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 33725

Article: 33725
Subject: Re: Does Flexlm Licensing Work on Windows 2000 Pro?
From: Russell Shaw <rjshaw@iprimus.com.au>
Date: Fri, 03 Aug 2001 11:33:50 +1000
Links: << >>  << T >>  << A >>
Go to:

  start|settings|control-panel|system|advanced|environment-variables

set the LM_LICENSE_FILE to point to c:\flexlm\license.dat

for user variables and system variables.


Install IPX/SPX network protocol in windows2k.


Dave Feustel wrote:
> 
> I am having *zero* success getting the Flexlm licensing process
> for Modelsim to work on my Windows 2000 Pro SP1 Dell
> system. This in spite of generous help from Model Technology.
> The first attempt to access Modelsim after each install results in a
> flexlm error message (latest is 'can't find license file') and an
> invalidated license.
> 
> This happens both with Webpack Modelsim and also the
> full version of Modelsim on CDROM delivered directly from Model.
> 
> To say that attempting to use Modelsim under these conditions
> is getting old fast would be a serious understatement.
> 
> Does the Flexlm licensing and license validation procedure actually
> *work* on Windows 2000?

--
   ___                                           ___
  /  /\                                         /  /\
 /  /__\ Russell Shaw, B.Eng, M.Eng(Research)  /  /\/\
/__/   / Victoria, Australia, Down-Under      /__/\/\/
\  \  /  http://home.iprimus.com.au/rjshaw    \  \/\/
 \__\/                                         \__\/

Article: 33726
Subject: Re: Alliance tools going away?
From: Phil Hays <spampostmaster@home.com>
Date: Fri, 03 Aug 2001 01:48:55 GMT
Links: << >>  << T >>  << A >>
Kamal Patel wrote:
> 
> That is correct Robert.  The ISE GUI ...

I hope that batch files still will work.


-- 
Phil Hays

Article: 33727
Subject: Re: Clock skew with Xilinx DLLs...
From: Ray Andraka <ray@andraka.com>
Date: Fri, 03 Aug 2001 02:18:12 GMT
Links: << >>  << T >>  << A >>
1ns sounds pretty high for just loading differences.  Make sure your clock is
clean and stable, also if you have lots of outputs toggling on the same bank as
the clock input, you might move them.  For experimentation purposes, perhaps you
can hold those outputs constant to see if the skew improves.  Also, what is the
resolution of your test setup, it might be exaggerating the skew.

Anyway, I have come to avoid depending on the edges being aligned (the case
where we had a problem had on the order of 500ps skew, partly because of greatly
unequal loading, partly because of clock jitter we could do nothing about).
Changing the UCF generally isn't going to help the situation.  The problem comes
about in the design because you are moving data between clock domains on edges
that are supposed to be aligned.  You'll have to change your design to make sure
data is stable on both sides of the transition where you cross clock domains.

There are a number of ways to do it.  You might refer to Peter Alfke's (Xilinx)
app note on synchronizing data.  It is intended for asynch interfaces, but the
mehtods apply here just as well.

Cary McCormick wrote:

> Hi folks,
>     I'm using a DLL in a SpartanII design and have discovered with lab
> experimentation that lo and behold, the /4 output lags the edge of the x1
> output by about 1ns. I'm certain that I'm using the DLL correctly (BUFGs on
> both outputs, feedback comes from BUFG'd x1 output) and I imagine that the
> phase difference is due entirely to loading differences since the /4 clock
> is *much* more heavily loaded than the x1 clock.
>     So, given that we're kind of stuck with this (what's the point of BUFG's
> anyway if this happens?) how can I design with this? Will the Design Manager
> (using 3.1) check for setup problems? Any design tricks that the gurus can
> share on this matter?? Safety precautions I can add to the UCF file??
> Thanks!!
>
> Cary McCormick

--
-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: 33728
Subject: Re: Xilinx Spartan XL length count question.
From: Philip Freidin <philip@fliptronics.com>
Date: Thu, 02 Aug 2001 20:28:48 -0700
Links: << >>  << T >>  << A >>
Maybe the following will give you some insight

    http://www.fpga-faq.com/archives/16150.html#16156

On 31 Jul 2001 11:54:54 -0700, kathy_n_brian@hotmail.com (Kathy & Brian) wrote:
>I simply want to use the Length Count value within the bitstream to
>determine exactly how many bits need to be clocked in to fully program
>the device.  I've thoroughly searched this group and Xilinx support
>for a clear explanation of the Length Count field, but I have yet to
>find a satisfactory explanation on the true meaning of Length Count.


