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 30375

Article: 30375
Subject: PC software for 10$/cd
From: <>
Date: 4 Apr 2001 22:47:07 +0200
Links: << >>  << T >>  << A >>
Send me an email
I'll send you a list
Thank you
software1for10@hotmail.com

Article: 30376
Subject: QPSK phase rotator implementation in FPGA ?
From: edlee@gpetech.com (Edward)
Date: Wed, 04 Apr 2001 20:48:53 GMT
Links: << >>  << T >>  << A >>
I have a QPSK receiver which has two sets of differential outputs, 
I & Q.  I'll need to compensate for phase rotation before I can 
convert the I & Q signals into serial data.  Where can I find
imformation on how to do this ?



Article: 30377
Subject: Re: some info. on FPGA
From: Brian Dipert <bdipert@NOSPAM.pacbell.net>
Date: Wed, 04 Apr 2001 21:43:10 GMT
Links: << >>  << T >>  << A >>
>Hello everyone
>
>I need some information on the basic introduction of FPGA.Please give me
>some advise .
>
>
I not-exactly-humbly suggest the resources at
http://www.ednmag.com/ednmag/verticalmarkets/PLD.asp
;-)

Brian Dipert
Technical Editor: Memory, Multimedia and Programmable Logic
EDN Magazine: http://www.ednmag.com
Contributing Editor, CommVerge Magazine: http://www.commvergemag.com
1864 52nd Street
Sacramento, CA   95819
(916) 454-5242 (voice), (916) 454-5101 (fax)
***REMOVE 'NOSPAM.' FROM EMAIL ADDRESS TO REPLY***
mailto:bdipert@NOSPAM.pacbell.net
Visit me at http://members.aol.com/bdipert

Article: 30378
Subject: Re: Timing Error
From: Rick Filipkiewicz <rick@algor.co.uk>
Date: Wed, 04 Apr 2001 23:43:05 +0100
Links: << >>  << T >>  << A >>


Qian wrote:

> I tried to implement my VHDL program by Xilinx Foundation 3.li
> however in my Place & Route step I got error information about timing constraint like
> "
> --------------------------------------------------------------------------------
>   Constraint                                | Requested  | Actual     | Logic
>                                             |            |            | Levels
> --------------------------------------------------------------------------------
>   TS_N870 = PERIOD TIMEGRP "N870"  20 nS    |            |            |
>   HIGH 10 nS                                |            |            |
> --------------------------------------------------------------------------------
> * TS_CLK = PERIOD TIMEGRP "CLK"  20 nS   HI | 20.000ns   | 24.742ns   | 18
>   GH 10 nS                                  |            |            |
> --------------------------------------------------------------------------------
>   TS_P2P = MAXDELAY FROM TIMEGRP "PADS" TO  | 20.000ns   | 11.374ns   | 3
>   TIMEGRP "PADS" 20 nS                      |            |            |
> --------------------------------------------------------------------------------
> * TS_0 = MAXDELAY FROM TIMEGRP "N870" TO TI | 10.000ns   | 27.188ns   | 17
>   MEGRP "PADS" 10 nS                        |            |            |
> --------------------------------------------------------------------------------
>   OFFSET = IN 20 nS  BEFORE COMP "CLK"      | 20.000ns   | 17.622ns   | 16
> --------------------------------------------------------------------------------
> * OFFSET = OUT 20 nS  AFTER COMP "CLK"      | 20.000ns   | 24.097ns   | 16
> --------------------------------------------------------------------------------
>
> 3 constraints not met.
> "
>
> Could anyone tell me how to solve that?
> Thanks a lot!

The timing of OFFSET contraints runs under the assumption that there is an external
device running off the same clock that is using your output signals for its inputs.

Basically what its saying is that the paths are so slow that even with 0 routing delay
the path delays are longer than the clock period, and the and therefore synchronous
operation is not possible. I've come across this before and although, strictly speaking,
it should be a warning, I'm quite happy that it comes out as an error. I'm not generally
in favour of tools that try to baby you in this way but in this case I/you are being told
that continuing this layout is a waste of time.

The ``PAR_ZEROFLOORS'' allows you to continue in the case that you need to analyse the
paths to see why they are so sloooooow. In this case I don't think you'll need that since
its clear that you will have to do some work to reduce the logic complexity - most likely
you will have to do some pipelining.

