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
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

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 67275

Article: 67275
Subject: Re: Release asynchrounous resets synchronously
From: Peter Alfke <peter@xilinx.com>
Date: Tue, 09 Mar 2004 09:59:03 -0800
Links: << >>  << T >>  << A >>
Silly me. I made a mistake. ;-(
Should have known better, will come up with another solution.
Peter Alfke


> From: TommyInTheNews@numba-tu.com
> Newsgroups: comp.arch.fpga
> Date: Tue, 09 Mar 2004 17:21:56 GMT
> Subject: Re: Release asynchrounous resets synchronously
> 
> Peter Alfke wrote:
>> Eh, what? Unfortunately anonymos...
> 
> Sorry, Tommy Thorn, tommy at numba-tu dot com.
> 
>> Each first level LUTs detects (output High) that exactly one of its inputs
>> is High.
>> The second tier LUT detects that exactly one of the first-tier LUT outputs
>> is high, which mans that there is exactly one High input.
>> Agreed ?
> 
> Did you actually read what I wrote?  It doesn't work.
> 
> Take for example 16'b1111_1110_1100_0001, only one of the nibbles has
> exactly one bit high, so the first tier LUTs goes 0 0 0 1, which has
> exactly one bit set and your scheme would say that's a valid state.
> 
> You need to distinguish 0, 1, and more, thus use roughly twice the LUTs.
> 
> Thanks,
> 
> Tommy
> 
> 


Article: 67276
Subject: Re: bit stream file examples ?
From: mike_treseler@comcast.net (Mike Treseler)
Date: 9 Mar 2004 10:15:12 -0800
Links: << >>  << T >>  << A >>
"Jack Moderatz" <jyhur@wanadoo.nl> wrote:

> Currently I need binary bit stream file(configuration file) examples (any
> applications are okay) for analysis.........Can somebody tell me how I can
> obtain them?

Connect a logic analyzer and boot the board.

   -- Mike Treseler

Article: 67277
Subject: Re: Can `protect-ed Verilog codes be synthesized with Xilinx XST?
From: Bassman59a@yahoo.com (Andy Peters)
Date: 9 Mar 2004 10:54:44 -0800
Links: << >>  << T >>  << A >>
"Kelvin @ SG" <kelvin8157@hotmail.com> wrote in message news:<404d756d@news.starhub.net.sg>...
> well, maybe that is why i couldnt simulate the Samsung SmartMedia model with
> Modelsim...
> How do I find out which is the right simulator?

Call Samsung?

--a

Article: 67278
Subject: Re: sorting need help as soon as possible
From: Ray Andraka <ray@andraka.com>
Date: Tue, 09 Mar 2004 13:59:58 -0500
Links: << >>  << T >>  << A >>
SRL16's are great for small reordering tasks too.

John_H wrote:

> konerusreeram@yahoo.com (sree) wrote in message news:<195eee74.0403090047.1e494af9@posting.google.com>...
> > Hi,
> > i am new to FPGA's and vhdl.i need to design a sorter of N numbers 32
> > bit each.i know sorting algorithms bubble sort.But i found that in
> > fpga's we can do bit comparision parallel and there are fast sort
> > methods i can use.i saw a couple but not sure how to implement on
> > virtex-II xc2v1000 which has a 100MHz clock.If some body have an
> > hardware discrption or digital schematic please let me know.i
> > appreciate their help
> > thank you
> > sree
>
> Do you need all 32 values at the same time such that you're processing
> N values on *every* clock expecting the sorted values with a few clock
> cycles latency, a new sorted list on every clock?
>
> Is your time less critical?  Do you need one value at a time from
> highest to lowest or vice-versa?  Do you just need the value or do you
> need to know which entry it was in the original list?
>
> As a for-instance, the bitonic sort will give a great parallel sort
> algorithm providing a list of N sorted elements every clock but with a
> huge amount of resources to accommodate the parallelism.
>
> BlockRAMs are great for doing manipulation of the large numbers for
> comparisons between 2 entries every clock cycle making them handy for
> some sort algorithms.
>
> If you read the values in one at a time, some of the sorting can be
> performed as you build up your array rather than waiting until it's
> fully in place.
>
> So - state your needs.

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 67279
Subject: Re: sorting need help as soon as possible
From: mike_treseler@comcast.net (Mike Treseler)
Date: 9 Mar 2004 11:01:59 -0800
Links: << >>  << T >>  << A >>
konerusreeram@yahoo.com (sree) wrote:

> i am new to FPGA's and vhdl.i need to design a sorter of N numbers 32
> bit each.i know sorting algorithms bubble sort.

Well, that's more fun than a vending machine.

