First off, I'll just mention that for the past seven years I've been designing RF chips for a variety of systems such as GSM, Bluetooth, 802.11 and other stuff as my real job.
Software defined radio is a big research topic in the commerical RF world. There are currently two big problems that people are trying to overcome. The first is the ADC requirements and the second is the MIPs requirements for any two way system are really beyond the ability of most current general purpose technolgy. It will get there but not yet.
I want to add my own thoughts to the questions and answers people asked. They are meant to point out how much work there is to do, and also some of the legal implications of trying to do them.
1) Hardware requirements
For GSM you will need about 13800ks/s if all the channel filtering is implemented in the radio. If not, reckon on about 16bits and 13MHz.
3) Winmodems etc.
Eric raises a hugly important point about the real time nature of many RF systems. For example, 802.11a requires that turn around time on ACKing packets is in the order of a few micro-seconds. This includes decoding that packet and generating the reply. It is hard enough to do this in custom hardware, and CPU interrupt latencies are just too high. Many 802.11a chipsets use two ARMs to meet these timing requirements, one running real-time code the other dealing with the protocol.
Oh yeah, a protocol stack for GSM is usually larger than a Linux/FreeBSD kernel in terms of lines of code...
BTW, if you hunt around the net, you can find the signal processing for GSM available as a matlab library. It includes all the vocoders and channel coding and viterbi's.
9) Interference etc.
Eric seems to not have much experience of type approval testing for devices such as cell phones. There is no way it will be legal to develop your own GSM cell phone on a PC and use it in Europe for example.
Every GSM handset must be type approved before it is allowed to be used on a network. This costs a good few kbucks. You must be licensed to transmit in the bands etc.
On the hardware side you have issues like the fact that it takes lots of RF design experience to build a circuit that meets the spurious emissions requirements, and standard PC hardware will not meet some of the basic timing requirements of 0.5ppm tracking to the BTS and that you must use the same crystal for frequency generation and timing.
10) Patents
Most protocols are covered in patents. GSM is a minefield and every cell phone has royalities paid on them. Qualcomm are the patent owners for IS95.
11) UWB
UWB will need samples rates >10Gs/s to do digitally. At present the Rake receviers are implemented as part of the RF.
Reasons why the band is filling up
on
Future of Wi-Fi
·
· Score: 1
A couple of posters have asked the question, if 802.11b has a range of 300m why are people worried about the band filling up?
It's all down to the formula posted earlier called Shannon's Law. This dictates that the amount of error free data sent through a particular bandwidth is related to the signal to noise ratio.
The unfortunate fact is that, as more and more 802.11 devices are used, along with Bluetooth, TV senders etc. the noise in a particular location will increase. Thus the error free data rate drops.
802.11g will help a little but all it does is to pack more data into the same bandwidth. In fact, 802.11g (and 802.11a which uses the same modulation) are approaching the practical limits of bits per second per Hertz (around 4 or 5) - i.e. the data rate for a given used amount of bandwidth. Here I am defining pratical as something that does not need an excessivly long forward-error correction scheme, stupid amounts of equalisation or excessive power. For those interested, the current state of the art in wireless data is turbo-coded data over orthogonal frequency division multiplex or quadrature amplitude modulation systems. 802.11a and g use orthogonal frequency division multiplex with convolutional encoding/forward error correction.
The up and coming star to get higher data rates (above the 108Mb/s that some 802.11a systems get) is Ultra-Wideband (UWB). The FCC has just (Feb/March) regulated this for indoor use in the states. This technique is the "short pulse" method mentioned in the poor journalistic piece we are talking about. At present UWB is being considered for 802.15.3 high rate personal area wireless networks http://www.ieee802.org/15/pub/TG3.html. This technology is mix of old and new. Old in that the ideas have been around for a while (ever seen the through the wall "radar" - it's really UWB) but for use in domestic/commerical/consumer applications there are a lot of fundimental questions to be answered - mainly around "how much data in a real life situation can we pump". The use of UWB for 802.15.3 is so early that people have only just got proposals in for the physical layer - this means products in the shops are three or four years away.
Re:ARRL Witch Hunt
on
Future of Wi-Fi
·
· Score: 2, Informative
Sorry, 802.11b falls under FCC Part 15, section 245, 247 etc. Take a look at the 802.11 specs. whwere in section 18.4.7.1 the TX power levels are given as defined in FCC Pt. 15.247.
Okay, so I'm a real EE who design in IBM SiGe processes 5HP and 6HP.
1) IBM did demonstrate a ring oscillator.
2) These are IBMs latest SiGe HBT transistors, targetted for the "8HP" process. At present, 5HP and 6HP are in production and producing ICs - a lot of GSM cell phones will have IBM silicon in them. 7HP is coming on line.
3) Yup - these process are not directly for PC processors. The processes are targetted at RF, electro-optical, high speed data etc. They have SiGe transistors and CMOS. The SiGe is typcially used as a front-end, e.g. 10gigabit mutliplexers and laser driver/demultiplexors and diode detectors for optical links and the CMOS does the back end processing - e.g. line equalization etc.
In addition, this is not the fastest semiconductor circuit. For many years people have been using semiconductors at tera-Hz for microwave stuff (granted maybe not ring oscillators but certainly parametric-active amplifiers). I worked on 94GHz radar systems over 10yrs ago that used active semiconductors (IMPATT and Gunn GaAs oscillators).