I have a rule-of-thumb for auto placed & routed designs coming from HDL synthesis. For a
-6 VirtexE I mentally allow for 1.5-2 nsec per level and another 2 for the sum of the
clock-to-out at the beginning of the path and the setup at the end. To do better than
this you need, as Ray has said, to start getting into the floorplanner
and/or use hard macros.


Article: 30379
Subject: Re: QPSK phase rotator implementation in FPGA ?
From: Ray Andraka <ray@andraka.com>
Date: Thu, 05 Apr 2001 02:25:10 GMT
Links: << >>  << T >>  << A >>
A simple way to do it is with differential detection.  Basically multiplying
your current complex sample by the complex conjugate of the previous sample will
net you the sine and cosine of the phase difference between the samples,
assuming the amplitude of the vector is more or less constant.  There is a foil
in my Modulation and Demodulation for FPGAs presentation (available on my
website) showing a simple 1/4pi DQPSK receiver using this technique.  The
example was an IS-94 cell modem implemented in a Xilinx 4013.

Edward wrote:
> 
> I have a QPSK receiver which has two sets of differential outputs,
> I & Q.  I'll need to compensate for phase rotation before I can
> convert the I & Q signals into serial data.  Where can I find
> imformation on how to do this ?

-- 
-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

Article: 30380
Subject: How to specify Spartan2 GSR/GTS for Synthesis
From: yuryws@banet.net
Date: Thu, 05 Apr 2001 05:14:14 GMT
Links: << >>  << T >>  << A >>
Would like to implement Power-On-Reset. Use no external Reset pin.
Asynchronous RST is used throughout the design on all flops/memories.

What is the procedure for tying GSR to RST during synthesis (FPGA
Express user)?

Thnx.





Article: 30381
Subject: Re: High Speed PLA/FPGA
From: hmurray-nospam@megapathdsl.net (Hal Murray)
Date: Thu, 05 Apr 2001 05:24:00 -0000
Links: << >>  << T >>  << A >>
>I have to map a clk phase generator, generating  8 clk phases
>@155MHz. The distance between phases is 800ps and high
>precision is required.
>Does anybody knows about some programmable (very)fast
>logic that can allow that? I don't know if there is some ECL
>(or PECL) programmable stuff anywhere....

Interesting problem.

PCB traces are pretty stable.  800 ns is ~5 inches.  That probably
isn't good enough for what you want.  (But keep it in mind.  You will
have to correct for it anyway.)  You could use coax instead and
adjust as needed.

Motorola  makes a chip that's a digitally controlled delay.
It's just a long chain of buffers with digital inputs selecting or
bypassing clumps of buffers.  You can adjust the delay in steps of
~20 ps up to ~2ns. E195/E196.

Motorola (and probably others) make PECL PLLs.  You could multiply
your clock by 4.  Now it's just "simple" logic to make the signals
you want.  Switch the differential pairs for the other 4 phases.


-- 
These are my opinions, not necessarily my employeers.  I hate spam.


Article: 30382
Subject: Re: High Speed PLA/FPGA
From: Jim Granville <jim.granville@designtools.co.nz>
Date: Thu, 05 Apr 2001 18:06:45 +1200
Links: << >>  << T >>  << A >>
Hal Murray wrote:
> 
> >I have to map a clk phase generator, generating  8 clk phases
> >@155MHz. The distance between phases is 800ps and high
> >precision is required.
> >Does anybody knows about some programmable (very)fast
> >logic that can allow that? I don't know if there is some ECL
> >(or PECL) programmable stuff anywhere....
> 
> Interesting problem.
> 
> PCB traces are pretty stable.  800 ns is ~5 inches.  That probably
> isn't good enough for what you want.  (But keep it in mind.  You will
> have to correct for it anyway.)  You could use coax instead and
> adjust as needed.
> 
> Motorola  makes a chip that's a digitally controlled delay.
> It's just a long chain of buffers with digital inputs selecting or
> bypassing clumps of buffers.  You can adjust the delay in steps of
> ~20 ps up to ~2ns. E195/E196.

 Has anyone tried a self-checking chain of the 40ps delays, in the 
newest FPGAs - you connect ~161 of them, and tap nominally every 20,
then
shuffle the groups Up/Dn under loop control to track the CLOCK.
( Peter A? )

Failing that, there is always a 1.24GHz source, and a PECL ring ctr,
that will be high precision, and without phase jumps.
On Semi's web site quotes 3GHz clocks for their QuadFF in PECL.

