GPRS info
snip - From GSM to GPRS
Before describing how a GPRS device uses the GSM network, it is important to review how a GSM mobile uses the same network. When a GSM user desires to make a call, they dial a number and push a button. The mobile negotiates with the network for access to a channel, and if granted, initiates communication with the person on the other end of the connection. Communication over the air interface occurs at a certain frequency, on one of eight timeslots on that frequency, and at a compliant power level. The GSM user solely owns this channel for as long as the conversation takes place. When the call is finished, the GSM mobile informs the network and the channel is released for use by someone else.
However, the GPRS device can use multiple timeslots and power levels can change between timeslots. These 'sessions' are random in nature resulting in a 'bursty' channel in stark contrast to the well-behaved GSM circuit switched session.
A further difference is the downlink and uplink configurations are no longer required to look identical. The mobile device can receive a different number of timeslots on the downlink than it transmits on the uplink. A common desire among today's GPRS designers is to have a device that can receive four downlink timeslots and transmit on only one uplink timeslot. This configuration lends itself well to applications like web browsing, as the user tends to ask for more data than they send.
In summary, with GPRS, the mobile can send and receive bursts at different power levels in the same frame; send and receive multiple timeslots per frame, and share the same physical channel with many users at the same time. For a complete understanding of the system, and a good starting point when approaching the standards, read GSM 03.64, entitled the Overall Description of the GPRS Radio Interface. The 3GPP standards committee provides GSM standards documents for no charge at 3gpp.org.
Power levels
The current system for establishing GSM mobile station transmit power has the base station assigning levels to each mobile so all signals received at the cell site are at about the same level. With GPRS, base stations and mobile devices use different, real-time power control algorithms that allow much lower levels during transmission. New to mobile station design are concepts of open loop and closed loop power control, much like CDMA today. GSM uses closed loop control to dictate mobile station power levels -- GSM mobiles do not vary their output power unless directed by the cell. Open loop is the new mode for GPRS mobile stations that has the mobile compare two possible values for power level, and transmit with the lower one. One value, PMAX, is a variable established by the base station and establishes the maximum power allowed for mobile transmission. The mobile, using network variables and cell power level as received by the mobile, calculates the other value. If the mobile station fails to evaluate this equation correctly, the resulting transmission could occur at either such a low level that the session drops, or such a high level that the air interface around the device becomes over-powered. Open loop control requires a good link between received signal and transmitted power, so it will likely require calibration, and possibly testing, during production. A fault during manufacturing could effect either up or down open loop transitions resulting in a single customer having sessions dropped through low power, or several customers losing service by a mobile transmitting too much power. Luckily, the standard gives a few examples of the many ways to calibrate and test power control features during production.
Multislot power transitions
A GPRS transmitter must be capable of transmitting multiple timeslots in every frame, with the dynamic range between adjacent timeslots limited only by the power class of the mobile station. A look at Figure 1 uncovers at least two new design challenges, and potential interference mechanisms, presented by multislot transmissions. First, the device must be capable of dissipating the additional heat generated by transmitting on more than one burst per frame. Second, ETSI has defined a new compliance mask for multislot transmissions that the entire burst profile must fit within to ensure GPRS mobiles do not interfere with each other. For the thermal challenge, there is no quick solution, but for the mask compliance challenge the standards give some leeway. GSM 11.10-1 describes a required test for power level and mask compliance. During the test, the mobile is set up on a call using the maximum number of uplink timeslots the device can transmit. The first timeslot is set to maximum power, the second to minimum power, and all subsequent timeslots set again to maximum. This presents a hurdle for the transmitter since it must be capable of producing large transitions over its entire range in a short time without violating the mask or splattering frequency components outside of the assigned channel. The standard provides some cushion during multislot transmissions by not demanding the mobile reduce power to zero between adjacent bursts. If we read the standard again more carefully we find, for example, that the transition from the low power to high power not only does not need to go down, but it can overshoot a little too!
snip
snip
BS power control
Two modes of power control are available to GPRS base stations: power control modes A and B. Mode A defines a conservative band for the downlink without regard to allocation mode. Mode B places few restrictions on the base station and, depending on allocation mode can result in very low downlink power levels. For most of a fixed allocation transfer, a base station in mode B need only transmit the minimum power required to maintain a link with the receiving mobile. Figure 2 shows a potential air interface for a mobile assigned to monitor four downlink timeslots for a pending GPRS session. The timeslots shown represent data transfers taking place with three other mobiles. Of interest is the transition between the power used for the blue mobile, which may be at the extremes of the cell and require a high broadcast power, and the orange mobile, which is in a fixed allocation and very close to the cell. This results in a situation where the decoding mobile must be able to receive and decode a very low power burst shortly after receiving and decoding bursts of a high power. The receiver designer now has an almost unlimited combination of power levels and multislot configurations to consider when mapping out a design strategy.
snip
snip
Verifying GPRS receivers
Traditional bit error ratio (BER) testing still applies to GPRS, but with standard test modes and multiple timeslots, testing is much faster than with GSM for the same number of bits. In addition to BER, a new feature provided by GPRS is block error ratio (BLER). BLER is a receiver-reported parameter that gives an indication of the number of bad blocks received over good blocks. BLER takes advantage of the lower layers of GPRS protocol and their use of block sequence numbers (BSN). Manipulating the flags used by the lower protocol layers provides engineers BLER testing in addition to BER for characterising mobile transceiver performance. Setting up a data transfer at a given low downlink power, and corrupting a certain number of BSNs, can verify operation. If the mobile transceiver works correctly, a query of BLER would return the correct number of errors. The steering bits -- called uplink state flags (USF) -- could be used in a similar fashion by granting the mobile a dynamic allocation using multiple slots, varying the power of the downlink timeslots, and limiting the correct USFs sent to a given number.
Comparing the number of times the mobile transmitted with the number of correct USFs could also be used to characterise transceiver performance. Testing BER and BLER during production verifies the device can actually deliver the high data rates advertised, which is key to successful implementation of GPRS systems. The potential test coverage when manipulating power level, the number of timeslots, and flag integrity is almost endless and limited only by the mobile station and test instrument capabilities.
GPRS is an impressive new technology that brings 2.5G performance to a stable GSM system in some very clever ways. Packet switching, multiple timeslots, and channel sharing will challenge engineers at all stages of product lifecycle with new hurdles to overcome. Luckily for GPRS designers, the standards provide many new ways to help characterise and verify system performance, allowing for testing of the very features like data speed and efficiency that make GPRS so exciting. Proper testing during production will provide top performing GPRS devices to the customer and high revenues to the network operators as they introduce a fully functional GPRS service. *
Richard Macquire, marketing product manager, Agilent Technologies |
