Hello all, What kind of modulation scheme that will suit high speed transmission (at least to 1 Gbps)? And obvious choice will be those highly spectral efficient modulation scheme like OFDM. However, if the linearity requirement and the sensitivity to phase noise are very high, then OFDM could be out of question especially for system that operate at very high frequency such as 40-70GHz . So what other modulation schemes that can give good trade off between linearity requirement, sensitivity to phase noise, equalization complexity, performance in such a high speed system? Any comments will be highly appreciated. Regards, Lindah
What is the best modulation scheme?
Started by ●October 26, 2005
Reply by ●October 26, 20052005-10-26
lindah74uk@yahoo.co.uk wrote:> Hello all, > > What kind of modulation scheme that will suit high speed transmission > (at least to 1 Gbps)? > And obvious choice will be those highly spectral efficient modulation > scheme like OFDM. However, if the linearity requirement and the > sensitivity to phase noise are very high, then OFDM could be out of > question especially for system that operate at very high frequency such > as 40-70GHz . > > So what other modulation schemes that can give good trade off between > linearity requirement, sensitivity to phase noise, equalization > complexity, performance in such a high speed system?You missed one critical specification - the bandwidth of the channel. With a millimetre wave carrier, the bandwidth might be quite large (GHz), allowing inefficient yet robust and simple to implement modulation schemes like BPSK. Regards, Allan
Reply by ●October 26, 20052005-10-26
On 25 Oct 2005 23:54:06 -0700, lindah74uk@yahoo.co.uk wrote:>What kind of modulation scheme that will suit high speed transmission >(at least to 1 Gbps)? >And obvious choice will be those highly spectral efficient modulation >scheme like OFDM. However, if the linearity requirement and the >sensitivity to phase noise are very high, then OFDM could be out of >question especially for system that operate at very high frequency such >as 40-70GHz . > >So what other modulation schemes that can give good trade off between >linearity requirement, sensitivity to phase noise, equalization >complexity, performance in such a high speed system?What are the channel characteristics? I often see OFDM touted as "spectrally efficient", but it's only efficient if the channel is frequency selective. If the channel is AWGN with a reasonable SNR then a high-order single carrier QAM system may be better. As you've already hinted the radio impairments matter as well. As does the expected range, power usage, etc. Eric Jacobsen Minister of Algorithms, Intel Corp. My opinions may not be Intel's opinions. http://www.ericjacobsen.org
Reply by ●October 27, 20052005-10-27
"Eric Jacobsen" <eric.jacobsen@ieee.org> wrote in message news:ogcvl1lvfbfviikp7vkfkqnrbnm9l3uo64@4ax.com...> On 25 Oct 2005 23:54:06 -0700, lindah74uk@yahoo.co.uk wrote: > > I often see OFDM touted as "spectrally efficient", but it's only > efficient if the channel is frequency selective. If the channel is > AWGN with a reasonable SNR then a high-order single carrier QAM system > may be better.Correct me if I am wrong, but isn't it the case that the symbol rate for QAM can be only as high as the carrier frequency? This limits the channel bandwidth you can cover, and hence the total data-rate you can achieve. For example, assume you have a flat fading channel that can pass frequencies from DC to 1MHz. If you place the QAM carrier at 500kHz, your symbol rate would have to be 500Ksymbols/s max. With OFDM you can fill up (almost) the entire 1MHz with carriers, hence reaching an aggregate symbol rate close to 1Msymbol/s. Is this correct...?
Reply by ●October 27, 20052005-10-27
> What are the channel characteristics? > > I often see OFDM touted as "spectrally efficient", but it's only > efficient if the channel is frequency selective. If the channel is > AWGN with a reasonable SNR then a high-order single carrier QAM system > may be better. > > As you've already hinted the radio impairments matter as well. As > does the expected range, power usage, etc. > > > Eric Jacobsen > Minister of Algorithms, Intel Corp. > My opinions may not be Intel's opinions. > http://www.ericjacobsen.orgHi Eric, The channel is freqeuncy selective as shown by many measurements. As pointed out by Allan that we have a huge bandwidth, then a less efficeint could be used. So the question is which modulation will give the best trade-off among the the constraints mentioned above. I guess it might not be so easy to answer unless simulation is performed in a realistic channel and circuit models. But I would be interested to hear some opinions and discussions from the experts in this newgroup about the suitability of type of modulation subject to the aforementioned problems. Comments are very much welcome. Regards, Lindah
Reply by ●October 27, 20052005-10-27
Snowball wrote:> "Eric Jacobsen" <eric.jacobsen@ieee.org> wrote in message > news:ogcvl1lvfbfviikp7vkfkqnrbnm9l3uo64@4ax.com... > >>On 25 Oct 2005 23:54:06 -0700, lindah74uk@yahoo.co.uk wrote: >> >>I often see OFDM touted as "spectrally efficient", but it's only >>efficient if the channel is frequency selective. If the channel is >>AWGN with a reasonable SNR then a high-order single carrier QAM system >>may be better. > > > Correct me if I am wrong, but isn't it the case that the symbol rate for QAM > can be only as high as the carrier frequency? This limits the channel > bandwidth you can cover, and hence the total data-rate you can achieve. For > example, assume you have a flat fading channel that can pass frequencies > from DC to 1MHz. If you place the QAM carrier at 500kHz, your symbol rate > would have to be 500Ksymbols/s max. With OFDM you can fill up (almost) the > entire 1MHz with carriers, hence reaching an aggregate symbol rate close to > 1Msymbol/s. Is this correct...?Each channel will have a bandwidth proportional to signaling speed and the channels must not overlap. A practical implementation won't turn out as optimistically as you seem to think. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 27, 20052005-10-27
"Jerry Avins" <jya@ieee.org> wrote in message news:ZZWdnfy7sspoe_3eRVn-rw@rcn.net...> > Each channel will have a bandwidth proportional to signaling speed and > the channels must not overlap. A practical implementation won't turn out > as optimistically as you seem to think. > > Jerry > --You're right, I guess I didn't express myself correctly. It was meant to be an idealized example to show that you can use more channel bandwidth with OFDM than with a single carrier scheme (this is only true in a channel with a lowpass response!). I hope I am not missing something obvious...