-jg

Article: 30383
Subject: How to combine bus in schematic
From: "Helen Long" <madisonfff@usa.net>
Date: Thu, 5 Apr 2001 01:14:12 -0500
Links: << >>  << T >>  << A >>
I just used DATAIN(23 downto 0)
however in schematic we only have IPAD(16) & IPAD(8)
I don't know how to combine them together
Thanks a lot!



Article: 30384
Subject: Re: DSP Volume-control in FPGA
From: Mark Sterk <void@strong.sjello.nl>
Date: Thu, 05 Apr 2001 09:43:08 GMT
Links: << >>  << T >>  << A >>


> But, I'm not quite sure how to implement a digital volume-control in the
> FPGA. I am using
>  PCM-code and have been dabbling with the idea to just divide the
> PCM-code with a user-defined number
> to get a "weaker" signal, but that would change the frequency of the
> signal somewhat.
> 

Making the pulses less wide won't infect the frequency of the signal
because the sampling frequency will/must be way beyond the maximal
frequency of the audio signal so the harmonics you generate by
shortening the PCM pulses will/must be filtered out in the last stage of
your amplifier.




Greetings,
Mark Sterk


If you want to mail me non-spam messages, don't send it into the void,
send it to: mailto:strong@sjello.nl

Article: 30385
Subject: Re: QPSK phase rotator implementation in FPGA ?
From: eml@riverside-machines.nospam.com (Evan Lavelle)
Date: Thu, 05 Apr 2001 10:26:45 GMT
Links: << >>  << T >>  << A >>
On Wed, 04 Apr 2001 20:48:53 GMT, edlee@gpetech.com (Edward) wrote:

>I have a QPSK receiver which has two sets of differential outputs, 
>I & Q.  I'll need to compensate for phase rotation before I can 
>convert the I & Q signals into serial data.  Where can I find
>imformation on how to do this ?

The idea is simple - you have to derotate your I and Q. This is simply
a complex conjugate multiply (think of vector rotation):

(A+iB).(C-iD) = (AC+BD) + i(BC-AD)

ie. 4 multiplies and 2 additions. The problem now is knowing what to
multiply by (ie. the values of C and D). I don't know about IS94
(Ray's post), but on the 3G stuff I've worked on the phase reference
changes so slowly (ie. C,D are effectively constant) that a DSP
calculates the required derotation from a pilot channel and supplies
the "constant" value to the hardware; you wouldn't bother to calculate
sample-to-sample rotation in hardware. Depends on what system you're
working on, and how clever your algorithms people are.

BTW, you don't have to serialise *after* the derotation. It generally
makes sense to serialise as early as possible; you can then optimise
your multiplications and additions.

Evan

Article: 30386
Subject: Re: How to combine bus in schematic
From: Dave Vanden Bout <devb@xess.com>
Date: Thu, 05 Apr 2001 06:58:05 -0400
Links: << >>  << T >>  << A >>
Helen Long wrote:

> I just used DATAIN(23 downto 0)
> however in schematic we only have IPAD(16) & IPAD(8)
> I don't know how to combine them together
> Thanks a lot!

I don't know what schematic editor you are using, but here is how to do
it with the editor in Xilinx Foundation:

1) Attach a bus to an instance of IPAD16 and name it DATAIN[15..0].
2) Attach another bus to an instance of IPAD8 and name it DATAIN[7..0].


--
|| Dr. Dave Van den Bout   XESS Corp.               (919) 387-0076 ||
|| devb@xess.com           2608 Sweetgum Dr.        (800) 549-9377 ||
|| http://www.xess.com     Apex, NC 27502 USA   FAX:(919) 387-1302 ||



Article: 30387
Subject: URGENT: Using SpartanII DLL to multiply clock freq
From: "Fuzesi Arnold" <arno@freemail.hu>
Date: Thu, 5 Apr 2001 19:23:28 +0200
Links: << >>  << T >>  << A >>
Hi,

How Can I use Xilinx Spartan II DLL to 2x multiple?
(~50% duty cycle needed)

I've tried to implement the schematic on page 12 at
http://www.xilinx.com/xapp/xapp174.pdf

But it doesn't work!

I've tried to modify CLKFB connect to CLK2X, but the result is similar.

