Design Article
Testing raises concerns over 802.11-based high-speed Bluetooth
Dr. Roberto Aiello and Siddharth Shetty, Staccato Communications
3/18/2008 9:00 AM EDT
This means that now WiMedia UWB and 802.11 are candidate alternate MAC/PHYs (AMPs) for the high-speed Bluetooth release. The principle is to allow the existing Bluetooth technology to be used in consumer devices while achieving faster throughput with the use of a secondary radio. However, many in the industry are concerned about the known interference issues that will result between the 802.11 radio in the Bluetooth device and other IMT-2000 services operating in adjacent frequencies, such as WiMAX, LTE, UMTS and WCDMA. The worry is that if consumers have a poor user experience with the initial implementation of high-speed Bluetooth using the 802.11 AMP, the long-term risks to the success of the technology could outweigh any short-term time-to-market gains. Bluetooth already has a high-profile position in the consumer marketplace (last year the SIG welcomed its 9000th member ), and rushing time-to-market with an interim technology is risky.
Usage models
While WLANs and IMT-2000 do not typically operate simultaneously because they both provide access to the network infrastructure. However, high-speed Bluetooth and IMT-2000 services will support independent applications, and they will often be operating simultaneously. This means that if the high-speed Bluetooth device is using the 802.11 AMP, it is likely to be running in an environment with IMT-2000 services operating in nearby frequency bands. To illustrate this point, consider the following usage models.
Figure 1 shows two multi-radio handsets located close to each other. One is making a voice call over WiMAX, the other is transferring a file to a PC with 802.11-based high speed Bluetooth. The voice over WiMAX call will drop when the other handset transfers a file to the PC even at a distance of several meters. Figure 2 shows a handset making a voice call with WiMAX and at the same time printing using high-speed Bluetooth. The handset needs to wait for the end of the call before printing without dropping the call. Alternatively, it would not be able to receive a call if the print job already started.
Click here for Figure 1.
Figure 1: One handset transfers a file to a laptop using 802-11-based high-speed Bluetooth. The other handset operates a WiMAX voice call. The two handsets will interfere with each other even at 8m separation.
Click here for Figure 2.
Figure 2: A handset with WiMAX and 802.11-based Bluetooth cannot operate both functions at the same time.
In both of these usage models, any interference between the Bluetooth system and the WiMAX or cellular services would be extremely detrimental to the end user's experience. The reality is that consumers expect to be able to use multiple technologies simultaneously, without interference.
Spectrum allocation
One of the initial reasons for concern regarding interference between the Bluetooth 802.11 AMP and IMT-2000 services is that they operate in adjacent parts of the spectrum (see Table 1). The concern is that the use of 802.11 as a high-speed option for Bluetooth would have severe detrimental effects on other services operating in adjacent licensed frequency bands.
Click here for Table 1.
Table 1. Spectrum allocations.
In addition, even though the Bluetooth SIG intends to limit the 802.11 AMP to file transfer applications, once this high-speed radio functionality exists, users may decide to use it for video streaming (as described by Bluetooth press materials ). Because of their continuous nature, these streaming applications over an 802.11 AMP will have an even higher potential for interference than file transfers.
|
|
|
|
|
robertoaiello
3/20/2008 5:25 PM EDT
Very good comments. Here are my answers:
1. We used conducted measurements to have a more controlled environment. Typical dipole antenna have only 0.1-0.2dB SWR loss at 2.3 or 2.6GHz. Their performance are pretty much equivalent to their in-band performances.
2. The out of band emission for this particular transmitter didnt change when we lowered the transmit power. That is probably because the gain is controlled at the baseband and the RF gain is the same. That means that the out of band interference effect is independent on the in band transmit power. You can see from the picture that the transmit power is less than +10dBm.
3. A different channel model would change the results beyond 5m, but it doesn't change the conclusion that interference is going to be a problem.
4. The interference level was adopted by the ECC based on their study. Either those limits are justified and a 802.11 AMP will interfere, or those limits are not justified and therefore the UWB emission in those bands can be raised. I don't think those limits can be right and wrong at the same time.