Reply by ●October 27, 20052005-10-27
Snowball wrote:> "Jerry Avins" <jya@ieee.org> wrote in message > news:ZZWdnfy7sspoe_3eRVn-rw@rcn.net... > >>Each channel will have a bandwidth proportional to signaling speed and >>the channels must not overlap. A practical implementation won't turn out >>as optimistically as you seem to think. >> >>Jerry >>-- > > > You're right, I guess I didn't express myself correctly. It was meant to be > an idealized example to show that you can use more channel bandwidth with > OFDM than with a single carrier scheme (this is only true in a channel with > a lowpass response!). I hope I am not missing something obvious...I don't think we can deduce channel capacity with arm waving. Factors to consider include robustness in the presence of noise and fading. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 27, 20052005-10-27
> > You're right, I guess I didn't express myself correctly. It was meant to be > an idealized example to show that you can use more channel bandwidth with > OFDM than with a single carrier scheme (this is only true in a channel with > a lowpass response!). I hope I am not missing something obvious...I don't see your point about OFDM ?? If you have 1 million carriers each 1 Hz wide, then each carrier can support a symbol rate of 1/2 sym/s. The total aggregate symbol rate is still 500ksym/s just like QAM? Mark
Reply by ●October 27, 20052005-10-27
This is a multi-part message in MIME format. ------=_NextPart_000_000E_01C5DB34.C68E6410 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable "Mark" <makolber@yahoo.com> wrote in message = news:1130430253.588590.265420@g43g2000cwa.googlegroups.com...>=20 > I don't see your point about OFDM ?? >=20 > If you have 1 million carriers each 1 Hz wide, then each carrier can > support a symbol rate of 1/2 sym/s. The total aggregate symbol rate =is> still 500ksym/s just like QAM? >=20 > Mark >Hi Mark, If the symbol rate, 1/T, on each of the subcariers in OFDM is equal to = the subcarrier separation, they are orthogonal over the symbol interval = T. Thus if you were to place 1 million carriers in 1MHz, they would be = spaced at 1Hz and their symbol rate would be 1symbol/s. I'm not sure = where you got 1/2 symbols/s. ------=_NextPart_000_000E_01C5DB34.C68E6410 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1522" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV><FONT face=3DArial size=3D2>"Mark" <</FONT><A=20 href=3D"mailto:makolber@yahoo.com"><FONT face=3DArial=20 size=3D2>makolber@yahoo.com</FONT></A><FONT face=3DArial size=3D2>> = wrote in=20 message </FONT><A=20 href=3D"news:1130430253.588590.265420@g43g2000cwa.googlegroups.com"><FONT= =20 face=3DArial=20 size=3D2>news:1130430253.588590.265420@g43g2000cwa.googlegroups.com</FONT=></A><FONT=20face=3DArial size=3D2>...</FONT></DIV> <DIV><FONT face=3DArial size=3D2>> <BR>> I don't see your point = about OFDM=20 ??<BR>> <BR>> If you have 1 million carriers each 1 Hz wide, then = each=20 carrier can<BR>> support a symbol rate of 1/2 sym/s. The total=20 aggregate symbol rate is<BR>> still 500ksym/s just like QAM?<BR>> = <BR>>=20 Mark<BR>></FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Hi Mark,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>If the symbol rate, 1/T, on each of the = subcariers=20 in OFDM is equal to the subcarrier separation, they are = orthogonal=20 over the symbol interval T. Thus if you were to place 1 million carriers = in=20 1MHz, they would be spaced at 1Hz and their symbol rate would be=20 1symbol/s. I'm not sure where you got 1/2=20 symbols/s.</FONT></DIV></BODY></HTML> ------=_NextPart_000_000E_01C5DB34.C68E6410--
