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 275

Article: 275
Subject: Xilinx configuration
From: bytheway@ced.utah.edu (Jared Bytheway)
Date: Tue, 11 Oct 1994 16:14:52 GMT
Links: << >>  << T >>  << A >>
Hey,
I am trying to get a Xilinx XC3064 fpga part up and running but I'm having some
configuration problems.  I'm trying to load the configuration data in master
parallel mode.  I can see the address lines changing, trying to look for the 
start bits.  But when I put the EPROM in, the Xilinx chip starts sucking large
amounts of current.  I have my PS limited to 150mA so the supply voltage drops
to about 2.8 volts.
	Any ideas?????
It seems to have something to do with the data lines.  I can put in an erased 
EPROM (all 1s) and it doesn't do it, but of course it won't configure.  For some
reason, the Xilinx wants to keep all of its data lines high all of the time.
Something else that I noticed is that the address lines are all high for the first
80mS or so after powerup.  Then the address lines start counting like their supposed
to but quickly decay down to 2.7V (because of current limited supply).

Help on this would be greatly appreciated!

Thanks

Jared Bytheway
Center for Engineering Design
University of Utah
bytheway@ced.utah.edu






Article: 276
Subject: Multipliers in FPGA's
From: lcsmph@sn520.utica.ge.com (Michael P. Hartnett)
Date: 11 Oct 1994 13:04:18 -0500
Links: << >>  << T >>  << A >>
Does anyone know what FPGA architectures lend themselves to arthimetic
operations such as a multiply?  How fast can I run a 16 x 16 multilplier?

_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/

			                       
            _/      _/  _/_/_/   _/    _/	Michael P. Hartnett  
           _/_/  _/_/  _/   _/  _/    _/	Martin Marietta
	  _/  _/  _/  _/_/_/   _/_/_/_/		525 French Road
         _/      _/  _/       _/    _/ 		Utica, NY 13502
	_/      _/  _/       _/    _/		USA
       _/      _/  _/       _/    _/		Phone: 315-793-7604
                                                Fax:   315-793-7637
                                                e-Mail:lcsmph@utica.ge.com


_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/





Article: 277
Subject: Re: Xilinx configuration
From: Mike Stump <mrs@cygnus.com>
Date: Tue, 11 Oct 1994 19:57:51 GMT
Links: << >>  << T >>  << A >>
In article <CxInss.5Hz@ced.utah.edu>,
>But when I put the EPROM in, the Xilinx chip starts sucking large
>amounts of current.  I have my PS limited to 150mA so the supply voltage drops
>to about 2.8 volts.
>	Any ideas?????

I have a 70A 5 volt supply in my closet that I use for just such an
occasion.  When connecting to your chip may sure you use a heat sink
and 1/2 inch wide power rails.  :-)

Doesn't the chip spec sheet say that 150mA isn't enough power?

Go to the store, and buy a cheap PC power supply.


Article: 278
Subject: Re: Xilinx configuration
From: devb@char.vnet.net (David Van den Bout)
Date: 11 Oct 1994 16:41:57 -0400
Links: << >>  << T >>  << A >>
In article <CxInss.5Hz@ced.utah.edu>,
Jared Bytheway <bytheway@ced.utah.edu> wrote:
>Hey,
>I am trying to get a Xilinx XC3064 fpga part up and running but I'm having some
>configuration problems.  I'm trying to load the configuration data in master
>parallel mode.  I can see the address lines changing, trying to look for the 
>start bits.  But when I put the EPROM in, the Xilinx chip starts sucking large
>amounts of current.  I have my PS limited to 150mA so the supply voltage drops

I knew someone who had exactly the same problem.  I think he found
out that he had the data lines hooked up in reverse order due to some
screw-up in the documentation he had.  Once he rearranged them, everything
went along just fine.  I could be mistaken about this since it was over a
year ago, but you could call your XILINX representative and ask about it.