In a nutshell:

For XC2K, 3K, 4K, and Spartan chips, the loading of the FPGA is
controlled primarily by the framing of the data. In the background, a
counter is counting CCLK cycles, and matching it against the Length
count field. A rather useful diagram that I would like you to look at
is not included in the Spartan data sheet that I have 

   http://www.xilinx.com/partinfo/ds060.pdf

but you could look at  Figure 48 in

   http://www.xilinx.com/partinfo/4000.pdf

which is the same logic that the Spartans use.

When the chip is "FULL" as a result of having received enough frames,
the "Full" signal goes high. Providing that the length count value is greater
than the number of CCLK cycles needed to get to the FULL state, when
the length count is matched by the CCLK clock counter, the signal
"Length Count" goes high. See fig 48, AND gate on left side. Then over the
next few CCLK cycles, the startup state machine releases GSR, enables
the I/Os, and issues DONE.

While not totally satisfying, may I suggest the following:

Take the PROMSIZE number from the data sheet:
   DS060.pdf, page 4-29, which is 249168.
   Add 2 bytes of 0xFF at the end of the bitstream
   When you do your configuration, supply enough CCLK
   cycles to clock in the 249168 bits, AND the extra 16 '1' bits.

   Check that Done is high.

>    For an XCS30XL running in Slave Serial mode, BitGen calculates a
>Length Count value of 249161.  This value places the "end" of the
>bitstream 2 bits beyond the Postamble.  I would expect Length Count to
>be either 249159 (number of bits from start to end of Postamble) or
>249163 (add 4 bits for Start-Up phase).  What is the purpose of this 2
>bit difference?

The Xilinx SW has had many attempts at getting these numbers right,
and there seems to always be a case where it is off by 1 or 2. The above
algorithm will always work, and what vetrans use (or something like it)
when they want to get on with their lives, and have had enough of trying
to figure out what the optimal number is.

FYI the underlying cause of the problem is that bitstreams can be loaded
either serially or in parallel, and you DO NEED at least 1 or 2 clocks after
DONE has gone high for config to have actually finished, for some startup
options, including the default.

Figure 47 in the 4000.PDF file shows this sequence, but this figure is missing
from the Spartan data sheet.  The default is shown as CCLK_NOSYNC, and
you can see that if you stop clocking CCLK as soon as you see DONE, (at
time C1), you end up with a chip that has DONE high, but I/Os are still
tri-stated, and GSR is still asserted.

Give it 1 extra clock, and you have a chip for which DONE is high, and I/Os
are active, and async paths through the chip work, but clocked stuff doesn't
because GSR is still asserted.  Two extra clocks and the chip is live. My
suggestion of 16 extra, is because I know it will work, and it doesn't hurt.

Another fly in this ointment is parallel load. For example if the FULL and
Length Count occur on the second last bit of the byte being shifted inside
the FPGA. You know you gave it all the data it needs, but infact it needs
one more byte of 0xFF so that it will generate enough clocks to finish the
startup sequence.

One extra Safety Note:  If you set Length Count to a value less than the
number of CCLKs that are needed to get to FULL, you will be disappointed,
because the comparator of Length Count to the CCLK counter will only
be valid for 1 CCLK cycle. Then the counter continues, and it will take
2^24 more clocks before it matches Length Count again. 

When you get to use Virtex chips, all this is irrelevant.

>    There is a support question (Answer Record #7912) that explains
>how the Length Count is calculated by BitGen, but it raises even more
>questions.  It states that the Length Count value = (PROM size - 7). 
>Why subtract 7?  Why derive Length Count from the PROM size at all
>(the PROM size includes an unknown quantity of fill bits!)?  The
>calculation presented does indeed match my bitstream file contents. 
>However, since the Length Count value does not have any useful
>relationship to the actual number of configuration bits (due to the
>unknown number of fill bits), it cannot be used by a microprocessor to
>clock in the correct number of bits!  The only solution that I see is
>to add some arbitrary number (7 would probably make the most sense) to
>Length Count to ensure that "enough" bits are clocked in.  This seems
>like such a hack.

Yep. Just what I am recommending  :-)

>    Does anybody know why Xilinx doesn't simply set Length Count equal
>to the exact number of bits necessary to completely program the
>device?  Or 4 more (for the Start-Up phase)?  This would seem to make
>so much more sense!

Hope the above shed some light on the issue.


Philip Freidin
Philip Freidin
Fliptronics

Article: 33729
Subject: Re: Does Flexlm Licensing Work on Windows 2000 Pro?
From: ronhui@ctimail3.com (ron)
Date: 2 Aug 2001 20:31:36 -0700
Links: << >>  << T >>  << A >>
I have painful experience when obtaining license from ModelSim and
Xilinx ISE Foundation, espeically Xilinx.

