Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarApr2017

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search

Messages from 30875

Article: 30875
Subject: Re: High resolution time measurement?
From: "Victor Schutte" <victors@mweb.co.za>
Date: Wed, 2 May 2001 09:55:59 +0200
Links: << >>  << T >>  << A >>
What do you want measure seconds, ms, us? This will determine the number of
bits in your counter. A 16 bit counter will only count from 0 to 65535.
Using a 10MHz clock you will get a range of about 6.5ms.  A 32bit counter at
10MHz will give about 430 seconds window, but will chew up more gates.

More accurate than 1/F?  You will never get better than 1/F accuracy.  To
get better than 1/F accuracy you will need multiple 1/Fs which will mean a
higher reference clock.  If you need better than 1/10E6 you can use a higher
clock, such as 50MHz. This will give a 1/50E6 which is 5 times better than a
10MHz clock.

FPGA? I don't think so.  Rather use a CPLD. One flipflop for start and stop
(counter enable signal), nbit counter with clear and registered outputs. The
clock freq. can be much higher than a FPGA. It is also cheaper and easier to
use than a FPGA. Even the slowest 10ns CPLDs should do the trick. Look at
the Altera (don't actually know much of the Xilinx products) 7000MAX series.
For jobs like these I would recommend the MAX7128SLC84 -10. It is fast
enough, contains a decent amount of gates, is in-circuit programmable and
comes in an easy to use PLCC package(plus you can get free software over the
internet (www.altera.com )and build your own downloadcable for the price of
a hamburger).




"Jonas Thor" <thor.NO@SPAM.sm.luth.se> wrote in message
news:p1dret0hssttb4uf9meii2t0998n35b1f3@4ax.com...
> Hello!
>
> This is a not well defined question, but I'll ask anyway... I want to
> measure, with high precision, the time between two rising edges of two
> pulses. I have a reference clock, frequency F, of about 5-10 MHz, but
> I need measurements much more accurate than 1/F. (The rising edges are
> of course asynchronous to the reference clock.)
>
> What can I do in a FPGA get the best precison? What's the best I can
> do without a DLL and with a DLL?
>
> Thanks for any advice!
>
> / Jonas Thor



Article: 30876
Subject: Re: ccd imaging with fpga
From: "Victor Schutte" <victors@mweb.co.za>
Date: Wed, 2 May 2001 10:02:44 +0200
Links: << >>  << T >>  << A >>
About a year ago I read an article on a new image device that looks like RAM
(manufactured by HP). You don't need to design high speed CCD circuitry and
device drivers for Windoze.  You can use something as simple as a 8051 uC,
scan the device and then send the data to the PC via RS232.

Unfortunately I do not know the part number or the price (which can probably
be very high). I know the part exists. If anyone has more details on this
please post.




"M.B." <martinb@magma.ca> wrote in message
news:3aeeb905.4091345@news.magma.ca...
> Hi I have a telescope and some basic experiance with VHDL, FPGAs.
> I would like to get a CCD (something common and cheap) and use a FPGA
> perhaps a xilinx or atmel with VHDL. and interface to my computer (win
> 98.
>
> Has anyone done somthing like this?
> Is there a project in any books or wed sites that you have seen that
> could get me started?
>
> Or is there a development kit from one of the CCD Manufacturers that I
> could look into?
>
>



Article: 30877
Subject: Newbie
From: "Noddy" <g9731642@campus.ru.ac.za>
Date: Wed, 2 May 2001 10:15:01 +0200
Links: << >>  << T >>  << A >>
Hi,

I've just started teaching myself how to program FPGA's...

What is the difference between a LogiBLOX module, and a Core module?

Thanks
adrian




Article: 30878
Subject: Re: USB CORE IN VHDL
From: Nicolas Matringe <nicolas.matringe@IPricot.com>
Date: Wed, 02 May 2001 10:28:17 +0200
Links: << >>  << T >>  << A >>
gr6@ukc.ac.uk wrote:
> 
> HELLO,
> I AM LOOKING FOR A FREE USB (v.2.0) CORE WRITTEN IN VHDL.
> REGARDS
> GR
> 

Have a look at opencores:
http://www.opencores.org/cores/usb/

-- 
Nicolas MATRINGE           IPricot European Headquarters
Conception electronique    10-12 Avenue de Verdun
Tel +33 1 46 52 53 11      F-92250 LA GARENNE-COLOMBES - FRANCE
Fax +33 1 46 52 53 01      http://www.IPricot.com/