-- 

||  Dave Van den Bout  ||
||  Xess Corporation   ||


Article: 279
Subject: Memory bus based FPGA's
From: vishin@devil.eng.sun.com (Sanjay Vishin)
Date: 12 Oct 1994 01:16:03 GMT
Links: << >>  << T >>  << A >>
Has anyone used FPGA's to interface to a standard memory
bus running at or above 40MHz ? (With the FPGA actually driving the bus lines)

Thanks.
Sanjay/

Actually most of these busses are spec'ed keeping in mind 
a fast technology. eg: Futurebus+ with BTL, VBus(from Vitesse?) 
with GTL etc for which a FPGA solution may not be viable,
but even for a 5v TTL memory buss like Sun's Mbus, it gets quite
tricky to drive the bus lines using FPGA's within the required spec. 

Thats why I was wondering if anyone has dealt with pure FPGA sols,
to driving fast busses.


vishin@devil.sun.com


Article: 280
Subject: CALL FOR PAPERS -- 1995 PLD DESIGN CONFERENCE
From: sgolson@trilobyte.com (Steve Golson)
Date: Wed, 12 Oct 1994 05:04:20 GMT
Links: << >>  << T >>  << A >>
CALL FOR PAPERS -- 1995 PLD DESIGN CONFERENCE

April 25-27, 1995
Santa Clara, California

At next spring's PLD Design Conference I am chairing a session devoted
to all aspects of state machine design for PLDs and FPGAs. Proposals
are hereby invited for papers relating to state machines. Suggested
topics include (but are not limited to):

    -- case studies (real world designs)
    -- design and implementation techniques
    -- theoretical aspects of state machines
    -- asynchronous state machines

If you are interested in presenting a paper, please contact me as soon
as possible. A brief abstract would be appreciated.

Deadline for proposals:     October 20
Notification of acceptance: November 21
Approval of handouts:       December 12
Final manuscripts due:      February 27

Thanks for your interest!

Steve Golson
Trilobyte Systems
33 Sunset Road
Carlisle MA 01741
phone: 508/369-9669
fax:   508/371-9964
email: sgolson@trilobyte.com


Article: 281
Subject: FLEX8000 architecture / routability with fixed pin-out
From: chris@labtam.labtam.oz.au (Chris Taylor)
Date: Wed, 12 Oct 1994 09:38:28 GMT
Links: << >>  << T >>  << A >>

QUERYING THE ALTERA FLEX8000 FPGA FAMILY ARCHITECTURE
=====================================================


The following assumes that you have a working knowledge of the
FLEX8000 architecture. 

My question basically concerns how dependant the routability of the
FLEX8000 is on the design pin-out.

Most FPGA software will advise that you should (always) let the software choose
the pinout in order to maximise routability.

In practice one reaches a point where the pinout must be frozen
to match a P.C.B design and one hopes that minor changes can still be
made to the FPGA design while keeping a static pinout.

To cope with practical design changes one expects the ROUTABILITY of
an FPGA architecture to be relatively insensitive to the pin-out.
A high level of routability should be achievable for an arbitrary pin-out
(or at least a wide range of arbitrary pin-outs),
despite the fact that the pin-out may result in poor timing performance
because signals have had to follow unnecessarily long paths.

My experience with the Xilinx LCA's seems to indicate that these devices
(despite having what appears to be a fairly restrictive routing interconnect)
display a good ability to make design changes with a fixed pin-out.


SIMPLIFIED FLEX ARCHITECTURE (EPF8452 device)
(Based on portrayal in data sheet)


--------------------------------------------|  168     21
                                            |---/------/--mux--IOE0
                                            |       |          
                                            |       |          etc 
                                            |       |  21
168                                         |       ---/--mux--IOE7
channel                                     |
row                                         |
           ?                      .         |
           |                   .  |         |
-----------|-------------------|--|---------|   
           |                   |  |
           |                   |  |
           / 24                |  |
           |                   |  |
           |  LAB              |  |
        --------------|-----|  |  |
        |             |LE0  |---  |
        |             |-----|     |
        |             | etc |     |
        | local       |_____|     |
        | interconnect|LE7  |------
        |_____________|_____|


LAB= logic array block
LE= logic element (i.e registered or combinational equation)
IOE= input/output element


This diagram shows one of the rows (simplified).