It is difficult to imagine we pay so much for the software in exchange
for such poor service

Mike Treseler <mike.treseler@flukenetworks.com> wrote in message news:<3B69B802.B5C06920@flukenetworks.com>...
> Dave Feustel wrote:
> 
>  
> > Does the Flexlm licensing and license validation procedure actually
> > *work* on Windows 2000?
> 
> I can testify that validation works on win2k.
> My license server is elsewhere on the network.
> I assume you have set a path for
> LM_LICENSE_FILE or MGLS_LICENSE_FILE 
> 
> (Start, Settings, ControlPanel, 
> System, Advanced, Environment . . .)
> 
> 
>  --Mike Treseler

Article: 33730
Subject: Re: finite defect statistics
From: "Austin Franklin" <austin@dar87kroom.com>
Date: Thu, 2 Aug 2001 23:52:49 -0400
Links: << >>  << T >>  << A >>
> Second, JUST BECAUSE THE SYSTEM IS DESIGNED TO RUN AT "ONLY" 30 MHZ DOES
> NOT MEAN YOU DO NOT HAVE A SIGNAL-INTEGRITY PROBLEM.  Read the Johnson
> book.  It's the edge rates that get you, not the clock speed.

You are absolutely right about that!  The CCLK has a very slow frequency,
8MHz, but the edge rate is sub-nanosecond!  I have run into quite a few PCB
layouts that have CCLK running all over the place, Ts etc....and the company
wonders why the boards don't reliably configure.  The other thing is, they
don't believe it when I tell them that is their problem...but a few
capacitors and resistors, cuts and jumps...and the thing configures
perfectly reliably.  I don't know if Xilinx has an application note on
exactly this, and CCLK routing, but if it doesn't, I'd strongly urge one!




Article: 33731
Subject: Re: finite defect statistics
From: "Austin Franklin" <austin@dar87kroom.com>
Date: Fri, 3 Aug 2001 00:04:44 -0400
Links: << >>  << T >>  << A >>
> I was merely making the observation that cost and size of company do not
> determine the quality of a design.

I would absolutely agree with that.

> I do
> realize the that large designs do not necessarily use many pips. In fact,
for
> high performance designs, I try to minimize the number of pips crossed.

One thing is nagging me about this...  I believe I was reading that the
intention of this "test" was to maximize the number of pips/resources used
on a single chip...and that seemed to induce failure.  I may not be saying
that correctly, but follow me...my head is still on vacation.  Pips are
quite resistive, and they also will degrade the signal...  I seem to
remember a time ago, when you could only go through so many pips without
having to be "re-powered".  Might there be an internal EMI issue, heat issue
or general signal integrity issue causing these failures?  I didn't read
that they were hard failures, or that they were even tested for hard
failures after a failure was found?  The tools might be causing the problem,
not taking into account actual signal integrity issues in the chip...  Just
a though...

The numbers just seem way too high to me too...and I'd have to say I've
shipped many tens of thousands of FPGAs, with VERY dense, fast designs, and
not seen anything remotely like this...




Article: 33732
Subject: 4 (8) bit Microporcessor Implementation
From: jaime.aranguren@ieee.org (Jaime Andres Aranguren Cardona)
Date: 2 Aug 2001 21:41:31 -0700
Links: << >>  << T >>  << A >>
Hi, gurus:

For a school project, I should implement within the next three months
a basic, but prefereably not very basic, CPU = microprocessor on a
PLCC84 cased Xilinx Spartan XCS05-3 running @ 10MHz, or even faster.

The design should meet, or exceed, the following requeriments:
- Speed
- Make it fit on the FPGA
- Should write our own compiler ("our" assembler -> bitsteram)
- Instructions: 8 bits long
- Data: 4 bits long

So, would some of you, please, help me with some starting points, as
links, textbooks, vhdl code, etc...?

I am sure most of you, guys, could give good advice on the subject.

Please reply to jaime.aranguren@ieee.org

Thanks a lot, in advance

Article: 33733
Subject: Re: May I connect two pins to the same net?
From: Falk <>
Date: Thu, 2 Aug 2001 23:41:45 -0700
Links: << >>  << T >>  << A >>
;-))
Iam sorry, I think I was a little bit impulsive. But there is (better was) a good reason for this. A guy in our company was trying to get a Flex10K working with a very simple design, but without success for one week. Then, he and me got together to fight the problem. After almost two days, we did, and 90% of the problem was our fault, but the behaviour of the undefined pins (DRIVING a signal) fooled us a lot. :-(

Anyway, every information is appreciated. Thank you.
Again, Iam sorrry about the LOUD reaction. 