Slashdot Mirror
First off, I'll just mention that for the past seven years I've been designing RF chips for a variety of systems such as GSM, Bluetooth,
802.11 and other stuff as my real job.
Software defined radio is a big research topic in the commerical RF world. There are currently two big problems that people are trying
to overcome. The first is the ADC requirements and the second is the MIPs requirements for any two way system are really beyond the ability of most current general purpose technolgy. It will get
there but not yet.
I want to add my own thoughts to the questions and answers people asked. They are meant to point out how much work there is to do, and also some
of the legal implications of trying to do them.
1) Hardware requirements
For GSM you will need about 13800ks/s if all the channel filtering is implemented in the radio. If not, reckon on about 16bits and 13MHz.
3) Winmodems etc.
Eric raises a hugly important point about the real time nature of many RF systems. For example, 802.11a requires that turn around time on ACKing packets is in the order of a few micro-seconds. This includes decoding that packet and generating the reply. It is hard enough to do this in custom hardware, and CPU interrupt latencies are just too
high. Many 802.11a chipsets use two ARMs to meet these timing requirements, one running real-time code the other dealing with the protocol.
Oh yeah, a protocol stack for GSM is usually larger than a Linux/FreeBSD kernel in terms of lines of code...
BTW, if you hunt around the net, you can find the signal processing for GSM available as a matlab library. It includes all the vocoders and channel coding and viterbi's.
9) Interference etc.
Eric seems to not have much experience of type approval testing for devices such as cell phones. There is no way it will be legal to develop your own GSM cell phone on a PC and use it in Europe for example.
Every GSM handset must be type approved before it is allowed to be used on a network. This costs a good few kbucks. You must be licensed to transmit in the bands etc.
On the hardware side you have issues like the fact that it takes lots of RF design experience to build a circuit that meets the spurious
emissions requirements, and standard PC hardware will not meet some of the basic timing requirements of 0.5ppm tracking to the BTS and that you must use the same crystal for frequency generation and timing.
10) Patents
Most protocols are covered in patents. GSM is a minefield and every cell phone has royalities paid on them. Qualcomm are the patent owners for IS95.
11) UWB
UWB will need samples rates >10Gs/s to do digitally. At present the Rake receviers are implemented as part of the RF.
A couple of posters have asked the question, if 802.11b has a range of 300m why are people worried about the band filling up?
It's all down to the formula posted earlier called Shannon's Law. This dictates that the amount of error free data sent through a particular bandwidth is related to the signal to noise ratio.
The unfortunate fact is that, as more and more 802.11 devices are used, along with Bluetooth, TV senders etc. the noise in a particular location will increase. Thus the error free data rate drops.
802.11g will help a little but all it does is to pack more data into the same bandwidth. In fact, 802.11g (and 802.11a which uses the same modulation) are approaching the practical limits of bits per second per Hertz (around 4 or 5) - i.e. the data rate for a given used amount of bandwidth. Here I am defining pratical as something that does not need an excessivly long forward-error correction scheme, stupid amounts of equalisation or excessive power. For those interested, the current state of the art in wireless data is turbo-coded data over orthogonal frequency division multiplex or quadrature amplitude modulation systems. 802.11a and g use orthogonal frequency division multiplex with convolutional encoding/forward error correction.
The up and coming star to get higher data rates (above the 108Mb/s that some 802.11a systems get) is Ultra-Wideband (UWB). The FCC has just (Feb/March) regulated this for indoor use in the states. This technique is the "short pulse" method mentioned in the poor journalistic piece we are talking about. At present UWB is being considered for 802.15.3 high rate personal area wireless networks http://www.ieee802.org/15/pub/TG3.html. This technology is mix of old and new. Old in that the ideas have been around for a while (ever seen the through the wall "radar" - it's really UWB) but for use in domestic/commerical/consumer applications there are a lot of fundimental questions to be answered - mainly around "how much data in a real life situation can we pump". The use of UWB for 802.15.3 is so early that people have only just got proposals in for the physical layer - this means products in the shops are three or four years away.
Sorry, 802.11b falls under FCC Part 15, section 245, 247 etc. Take a look at the 802.11 specs. whwere in section 18.4.7.1 the TX power levels are given as defined in FCC Pt. 15.247.
Okay, so I'm a real EE who design in IBM SiGe processes 5HP and 6HP.
1) IBM did demonstrate a ring oscillator.
2) These are IBMs latest SiGe HBT transistors, targetted for the "8HP" process. At present, 5HP and 6HP are in production and producing ICs - a lot of GSM cell phones will have IBM silicon in them. 7HP is coming on line.
3) Yup - these process are not directly for PC processors. The processes are targetted at RF, electro-optical, high speed data etc. They have SiGe transistors and CMOS. The SiGe is typcially used as a front-end, e.g. 10gigabit mutliplexers and laser driver/demultiplexors and diode detectors for optical links and the CMOS does the back end processing - e.g. line equalization etc.
In addition, this is not the fastest semiconductor circuit. For many years people have been using semiconductors at tera-Hz for microwave stuff (granted maybe not ring oscillators but certainly parametric-active amplifiers). I worked on 94GHz radar systems over 10yrs ago that used active semiconductors (IMPATT and Gunn GaAs oscillators).
Bluetooth and 802.11-FH seem to be okay at work. We develop both and have a deployed 802.11 network throught the building.