A vhdl prototype might include:
An array of constants for input,
An array of variables for the sort process.
A rollover counter variable to loop through the array until done.
A procedure to swap array values-- called once per clock.
A done boolean to remember if no swaps were need on the last pass.


 --Mike Treseler

Article: 67280
Subject: Re: sorting need help as soon as possible
From: Eric Crabill <eric.crabill@xilinx.com>
Date: Tue, 09 Mar 2004 11:02:45 -0800
Links: << >>  << T >>  << A >>

Hi,

Go read http://www.engr.sjsu.edu/crabill/module06.pdf

I've got an example of a bubble sort, written in Verilog.
This was meant to illustrate how you can take sequential
algorithms from a CS text and implement them in hardware.
And get lackluster results.

Then, there is an example of an odd-even transposition
sort, which is a parallel algorithm for sorting.  The
result is compared to the bubble sort in terms of fmax.

Hope that helps,
Eric

sree wrote:
> 
> Hi,
> i am new to FPGA's and vhdl.i need to design a sorter of N numbers 32
> bit each.i know sorting algorithms bubble sort.But i found that in
> fpga's we can do bit comparision parallel and there are fast sort
> methods i can use.i saw a couple but not sure how to implement on
> virtex-II xc2v1000 which has a 100MHz clock.If some body have an
> hardware discrption or digital schematic please let me know.i
> appreciate their help
> thank you
> sree

Article: 67281
Subject: Re: Release asynchrounous resets synchronously
From: Ray Andraka <ray@andraka.com>
Date: Tue, 09 Mar 2004 14:04:12 -0500
Links: << >>  << T >>  << A >>
Peter,

Consider the case where you have 1 input high to one of your luts, and 2 inputs
high to another lut.  It is clearly an illegal state, as it has two extra '1' bits,
but it will be detected as OK by your circuit because exactly one LUT is indicating
one input on.  Each group of four requires two outputs to distinguish 0,1 or more
than 1 input on.

Peter Alfke wrote:

> Eh, what? Unfortunately anonymos...
>
> Each first level LUTs detects (output High) that exactly one of its inputs
> is High.
> The second tier LUT detects that exactly one of the first-tier LUT outputs
> is high, which mans that there is exactly one High input.
> Agreed ?
> Peter Alfke
>
> > From: user@domain.invalid
> > Newsgroups: comp.arch.fpga
> > Date: Tue, 09 Mar 2004 05:24:26 GMT
> > Subject: Re: Release asynchrounous resets synchronously
> >
> > Peter Alfke wrote:
> >> LUTs are very efficient "illegal state" detectors.
> >> Let's say you have a 16-state one-hot machine. Four LUTs can each detect
> >> "exactly one of my inputs is High", and a fifth LUT does the same with the
> >> four LUT outputs. So 5 LUTs can detect any illegitimate 16-bit code. Take it
> >> from there...
> >
> > Eh, what?  So the first tier LUT compute f, where
> > f(a,b,c,d) = 1 iff a+b+c+d = 1, else 0.
> > For the 5th LUT we have the same property that a legal 16-state would
> > map exactly one of the four first tier LUTs to 1, thus it sounds like
> > what you have in mind is something like this:
> >
> > f({f(s[3:0]), f(s[7:4]), f(s[11:8]), f(s[15:12])})
> >
> > but this could accept states like 16'b1111_1110_1100_0001.
> >
> > I don't see how you can detect legal states with only five four-input LUTs.
> >
> >
> > Peter, the FPGA reset question has come many times.  What does Xilinx
> > recommend in general?  Async-reset+Sync-release, all-sync, or all-async?
> > Which uses fewest resources?
> >
> > Thanks,
> >
> > Tommy
> >

--
--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930     Fax 401/884-7950
email ray@andraka.com
http://www.andraka.com

 "They that give up essential liberty to obtain a little
  temporary safety deserve neither liberty nor safety."
                                          -Benjamin Franklin, 1759



Article: 67282
Subject: Re: copy protection on FPGA using embedded serial number
From: gregs@altera.com (Greg Steinke)
Date: 9 Mar 2004 11:13:44 -0800
Links: << >>  << T >>  << A >>
"Kenneth Land" <kland1@neuralog1.com1> wrote in message news:<104rmkoja30hm04@news.supernews.com>...
> Altera has a simple "Security Bit" that is supposed to keep other
> programmers from being able to read your config from either the FPGA or its
> config device.  I have no idea how secure this actually is in practice.  I
> guess it might at least throw up a hurdle to to some?
> 
> Ken
> 
> "millim" <merlin1974@gmx.at> wrote in message
> news:8e53c613.0403080739.aa1236b@posting.google.com...
> > I am trying to find an efficient and simple way to protect
> > prototypes from cloning.
> >
> > One idea relies on using the serial number of the FPGA device,
> > because in any prototyping phase only a hand full devices are
> > present. But is there an embedded hardware register in the FPGA device,
> > which contains a unique serial number?
> >
> > regards

Over the years a few customers have claimed that they could break the
security bit on MAX devices. We tested them by giving them secured
devices and asked them to give us back the programming file. None were
successful.