At one end of the row are eight IOE's (only IOE0 and IOE7 shown).

One of the 21 LAB's is shown.
In this LAB is eight LE's (only LE0 and LE7 shown)
each of which drive a single line back into the row.
There are 24 inputs from the row into the LAB.


The SIMPLEST interperatation of this architecture suggests the following:

 - each LE is allocated a SINGLE channel which it can drive.

 - each IOB is allocated 21 channels from which it can choose one.

***********************************************************************
Note:
   21 LABs  x  8 LE's  = 168 channels
   21 LABs  x  8 IOE's  = 168 channels

This relationship also holds for the 27 column, 216 row channels device 
- however for the 13 column, 168 row channels device there appears to
be a surplus of channels.
***********************************************************************


Now for example this could mean that:
LE0 in LAB0 can only drive channel0,
LE1 in LAB0 can only drive channel1,
and so on.

IOE0 can only choose to connect to one of channels 0 to 20,
IOE1 can only choose to connect to one of channels 21 to 41,
and so on.

Therefore IOE0 can only be directly driven by 
  LE 0 to 7 from LAB0
  LE 0 to 7 from LAB1
  LE 0 to 4 from LAB2
  (i.e first 21 LE's in that row)

Of course the specific connectivety given here is just an example,
but the point is that each IOE can only be directly driven by quite
a small sub-set of the LE'S. (i.e 21 out of 168).

This means that once an IOE is assigned to a specific output signal, then
the logic that generates this signal becomes physically constrained to a 
particular sub-set of the LE's.

Now this constraint is offset by that fact that the signal does not 
necessarily have to follow the most direct path from logic to IOE.
It can find numerous other pathways through unused LE's, switching from row
to column and column to row. However this introduces the possibility of
degrading the timing excessively and it reduces the routing paths available
for other signals. (These effects compound greatly when a large number of
such paths need to be routed)


I would expect this architecture would make the routability of FLEX8000
extremely sensitive to the pin-out, more so that devices like the Xilinx LCA's.  
That is just a gut-estimate however.
If you had to make a minor modification to the logic of a highly utilized 
FPGA design with the pin-out fixed, I would expect the following.
A Xilinx LCA would be more likely to be able to find a routable path
but would generally give poorer timing. 
A FLEX device would generally give superior timing when routable,
but would be less likely to find a routable path.
In other words the FLEX is more likely to run into unresolvable
routing restrictions if the pin-out is fixed. 


I emphasize again that this is only a gut-feeling based on the above
architecture description.
I could have misjudged the FLEX architectures' ability to get from arbitrary
LE to arbitrary IOE in a highly utilized design.
(Equally I could have misjudged the FLEX's ability to route tricky paths
more efficiently that the Xilinx does on average - if one needs to go through
several unused LE's and swap between row and columns several times in order
to make a route then the FLEX fast-track starts to look as bad as the segmented
Xilinx interconnect.)

I have heard it suggested that one should leave 20% spare LE's and
20% spare IOE's in order to allow for design changes with fixed pin-out. 
In practice our designs tend to use 99% of the available pins.


ALTERNATIVELY
-------------

If the FLEX architecture is actually slightly different to that implied
above, the routability (with pins fixed) could be much better.

If instead of each LE only connecting to a SINGLE channel it actually
has the ability to connect to one of several channels then the
possible routing paths are increased.

For example, if LE0 in LAB0 can connect to channels 0,21,41 (but perhaps only
one at a time), this would mean that LE0 in LAB0 could directly connect to 
IOE0,1 or 2 rather than only IOE0.
If each LE had the ability to drive one out of a selection of eight
channels then any LE could directly drive any IOE the same row.

This is a relatively small difference in terms of architecture
(adds switch logic and increases loading) but the routability should
increase dramatically.

A similar result can be attained by adding more flexability at inputs to the
IOE's rather then the outputs of the LE's.


 
SUMMARY
-------

To re-state my basic query,
is the FLEX8000 architecture as I initially described it
(i.e Each LE output is dedicated to a single row channel)
or does it have enhancements like I have suggested
(i.e Each LE output can drive one of several row channels)?