5. It may be true that a Wi-Fi transmitter could be harmful to a WiMAX receiver, however adding more devices is only going to make the interference problem worse. Also keep in mind that high speed Bluetooth will operate at a much higher duty cycle than a Wi-Fi network because of the usage scenarios.
6. Any radio that operates farther away from 2.3-2.6GHz will reduce the interference issue. We have limited the analysis to UWB above 6GHz and to 802.11g because they are currently considered by the Bluetooth SIG: UWB because of it is expected to provide good performance and 802.11g because it is expected to be deployed in handsets for other reasons. UWB at 3.1-4.8GHz could also be a viable solution but it is not considered by the Bluetooth SIG.
Roberto
Sign in to Reply
pm01
3/23/2008 4:51 PM EDT
Hello,
After reading the article by Jack Shandle (jshandle@cmp.com)
Sparks fly over Wi-Fi/Bluetooth combo
in
Wireless Net DesignLine
concerning possible interference issues arising from the use of 802.11g until another standard such as 6GHz USB became available, I was wondering:
Why you couldn't further limit interference issues by a "meta-standard", (in-between a "one-size fits--all" standard and a fully "software-defined radio" one)
to:
1. determine the specifications (e.g. bandwidth, error-tolerance, etc.) required by a set of (2 or more) emitters/receivers/transceivers, one based on those specs. and then
2. the modulation scheme which will best meet those specs depending on the context/conditions under which the emitters/receivers are placed (signal/noise ratio, etc.)
For example:
1. First, a general modulation scheme, for example:
a low-power broadband signal
is sent from a transmitter to a receiver to determine the relative distance and location of the latter with respect to the former. Let's say it determines that:
1.1: The receiver (earpieces) are very close to the transmitter, that
1.2: the bandwidth requirements are not very great and that
1.3: there is a minimum latency requirement to be met
(the nature of the content is medium-fidelity mixed music/speech)
2: Based on this information, the transmitter will now switch to
a particular modulation scheme: for instance,
a medium-power narrowband signal
(using a highly directional antenna aimed at the earpieces)
3: The transmitter can
(using the same general modulation scheme use previously)
periodically:
3.1 check to verify that
3.1.1: the specifications of the information to be transmitted (bandwidth, etc.)
-specs determined (for example) by
3.1.1.1: a predetermined value or say
3.1.1.2: a function which depends on the (varying) nature
of the information being transmitted
(e.g. voice vs. video)
and,
3.1.2: the conditions which will allow those specs to be met
- conditions determined (for example) by
3.1.2.1: the maximum variance of the (relative) positions of the
receivers it's targeting
3.1.2.2: obstacles between the emitter and receiver changing
the signal/noise ratio
are (3.1.1 and 3.1.2) still coherent, and if not
3.2: switch to a different particular modulation scheme:
for instance,
a higher-power multi-cast (not narrowband) signal
(less-direction-specific)
On the surface, it would seem that a meta-protocol like this would provide the best of both worlds.
Thank you in advance for any (constructive) feedback you might have on the issue.
-pm01
Sign in to Reply
BOMBOVA
4/17/2008 5:53 PM EDT
It seems to me, that the author has posted a premise for discussion on rf nearfield potential interference.
i asume his intent is to champion the best data rate possible, (highest effective rate) for wireless bluetooth.
The discussion so far seems to have gone from an alert to an alarmist reaction.
Yet understanding this near field model is the objective.
The simplified mathematical model would contain the source, the path loss in wavelengths, etc along with
the coupling effiency.
In addition the model might be, Three users on a bench seat in a mall using their wireless,
the stationary kind; (the mostdebilitating , in my opinion) The moving ones separated by near field
are transmitory and in light systems load are invissible to systems performance.
The static ones (wifi payphones) and other fixed nearfield apparatus, are the most harmful.
The mathmodel, the conducted model and even the radiated model all concur.
Only after this conclusion is reached does the application survive its introduction.
( one thing does puzzle me though, is that the radiating device when reduced from its'
nominal Po. , reducing power, does not reduce spurious, by a move significant amount
other than linear. It may be that a signal generator was used for the real world device...
So if that is valid, then the interference in fianal Po. and radiation is more dirty and harmful
than presented... " my interest, is only in the application and a best of model outcome. " rtg.
Sign in to Reply