For the specific question of an embedded serial number -
Altera has just introduced a new line of CPLD called MAX II. This
device family has a User Flash Memory of up to 8 kilobits, enough to
store serial numbers and plenty more.

For details on the family and the User Flash Memory, check out the
technical literature on:
http://www.altera.com/literature/lit-max2.jsp

Sincerely,
Greg Steinke
gregs@altera.com
Altera Corporation

Article: 67283
Subject: Re: copy protection on FPGA using embedded serial number
From: nweaver@ribbit.CS.Berkeley.EDU (Nicholas C. Weaver)
Date: Tue, 9 Mar 2004 19:23:46 +0000 (UTC)
Links: << >>  << T >>  << A >>
In article <8e53c613.0403080739.aa1236b@posting.google.com>,
millim <merlin1974@gmx.at> wrote:
>I am trying to find an efficient and simple way to protect 
>prototypes from cloning.
>
>One idea relies on using the serial number of the FPGA device, 
>because in any prototyping phase only a hand full devices are
>present. But is there an embedded hardware register in the FPGA device,
>which contains a unique serial number?

If you are using a Xilinx, the V2/V2Pro bitfile encryption is a VERY
high barrier for an attacker to cross: I'd comfortably protect a $10K
secret, and uncomfortably protect a $100K secret from a malicious,
capable adversary.
-- 
Nicholas C. Weaver                                 nweaver@cs.berkeley.edu

Article: 67284
Subject: Re: Altera ACEX chip wide reset
From: gregs@altera.com (Greg Steinke)
Date: 9 Mar 2004 11:26:03 -0800
Links: << >>  << T >>  << A >>
rickman <spamgoeshere4@yahoo.com> wrote in message news:<4044BE39.A2074EB8@yahoo.com>...

> 
> This is exactly what I am talking about.  Even though the synthesis tool
> is capable of inverting a FF input and output in the logic equations so
> that a reset to zero operates as a preset to one, there has to be a way
> to tell the tool in your HDL that it needs to do this on a given FF.  
> 
> *That* is what I am asking about.  
> 
> -- 
> 
> Rick "rickman" Collins
> 
> rick.collins@XYarius.com

Rick,

The way to do this (set a certain register to appear to power up high,
on power up or on chip wide reset) is to use the POWER_UP_LEVEL=HIGH
assignment.
This creates a DFFE that will appear to power up to high after a
chip-wide reset. As we do not have hardware to implement this option
for core FFs on ACEX parts it will be implemented by inverting the
input and output to the DFF. The altera_attribute syntax will only
work in Quartus Integrated Synthesis. Synplify and Leonardo may have a
similar attribute; otherwise, you can make the POWER_UP_LEVEL
assignment from the Assignment Editor in the Quartus UI.

Here's a code example that shows how to do this:

LIBRARY ieee;
USE ieee.std_logic_1164.all;

LIBRARY wysiwyg;
USE wysiwyg.stratix_components.all;

ENTITY attr2 IS 
  PORT (
    data, clk, ena : IN std_logic;
    o : OUT std_logic
    );
END ;

ARCHITECTURE rtl OF attr2 IS

  SIGNAL tmp : STD_LOGIC;
  
  ATTRIBUTE ALTERA_INTERNAL_OPTION : string;
  ATTRIBUTE ALTERA_INTERNAL_OPTION OF tmp : SIGNAL IS "POWER_UP_LEVEL
= HIGH";

  
BEGIN
    

  PROCESS
  BEGIN
    WAIT UNTIL clk = '1';
    IF ena = '1' THEN
      tmp <= data;
    END IF;
  END process;

  o <= tmp;
  
END ;


There is no other good way of doing this in your VHDL: the chip-wide
reset is not specified as part of the VHDL code you write, it is
hardware that comes "free" with every FF.