Article: 30879
Subject: FPGA SPROM boot problem
From: oliver.amft@ch.abb.com.nospam
Date: Wed, 02 May 2001 10:32:48 +0200
Links: << >>  << T >>  << A >>
Xilinx FPGA (XC2S100-FG256) load via JTAG _and_ FPGA cable works fine,
but SPROM (XC18V01) boot doesn't. I've checked all signals (DIN, CCLK,
/INIT, /PROG): CCLK is glitch-free, DIN has DATA (from SPROM). After
about 70% of the bits /INIT went low; DONE is stays low. (This might be
the first CRC check.)

I've verified pullups on DONE, /PROG, /INIT and GND on /CS and /WRITE.
I've would like to try the bitgen "-g DebugBitstream:Yes" option to see
if the FPGA syncs but the unfortunately the SPROM is too small. I'm even
not sure if this makes any sense at all to check the sync. Beside this,
has anyone an idea what else might go wrong there?
Oliver


Article: 30880
Subject: Re: ccd imaging with fpga
From: Michael Strothjohann <strothjohann@rheinahrcampus.de>
Date: Wed, 02 May 2001 12:35:31 +0100
Links: << >>  << T >>  << A >>
Probably the device is like the one in an
old BYTE-article ( over one decade ago ):
an DRAM was opened and the
precharged C's are decharged by the incoming
photons. The picture reflects the chip-structure,
but the low-cost ('home-made') device realy works.
michael strothjohann


Victor Schutte schrieb:

> About a year ago I read an article on a new image device that looks like RAM
> (manufactured by HP). You don't need to design high speed CCD circuitry and
> device drivers for Windoze.  You can use something as simple as a 8051 uC,
> scan the device and then send the data to the PC via RS232.
>
> Unfortunately I do not know the part number or the price (which can probably
> be very high). I know the part exists. If anyone has more details on this
> please post.
>
> "M.B." <martinb@magma.ca> wrote in message
> news:3aeeb905.4091345@news.magma.ca...
> > Hi I have a telescope and some basic experiance with VHDL, FPGAs.
> > I would like to get a CCD (something common and cheap) and use a FPGA
> > perhaps a xilinx or atmel with VHDL. and interface to my computer (win
> > 98.
> >
> > Has anyone done somthing like this?
> > Is there a project in any books or wed sites that you have seen that
> > could get me started?
> >
> > Or is there a development kit from one of the CCD Manufacturers that I
> > could look into?
> >
> >


Article: 30881
Subject: failed to configure virtex
From: rebai chiheb <rebai@ixl.u-bordeaux.fr>
Date: Wed, 02 May 2001 13:46:32 +0200
Links: << >>  << T >>  << A >>
Hi,

I have some problems in the configuration of virtex FPGA :

I try to configure my xc300 FPGA in master serial mode with a 18v02
serial PROM  in AFX-PQ-240-100 prototype board.
I use the xilinx fundation software to generate the mcs file foe EPROM.
The bitsteam is dowloaded in the FPGA, but no result appear !!!
What happened ???

I'm completely helpless!

Can anybody help me

Chiheb REBAI.
IXL laboratory
University of Bordeaux
FRANCE.



Article: 30882
Subject: Re: VHDL coding question.
From: "auliya" <murrayal@yahoo.com>
Date: Wed, 2 May 2001 22:38:40 +1000
Links: << >>  << T >>  << A >>
Hi,

With this you have 2 options:
1. Use a generic to pass in a parameter which will specify which of B1 or B2
to use.  In this case you need to use 'if...generate' statements to
instantiate each option.  You should the B1 and B2 entity/architecture pairs
in separate files and use B as a 'wrapper' to choose which one to use.

2. Use a configuration declaration to specift which of B1 or B2 to use.  For
this you need an entity declaration (B) and the two architectures (B1 and
B2).  Basically, a configuration allows you to bind a specific architecture
(out of many options) to an entity.