Any idea?

Arnold
Ps: Please mail me the answares too: mailto:arno@freemail.hu





Article: 30388
Subject: Re: RC4/ARC4 on an FPGA.
From: "luigi funes" <fuzzy8888@hotmail.com>
Date: Thu, 05 Apr 2001 17:26:07 GMT
Links: << >>  << T >>  << A >>

Matt Hayes ha scritto nel messaggio <3ac9ba09$1@news.star.co.uk>...
>I'm aware that quite a number of people believe that the RC4 algorithm
isn't
>particularly suited to hardware but the idea interests me nonetheless.

The question is interesting and I hope to hear some guys more
expert than me!
It seems me RC4 is easy to implement in a PLD or FPGA with good
performances. It requires only three additions and a swap for each
byte to encrypt and I think it can be done with less than hundred cells
and a 256 x 8 ram.
Perhaps it has the best performance to cost ratio of all the well-known
encryption algorithms!

>
>I have performed a few websearches but can't really find the answers I am
>looking for.
>
>In particular, I would like to know:
>a) what rates of data throughput have been achieved by RC4 implementations
>on an FPGA? what is the fastest ever?

I don't know the fastest implementation.
I think it can be about 20-50 MBytes/s in a low cost PLD.
The bottleneck is the RAM cycle time, and perhaps this
prevent significant enhacements by pipelining.

Luigi

>b) if it is possible to purchase a fast RC4/ARC4 IP Core and what
throughput
>rates can be expected.
>
>Thanks,
>
>   Matt.
>
>









Article: 30389
Subject: Altera 20k programming
From: "Andrew Hosmer" <ahosmer@scoundrelz.com>
Date: Thu, 05 Apr 2001 18:17:06 GMT
Links: << >>  << T >>  << A >>
I just started integration with my new boards and we're having trouble
programming the FPGA's.  This board is an IP format board on an Acromag
carrier (PCI) on a PC running Windows 2k pro.  It has 1 7k series and 2
20k200E's.  The 7k is interface and control logic, etc. and is programmed
via JTAG.  The 2 20k's are programmed via Passive Serial w/ no device (just
writing through the 7k).

The 7k is fine,  no problem, talks to the PC great.  Programming the 20k's
starts fine, the logic analyzer shows the proper waveform (per AN116), but
CONF_DONE never goes high.  I'm downloading .rbf files which are fixed size
and big enough to generate the extra DCLK's necessary for INIT_DONE to go
high also, but it doesn't.  Hmmm... just to check things, I tried to program
all 3 chips via JTAG chain (using .sof files for the 20k's).  It sets up
fine, downloads data, but no CONF_DONE so it fails at the end.  Hmmmm,
again...... pull-up resister ok, no short to ground, no short to another
signal driving low, no noisy TCK, etc...

I'm using MAX PLUS v10 for the 7k and latest Quartus for the 20k's.  I've
tried downloading the JTAG chain from both (skipping the 7k in Quartus).
I've also tried using the chain, but programming each device individually.


Questions:
Should JTAG chain work?  Since the 20k's were set up for passive serial, the
CONF_DONE's are tied together.  Since CONF_DONE powers up low, will chip2
keep chip1 from driving CONF_DONE hi after programming (and vice versa)?

Any other ideas?


Thanks much for all help and consideration,

Andrew



Article: 30390
Subject: Re: How to specify Spartan2 GSR/GTS for Synthesis
From: Ray Andraka <ray@andraka.com>
Date: Thu, 05 Apr 2001 19:45:07 GMT
Links: << >>  << T >>  << A >>
The power on reset happens whether you specify a global reset or not, and
regardless of whether or not your logic has async resets on it or not.  For
synthesis, you needn't do a thing if you don't have a reset pin on your design.
If you want to simulate it however, then use the ROC component in the unisim
library.  It gets connected to your Global reset net.  You can leave it in there
if you want.  Synthesis will put the ROC in your design as a black box, and the
xilinx software will ignore it.

yuryws@banet.net wrote:
> 
> Would like to implement Power-On-Reset. Use no external Reset pin.
> Asynchronous RST is used throughout the design on all flops/memories.
> 
> What is the procedure for tying GSR to RST during synthesis (FPGA
> Express user)?
> 
> Thnx.