Sincerely,
Greg Steinke
gregs@altera.com
Altera Corporation

Article: 67285
Subject: Re: Using ALTPLL
From: gregs@altera.com (Greg Steinke)
Date: 9 Mar 2004 11:34:10 -0800
Links: << >>  << T >>  << A >>
"SneakerNet" <nospam@nospam.org> wrote in message news:<Qea3c.6567$Nc3.108000@news.xtra.co.nz>...
> Hi Subroto
> 
> Thanks for the clarification.. This means I must have used my own code to
> generate the 12.5MHz. Oh well back to using the logic.
> Thanks again
> 
> Regards
> 
> "Subroto Datta" <sdatta@altera.com> wrote in message
> news:41a3c.22164$cY5.21375@newssvr31.news.prodigy.com...
> > Hi Sneakernet,
> >
> > You cannot implement their desired output frequency of 12.5 Mhz using the
> > Cyclone PLL without going out of spec.
> >
> > The Cyclone PLL does have a minimum output frequency limit, which is a
> > function of the PLL's minimum VCO frequency.  The minimum VCO frequency is
> > 491Mhz, and the maximum output divider counter size is 32, which results
>  in
> > a minimum output frequency of approximately 15.3 Mhz (i.e.491/32).
> >
> > Page 4-31 of the device handbook gives the PLL frequency specs.  The input
> > frequency has to be between 15.625 MHz and 464 MHz for the
> > fastest (-6) speed grade, and between 15.625 MHz and 387 MHz for the
>  slowest
> > (-8) speed grade.  See
> > http://www.altera.com/literature/hb/cyc/cyc_c51004.pdf  for details.
> >
> > - Subroto Datta
> > Altera Corp.
> >
> >
> > "SneakerNet" <nospam@nospam.org> wrote in message
> > news:Ow43c.6282$Nc3.101499@news.xtra.co.nz...
> > > Hello Ppl
> > >
> > > I have actually already asked this question before but I need help
>  again.
> > >
> > > I'm using a Nios Development Board (50MHz Oscillator, FPGA = Cyclone
> > > EP1C20F400C7). I'm using Quartus II v4.0 Build 190
> > > I need to generate a clock speed of 12.5MHz. While I have my own vhdl
>  code
> > > that can do this, I can't use the ALTPLL to do this for me.
> > >
> > > How can i do this? When I try to generate a speed of 12.5MHz using
> > > MegaFunction Wizard in Quartus it complains ('Post divider max count
> > > exceeded')
> > > Does this mean there is a lower limit? Is there a way to achieve this
>  speed
> > > using ALTPLL and the above specs?
> > >
> > > Thanks
> > >
> > >
> >
> >


SneakerNet,
You can take the 50 mhz clock into a 2 bit counter and take the output
of the MSB for 12.5 MHz. This can drive a global clock tree within the
device to minimize skew between all registers driven by the 12.5 MHz.
The only drawback is that you will have some skew between the 50 mhz
and 12.5 mhz domains, so you need to be careful with data transfers
across the domains.

Or you could take a 25 MHz PLL output and divide by 2 with a TFF. A
variety of options here.

Sincerely,
Greg Steinke
gregs@altera.com
Altera Corporation

Article: 67286
Subject: In-system configuration through JTAG on Spartan-3
From: qiufeng_song@yahoo.com (Alex)
Date: 9 Mar 2004 11:35:40 -0800
Links: << >>  << T >>  << A >>
Hi,

I'm trying to configure the Spartan-3 FPGA in-system. I ported the
source code that's in the application note xapp058 from Xilinx. The
entry fucntion xsvfExecute is called and returns 0. But the FPGA isn't
working and the clock signal we're expecting is not there. The same
FPGA code is flashed into a serial flash and it works in this way.

Anyone has some idea?

Your help is really appreciated.

Alex

Article: 67287
Subject: ISE 6.2 issues
From: do_not_reply_to_this_addr@yahoo.com (Sumit Gupta)
Date: 9 Mar 2004 11:48:37 -0800
Links: << >>  << T >>  << A >>
My design which used to work fine with ISE 4.2 is not working at all
with ISE 6.2. While I debug it to figure out what went wrong, is it
possible for people to download old version of ISE webpack from
somewhere ?

Sumit
------
http://www.c-nit.net

Article: 67288
Subject: Re: HOW to Increase jitter in ALTERA PLL ?
From: gregs@altera.com (Greg Steinke)
Date: 9 Mar 2004 12:10:05 -0800
Links: << >>  << T >>  << A >>
cdufourfour@yahoo.ca (john) wrote in message > I know I can change the counters values but I don't know the out-spec
> limits of the PLL such as :
> -input frequency min max (out-spec)
> -minimum frequency of VCO (out spec)
> -Specifications of the loop-filter inside the VCO
> 
> The altpll primitive is the same for ALTERA Stratix and Cyclone
> devices. I use Cyclone FPGA and I don't know what parameters concern
> Cyclone (cauz Stratix PLLs are much more parameterable) ...