Bear in mind that configurations may not be fully supported by your VHDL
simulator, and by your synthesis tools....this leaves passing a generic as
probably the way to go

cheers,
Chris

"mok" <mok_3001@yahoo.com> wrote in message
news:9c8v5f$hr2$1@news.qub.ac.uk...
> Hi Folks,
>
> I would like to describe the following in VHDL:
>
>       -----        -----
>      |       |      |       |
> -->|  B  |--->|  C   | --->
>      |      |       |       |
>       -----        -----
>
> where B can be either B1 or B2.
> I have VHDL code for C, B1 and B2. B1 and B2 describe two different logic
> blocks.
> I would like to write one VHDL file which can describe both cases,
> and choose between B1 and B2 by testing a 'generic' constant or something
> like
> that.
> In short I do not want to write two seperate file for each case,
> but one file which can describe both according to a particular
> input data (either B1 or B2).
> First, is using a 'generic' the way to go?if yes, How can it be
> done? if not, how could i reach my goal then? Is there any other
> way?
>
> Cheers.
>
> PS. I have not got much VHDL knowledge.
>
>



Article: 30883
Subject: Re: Multiple state machines in altera AHDL
From: Greg Deuerling <egads@fnal.gov>
Date: Wed, 2 May 2001 07:51:34 -0500
Links: << >>  << T >>  << A >>
In article <3AEEE5FD.30C70EA1@sqf.hp.com>, nials@sqf.hp.com says...
> Greg Deuerling wrote:
> 
> > I plan to use Foundation and a SpartanII comming up.  Altera really
> > lost me as a high end FPGA customer when they put out that load of
> > crap called Quartus.  They should have left MaxPlusII alone, it was
> > simple, fast, and easy to use.  I'll get used to using VHDL, but I'll
> > NEVER like it, and wish I was using AHDL.
> 
> 
> Greg,
> 
> You must have had one of the early releases of Quartus, I think it's
> much more stable now. I haven't used it for about a year now, but
> even then I had no problems (it was on an NT box).
> 
> The Xilinx tools are a bit more mature, but I'm sure they'd problems
> initially.

Nope, I'm running QuartusII.  I can pretty much crash Quartus at will.
All you have to do is switch between compile mode to simulation mode
several times and sooner or later it will crash.  Altera had a VHDL
class here and I was complaining how slow and buggy Quartus was and
they did not beleive me.  I crashed 4 of their computers in several
minutes with out breaking a sweat.  When I started using Foundation
it was also very slow and buggy, but now its pretty nice and a whole
lot better than Quartus.  I'll stick to MaxPlusII for small designs
and Foundation and Spartan's for big designs...

By the way, I use Win2000...

-- 
                                                                     gad
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
=          Greg Deuerling, Fermi National Accelerator Laboratory          =
= P.O.Box 500 MS368  Batavia, IL 60510  (630)840-4629, FAX  (630)840-5406 =
=                  Electronic Systems Engineering Group                   =
=            Work: egads@fnal.gov       Personal: gad@elnet.com           =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Article: 30884
Subject: Re: failed to configure virtex
From: Oliver Amft <oliver.amft@nospam.ch.abb.com>
Date: Wed, 02 May 2001 15:36:41 +0200
Links: << >>  << T >>  << A >>
Well, thats about the same situation as I have (see FPGA SPROM boot problem
). In my case direct JTAG and FPGA load works fine. Have you checked your
ConfigClk:CCLK for bitgen? Does /INIT stays asserted for the entire loading
time?
Cheers Oliver


rebai chiheb wrote:

> Hi,
>
> I have some problems in the configuration of virtex FPGA :
>
> I try to configure my xc300 FPGA in master serial mode with a 18v02
> serial PROM  in AFX-PQ-240-100 prototype board.
> I use the xilinx fundation software to generate the mcs file foe EPROM.
> The bitsteam is dowloaded in the FPGA, but no result appear !!!
> What happened ???
>
> I'm completely helpless!
>
> Can anybody help me
>
> Chiheb REBAI.
> IXL laboratory
> University of Bordeaux
> FRANCE.


Article: 30885
Subject: Final Program CHES 2001
From: Christof Paar <christof@ece.wpi.edu>
Date: Wed, 2 May 2001 09:41:58 -0400
Links: << >>  << T >>  << A >>
The Final Program for CHES 2001 in Paris can be found at

   http://www.chesworkshop.org

For registration and hotel information, please also check the web site
above.

