Time Division Multiplexing (TDM) is one tool used against the front end overload-type interference which cannot be managed by AFH. This approach was originally introduced to protect 802.11b/g transmissions from Bluetooth interference rather than vice versa and works by shutting down all Bluetooth transmissions except those which are high priority when the 802.11b/g radio is active on the ISM band.
This approach, like channel skipping, sacrifices part of the Bluetooth bandwidth but the amount of bandwidth sacrificed is proportionate to the 802.11b/g duty cycle. Therefore, if the 802.11b/g is idle, the link maintenance traffic may lead to a small 2 to 3% bandwidth degradation, which is impossible for a user to detect.
Click here for Figure 2
Figure 2: Time Division Multiplexing (TM).
In order to enhance the effectiveness of TDM, it is necessary to have accurate information regarding the activity of the 802.11b/g radio. One wireless technology company, CSR, uses a WLAN_Active hardware signal to ensure that the b/g signal is protected when the radio becomes active.
However, there are times when degradation of the Bluetooth signal is to be protected from 802.11b/g interference, and so CSR developed BT_Priority, an optional signal which indicates when an important Bluetooth packet is being transmitted or received.
This signal can be used to protect SCO audio using HV3 packets, the most common form of streaming audio to and from a mono headset.
Channel Quality-Driven Data Rate (CQDDR)
Two forms of data packets exist, DH and DM, which use high and medium bandwidth respectively. DH packets can transmit more data within the packets but if a part of the packet is corrupted the entire packet must be retransmitted to recover the data.
The DM packets include forward error correction (FEC) code which takes up a third of the payload: for every 10 bits of data, a 5-bit forward error correction code is added, allowing the correction of up to two bit errors in each 15-bit data/FEC block.
This data packet format may reduce maximum data rate but is more robust than the DH packets which have no error correction included. It allows a receiving device to negotiate with a transmitter to agree which packet format is used according to ambient interference.
For example, if one device believes it is receiving packets with many errors, it tells the transmitter to send the data in DM packets. If the link clears up, it allows the other side to switch back to DH packets. Figure 3 illustrates this communication interchange.
Click here for Figure 3
Figure 3: Channel Quality-Driven Data Rate (CQDDR).
CQDDR remains an optional addition to a Bluetooth link and is not required by any Bluetooth specification.
Therefore, for example, when a CSR's BlueCore-enabled Bluetooth device sends data to a non-CQDDR-enabled device, CSR has developed an algorithm to estimate the performance of the link and to modify the type of data packets sent in accordance with the ratio of acknowledged packets (ACKs) to not-acknowledged packets (NACKs).
However, when receiving information from a non-CQDDR-enabled device, BlueCore cannot implement any such countermeasures if the data packets are corrupted.
Extended Synchronous Connection Oriented Channels (eSCO)
eSCO are error-checking voice channels that allow the retransmission of corrupted voice data. Each packet has a CRC (Cyclic Redundancy Check) so the receiver can check that packets have been received correctly.
Packets which are received with errors or lost altogether are negatively acknowledged. Retransmission windows allow retransmission of unacknowledged packets. eSCO was introduced with v1.2 of the Bluetooth specification.
Version 1.1 SCO used in earlier versions of Bluetooth only used single slot packets. Extended SCO (eSCO), on the other hand, allows the use of 3 slot packets for synchronous voice or data.
This means it is possible to get >100 kbps connection compared with the fixed 64 kbps from version 1.1. This is possible due to link capacity being lost in the case of single slot packets, to gaps between packets while the radio changes frequencies.
Click here for Figure 4
Figure 4: Extended Synchronous Connection Oriented Channels (eSCO).
At each eSCO instant the master transmits an eSCO packet, the slave responds using the normal SCO rules (the slave is allowed to respond even if it doesn't receive the master's packet). Then the differences from SCO become apparent: there is a retransmission window during which unacknowledged packets can be resent until acknowledged. The spacing of the eSCO instant is negotiable.
With version 1.1 SCO there was a choice of 3 different packet spacings all giving the same 64kb/s. With extended SCO, both packet length and intervals can be negotiated in both directions of the link allowing asymmetric traffic.
Although eSCO channels do not actively handle or avoid interference, the retransmission of corrupt data packets ensures that audio quality is less affected by interference from other radio devices than before.