The enhancements need not necessarily be on the LE outputs, 
they could be on IOE inputs or on column to row connections.
But basically they provide the same effect of increased routing paths.

Note that the inputs to the LAB's must have just this sort of
enhancement since there are 24 inputs x 21 LAB's = 504 connections
but only 168 channels and this suggests that (at minimum) each LAB input
connects to one of three possible channels. (504/168=3).


If anyone knows an E-mail address where I may get a definitive
answer to such a query (i.e an Altera technical query hotline) I would
appreciate it.

Also any success or horror stories concerning the fittability
of designs into FLEX8000 devices with fixed pin-out constraints
may be enlightening.


chris@labtam.oz.au


Article: 282
Subject: Re: Xilinx configuration
From: derek@redstar.dcs.qmw.ac.uk (Derek Coppen)
Date: 12 Oct 1994 13:39:48 GMT
Links: << >>  << T >>  << A >>
Jared Bytheway (bytheway@ced.utah.edu) wrote:
: Hey,
: I am trying to get a Xilinx XC3064 fpga part up and running
: but I'm having some configuration problems.

<snip>

You may be loading the data in reverse order! There's  master
high and master low modes. If this is the case you don't need
to reblow the eprom, just flip the state of the M1 input.



Article: 283
Subject: Re: Multipliers in FPGA's
From: satwant@regulus (Satwant Singh)
Date: Wed, 12 Oct 1994 15:59:42 GMT
Links: << >>  << T >>  << A >>
Michael P. Hartnett (lcsmph@sn520.utica.ge.com) wrote:
: Does anyone know what FPGA architectures lend themselves to arithmetic
: operations such as a multiply?  How fast can I run a 16 x 16 multiplier?

I know a guy who was able to hook Twelve  16-bit multipliers on
to a 32-bit tri-state bus in  ATT2C26 (the 26,000 gates ORCA 
FPGA from AT&T).  I think, the speed obtained was about 35 Mhz.

Each instance was a special integer multiplier where you multiply 
a variable by a constant and updating the constant requires a 
multiplier to pause for sometime. (eg, see page 80 of EDN, May 1994)

Satwant.



Article: 284
Subject: ALTERA EPLDs
From: cam@grenoble.hp.com (Daniele Beccari)
Date: Wed, 12 Oct 1994 16:59:01 GMT
Links: << >>  << T >>  << A >>
Hello, I have some questions about the best ways to exploit ALTERA
EPLDs, but I don't know if this is the appropriate newsgroup (fpga != eplds...)

Thanks for suggestions...

Dan



Article: 285
Subject: Re: Any documentation for Xilinx XNF file format?
From: carr@procyon.pmc-sierra.bc.ca (Larrie Carr)
Date: Wed, 12 Oct 1994 09:02:09 -0800
Links: << >>  << T >>  << A >>
In article <37bon3$pdl@news.tv.tek.com>, bobe@soul.tv.tek.com (Bob Elkind)
wrote:

> 
> I don't know if Xilinx considers their .XNF file format (which is
> entirely ASCII text) proprietary or not.
> 
> If not, then they should be happy to provide you with a format
> definition doc.  If they consider it proprietary so as to protect it
> from *others'* enhancements (e.g. the PC-ISA bus), then they should be
> happy to give you a copy after you sign a non-disclosure agreement.
> 
> If they consider it proprietary in the sense of *trade secret*, then
> you have to sign a non-disclosure agreement or be a big-bucks customer
> or both; or you can just decipher the "language" yourself and learn
> from the error messages you get from the "other tools".
> 

I believe that they *do* consider their .XNF file format to be proprietary
(at least thats what the rep said when I talked to him earlier this year).

However, they seem very reasonable to sign agreements with people from
universities.  Don't know exactly how much information you must supply.

-- 
Larrie Carr
Product Designer
PMC-Sierra, Inc.
Burnaby, B.C.


Article: 286
Subject: Re: Xilinx configuration
From: MoellerInc <mmoeller@delphi.com>
Date: Wed, 12 Oct 94 22:21:14 -0500
Links: << >>  << T >>  << A >>
I agree totally with chceking address lines, also rechekc that the
file istself is ok, use the downlaod cable to laod it if you can temporarily
reconfigure the circuit.
 
