Both manufacturers and test labs must decide on the proper test procedures for DFS products.
Since July 20, 2006, systems operating under FCC Part 15.407, specifically in the 5250- to 5350-MHz and the 5470- to 5725-MHz bands, are required to support Dynamic Frequency Selection (DFS) before the products can be certified. As part of the approval process, the FCC must carry out a 100-percent pre-audit test procedure on the DFS master devices. Now that the FCC is in the second year of the certification program, the program is no closer to being transferred to the Telecommunication Certified Bodies than it was at the outset. Further, those expecting a quick review time find themselves in a program that can take up to five months or more. There are several reasons for this, and most of the issues are not FCC related, but test lab and manufacturer related. This article will discuss some of the issues that have helped slow the program down.
SHORT HISTORY OF DFS
The requirements for DFS can be found in ITU-R Resolution 229, which addressed allocating the 5150- to 5350-MHz and the 5470- to 5725-MHz bands for mobile systems. To facilitate band sharing, two cognitive radio techniques were adopted. The first is TPC (Transmitter Power Control); the second is Dynamic Frequency Selection.
As defined, the master (control system) must have radar detection capabilities—i.e., DFS. Client cards that support peer-to-peer must also be DFS capable.
Basically, DFS functionality performs a channel availability check before using the channel and then listens in the quiet periods to determine if a radar signal is present. If one is present on a given channel, the master that has DFS capabilities issues a control signal and then moves the clients off the channel that has the radar signal—thus averting any interference with the radar. As part of DFS, the systems also make sure that the channels are loaded evenly.
DFS PRE-GRANT AUDIT TESTING ISSUES
What causes the lengthy turnaround times? Actually, there are a number of reasons for the sluggish pace. First, any audit program is usually slow, and the FCC is constrained by budget and resource limitations and can only audit the products serially. A second reason is that the National Telecommunications and Information Administration (NTIA) insists that the FCC lab perform the complete DFS test, a process that can take several days even if no problems exist.
One of the major reasons for the lengthy audit period is unrelated to any lack of FCC resources or NTIA’s insistence on a 100- percent audit. Some problems are caused by the test labs and/or manufacturers who fail to understand some crucial issues. For example, one issue is that manufacturers do not ship everything needed for the audits or fail to enclose complete instructions. The FCC lab must then spend additional time getting the unit to operate correctly or must wait for additional equipment to complete the setup. Moreover, the FCC will not discover these omissions until the unit is ready for audit. Some of the specific issues include: the lack of setup instructions, no test mode, an insufficient test mode, no test mode instructions, and missing peripheral equipment. Other problems include devices that have been setup only for conducted testing and failure to supply antennas with the device. By far, the failure to indicate a test mode is the most significant problem.
In some cases, the systems sent simply do not work, or have other technical compliance problems that delay the audit test and hold up the next unit in the queue. Sometimes, the manufacturers test software will not turn off the 30-minute channel lock out-time, or it will turn off coming up on a random channel—thus making it impossible to select the same the channel again for testing. (These problems are related to test mode issues.)
The biggest “show-stopper” according to anecdotal information, is the high percentage of failures that occur during the initial DFS pre-grant audits. Certainly, an occasional failure may occur; but when a large percentage of products—perhaps 35 to 40 percent— of the systems has problems, there is a major cause for concern. As discussed in a recent Telecommunications Certified Body training class, such failures include, but are not limited to, the following:
- Radiated Testing vs. Conducted Testing
- Threshold not correct
- Must account for cable losses
- Manufacturers tolerance in antenna gain
- Include minimum gain antenna w/sample requests
- Non-linearity in statistical tests across the radar detection BW
- 40-MHz BW (Channel Bonding) compliance issues – statistical tests differ greatly when compared to 20- MHz BW mode
- Channel Move Time –packet transmission to stop within 200 ms
- 30 Minute Non-occupancy – no emissions permitted
NTIA ISSUES
Any possibility of changing and streamlining the FCC audit program lies, at least in part, with the NTIA since this agency is charged with protecting the government radar signals. To safeguard radar, the NTIS continues its practice of monitoring available DFS devices bought directly off the shelf. The outcome of these NTIA audits could have a big impact— either positive or negative—depending on how the chosen products comply. Also, further FCC testing of some units has resulted in a clarification by the FCC on client card testing requirements.
ADDRESSING THE ISSUES
In regards to the entire telecommunications industry, both the manufacturers and test labs should acquire a better understanding of the issues and should decide on the proper test procedures for these products. For example, the FCC performs radiated—not conducted— testing. If a lab carries out conductive testing and the DFS threshold is set to detect with a certain gain antenna, the tester needs to make sure that the antenna cable loss is taken into account and must verify that the antenna used actually has the correct gain. This verification can be achieved by performing an over-the-air test (radiated) to see if it actually detects at that threshold.
Test labs must be aware of changes to the procedures as well. Specifically, as technologies evolve, so must the FCC test procedures. The FCC and industry have discovered a small problem with the detection of certain radar signals. In the 40-MHz 802.11 channels, problems have arisen at the band edge. Consequently, there is now an FCC guidance note that amends the test procedure so as to verify that the systems do detect the radar signals in this mode.
CONCLUSION
The simple truth is that the testing is complicated, and additional industry training is needed that focuses on the accurate performance of these tests. This issue has been discussed with the FCC. It has been proposed that the FCC should hold a DFS training session for test labs similar to that which was held for labs performing SAR testing. Until such time as the telecom industry can resolve the high failure rate of initial audits, and the NTIA audits demonstrate that the systems can pass the tests, delays in the approval process are inevitable.
END NOTES
1. D. Case, “The FCC Dynamic Frequency Selection Test Procedure,” Interference Technology
Design Guide and Directory, 2006.
2. Costs of the test set up can range up to $100,000 or more.
3. FCC KDB #: 594340, Dated 03/05/07
4. October 2007 TCB workshop slides 8 and 9 (client Beacon Test or slides 10 – 15 (Test Issues).
5. October 2007 KDB Workshop, slide 12.
BIBLIOGRAPHY
47 Code of Federal Regulations Part 15.407 (2005)
ITAC-R US DFS Test procedure Rev 7
ITU-R Resolution 229
ITU-R M-1652