John,
For Cyclone PLL, the input frequency limits vary by speed grade:

Speed grade    Min      Max
-6            15.625    464
-7            15.625    428
-8            15.625    387

VCO range is 500 to 1000 MHz.

For more details, please see the Cyclone Handbook:
Chapter 4: DC & Switching Characteristics 
http://www.altera.com/literature/hb/cyc/cyc_c51004.pdf

and 

Chapter 6: Using PLLs in Cyclone Devices
http://www.altera.com/literature/hb/cyc/cyc_c51006.pdf

Sincerely,
Greg Steinke
gregs@altera.com
Altera Corporation

Article: 67289
Subject: Re: Using ALTPLL
From: Peter Alfke <peter@xilinx.com>
Date: Tue, 09 Mar 2004 12:12:40 -0800
Links: << >>  << T >>  << A >>
In the spirit of friendly competition, I want to point out that Xilinx DCMs
have no problem whatsoever dividing 50 MHz by 4 and maintaining phase
coherence.
Please no flames...
Peter Alfke

> From: gregs@altera.com (Greg Steinke)
> Organization: http://groups.google.com
> Newsgroups: comp.arch.fpga
> Date: 9 Mar 2004 11:34:10 -0800
> Subject: Re: Using ALTPLL
> 
> "SneakerNet" <nospam@nospam.org> wrote in message
> news:<Qea3c.6567$Nc3.108000@news.xtra.co.nz>...
>> Hi Subroto
>> 
>> Thanks for the clarification.. This means I must have used my own code to
>> generate the 12.5MHz. Oh well back to using the logic.
>> Thanks again
>> 
>> Regards
>> 
>> "Subroto Datta" <sdatta@altera.com> wrote in message
>> news:41a3c.22164$cY5.21375@newssvr31.news.prodigy.com...
>>> Hi Sneakernet,
>>> 
>>> You cannot implement their desired output frequency of 12.5 Mhz using the
>>> Cyclone PLL without going out of spec.
>>> 
>>> The Cyclone PLL does have a minimum output frequency limit, which is a
>>> function of the PLL's minimum VCO frequency.  The minimum VCO frequency is
>>> 491Mhz, and the maximum output divider counter size is 32, which results
>> in
>>> a minimum output frequency of approximately 15.3 Mhz (i.e.491/32).
>>> 
>>> Page 4-31 of the device handbook gives the PLL frequency specs.  The input
>>> frequency has to be between 15.625 MHz and 464 MHz for the
>>> fastest (-6) speed grade, and between 15.625 MHz and 387 MHz for the
>> slowest
>>> (-8) speed grade.  See
>>> http://www.altera.com/literature/hb/cyc/cyc_c51004.pdf  for details.
>>> 
>>> - Subroto Datta
>>> Altera Corp.
>>> 
>>> 
>>> "SneakerNet" <nospam@nospam.org> wrote in message
>>> news:Ow43c.6282$Nc3.101499@news.xtra.co.nz...
>>>> Hello Ppl
>>>> 
>>>> I have actually already asked this question before but I need help
>> again.
>>>> 
>>>> I'm using a Nios Development Board (50MHz Oscillator, FPGA = Cyclone
>>>> EP1C20F400C7). I'm using Quartus II v4.0 Build 190
>>>> I need to generate a clock speed of 12.5MHz. While I have my own vhdl
>> code
>>>> that can do this, I can't use the ALTPLL to do this for me.
>>>> 
>>>> How can i do this? When I try to generate a speed of 12.5MHz using
>>>> MegaFunction Wizard in Quartus it complains ('Post divider max count
>>>> exceeded')
>>>> Does this mean there is a lower limit? Is there a way to achieve this
>> speed
>>>> using ALTPLL and the above specs?
>>>> 
>>>> Thanks
>>>> 
>>>> 
>>> 
>>> 
> 
> 
> SneakerNet,
> You can take the 50 mhz clock into a 2 bit counter and take the output
> of the MSB for 12.5 MHz. This can drive a global clock tree within the
> device to minimize skew between all registers driven by the 12.5 MHz.
> The only drawback is that you will have some skew between the 50 mhz
> and 12.5 mhz domains, so you need to be careful with data transfers
> across the domains.
> 
> Or you could take a 25 MHz PLL output and divide by 2 with a TFF. A
> variety of options here.
> 
> Sincerely,
> Greg Steinke
> gregs@altera.com
> Altera Corporation


Article: 67290
Subject: Re: Using ALTPLL
From: "SneakerNet" <nospam@nospam.org>
Date: Wed, 10 Mar 2004 09:35:14 +1300
Links: << >>  << T >>  << A >>
Hi Greg