-- 
-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

Article: 30391
Subject: Re: URGENT: Using SpartanII DLL to multiply clock freq
From: Ray Andraka <ray@andraka.com>
Date: Thu, 05 Apr 2001 19:55:53 GMT
Links: << >>  << T >>  << A >>
That schematic works fine.  If you are not using the 1x clock, you can take the
feedback from either the 1x or 2x outputs (in that case, I'd favor the 2x
because it will take into account your circuit loading and won't use up the
other BUFG.  If you are using both the 1x and 2x clock, you must take the
feedback off the 1x or you wind up with phase ambiguity on the generated 1x
clock (and it usually seems to come up opposite what you wanted only when you
aren't looking).

If it is not owrking in simulation, the reason may be that your simulator
resolution is too coarse.  Set the simulator resolution to 100ps or finer.




Fuzesi Arnold wrote:
> 
> Hi,
> 
> How Can I use Xilinx Spartan II DLL to 2x multiple?
> (~50% duty cycle needed)
> 
> I've tried to implement the schematic on page 12 at
> http://www.xilinx.com/xapp/xapp174.pdf
> 
> But it doesn't work!
> 
> I've tried to modify CLKFB connect to CLK2X, but the result is similar.
> 
> Any idea?
> 
> Arnold
> Ps: Please mail me the answares too: mailto:arno@freemail.hu

-- 
-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

Article: 30392
Subject: Spartan II Configuration
From: Dean Armstrong <daa1@cs.waikato.ac.nz>
Date: Fri, 06 Apr 2001 09:23:34 +1200
Links: << >>  << T >>  << A >>
Hi All,

I am having trouble configuring a Spartan II chip via the JTAG
interface.

The chip is an XC2S100-5CES in the PQ208 package.

This is what I have done:

I have a simple VHDL design, which I synthesise using Synplify Pro.
I enable Readback and Reconfiguration in the Design Manager (To turn off
the security flag).
I then use the Xilinx Design Manager to generate a bitstream for this
device.
All appears to go well up to here.

I enter the JTAG programmer. Connect to the cable and initialise the
chain.
The device is detected as being a Virtex XCV100.
I assign the bitstream file I have created to the device. (which is
targeted to an XC2S100_PQ208)
I can get the ID number and the signature/usercode number from the
device.

Now when I try to program the device (without verify) this is the
feedback I get.

'cpu_wrapper(Device1)': Checking boundary-scan chain integrity...done.
'cpu_wrapper(Device1)': Reading bit-stream file...done.
'cpu_wrapper(Device1)': Programming device.....done.
'cpu_wrapper(Device1)': If the security flag is turned on in the
bitstream,
programming status can not be confirmed; otherwise,
programming terminated due to error.

I have a suspicion that this is somehow due to the device being detected
as a Virtex, and this is something to do with the fact that the chip is
an Engineering Sample, but really I'm only guessing.

Does anyone have any ideas about what is happening?

Thanks very much.
Dean Armstrong


Article: 30393
Subject: Re: RC4/ARC4 on an FPGA.
From: "rws" <semenoff@yahoo.com>
Date: Fri, 06 Apr 2001 01:54:03 GMT
Links: << >>  << T >>  << A >>

Matt Hayes <matthewhayes9000@hotmail.com> wrote in message
news:3ac9ba09$1@news.star.co.uk...
> I'm aware that quite a number of people believe that the RC4 algorithm
isn't
> particularly suited to hardware but the idea interests me nonetheless.

RC4 is not particularly compute intensive as encryption algorithms go,
but being a symmetric algorithm, usually requires a key exchange mechanism
such as Diffie-Hellman which is what really kills throughput.

> I have performed a few websearches but can't really find the answers I am
> looking for.
>
> In particular, I would like to know:
> a) what rates of data throughput have been achieved by RC4 implementations
> on an FPGA? what is the fastest ever?
> b) if it is possible to purchase a fast RC4/ARC4 IP Core and what
throughput
> rates can be expected.

I don't know about cores, but you might check rainbow.com and ncipher.com
for encryption specific chipsets.



Article: 30394
Subject: Modlesim5.5
From: W.Turk <kgut@dotuy.com>
Date: Thu, 5 Apr 2001 19:51:35 -0800
Links: << >>  << T >>  << A >>
Hi Gang:
   Now ,i use Modelsim5.5.When i load or run a large design,it will 
always close automatic.
   Why?