Regards,

Christof

! WORKSHOP ON CRYPTOGRAPHIC HARDWARE AND EMBEDDED SYSTEMS (CHES 2001) !
!                  Paris, France, May 13-16, 2001                     !
!                       www.chesworkshop.org                          !

***********************************************************************
                 Christof Paar,  Associate Professor
          Cryptography and Information Security (CRIS) Group
      ECE Dept., WPI, 100 Institute Rd., Worcester, MA 01609, USA
fon:(508) 831 5061  email: christof@ece.wpi.edu
fax:(508) 831 5491  www: http://www.ece.wpi.edu/People/faculty/cxp.html
***********************************************************************



Article: 30886
Subject: Re: Shannon Capacity
From: nemo@dtgnet.com (Nemo)
Date: Wed, 02 May 2001 13:47:50 GMT
Links: << >>  << T >>  << A >>
You may want to get ahold of Shannon's paper, it explains the details better
than I can.   C refers the the capacity of a physical band-limited channel.  By
definition, C includes *all* the data carried by the channel. This includes the
payload data and any overhead (parity bytes, etc) needed for the FEC layer if
used.  In-line comments below...

"Kevin Neilson" <kevin_neilson@yahoo.com> wrote:

>There's something I've never quite understood about Shannon's Law.  It
>states something like this:
>
>C = W*logbase2(1 + S/N)
>
>where C is channel capacity, bits/s
>    W = bandwidth, Hz
>    S/N = signal-to-noise ratio
>
>Theoretically, if using enough error-correction, you can transmit C bits/s
>on a channel error-free.

Yes, exactly as it says -  bits are bits, and parity bits are information too.
>
>Say you determine that the channel capacity for a particular channel is
>1Mbit/s.  Does that include the error correction bits or not?  

Yes it includes error correction.

>What if you
>have a 1/2 rate code, where half of the bits are parity bits.  Does that
>mean you transmit 2Mbit on that channel (with 1Mbit capacity) for a net data
>throughput of 1Mbit error-free?  

No, if the Shannon limit is 1Mbps, then thats all you get *including* parity.

>Or does the 1Mbit include the partiy bits?
Yes.

>
>-Kevin
>


Hope that helps.

Nemo

Article: 30887
Subject: Re: failed to configure virtex
From: Tom Fischaber <Tom.Fischaber@xilinx.com>
Date: Wed, 02 May 2001 08:01:46 -0600
Links: << >>  << T >>  << A >>
A couple of things to check would be to monitor to see if DONE goes high,
or INIT goes low. Also, do you see CCLK toggling at all? Is PROG being
released properly? I would recommend walking through the Xilinx
Configuration Problem Solver, it might help you get closer to the root of
the problem.

http://service.xilinx.com/support/cgi-bin/webcgi.exe?New,KB=config

Cheers,
Tom

rebai chiheb wrote:

> Hi,
>
> I have some problems in the configuration of virtex FPGA :
>
> I try to configure my xc300 FPGA in master serial mode with a 18v02
> serial PROM  in AFX-PQ-240-100 prototype board.
> I use the xilinx fundation software to generate the mcs file foe EPROM.
> The bitsteam is dowloaded in the FPGA, but no result appear !!!
> What happened ???
>
> I'm completely helpless!
>
> Can anybody help me
>
> Chiheb REBAI.
> IXL laboratory
> University of Bordeaux
> FRANCE.


Article: 30888
Subject: Re: VHDL coding question.
From: steve (Steve Rencontre)
Date: Wed, 2 May 2001 15:36 +0100 (BST)
Links: << >>  << T >>  << A >>
In article <X1TH6.16348$482.81848@newsfeeds.bigpond.com>, 
murrayal@yahoo.com (auliya) wrote:

> Hi,
> 
> With this you have 2 options:
> 1. Use a generic to pass in a parameter which will specify which of B1 
> or B2
> to use.  In this case you need to use 'if...generate' statements to
> instantiate each option.  You should the B1 and B2 entity/architecture 
> pairs
> in separate files and use B as a 'wrapper' to choose which one to use.
> 
> 2. Use a configuration declaration to specift which of B1 or B2 to use. 
>  For
> this you need an entity declaration (B) and the two architectures (B1 
> and
> B2).  Basically, a configuration allows you to bind a specific 
> architecture
> (out of many options) to an entity.