Thanks for the contribution. I have something similar going right now.
Cheers


"Greg Steinke" <gregs@altera.com> wrote in message
news:5c1de958.0403091134.416775f3@posting.google.com...
> "SneakerNet" <nospam@nospam.org> wrote in message
news:<Qea3c.6567$Nc3.108000@news.xtra.co.nz>...
> > Hi Subroto
> >
> > Thanks for the clarification.. This means I must have used my own code
to
> > generate the 12.5MHz. Oh well back to using the logic.
> > Thanks again
> >
> > Regards
> >
> > "Subroto Datta" <sdatta@altera.com> wrote in message
> > news:41a3c.22164$cY5.21375@newssvr31.news.prodigy.com...
> > > Hi Sneakernet,
> > >
> > > You cannot implement their desired output frequency of 12.5 Mhz using
the
> > > Cyclone PLL without going out of spec.
> > >
> > > The Cyclone PLL does have a minimum output frequency limit, which is a
> > > function of the PLL's minimum VCO frequency.  The minimum VCO
frequency is
> > > 491Mhz, and the maximum output divider counter size is 32, which
results
> >  in
> > > a minimum output frequency of approximately 15.3 Mhz (i.e.491/32).
> > >
> > > Page 4-31 of the device handbook gives the PLL frequency specs.  The
input
> > > frequency has to be between 15.625 MHz and 464 MHz for the
> > > fastest (-6) speed grade, and between 15.625 MHz and 387 MHz for the
> >  slowest
> > > (-8) speed grade.  See
> > > http://www.altera.com/literature/hb/cyc/cyc_c51004.pdf  for details.
> > >
> > > - Subroto Datta
> > > Altera Corp.
> > >
> > >
> > > "SneakerNet" <nospam@nospam.org> wrote in message
> > > news:Ow43c.6282$Nc3.101499@news.xtra.co.nz...
> > > > Hello Ppl
> > > >
> > > > I have actually already asked this question before but I need help
> >  again.
> > > >
> > > > I'm using a Nios Development Board (50MHz Oscillator, FPGA = Cyclone
> > > > EP1C20F400C7). I'm using Quartus II v4.0 Build 190
> > > > I need to generate a clock speed of 12.5MHz. While I have my own
vhdl
> >  code
> > > > that can do this, I can't use the ALTPLL to do this for me.
> > > >
> > > > How can i do this? When I try to generate a speed of 12.5MHz using
> > > > MegaFunction Wizard in Quartus it complains ('Post divider max count
> > > > exceeded')
> > > > Does this mean there is a lower limit? Is there a way to achieve
this
> >  speed
> > > > using ALTPLL and the above specs?
> > > >
> > > > Thanks
> > > >
> > > >
> > >
> > >
>
>
> SneakerNet,
> You can take the 50 mhz clock into a 2 bit counter and take the output
> of the MSB for 12.5 MHz. This can drive a global clock tree within the
> device to minimize skew between all registers driven by the 12.5 MHz.
> The only drawback is that you will have some skew between the 50 mhz
> and 12.5 mhz domains, so you need to be careful with data transfers
> across the domains.
>
> Or you could take a 25 MHz PLL output and divide by 2 with a TFF. A
> variety of options here.
>
> Sincerely,
> Greg Steinke
> gregs@altera.com
> Altera Corporation



Article: 67291
Subject: ACEX: max current per pin
From: Pawel Kolodziej <P.Kolodziej@elka.pw.edu.pl>
Date: Tue, 9 Mar 2004 22:43:13 +0100
Links: << >>  << T >>  << A >>
Hi,
In Altera's ACEX 1k family datasheet there is an absolute maximimum rating:
Iout output current, per pin min = ­25; max = 25 mA

Does it apply only to IO pin's or as well to power/gnd pins ?
If it does not applay to VCC/GND pins what are the constrains for them?

I drive 16 LED's, 10mA per each. I don't know if i need any external
dirver (some HC chip) or not. 

Best regards,
Pawel

Article: 67292
Subject: Forcing FSM Encoding to be user in ISE 6.2
From: "Sumit Gupta" <do_not_reply_to_this_addr@yahoo.com>
Date: Tue, 09 Mar 2004 22:22:32 GMT
Links: << >>  << T >>  << A >>
It seems ISE 6.2 uses one hot encoding for state machines even if I specify
fsm encoding type to be user in synthesis options. This is really odd. Why
would Xilinx do that ? Also How do I force FSM encoding type to be user.