Article: 30395
Subject: Re: High Speed PLA/FPGA
From: Peter Alfke <palfke@earthlink.net>
Date: Fri, 06 Apr 2001 05:13:32 GMT
Links: << >>  << T >>  << A >>
The way I interprete your question, you need to divide a 155 MHz clock
perid into eight parts. (Right??)
The Virtex and Virtex-E devices give you four clock quadrants, which
seems to meet half your requirements.
Virtex-II goes further than that, it allows you to define the phase
delay of any clock in increments of 1/256 of the clock period,
implemented with stable delays of 50 ps increment. And there are 16
low-skew clock networks...

So, Virtex-E gets you halfway to your goal, and Virtex-II easily meets
all your requirements, even if it were *NOT* an even division of the
clock period. And you get hundreds or thousands of flip-flops and LUTs
plus a bunch of BlockRAMs and multipliers as a bonus. Should be hard to
resist  :-)

Peter Alfke, Xilinx Applications


Riccardo Zambon wrote:

> Hi all,
> I have to map a clk phase generator, generating  8 clk phases
> @155MHz. The distance between phases is 800ps and high
> precision is required.
> Does anybody knows about some programmable (very)fast
> logic that can allow that? I don't know if there is some ECL
> (or PECL) programmable stuff anywhere....
>
> Forgetting ECL, has anybody experienced VirtexII family at very high
> system speed?
>
> Thanks for help,
> Riccardo
>
> --
> Posted from david.siemens.it [192.109.0.136]
> via Mailgate.ORG Server - http://www.Mailgate.ORG


Article: 30396
Subject: Re: Modlesim5.5
From: Utku Ozcan <ozcan@netas.com.tr>
Date: Fri, 06 Apr 2001 08:34:27 +0300
Links: << >>  << T >>  << A >>
"W.Turk" wrote:
> 
> Hi Gang:
>    Now ,i use Modelsim5.5.When i load or run a large design,it will
> always close automatic.
>    Why?

  How much RAM do you have?

  Utku

Article: 30397
Subject: Re: Altera 20k programming
From: "Vitaliy Tkachenko" <vit@telus.net>
Date: Fri, 06 Apr 2001 07:25:19 GMT
Links: << >>  << T >>  << A >>
If the 20ke's are set up for passive serial configuration, TRST pins should
be tied to GND. To be able to configure the devices via the JTAG chain you
need to connect the pins to I/O Vcc.

Vitaliy




Article: 30398
Subject: Re: Modlesim5.5
From: Nicolas Matringe <nicolas.matringe@IPricot.com>
Date: Fri, 06 Apr 2001 09:38:55 +0200
Links: << >>  << T >>  << A >>
"W.Turk" wrote:
> 
> Hi Gang:
>    Now ,i use Modelsim5.5.When i load or run a large design,it
> will always close automatic.
>    Why?

I have the same problem and I think, as Utku suggested, that it is RAM
related.
It usually happens when I restart the simulation. It works fine after
freeing some RAM.

-- 
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: 30399
Subject: Re: High Speed PLA/FPGA
From: "Meelis Kuris" <matiku@hot.ee>
Date: Fri, 6 Apr 2001 13:02:32 +0300
Links: << >>  << T >>  << A >>
I have a similar problem, I need 4 phases of 270MHz clock.
But in Virtex-II, in low frequency mode I get only 210MHz from DCM,
in high frequency mode clk90 output is not available.
Will it work if I first create two 135MHz clocks with 45 degrees phase shift
in low freq mode and then multiply both by 2 in high freq mode?
The phase shift doesn't have to be very precise, -/+10 degrees would be ok
I think.

Mexx.

"Peter Alfke" <palfke@earthlink.net> wrote in message
news:3ACD5062.7425E890@earthlink.net...
..
> Virtex-II goes further than that, it allows you to define the phase
> delay of any clock in increments of 1/256 of the clock period,
> implemented with stable delays of 50 ps increment. And there are 16
> low-skew clock networks...
>
> So, Virtex-E gets you halfway to your goal, and Virtex-II easily meets
> all your requirements, even if it were *NOT* an even division of the
> clock period. And you get hundreds or thousands of flip-flops and LUTs
> plus a bunch of BlockRAMs and multipliers as a bonus. Should be hard to
> resist  :-)
>
> Peter Alfke, Xilinx Applications








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