You say you can do all designs with Quartus? Also Flex10k?? AFAIK this is not possible?

Regards
Falk

Article: 33734
Subject: Re: Duty cycle problem with Virtex-II
From: "Sune G. Krohn" <sgk@exbit.com>
Date: Fri, 3 Aug 2001 08:54:07 +0200
Links: << >>  << T >>  << A >>
The load is a track (2.5cm) to a connector with a FET probe (0.7pF||100k
ohm).
We have tested with different frequency, so it is not the reflection.


"Andy Peters <andy [@] exponentmedia com >" <".> wrote in message
news:FFga7.1224$B.115611@newsread1.prod.itd.earthlink.net...
> "Sune G. Krohn" wrote:
> >
> > I can't get a signal out of Xilinx Virtex-II 2V100 and 2V40 with a
correct
> > duty cycle.
> >
> > I only see this problem in 1.5 and 2.5 voltages mode.
> >
> > I also see the problem on Xilinx Virtex-II Evaluation Kit with a 2V40.
> >
> > As output I use OBUF_LVCMOS15_F_16 for 1.5 V and OBUF_LVCMOS33_F_16.
> >
> > With a frequency about 100MHz is the duty cycle about 35/65. In the test
I
> > run the clock through a FF to make a 50/50 duty cycle and with no luck.
> >
> > It is always the high pulse that a shorter than the low, even if I
invert
> > the signal.
> >
> > We have asked Xilinx's Technical Support Office United Kingdom every day
for
> > two weeks and they can't answer the question they just ask irrelevant
> > questions. For instance, they ask my to do an IBIS simulation on their
> > Evaluation Kit with their chip.
>
> What's the signal's load?  That's why they want you to do an IBIS sim...
>
> -andy



Article: 33735
Subject: Re: In-Circuit Power Supply Verification of Xilinx Chips
From: hmurray-nospam@megapathdsl.net (Hal Murray)
Date: Fri, 03 Aug 2001 07:45:27 -0000
Links: << >>  << T >>  << A >>
In article <3B5F015E.E3DBFF54@xilinx.com>,
 Peter Alfke <peter.alfke@xilinx.com> writes:
>This is a tricky problem, and it is common to all BGAs not just Xilinx devices,
>but programmable logic may offer a solution:
>You want to have significant Icc, and then you want to observe whether there is
>a Vcc gradient on the internal supply.
>I would reconfigure the chip with a bunch of distributed local oscillators ( odd
>number of inverting CLBs feeding back in a loop), and you can then observe the
>relative frequency of these oscillators. If one is significantly slower than the
>others, it indicates a missing Vcc connection.

Interesting suggestion.  But now I'm curious.  How are BGAs actually built?

I see what looks like a small PCB.  Is that a 1 layer board with simple
traces from each ball to the correponding bonding pad?

Or are there power and ground planes to reduce lead inductance?  If
so, wouldn't that mostly mask a missing pwr/gnd connection?


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


Article: 33736
Subject: Re: Spartan II and asynchronous memory interface
From: Steven Derrien <sderrien@irisa.fr>
Date: Fri, 03 Aug 2001 10:07:50 +0200
Links: << >>  << T >>  << A >>


Ray Andraka wrote:
> 
> Yes.  The problem is in the di/dt caused by very fast edges on the signals.  You
> can help it out some be selecting the low slew rate drive and setting the output
> currents at the lowest possible values consistent with what you are driving.  You
> can also intentionally skew outputs if you have a higher clock available to
> minimize the number of outputs changing at once.

Isn't it possible to add some kind of passive low-pass filter to
copensate for these 
fast edges ?

And BTW, would you have some reference about suche PCB issues (a book
for example).

Thank you for your help.

Steven


> 
> Steven Derrien wrote:
> 
> > Is this also a strong issue when the system is clocked below 1MHz ?
> > (sorry I have very little knowledge of PCB layout issues)
> >
> 
> --
> -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: 33737
Subject: Re: 4 (8) bit Microporcessor Implementation
From: CBFalconer <cbfalconer@yahoo.com>
Date: Fri, 03 Aug 2001 08:24:50 GMT
Links: << >>  << T >>  << A >>
Jaime Andres Aranguren Cardona wrote:
> 
> For a school project, I should implement within the next three months
> a basic, but prefereably not very basic, CPU = microprocessor on a
> PLCC84 cased Xilinx Spartan XCS05-3 running @ 10MHz, or even faster.
> 
> The design should meet, or exceed, the following requeriments:
> - Speed
> - Make it fit on the FPGA
> - Should write our own compiler ("our" assembler -> bitsteram)
> - Instructions: 8 bits long
> - Data: 4 bits long
> 
> So, would some of you, please, help me with some starting points, as
> links, textbooks, vhdl code, etc...?
> 
> I am sure most of you, guys, could give good advice on the subject.
> 
> Please reply to jaime.aranguren@ieee.org

No - post here, read here.