Sumit
------
http://www.c-nit.net



Article: 67293
Subject: Re: HOW to Increase jitter in ALTERA PLL ?
From: Jim Granville <no.spam@designtools.co.nz>
Date: Wed, 10 Mar 2004 11:42:51 +1300
Links: << >>  << T >>  << A >>
john wrote:
> Thanks ...
> 
> My purpose is to implement a True Random Number Generator like Viktor
> Fischer in his article "True Random Number Generator in Reconfigurable
> Devices" http://www.kemt.fei.tuke.sk/publication/Drutarovsky/ches2002.pdf
> .
> 
> I know I can change the counters values but I don't know the out-spec
> limits of the PLL such as :
> -input frequency min max (out-spec)
> -minimum frequency of VCO (out spec)
> -Specifications of the loop-filter inside the VCO
> 
> The altpll primitive is the same for ALTERA Stratix and Cyclone
> devices. I use Cyclone FPGA and I don't know what parameters concern
> Cyclone (cauz Stratix PLLs are much more parameterable) ...

  You could look at some of the spread spectrum clock generators, from
companies like Cypress, or Maxim ?
  On these you can control the spread to some extent, and they are RC in
nature, so that should help them be gaussian relative to the FPGA crystal.
  One real problem with an on-chip PLL that you are trying to wiggle
deliberately, is it will tend to noise-lock to any comb harmonics
from the digital portions - so smooth freq control is unlikely, and
quasi random is the best you could expect.

-jg


Article: 67294
Subject: Re: Release asynchrounous resets synchronously
From: Jim Granville <no.spam@designtools.co.nz>
Date: Wed, 10 Mar 2004 12:03:48 +1300
Links: << >>  << T >>  << A >>
Ray Andraka wrote:

> Peter,
> 
> Consider the case where you have 1 input high to one of your luts, and 2 inputs
> high to another lut.  It is clearly an illegal state, as it has two extra '1' bits,
> but it will be detected as OK by your circuit because exactly one LUT is indicating
> one input on.  Each group of four requires two outputs to distinguish 0,1 or more
> than 1 input on.

  I think Peter was partly correct. You can protect/correct a 16 stage 
One-Hot engine against illegal states with 5 LUTs, but it will not
recover in a single clock cycle.
  Simplest topology is to have 15 shifters, and #16 loads a HI ONLY if
all Prev15 are 000000000000000, if not, it simply waits until
the bogus ones ripple out.

  -jg


> 
> Peter Alfke wrote:
> 
> 
>>Eh, what? Unfortunately anonymos...
>>
>>Each first level LUTs detects (output High) that exactly one of its inputs
>>is High.
>>The second tier LUT detects that exactly one of the first-tier LUT outputs
>>is high, which mans that there is exactly one High input.
>>Agreed ?
>>Peter Alfke
>>
>>
>>>From: user@domain.invalid
>>>Newsgroups: comp.arch.fpga
>>>Date: Tue, 09 Mar 2004 05:24:26 GMT
>>>Subject: Re: Release asynchrounous resets synchronously
>>>
>>>Peter Alfke wrote:
>>>
>>>>LUTs are very efficient "illegal state" detectors.
>>>>Let's say you have a 16-state one-hot machine. Four LUTs can each detect
>>>>"exactly one of my inputs is High", and a fifth LUT does the same with the
>>>>four LUT outputs. So 5 LUTs can detect any illegitimate 16-bit code. Take it
>>>>from there...
>>>
>>>Eh, what?  So the first tier LUT compute f, where
>>>f(a,b,c,d) = 1 iff a+b+c+d = 1, else 0.
>>>For the 5th LUT we have the same property that a legal 16-state would
>>>map exactly one of the four first tier LUTs to 1, thus it sounds like
>>>what you have in mind is something like this:
>>>
>>>f({f(s[3:0]), f(s[7:4]), f(s[11:8]), f(s[15:12])})
>>>
>>>but this could accept states like 16'b1111_1110_1100_0001.
>>>
>>>I don't see how you can detect legal states with only five four-input LUTs.
>>>
>>>
>>>Peter, the FPGA reset question has come many times.  What does Xilinx
>>>recommend in general?  Async-reset+Sync-release, all-sync, or all-async?
>>>Which uses fewest resources?
>>>
>>>Thanks,
>>>
>>>Tommy
>>>
> 
> 
> --
> --Ray Andraka, P.E.
> President, the Andraka Consulting Group, Inc.
> 401/884-7930     Fax 401/884-7950
> email ray@andraka.com
> http://www.andraka.com
> 
>  "They that give up essential liberty to obtain a little
>   temporary safety deserve neither liberty nor safety."
>                                           -Benjamin Franklin, 1759
> 
> 


Article: 67295
Subject: xilinx configuration problem
From: topfuel1015@comcast.net (warren)
Date: 9 Mar 2004 15:06:03 -0800
Links: << >>  << T >>  << A >>
I have been going nuts with a Spartan 2 and 18v04 during configuration
with a JTAG cable (III) and impact 6.1.