DO NOT up the current limit, 150 ma should be more than plenty. You can
destroy chips by bad config so caution is wise.   Always keep my finger
on the chip first time I trie a new board or file, might warm up
just a bit but if is geeting hot you have about  5 seconds to kill
the power.
 
Any ham operator know the time constant, just about the same amount
of time a plate glowing bright orange takes to melt and destroy
the final ouput tube....
 
Also double check the config mode pins, they can also cause problems
if floating.
 
Pardons typos, in a non editing enviromnet.
 
Martin Moeller
mmoeller@delphi.com
 
We do Video and Xilinx.


Article: 287
Subject: Re: Xilinx configuration
From: dfowler@bga.com (David Fowler)
Date: Wed, 12 Oct 1994 23:31:02
Links: << >>  << T >>  << A >>
In article <CxInss.5Hz@ced.utah.edu> bytheway@ced.utah.edu (Jared Bytheway) writes:
>Hey,
>I am trying to get a Xilinx XC3064 fpga part up and running but I'm having some
>configuration problems.  I'm trying to load the configuration data in master
>parallel mode.  I can see the address lines changing, trying to look for the 
>start bits.  But when I put the EPROM in, the Xilinx chip starts sucking large
>amounts of current.  I have my PS limited to 150mA so the supply voltage drops
>to about 2.8 volts.
>        Any ideas?????

We had a simular problem with the AT&T version of the Xilinx FPGAs.  We were 
not configuring them them the same way (we used serial from the host 
processor).  The problem had to do with the write pulse timming latching up 
the part internaly... we did not figure it out until a we got a new data 
book... they changed the timming requirements...  Anyway, the part got very 
hot because of the excess current draw.  Maybe you have a simular problem..


Article: 288
Subject: Re: Xilinx FPGA Pics: LCA, CLB etc.?
From: fliptron@netcom.com (Philip Freidin)
Date: Thu, 13 Oct 1994 06:57:36 GMT
Links: << >>  << T >>  << A >>
In article <CxC9E6.96B@latcs1.lat.oz.au> fulton@latcs1.lat.oz.au (John R Fulton) writes:
>I require a few pics of the Xilinx LCA, CLB etc for an overview
>of their architecture in my final year thesis.
>
>If anyone has any, or could tell me where to get some I would
>be grateful. Otherwise I'll just have to draw them myself :(.
>
>Regards 
>John
>
>--
>John Fulton, STUD IEAust                            | Computer Science
>Email:   fulton@lucifer.latrobe.edu.au              | and Electronic Eng
>TEL:     853 9068                                   | La Trobe University
>

The Xilinx databook chapter 2 is full of diagrams on CLBS, routing and
arrays. Is this what you want, or is it photomicrographs?

All the best
		Philip Freidin



Article: 289
Subject: Re: Xilinx configuration
From: fliptron@netcom.com (Philip Freidin)
Date: Thu, 13 Oct 1994 08:06:13 GMT
Links: << >>  << T >>  << A >>
In article <CxInss.5Hz@ced.utah.edu> bytheway@ced.utah.edu writes:
>Hey,
>I am trying to get a Xilinx XC3064 fpga part up and running but I'm having some
>configuration problems.  I'm trying to load the configuration data in master
>parallel mode.  I can see the address lines changing, trying to look for the 
>start bits.  But when I put the EPROM in, the Xilinx chip starts sucking large
>amounts of current.  I have my PS limited to 150mA so the supply voltage drops
>to about 2.8 volts.
>	Any ideas?????
>It seems to have something to do with the data lines.  I can put in an erased 
>EPROM (all 1s) and it doesn't do it, but of course it won't configure.  For some
>reason, the Xilinx wants to keep all of its data lines high all of the time.
>Something else that I noticed is that the address lines are all high for the first
>80mS or so after powerup.  Then the address lines start counting like their supposed
>to but quickly decay down to 2.7V (because of current limited supply).
>
>Help on this would be greatly appreciated!
>
>Thanks
>
>Jared Bytheway
>Center for Engineering Design
>University of Utah
>bytheway@ced.utah.edu


