Time Division Multiple Access or TDMA is a method used to enable multiple earth stations or VSAT terminals to transmit intermittently on the same frequency, but with the timing of their transmissions so arranged that the bursts do not overlay when they arrive at the satellite but arrive in sequence and thus are all successfully received by the teleport hub modem burst demodulator.
The operation of TDMA requires an outlink control to all the remote sites which contains some control information. This outlink carrier also had a frame structure that provides accurate timing information for all the remote sites. The teleport hub equipment computer tells each VSAT site what particular time slot to use in the TDMA frame and this time plan information is broadcast to all sites periodically. The burst time plan may be fixed, so as to allocate each site a particular proportion of the total TDMA frame time or is may be dynamic, whereby the time slot allocated is adjusted in response to the traffic needs of each site.
This image below shows a sequence of two successive TDMA frames passing through the satellite. The carrier bit rate is 250 kbit/s
Explanation: Site 1 transmits a burst, starting at the beginning of each TDMA frame. The burst lasts 180 mS, so at a rate of 250kbit/s site 1 sends 45,000 bits per burst, or 45,000 bits per second. Site 2 transmits a burst, timed to arrive at the satellite just after the end of burst 1. The red, second, burst lasts 80 mS, so at a rate of 250kbit/s, site 2 sends 20,000 bits per burst, or 20,000 bits per second. The diagram shows a fixed time plan, where each VSAT has been allocated a predetermined portion of the total time.
There is designed in 20mS guard period between each burst. This allows for slight mistiming in the transmission of the bursts. Severe mistiming would cause bursts to arrive overlapping or on top of each other, causing loss of service to both sites involved in the mutual interference. The long 20mS guard period is illustrative only, so you can see the white space in the figure above. In actual TDMA systems the guard band may be very much less and there may be very many more bursts per frame.
The above is just an example. TDMA frame length may be as short as 2000 microseconds or as long as 1 second, as in the example above. The shortest TDMA frame periods are associated with the highest speed TDMA systems, operating at say 120.832 Mbit/s. On low speed 250kbit/s VSAT return links, with perhaps 2 to 50 sites sharing, and used for internet browsing and emails, the TDMA frame period is typically 500mS.
The allocation of bursts to time slots within the TDMA frame is the burst time plan. As shown above the plan is:
VSAT site 1 start 0 mS, time allocated 180 mS
VSAT site 2 start 200 mS, time allocated 80 mS
VSAT site 3 start 300 mS, time allocated 180 mS
VSAT site 4 start 500 mS, time allocated 280 mS
VSAT site 5 start 800 mS, time allocated 180 mS
This information is broadcast to all sites, which then follow the timing instructions. This burst time plan might be applied unchanged for several days or weeks or it might be changed every few seconds or minutes according to the traffic demand.
The time of arrival of each burst at the satellite is critical and to get this right each site is told when to transmit.
This instruction is based on the burst time plan start time value plus also a time delay based on the range of VSAT site to the satellite. The range is initially calculated based on the latitude and longitude of the earth station and the orbital position of the satellite. The amount of time to allow for, due to range, is calculated using the speed of light, which is 300m per microsecond.
For a new site to be introduced into the burst time plan, a slot is prepared for it by the hub staff. Using the site lat and long, an approximate timing value is estimated. When the site is activated the burst should appear in the correct place. Fine burst timing adjustment may then be possible to fit the burst exactly without wasting too much space in the guard times either side. Time wasted in the guard times is effectively capacity lost out of the total 250kbit/s available.
VSAT terminals further away from the satellite, like site 1 on the left, need to transmit earlier than site 2 which is already nearer the satellite.
This is an explanation why you need to tell the VSAT hub what is your latitude and longitude before starting first transmissions to the teleport hub.
This image represents a burst, whose time length may be 180 mS. At the start of the burst the bit sequence follows a predetermined burst preamble or header designed to help the hub receiver demodulator lock onto the carrier. This is described as the "carrier and bit timing recovery sequence and unique word". A brief period of un-modulated CW carrier is a good start. This helps the demodulator automatic frequency control (AFC) adjust its centre frequency and also get the automatic gain control (AGC) so that the level of the carrier is made nominal. Next a period of alternate ones and zeros helps the demodulator to set the exact bit rate e.g. 249,999 bit/sec. Following this is the unique word. This is a pre-selected complex sequence of ones and zeros which the receiver attempts to match against two opposite matching patterns. When a hit is obtained the demodulator knows two things. It can now distinguish a one from a zero and it know its exact position in the burst. At the end of the unique word the User data message starts. Right at the end may be a cyclic redundancy (CRC) checksum to check if there have been any errors during the burst. Additional bits for Forward Error Correction (FEC) are normally added also. These may be distributed along the packet or added as a extra group of bits at the end. The CRC checksum simply detects if there have been any errors. FEC will put right and repair errors, provided there are not too many.
Allowing for all the guard time slots between the bursts and the burst preambles the overall efficiency of use is typically 90%. The situation improves for fewer, longer bursts and deteriorates for large numbers of short bursts.
In a random demand TDMA system the actual fill efficiency will fall to 0% when nobody is active, such as night. During busy times, efficiency of fill may still need to be kept very low, such as 35% to avoid of burst collisions and congestion. Much effort has gone into the design of demand assigned TDMA systems where earth stations are temporarily assigned to certain sequence time slots for the duration of transmission of a large file for example. This improved the fill factor significantly.
Page started 20 Nov 2006, amended 18 July 2021