3. Run it through a C preprocessor first :-)

That's not an entirely frivolous suggestion, either. For all its faults, 
the C preprocessor lets you do useful things that can't really be done in 
any other way, and I miss it badly in other languages.

--
Steve Rencontre		http://www.rsn-tech.co.uk
//#include <disclaimer.h>


Article: 30889
Subject: Re: Shannon Capacity
From: steve (Steve Rencontre)
Date: Wed, 2 May 2001 15:36 +0100 (BST)
Links: << >>  << T >>  << A >>
In article <3AEF11CD.CEA6992D@xilinx.com>, austin.lesea@xilinx.com (Austin 
Lesea) wrote:

> See what I mean?
> 
> Still confused after all these years.
> 
> If the CRC is not information, then you don't need it.
> 
> If the CRC is information, then you need to send it.
> 
> I rest my case.
> 
> Austin

I disagree. The CRC (or whatever) is not information, but you do need to 
send it! The total message of N+CRC physical bit symbols only carries 
N bits of /information/, and it's that which Shannon's theorem applies to.

I think the confusion arises because we automatically think that a 1 or 0 
in the digital domain is a 'bit' in the Shannon sense, but it's not - it's 
just a symbol which can be used to encode bits. Adding a CRC increases the 
predictability of lost symbols, which means each 0/1 symbol carries less 
than one actual bit.

Shannon tells us how many bits of genuine information we can transmit, but 
we're free to send as many 
symbols-which-superficially-look-like-bits-but-aren't as we like :-)

--
Steve Rencontre		http://www.rsn-tech.co.uk
//#include <disclaimer.h>


Article: 30890
Subject: Re: Shannon Capacity
From: steve (Steve Rencontre)
Date: Wed, 2 May 2001 15:36 +0100 (BST)
Links: << >>  << T >>  << A >>
In article <3af20db4.2149921@news.dtgnet.com>, nemo@dtgnet.com (Nemo) 
wrote:

> C refers the the capacity of a physical band-limited 
> channel.  By definition, C includes *all* the data carried by the
> channel. This includes the payload data and any overhead (parity bytes,
> etc) needed for the FEC layer if used.

Sorry, but that doesn't add up. Shannon's Theorem says that you can carry 
C bps /without/ error. In that case, why have you included error detection 
and correction bits with the payload data?!

--
Steve Rencontre		http://www.rsn-tech.co.uk
//#include <disclaimer.h>


Article: 30891
Subject: Looking for a prototyping board
From: "Rotem Gazit" <rotemg@mysticom.com>
Date: Wed, 2 May 2001 18:27:56 +0300
Links: << >>  << T >>  << A >>
I am looking for a prototyping board with the following characterizations:

1) TI 5408 or 548 or 5409 or 549 DSP.

2) Xilinx Virtex or Spartan-II FPGA.

(Didn't find any in Optimagic or Xilinx WEB sites).

Thanks for your help,

Rotem.






Article: 30892
Subject: Re: Shannon Capacity
From: Austin Lesea <austin.lesea@xilinx.com>
Date: Wed, 02 May 2001 08:29:01 -0700
Links: << >>  << T >>  << A >>
Wrong.

Go back and re-read his paper.

This is really funny, and sad.

Austin

Steve Rencontre wrote:

> In article <3AEF11CD.CEA6992D@xilinx.com>, austin.lesea@xilinx.com (Austin
> Lesea) wrote:
>
> > See what I mean?
> >
> > Still confused after all these years.
> >
> > If the CRC is not information, then you don't need it.
> >
> > If the CRC is information, then you need to send it.
> >
> > I rest my case.
> >
> > Austin
>
> I disagree. The CRC (or whatever) is not information, but you do need to
> send it! The total message of N+CRC physical bit symbols only carries
> N bits of /information/, and it's that which Shannon's theorem applies to.
>
> I think the confusion arises because we automatically think that a 1 or 0
> in the digital domain is a 'bit' in the Shannon sense, but it's not - it's
> just a symbol which can be used to encode bits. Adding a CRC increases the
> predictability of lost symbols, which means each 0/1 symbol carries less
> than one actual bit.
>
> Shannon tells us how many bits of genuine information we can transmit, but
> we're free to send as many
> symbols-which-superficially-look-like-bits-but-aren't as we like :-)
>
> --
> Steve Rencontre         http://www.rsn-tech.co.uk
> //#include <disclaimer.h>


Article: 30893
Subject: Re: Shannon Capacity
From: Austin Lesea <austin.lesea@xilinx.com>
Date: Wed, 02 May 2001 08:29:44 -0700
Links: << >>  << T >>  << A >>
Klaus,

Ahhh, finally, someone who understands!