Almost certainly what is wrong is that you are not giving the chip a
valid bitstream. This is either because the data you are giving it
is wrong because of address line or data line confusion, or because
the data is not in the PROM the right way. The 80mS delay is normal,
it is called housecleaning, and is the chip scrubbing the memory to
a known state before programming. during this time the address lines
are not driven, but have a light pullup resistor enabled. The chip
looks for some leading '1' bits, followed by a bit pattern of 0010
followed by a length count. the 0010 is read left to right, in terms
of what the chip is looking for, and the length count is expected
msb first. See page 2-120 of the data book (1994). I am assuming that
you are selecting parallel up mode, so the addreses start at 0. for
master down, the data is the same, just the address sequence starts 
at ffff and counts down. 

Assuming you have a length count of 46112 (decimal), 0x00B420 (hex)
here is what the start of your PROM  MUST look like:

	msb            lsb      HEX
addr 0  1 1 1 1  1 1 1 1        FF
addr 1  0 0 0 0  0 1 0 0	04  (the 0 is the top 4 bits of LC, the 4
					is the 0010)
addr 2  1 1 0 1  0 0 0 0	D0  (the 0 is the next 4 bits of LC, and the
					D is the B of the LC)
addr 3  0 1 0 0  0 0 1 0	42  ( the 2 is the 4 of LC, and the 4 is the
					2 of the LC)
addr 4  1 1 1 1  0 0 0 0	F0  ( the 0 is the lest sig nibble of LC, and
					the F is the dummy bits before the
					first data frame)
addr 5  x x x x  x x x 0	Xx  ( the leat sig bit is the start bit of
				        the first data frame, the rest is
					design dependant.)
etc...

As you can see, the data is sent MSB first, but is shifted from the LOW
order end of the byte wide data comming from the EPROM. I.E. the first bit
seen by the config logic is the LSB of ADDR 0.

If data is not sent in correctly, then the chip miss configures, and
contention usually occurs. This leads to high current drawn. Dont put
a bigger power supply on this, fix the bitstream. 

At the start of configuration, all of the header upto and including the
first dummy bit that follows the length count comes out of the DOUT pin.
If you watch this pin amd the CCLK pin with a storage scope, or analyser,
you should see the header in the sequence of 0010, 24 bit LC (MSB first),
then constant high. If you see anything else, you have a problem. If the
LC coming out is not what you expect, then check that you have the data
lines connected up in the right order.

Why have I written all this rather than pointing you at a manual you
might ask??? Well it's because this is not documented well in any of
Xilinx's current docs. To say this sucks would be being generous. This
used to be well documented in the user guide and tutorial book, but
is not in any of the current ones.

Hopefully this will be enough to get you going.

All the best,
		Philip Freidin



Article: 290
Subject: Re: Xilinx configuration
From: evans@rd.eng.bbc.co.uk (Richard Evans)
Date: Thu, 13 Oct 1994 11:38:55 GMT
Links: << >>  << T >>  << A >>
Jared Bytheway (bytheway@ced.utah.edu) wrote:
: Hey,
: I am trying to get a Xilinx XC3064 fpga part up and running but I'm having some
: configuration problems.  I'm trying to load the configuration data in master
: parallel mode.  I can see the address lines changing, trying to look for the 
: start bits.  But when I put the EPROM in, the Xilinx chip starts sucking large
: amounts of current.  I have my PS limited to 150mA so the supply voltage drops
: to about 2.8 volts.
: 	Any ideas?????
: It seems to have something to do with the data lines.  I can put in an erased 
: EPROM (all 1s) and it doesn't do it, but of course it won't configure.  For some
: reason, the Xilinx wants to keep all of its data lines high all of the time.
: Something else that I noticed is that the address lines are all high for the first
: 80mS or so after powerup.  Then the address lines start counting like their supposed
: to but quickly decay down to 2.7V (because of current limited supply).