1. The mode pins are set for boundary scan.
2.  If I run the cable with a 5v reference:
   a. Chain initialization will succeed 99% of the time.
   b. EPROM programming/verify will succeed 95% of the time.
   c. FPGA programming/verify will fail verify 100% of the time and
the FPGA is left with what appears to be wrong IOB configuration;
starts drawing a great deal of power.
   d. If I run IDCODE looping, it will fail before my defined 1000
loops are completed.
3. If I run the cable with 3.3v reference (which it should be anyway),
I cannot get through the Chain initialization 100% of the time. I get
up to 10 unknown devices reported.

I have pulled out a previously designed board and reprogrammed it
without any problems, to verify the software/drivers/cables.

I have checked for noise on the power and JTAG lines, didn't find
anything alarming.

If anyone has any suggestions, your input would be GREATLY
appreciated.


Thanks!

Article: 67296
Subject: Re: lattice metastable info
From: Peter Alfke <peter@xilinx.com>
Date: Tue, 09 Mar 2004 15:11:21 -0800
Links: << >>  << T >>  << A >>
Here is the URL for my evaluation of metastability in Virtex-II Pro
flip-flops:
http://www.xilinx.com/xlnx/xweb/xil_tx_display.jsp?multPartNum=1&sTechX_ID=p
a_metastability

Peter Alfke, Xilinx Applications
============================
> From: Jim Granville <no.spam@designtools.co.nz>
> Organization: TelstraClear
> Newsgroups: comp.arch.fpga
> Date: Mon, 08 Mar 2004 18:10:44 +1300
> Subject: lattice metastable info
> 
> I see lattice have just posted a Metastable document.
> 
> I recall Peter A. made some Virtex measurements, but
> looking back at the Xilinx web, I can only find the venerable XAPP 096.
> - is there anything more up to date / based on newer silicon ?
> 
> -jg
> 


Article: 67297
Subject: Re: copy protection on FPGA using embedded serial number
From: Eric Smith <eric-no-spam-for-me@brouhaha.com>
Date: 09 Mar 2004 17:49:50 -0800
Links: << >>  << T >>  << A >>
nweaver@ribbit.CS.Berkeley.EDU (Nicholas C. Weaver) writes:
> If you are using a Xilinx, the V2/V2Pro bitfile encryption is a VERY
> high barrier for an attacker to cross: I'd comfortably protect a $10K
> secret, and uncomfortably protect a $100K secret from a malicious,
> capable adversary.

I'd expect the cost of extracting a valid decrypted bitstream from the
part to be even higher than that.  You have to do it without losing the
backup power to the part long enough for cells to change state.  This
makes a lot of the otherwise promising avenues of attack difficult if
not impossible.  For instance, you're not likely to be able to decap
the part and microprobe it.  (Microprobing probably wouldn't work anyhow,
as the signals you need for bistream extraction are likely below metal.)

It is more likely that a successful attack would result from some sort
of "social engineering", and even the fanciest protection features built
into the FPGA aren't going to protect from that.

Article: 67298
Subject: Re: novice for FPGA
From: "Kelvin @ SG" <kelvin8157@hotmail.com>
Date: Wed, 10 Mar 2004 10:02:05 +0800
Links: << >>  << T >>  << A >>
it seems your professor is fooling you...i guess designing a decent FPGA
chip will require a big team of
engineers and top of the line Cadence & Mentor softwares...be prepared to
pay hundreds of thousands US$
for the licenses alone...besides, it seems u r from a Chinese university...







Tong <nospam@cuhk.edu.hk> wrote in message
news:c2km7a$ko9$1@justice.itsc.cuhk.edu.hk...
> hi all,
>
> I am a undergraduate studying VLSI and Magic manul layout. My professor
have
> asked us to design an FPGA chip using Magic.
> I am confused in how to build the switch box and look up table. Could
anyone
> kindly recommend some website or books that gives detail explanation on
> these stuff.
>
> Thanks a lot.
>
> Regards,
> Tong
>
>



Article: 67299
Subject: Re: 66B mode of VirtexII-ProX Rocket I/O
From: Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid>
Date: Wed, 10 Mar 2004 13:33:19 +1100
Links: << >>  << T >>  << A >>
On Tue, 09 Mar 2004 15:14:41 +0100, Magnus Danielson
<magda@netinsight.net> wrote:
>Cheers,
>Magnus - who defeated the SDH/SONET scrambler with a customized ping!

OOOhhhh!  What protocol was running over the SONET?  Was it POS?

Allan.



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
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

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