Austin

Klaus Falser wrote:

> On Mon, 30 Apr 2001 17:03:09 GMT, "Kevin Neilson"
> <kevin_neilson@yahoo.com> wrote:
>
> >There's something I've never quite understood about Shannon's Law.  It
> >states something like this:
> >
> >C = W*logbase2(1 + S/N)
> >
> >where C is channel capacity, bits/s
> >    W = bandwidth, Hz
> >    S/N = signal-to-noise ratio
> >
> >Theoretically, if using enough error-correction, you can transmit C bits/s
> >on a channel error-free.
> >
> >Say you determine that the channel capacity for a particular channel is
> >1Mbit/s.  Does that include the error correction bits or not?  What if you
>
> No, the 1 MBit/s its the rate for usable information.
>
> >have a 1/2 rate code, where half of the bits are parity bits.  Does that
> >mean you transmit 2Mbit on that channel (with 1Mbit capacity) for a net data
> >throughput of 1Mbit error-free?  Or does the 1Mbit include the partiy bits?
> >
> >-Kevin
> >
>
> You must not think in terms of parity and CRC, Shannons law applies to
> a much more generic principle.
> It states, that given a noisy channel (where the signals at the end
> of the channel are not in their ideal form anymore, but there is some
> added noise), you are not able to exceed the information rate C.
>
> The used information coding scheme in the channel it's at your
> decision and you can and will use any trick you want.
> (AM, FM, QA[Modulation], Viterbi coding ....).
>
> Best regards.
> Klaus
>
> Falser Klaus
> R&D Electronics Department
> Company : Durst Phototechnik AG
>         Vittorio Veneto Str. 59
>         I-39042 Brixen
> Voice   : +0472/810235
>         : +0472/810111
> FAX     : +0472/830980
> Email   : kfalser@IHATESPAMdurst.it


Article: 30894
Subject: Re: Shannon Capacity
From: Austin Lesea <austin.lesea@xilinx.com>
Date: Wed, 02 May 2001 08:30:57 -0700
Links: << >>  << T >>  << A >>
Hey!

Thats two (besides me) who understand...getting better.

Austin

Nemo wrote:

> You may want to get ahold of Shannon's paper, it explains the details better
> than I can.   C refers the the capacity of a physical band-limited channel.  By
> definition, C includes *all* the data carried by the channel. This includes the
> payload data and any overhead (parity bytes, etc) needed for the FEC layer if
> used.  In-line comments below...
>
> "Kevin Neilson" <kevin_neilson@yahoo.com> wrote:
>
> >There's something I've never quite understood about Shannon's Law.  It
> >states something like this:
> >
> >C = W*logbase2(1 + S/N)
> >
> >where C is channel capacity, bits/s
> >    W = bandwidth, Hz
> >    S/N = signal-to-noise ratio
> >
> >Theoretically, if using enough error-correction, you can transmit C bits/s
> >on a channel error-free.
>
> Yes, exactly as it says -  bits are bits, and parity bits are information too.
> >
> >Say you determine that the channel capacity for a particular channel is
> >1Mbit/s.  Does that include the error correction bits or not?
>
> Yes it includes error correction.
>
> >What if you
> >have a 1/2 rate code, where half of the bits are parity bits.  Does that
> >mean you transmit 2Mbit on that channel (with 1Mbit capacity) for a net data
> >throughput of 1Mbit error-free?
>
> No, if the Shannon limit is 1Mbps, then thats all you get *including* parity.
>
> >Or does the 1Mbit include the partiy bits?
> Yes.
>
> >
> >-Kevin
> >
>
> Hope that helps.
>
> Nemo


Article: 30895
Subject: Re: Shannon Capacity
From: Kolja Sulimma <kolja@prowokulta.org>
Date: Wed, 02 May 2001 17:55:16 +0200
Links: << >>  << T >>  << A >>
>
> If the CRC is not information, then you don't need it.
> If the CRC is information, then you need to send it.

Shannon states a theoretical bound. You can not expect to approach it with a CRC.