You should first decide on the size of memory available, i.e. how
many words can be addressed as instruction, how many as data, are
they separate spaces (must be if you are making data memory 4 bit
units and code memory 8 bit units).

Look up the instruction sets and architecture of some classic
simple computers.  The 8008 will be instructive, as will the 8080,
6502, 6800.  Don't forget the PDP8.  The 8080 was the first real
general purpose CPU.  The 6502 architecture was better suited to
controllers than computing.  The 6800 was also a good
architecture, but less useful than the 8080 IMO.  The 8008 was
effectively a crippled subset of the 8080.

You can do a lot with a 4 bit instruction word - I built such a
machine in 1964.  If you don't want to implement multi-word
instructions you need to make the instructions context sensitive,
but that leads to complexity.

-- 
Chuck F (cbfalconer@yahoo.com) (cbfalconer@XXXXworldnet.att.net)
   (Remove "XXXX" from reply address. yahoo works unmodified)
   mailto:uce@ftc.gov  (for spambots to harvest)



Article: 33738
Subject: Re: Spartan II and asynchronous memory interface
From: "Tim" <tim@rockylogic.com.nospam.com>
Date: Fri, 3 Aug 2001 09:32:28 +0100
Links: << >>  << T >>  << A >>

"Steven Derrien" <sderrien@irisa.fr> wrote in message
news:3B6A5BD6.C5208CDE@irisa.fr...
>
>
> Ray Andraka wrote:
> >
> > Yes.  The problem is in the di/dt caused by very fast edges on the signals.
You
> > can help it out some be selecting the low slew rate drive and setting the
output
> > currents at the lowest possible values consistent with what you are driving.
You
> > can also intentionally skew outputs if you have a higher clock available to
> > minimize the number of outputs changing at once.
>
> Isn't it possible to add some kind of passive low-pass filter to
> copensate for these
> fast edges ?
>
> And BTW, would you have some reference about suche PCB issues (a book
> for example).
>

The standard text is Johnson and Graham's 'High-Speed Digital Design'.

There is also a new text, whose title escapes me at the moment.





Article: 33739
Subject: Re: Clock skew with Xilinx DLLs...
From: Falk <>
Date: Fri, 3 Aug 2001 01:44:22 -0700
Links: << >>  << T >>  << A >>
How can 500ps of clock skew cause this kind of trouble? The clock to out time plus some routing should always be greater than this? How did you measure the skew (skew matching of IO cells, equal routing to IO cells)

Regards
Falk

Article: 33740
Subject: Re: Spartan II and asynchronous memory interface
From: Steven Derrien <sderrien@irisa.fr>
Date: Fri, 03 Aug 2001 11:11:08 +0200
Links: << >>  << T >>  << A >>


"Andy Peters
> 
> david garnett wrote:
> >
> > The BurchED Spartan board is only two layer,
> 
> That's good to know.
> 
> If it were me, I'd do at least four layers - top and bottom for signal,
> middle two for VCC and GND.

BTW, just to get an idea, let's say I m an inexperienced PCB designer
and 
that I want to have my own PCB board for a specific app. (let's say a 4 
layers board with a SPII-pq208, SRAM, and and ethernet controller + 
transceiver) i'd need around 100 sample of them (so it's very low
volume)

1) I could can ask for PCB designer services, to design the layout and 
handle the production but then it's likely to be expensive (price range
?)

2) I can do the layout on my own using for ex Orcad, since i am
inexperienced 
do I have a chance to succeed (I mean to get a working board) and if so
how 
long will it take ? and in such a case how much it would cost ?

Thanks,

steven




app an 
> 
> -andy

Article: 33741
Subject: Re: May I connect two pins to the same net?
From: martin.j.thompson@trw.com
Date: 03 Aug 2001 10:14:06 +0100
Links: << >>  << T >>  << A >>
"Peter Ormsby" <faepetedeletethis@mediaone.net> writes:

> Wow.  Bite my head off.
> 

Sorry!  Didn't mean to come across like that - just wanted to point
out that there was more than one of us who'd been caught out!