How much is your Xilinx taking and how much is the rest of the circuit taking?
May be your Xilinx is pushing your PSU over the limit during configuration. It
would be normal for the PSU to drop the o/p voltage once it reaches current 
limiting. May be if you gave the circuit the current it requires you would be
OK. This would be the first thing I'd check.

Apart from that, make sure you are counting in the correct direction, ie if
the mode pins are selecting MASTER PARALLEL UP then your data in the EPROM 
must match. Check the count direction using a logic analyser on the address
pins for absolute confidence. Also check that none of the address/data lines
are open circuit or short circuit.




richard

My views not those of my employer



Article: 291
Subject: PALASM versions?
From: seamang@westminster.ac.uk (Graham Seaman)
Date: Thu, 13 Oct 1994 14:16:34 GMT
Links: << >>  << T >>  << A >>
Does anyone know what the latest version of PALASM is and if it
is public domain? I'm unsure of the status of this and related
software (eg. ABEL/CUPL) - are they p.d., shareware, proprietary
or what? Any general info or ftp sites would be much appreciated!

Thanks
Graham
-- 
--------------------------------------------------------------------
Graham Seaman, School of Computer Science, 
University of Westminster, 115 New Cavendish St.  London W1M 8JS 
email:  seamang@wmin.ac.uk


Article: 292
Subject: Re: PALASM versions?
From: belanger002@wcsub.ctstateu.edu
Date: 13 Oct 94 12:35:17 EST
Links: << >>  << T >>  << A >>
In article <CxM7nn.207@westminster.ac.uk>, seamang@westminster.ac.uk (Graham Seaman) writes:
> Does anyone know what the latest version of PALASM is and if it
> is public domain? I'm unsure of the status of this and related
> software (eg. ABEL/CUPL) - are they p.d., shareware, proprietary
> or what? Any general info or ftp sites would be much appreciated!

The current version of AMD's PALASM is PALASM IV , rev 1.5a.  This is
AMD Copyrighted software.  The sell it in the us through their industrial
distributors for $125us.  If you have a good relationship (Purchasing wise)
with your AMD distributor or local AMD FAE, you can sometimes work out a
better arrangement.

Gerry Belanger



Article: 293
Subject: Re: PALASM versions?
From: mauro@cher.stanford.edu (Mauro Cerisola)
Date: Thu, 13 Oct 1994 17:20:54
Links: << >>  << T >>  << A >>
In article <CxM7nn.207@westminster.ac.uk> seamang@westminster.ac.uk (Graham Seaman) writes:

>Does anyone know what the latest version of PALASM is and if it
>is public domain? I'm unsure of the status of this and related
>software (eg. ABEL/CUPL) - are they p.d., shareware, proprietary
>or what? Any general info or ftp sites would be much appreciated!

>Thanks
>Graham

The latest CUPL for PC is called TotalDesigner 4.5A I received this version 
update last week.
By the way, the experience of my group with Logical Devices (we use CUPL and 
their logic programmer ALLPRO 88XR with fast PALCE and MACH devices) is a 
nightmare. Stay away from them! They are expensive, the CUPL has plenty of 
bugs, some of the manuals they ship with CUPL are years old, they know this 
and don't plan any update (sometimes they refer to non-existing chapters in 
other manuals). The ALLPRO 88XR comes with no manual, just a 30-pages handout.

What's worst, their technical support is really lousy, I NEVER got an answer 
from them better than what I already guessed myself.

DON'T MAKE OUR SAME MISTAKE! :-)

>--------------------------------------------------------------------
>Graham Seaman, School of Computer Science, 
>University of Westminster, 115 New Cavendish St.  London W1M 8JS 
>email:  seamang@wmin.ac.uk

Mauro Cerisola,
OCRL Stanford University


Article: 294
Subject: Re: Xilinx configuration
From: lalone@attila.mc.com (Jim LaLone)
Date: 14 Oct 1994 09:36:12 -0400
Links: << >>  << T >>  << A >>
Having just debugged a similar problem, the cause was that the power
supply rise time was sufficiently slow to cause the Xilinx internal
power-up initialization circuitry to get confused.  Quoting from the
data book (fine print in a footnote!):

	"At power-up, