I am trying to understand how information is encoded within a GSM signal. All I have found is as summarized in the link below.
Can anyone please explain briefly how and where the voice is inserted?
Thank you for any insights.
Bob Graham
I am trying to understand how information is encoded within a GSM signal. All I have found is as summarized in the link below.
Can anyone please explain briefly how and where the voice is inserted?
Thank you for any insights.
Bob Graham
"This digital signal is modulated onto the analog carrier frequency using Gaussian-filtered Minimum Shift Keying (GMSK)"
Read that and still a bit hazy.
Is the following interpretation correct please?
The "900MHz" analog carrier is actually one of the 124 200KHz-spaced FDMA frequencies.
As I understand, these are hopped to minimize interference. BTW what is the usual hopping rate?
For TDMA, the selected carrier is pulsed at a 217Hz rep rate, with every 25th and 26th pulse omitted.
The pulse duration is 0.5ms, meaning 8 separate real time messages can be fit into one TDMA slot. And riding within that time slot is the GMSK.
Is that about it?
Just wondering ... if all eight 0.5ms time slots are filled, presumably by 8 transmitters does the transmitted signal more or less resemble a continuous 900MHz carrier transmision? MSK aside of couse.
Bob Graham
All GSM standards can be found at
Wim Ton
The GSM specification is available on the net. A careful reading will reveal all.
Yes. Stated frequencies are often frequency bands. Except when they're not.
Can't remember. The spec is on the net.
I wouldn't say 'riding within that slot', I'd say 'during that slot, GMSK is transmitted'.
If it's eight different transmitters, then the amplitude and delay within each slot will vary. If it's eight different sources all going through the same transmitter, then I think there'll still be 'holes' for the guard times in TDMA.
The spec is on the net.
-- Tim Wescott Wescott Design Services
Roughly-speaking, yes... but note that in some cases people just round to some "nice" number and the frequency specified isn't the exact center frequency of where the systems operates. See:
Apparently it's often at the same ~217Hz rate as the burst (repetition) rate... although I wouldn't be surprised if this is controllable via the software/cell tower's configuration.
Those I don't know.
Yes -- pretty much any modern (TDMA, CDMA, OFDM, etc.) digital communication system today looks a lot like a "brick wall" on a spectrum analyzer.
---Joel
I hadn't read before that each transmitter could be assigned a space WITHIN one of the 8 slots, eg. by differing delay and amplitude as you mention.
I thought each BTS just had a different slot, and the point of the low duty cycle time was to maintain amplitude, being signal strength..
It would be appreciated if you can please explain your comment further so I do not misunderstand.
Bob Graham
Certainly:
The complete specification is on the web, for you to read. It will clarify all.
-- Tim Wescott Wescott Design Services
I looked and could not find anything about delays or amplitude variations within a _single_ TDMA time slot that you seem to have referred to.
Any chance of a more definitive answer?
Bob Graham
The timeslots at the base station are fixed, but the handsets have adjustable delays such that regardless of how far away they are (within specified limits) their transmissions arrive at just the right time to fit into the timeslot in which they are expected.
If the base station is transmitting on all the timeslots then the transmission is almost continuous. The signal level from the base station is normally at constant level, but the handsets vary their transmitter power so as to maintain an adequate noise margin yet avoid running down the battery unnecessarily.
The base station has filters which allow simultaneous transmission and reception (on widely separated frequencies), whereas the handset has simpler filters and transmits and receives on different timeslots. Frequency hopping is optional and is often not used. There is a lot more going on, including pulse-by-pulse equalisation of multipath effects using a training sequence in the middle of each pulse and keeping track of the availability of alternative base stations.
Nearly everything is in the published standards, although they don't make easy reading.
John
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.