> I think there's some confusion on the tools.  I'm sorry I didn't
> specify I waas speaking of Quartus.  Unless you're designing MAX3000
> or MAX7000, there's no reason to be using MAX+PLUS II anymore.
> 

Or Flex10K, Acex 1K, (or indeed Max 9000 or any of the older stuff).
Max is still used for highish density designs, although the fitter has
been ported from Quartus.

> In any case, the default behaviour "out-of-the-box" for Quartus is to
> have the unused pins configured as inputs (tristated).  You don't need
> to delve into the menus if you don't want to.  Is that nice enough
> default behaviour?
> 

Just right!

> I was just trying to show you how to change the default settings if
> you wanted to.  Too much information, I guess.
> 

Fair enough, I just wanted to make sure the Max Plus II *default* behaviour was
clear to people.

Cheers,
Martin

Article: 33742
Subject: Re: Spartan II and asynchronous memory interface
From: martin.j.thompson@trw.com
Date: 03 Aug 2001 10:22:16 +0100
Links: << >>  << T >>  << A >>
Steven Derrien <sderrien@irisa.fr> writes:

<snip>
> And BTW, would you have some reference about suche PCB issues (a book
> for example).
> 

You could have a look at some of the docs written by Doug Brooks at 
http://www.ultracad.com/articles.htm
There is also a signal integrity forum on freelists.org... and many
more useful documents scattered around the web.

> Thank you for your help.
> 

Not at all!
Cheers,
Martin

> Steven
> 
> 
> >  Steven Derrien wrote:
> > 
> > > Is this also a strong issue when the system is clocked below 1MHz
> > >?  (sorry I have very little knowledge of PCB layout issues)
> > >

Article: 33743
Subject: Re: 4 (8) bit Microporcessor Implementation
From: Mike <mikerj@removeme.clara.net>
Date: Fri, 03 Aug 2001 10:33:03 +0100
Links: << >>  << T >>  << A >>
On Fri, 03 Aug 2001 08:24:50 GMT, CBFalconer <cbfalconer@yahoo.com>
wrote:

>Jaime Andres Aranguren Cardona wrote:
>> 
>> For a school project, I should implement within the next three months
>> a basic, but prefereably not very basic, CPU = microprocessor on a
>> PLCC84 cased Xilinx Spartan XCS05-3 running @ 10MHz, or even faster.
>> 
>> The design should meet, or exceed, the following requeriments:
>> - Speed
>> - Make it fit on the FPGA
>> - Should write our own compiler ("our" assembler -> bitsteram)
>> - Instructions: 8 bits long
>> - Data: 4 bits long
>> 
>> So, would some of you, please, help me with some starting points, as
>> links, textbooks, vhdl code, etc...?
>> 
>> I am sure most of you, guys, could give good advice on the subject.
>> 
>> Please reply to jaime.aranguren@ieee.org
>
>No - post here, read here.
>
>You should first decide on the size of memory available, i.e. how
>many words can be addressed as instruction, how many as data, are
>they separate spaces (must be if you are making data memory 4 bit
>units and code memory 8 bit units).
>
>Look up the instruction sets and architecture of some classic
>simple computers.  The 8008 will be instructive, as will the 8080,
>6502, 6800.  Don't forget the PDP8.  The 8080 was the first real
>general purpose CPU.  The 6502 architecture was better suited to
>controllers than computing.  The 6800 was also a good
>architecture, but less useful than the 8080 IMO.  The 8008 was
>effectively a crippled subset of the 8080.
>
>You can do a lot with a 4 bit instruction word - I built such a
>machine in 1964.  If you don't want to implement multi-word
>instructions you need to make the instructions context sensitive,
>but that leads to complexity.


A micro with a 4 bit instruction and 8 bit data sounds very much like
a low end (12 bit) PIC, such as the 16C5x or 12C508 etc.
A free VHDL  core exists for a basic PIC, a quick serach should find
it.

Mike

Article: 33744
Subject: Re: Spanning the heirarchy
From: Brian Dickinson <Brian@nospamesperan.dotcom>
Date: Fri, 03 Aug 2001 11:48:18 +0100
Links: << >>  << T >>  << A >>

"Rick Collins" <spamgoeshere4@yahoo.com> wrote in message
 news:3B67A4B4.29694DFF@yahoo.com...
> I am adding some code to a verilog design for debug and I need to access
> signals in a remote portion of the design. I have been told that there
> is a way to do this in the form of
> "top_level.mid_level.low_level.signal_name" where the level names are
> module instance names. This works ok in simulation, but I can't get it
> to work in synthesis. We are using Synplify. Is this not supported by
> this tool? Is this not supported by any synthesis tool?