And you will not approach it with superior schemes like reed solomon or expander
codes.

I believe that this line contains the source of the misconceptions in this
thread:
> > > Theoretically, if using enough error-correction, you can transmit C bits/s

Not "enough" error-correction, but a "good enough" error correction.
This means an error correction scheme that introduces the least amount of
overhead for a given error rate.

Shannon states, that there will never be an error correction scheme that can
transport more than C usable data bits over a channel that can transport W bits
if there is no noise present.

> > > C = W*logbase2(1 + S/N)
> > > where C is channel capacity, bits/s
> > >     W = bandwidth, Hz
> > >     S/N = signal-to-noise ratio

Kolja Sulimma


Article: 30896
Subject: Re: High resolution time measurement?
From: Falk Brunner <Falk.Brunner@gmx.de>
Date: Wed, 02 May 2001 18:50:45 +0200
Links: << >>  << T >>  << A >>
Victor Schutte schrieb:
> 
> 
> FPGA? I don't think so.  Rather use a CPLD. One flipflop for start and stop
> (counter enable signal), nbit counter with clear and registered outputs. The
> clock freq. can be much higher than a FPGA. It is also cheaper and easier to

??? Have a look at todays FPGAs. Virtex-E reaches 300 MHz and its cheap.
Virtex II even reaches 420MHz.
With the clock doubling (can be cascaded to 8x clock multiplication) in
Virte-(E) you can apply a low speed clock outside the FPGA and have a
nice high speed clock inside the FPGA. With Virtex II you even get a
frwquncy synthesiser (DCM) with much more options.

-- 
MFG
Falk



Article: 30897
Subject: Re: Serial UART
From: "Kevin Neilson" <kevin_neilson@yahoo.com>
Date: Wed, 02 May 2001 17:00:51 GMT
Links: << >>  << T >>  << A >>
Insight Design Services has Verilog source that functions similarly to the
Mot MC145428, which is a serially-interfaced UART.  Not free, though.

"Mark Walter" <maw@nospam.com> wrote in message
news:1ALH6.226933$GV2.47295492@typhoon.san.rr.com...
> Does anyone know of where I can find a free Verilog source code for a
Serial
> UART that can be used in a Xilinx XC4005XL FPGA? I am in need of such code
> to add to a FPGA design I am making for remote access to the FPGA CPU...
>
> Thanks,
> Mark
>
>
>



Article: 30898
Subject: Re: Serial UART
From: Kolja Sulimma <kolja@prowokulta.org>
Date: Wed, 02 May 2001 20:00:23 +0200
Links: << >>  << T >>  << A >>
I had two students build a fixed frequency UART as a one week homework.
Including 16 times oversampling.

If you do not need any fancy functionality, just build one yourself.

Or you can download the code of my students.
http://bounty.em.informatik.uni-frankfurt.de/~prak/ws0001/projekte/jan/index.html

(page in german, download at the bottom)

It not very good code, it had better students then them, but the UART work in
its first version.

Kolja Sulimma

Kevin Neilson wrote:

> Insight Design Services has Verilog source that functions similarly to the
> Mot MC145428, which is a serially-interfaced UART.  Not free, though.


Article: 30899
Subject: Re: Newbie
From: Tim Jaynes <tim.jaynes@xilinx.com>
Date: Wed, 02 May 2001 11:03:41 -0700
Links: << >>  << T >>  << A >>
At a certain level of abstraction, LogiBLOX and Coregen modules are
essentially the same thing- pieces of customizable ip that a FPGA
designer can drop into their design who's implementation details are not
visible to the user.
Their differences are the following:
-delivery method (Coregen gives you edif cores, LogiBLOX delivers ngc
files- binary netlists).
-Parts supported (LogiBLOX only supports up through the 4000 family and
derivatives).
-Interface method (Coregen uses a slick java interface- Logiblox a
simple GUI w/ pull-down menus etc.)

One other thing to note is that, going forward, Coregen will continue to
be a part of the Xilinx toolset and LogiBLOX will be phased out.
Hope this helps,
Tim Jaynes
CAE



Noddy wrote:

> Hi,
>
> I've just started teaching myself how to program FPGA's...
>
> What is the difference between a LogiBLOX module, and a Core module?
>
> Thanks
> adrian




Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarApr2017

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search