sector. In practice, there may be a continuing need for detailed scrutiny of public sector holding and the application of pressure to hand back unneeded spectrum. The goals of spectrum management are broadly the same in developing as in developed countries, and so are the available instruments. However, their governments and regulators may face less congestion and have an incentive to speed up the assignment process in order to buy services for their citizens. This may imply either use of fast-track administrative methods or the development of a simple to administer market process. Stimulating the market Finally, there are some tools that might assist the development of the market, or help with transitional issues. These include band managers and spectrum pricing. There are many potential types of commercial band manager. Under an environment where spectrum is tradable, the key to successful operation will be the ability to make use of the spectrum more efficiently than the regulator. However, this is difficult to do, making the economics of band management problematic. If regulators wish to encourage band managers they may need to look into specific policy measures to do so. A spectrum management agency can use prices to achieve efficiency, which will generally lead to superior outcomes compared to pricing on the basis of cost recovery. Charging annual fees can encourage current and prospective holders to make the right decisions to ensure efficient use of the spectrum. Many holders of spectrum are not in a position to make rapid changes to their use of spectrum in response to the application of pricing, but in practically every case the holders of spectrum have opportunities to change their use of spectrum in the longer term. In our book we discuss each of these areas in substantially more detail, building a strong and complete case for modern spectrum management being based upon economic and market principles and showing how these can be deployed in practice. We expect to see this approach increasingly deployed by regulators around the world over the coming years. • “Essentials of Modern Spectrum Management” by Martin Cave, Chris Doyle and William Webb is published by Cambridge University Press ISBN-13: 9780521876698 About the authors: Professor Martin Cave was the author of the UK treasury report which paved the way for liberalisation in the UK, Professor William Webb is Ofcom’s Head of Research and Development and Dr Chris Doyle of the University of Warwick is the author of many papers on spectrum economics.

700 MHz Public Safety/Private Partnership - FCC Votes to Implement Marcus "Interruptible Spectrum" Proposal


Yesterday's Commission meeting got lots of press coverage in the national media. Almost all of it focusing on Google's request open access to part of the 700 MHz band. This post deals with another issue that got less press coverage.

Here are some key quotes from the FCC Press Release:

Today's Order establishes a framework for a 700 MHz Public Safety/Private Partnership between the licensee for one of the commercial spectrum blocks and the licensee for the public safety broadband spectrum. As part of the Partnership, the commercial licensee will build out a nationwide, interoperable broadband network for the use of public safety. This network will facilitate effective communications among first responders not just in emergencies, but as part of cooperative communications plans that will enable first responders from different disciplines, such as police and fire departments, and jurisdictions to work together in emergency preparedness and response. Under the Partnership, the Public Safety Broadband Licensee will have priority access to the commercial spectrum in times of emergency, and the commercial licensee will have preemptible, secondary access to the public safety broadband spectrum. Many national and local public safety organizations have expressed support for a public safety/private partnership approach. Providing for shared infrastructure will help achieve significant cost efficiencies while maximizing public safety's access to interoperable broadband spectrum.

Public Safety/Private Partnership

• The Upper D Block commercial licensee and the Public Safety Broadband Licensee will form a Public Safety/Private Partnership to develop a shared, nationwide interoperable network for both commercial and public safety users.

The terms of the Partnership will be governed both by FCC rules and by the details of the Network Sharing Agreement (NSA) to be negotiated by the Upper D Block commercial licensee and the Public Safety Broadband Licensee. The NSA is subject to FCC approval, and must contain certain provisions such as service fees and a detailed build-out schedule for the network.

Chairman Martin issued the following on this issue,

While I also would have supported a network exclusively for the use of public
safety, the simple reality is that there currently is no way to fund such an enterprise. The use of a public safety-private partnership, however, creates an opportunity to provide state-of-the-art technologies to our Nation’s first responders in a timely and affordable manner. Many national and local public safety organizations have expressed support for a public-private partnership approach as their last, best chance to make this network a reality. We cannot afford to let the opportunity that the 700 MHz band offers for public safety pass us by.

The adoption of a National Public Safety Broadband Licensee to be a part of this partnership is also the best way to establish a truly interoperable network. The local licensing regime that has been used to date has resulted in a patchwork of networks that do not talk to each other. We cannot keep licensing public safety spectrum in the same manner as before and expect a different result. A National Public Safety Broadband Licensee will facilitate a unified national approach to the use of this spectrum, finally enabling all public safety users to talk to each other during a crisis. I therefore wholly support the public safety-private partnership adopted in today’s order.

The diagram at the top of this post comes from a 2003 internal FCC paper I wrote as part of the Spectrum Policy Task Force (SPTF) deliberations. (I have now posted this paper on my website.) The diagram below comes from the FCC band plan that has been released. There are certain similarities and it is pleasing to see that the concept has been adopted.




In view of the implicit and unannounced abolition of the SPTF, the Commission probably doesn't want to be reminded that this was actually SPTF Recommendation #27:

27. Address additional public safety needs through alternative “safety valve” mechanisms
to make spectrum is available to public safety in emergency situations when more
capacity is needed.

a. Because some public safety spectrum use is characterized by intermittent
“spikes,” public safety users should have flexibility to lease spectrum capacity
that is available during lower-use periods to commercial users with a “take-back”
mechanism when public safety use increases.

b. For major regional or national emergencies, additional public safety spectrum
needs should be addressed through enhanced easement rights to non-public safety
spectrum.

The SPTF paper referenced above dealt with both non-Federal Government public safety spectrum, e.g. state and local police and fire users, and Federal Government spectrum regulated by NTIA pursuant to 47 USC 305 that has the same high peak, low average utilization as state and local public safety. NTIA staff has a "temper tantrum" on this issue and threatened to have the head of NTIA complain to the FCC Chairman if the SPTF report even mentioned the possibility that NTIA-regulated spectrum might be shared in such a mode with civil users. Hopefully, now that FCC has taken this bold action, NTIA will at least tolerate public discussion of the topic with respect to the spectrum they control. Chairman Martin's above insightful comments really apply to all spectrum used for public safety purposes.

My paper for the 2005 IEEE DySPAN conference also explores the same issue and has been publicly available for a while. I would be glad to help parties preparing for the 700 MHz auction understand the options and opportunities that result from the FCC's bold action yesterday.