Don't know of a synthesis tool which supports this...
If you want to access the signal in hardware, you need to make an
explicit
connection through the module ports.
If the signal access is just for debug, then you can hide the code from
the synthesis tool by using synthesis directives (comments which control
synthesis) e.g. enclose the remote signal access with the following
comments 

/* synthesis translate_off */
<unsynthesisable debug code>
/* synthesis translate_on */

B
 

        HDL, FPGA, PCB, Perl and Tcl training from Esperan
       The World's Leading Methodology Training Company for 
                       Electronic
Design                                        
                     http://www.esperan.com

Article: 33745
Subject: Re: 4 (8) bit Microporcessor Implementation
From: Falk <>
Date: Fri, 3 Aug 2001 03:51:12 -0700
Links: << >>  << T >>  << A >>
Have a look here

http://www.mindspring.com/~tcoonan/

Regards
Falk

Article: 33746
Subject: Re: Spartan II and asynchronous memory interface
From: "Martin Schoeberl" <martin.schoeberl@chello.at>
Date: Fri, 03 Aug 2001 12:12:37 GMT
Links: << >>  << T >>  << A >>
Hi Steven,

I had a simmilar problem with my own board (for Altera ACEX) also with
bad ground (because of saving money for two layer... the wrong decision).
I connected the printer port (long cable) to communicate via ECP with
a PC and got wrong data when to many bits switched. I think this led
to a 'ground bounce'.
One dirty solution (for me) was: Becaus I worked with a clock at 24 MHz
and ECP is below 1 MHz I switched the data lines one after the other.
If your clock is higher than the access time of the RAM you could try to
split the switching outputs.

Martin
--
Whant to see the evolution of a Java processor?

         http://www.jopdesign.com

"Steven Derrien" <sderrien@irisa.fr> schrieb im Newsbeitrag
news:3B697EF9.E5BDF155@irisa.fr...
> Hi,
>
> We are curently trying to port the XR16/Xsoc project (www.fpgacpu.org)
> to a VHDL targeted to the BurchEd Spartan II board
> (http://www.burched.com)
>
> We plan to make our work freely available, but are currently stuck on
> a problem. The design is a 16 CPU-SOC which interfaced to a parallel
> port.
>
> We have somes on-chip blockrams which serves as ROM, and off-chip
> asynchronous
> SRAM whiwh serves as main memory. Our problem is that we get frequent
> errors when accessing the off-chip SRAM banks. Generally a single bit
> wrong in a 16 bit data word every 200-300 access.
>
> All simulation (RTL,gate-level,post place and route) went fine.
> Right now, our system is clocked at 1Mhz far below its maximum
> frequency.
> Besides, the SRAM Write Enable command output signal is registered
> (although not in a IOB register) to avoid glitches which could cause
> wrong write operations.
>
> All IOB are configured with SLOW slew-rate and drive 12mA (default IOB
> config)
>
> We have been beating our heads on this problem for almost a week now,
> are there any experts around there to offer some tips/ideas/advices ?
>
> Thanks,
>
> Steven



Article: 33747
Subject: Re: May I connect two pins to the same net?
From: "Peter Ormsby" <faepetedeletethis@mediaone.net>
Date: Fri, 03 Aug 2001 12:38:42 GMT
Links: << >>  << T >>  << A >>
Falk,

No problem.  I understand the frustration of problems like that.  I've been
there myself.

As far as device support, the latest release of Quartus supports almost
everything except MAX3000 and MAX7000 (yes, it supports FLEX 10K and ACEX
1K).  Here's the link (device support is almost at the bottom of this page):

http://www.altera.com/products/software/quartus2/qts-index.html

Here's the fancy marketing stab at the same information.

http://www.altera.com/literature/sg/dtsg.pdf

The reason I'm including this second link is that the brochure indicates
that there is going to be a "Web Edition" of Quartus that's free (like the
MAX+PLUS II baseline).  Altera is also adding support for some small APEX
20KE devices to the free software too.

-Pete-

<Falk> wrote in message news:ee71bc7.7@WebX.sUN8CHnE...
> ;-))
> Iam sorry, I think I was a little bit impulsive. But there is (better was)
a good reason for this. A guy in our company was trying to get a Flex10K
working with a very simple design, but without success for one week. Then,
he and me got together to fight the problem. After almost two days, we did,
and 90% of the problem was our fault, but the behaviour of the undefined
pins (DRIVING a signal) fooled us a lot. :-(
>
> Anyway, every information is appreciated. Thank you.
> Again, Iam sorrry about the LOUD reaction.
>
> You say you can do all designs with Quartus? Also Flex10k?? AFAIK this is
not possible?
>
> Regards
> Falk



Article: 33748
Subject: Re: finite defect statistics
From: Ray Andraka <ray@andraka.com>
Date: Fri, 03 Aug 2001 12:47:15 GMT
Links: << >>  << T >>  << A >>


Austin Franklin wrote:

> > I was merely making the observation that cost and size of company do not
> > determine the quality of a design.
>
> I would absolutely agree with that.
>
> > I do
> > realize the that large designs do not necessarily use many pips. In fact,
> for
> > high performance designs, I try to minimize the number of pips crossed.
>
> One thing is nagging me about this...  I believe I was reading that the
> intention of this "test" was to maximize the number of pips/resources used
> on a single chip...and that seemed to induce failure.  I may not be saying
> that correctly, but follow me...my head is still on vacation.  Pips are
> quite resistive, and they also will degrade the signal...  I seem to
> remember a time ago, when you could only go through so many pips without
> having to be "re-powered".  Might there be an internal EMI issue, heat issue
> or general signal integrity issue causing these failures?  I didn't read
> that they were hard failures, or that they were even tested for hard
> failures after a failure was found?  The tools might be causing the problem,
> not taking into account actual signal integrity issues in the chip...  Just
> a though...

This is what I was trying to get at.  If I were seeing this high a rate of
failures, I'd want to know why and I want to know exactly what the failure was
so that I could be absolutely sure I wasn't the cause of my own pain.  I was
trying to get him to state whether these were hard fails or soft errors, but it
is apparent that once they isolate to a chip they replace it without ever
digging further.  We are also not privy to what if anything the FPGA vendor has
found.

I also recall there being a limit to the number of pips you can transit before
redriving a signal, at least for the 3K and 4K parts.  I don't remember if the
virtex switch boxes repower the signal or not (the tristate long lines do, where
they didn't in 4K).  IIRC, the software normally will keep you from going
through too many without a redrive.  If he is creating the test patterns in FPGA
editor, I could see where he might be exceeding that limit and thus causing his
own pain.

Also, there has been mention of signal integrity issues in FPGAs with smaller
geometries.  I know tht several of the vendors have been looking at the routing
software to arrange the routing to reduce crosstalk between runs.  Crosstalk has
been observed in commercial FPGAs (I believe there was a paper at FPGA'2000
documenting this and discussing routing algorithms to reduce it), but those were
on designs specifically tailored to induce crosstalk.  Perhaps a densely routed
test pattern might also induce the crosstalk especially if there are long
parallel runs.

Without knowing more of the details, which are apparently proprietary to his
customer, all we can do is to speculate on why he is experiencing such high
failure rates.  If the problems were hard defects and occurred at the claimed
frequency, I still think all of us who do FPGA designs day in day out would be
tripping over the issue multiple times.  The fact is we are not.  If it were my
problem, I'd want to get to the bottom of it to know why I am seeing such a high
failure rate when others are apparently not.  That means determining if it is a
hard defect (which means pinpointing the defect to a particular node), and if
not determining exactly what conditions are causing a soft error.  I'd certainly
enlist the vendor's help, but I wouldn't throw it over the wall and just write
it off as poor yield until I knew what was going on.



>
>
> The numbers just seem way too high to me too...and I'd have to say I've
> shipped many tens of thousands of FPGAs, with VERY dense, fast designs, and
> not seen anything remotely like 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: 33749
Subject: Re: 4 (8) bit Microporcessor Implementation
From: Veronica Merryfield <Veronica.Merryfield@tesco.net>
Date: Fri, 03 Aug 2001 14:02:56 +0100
Links: << >>  << T >>  << A >>
As I have posted before

Take a look at http://www.fpgacpu.org - good source of material.

Veronica

Jaime Andres Aranguren Cardona wrote:

> Hi, gurus:
>
> For a school project, I should implement within the next three months
> a basic, but prefereably not very basic, CPU = microprocessor on a
> PLCC84 cased Xilinx Spartan XCS05-3 running @ 10MHz, or even faster.
>
> The design should meet, or exceed, the following requeriments:
> - Speed
> - Make it fit on the FPGA
> - Should write our own compiler ("our" assembler -> bitsteram)
> - Instructions: 8 bits long
> - Data: 4 bits long
>
> So, would some of you, please, help me with some starting points, as
> links, textbooks, vhdl code, etc...?
>
> I am sure most of you, guys, could give good advice on the subject.
>
> Please reply to jaime.aranguren@ieee.org
>
> Thanks a lot, in advance




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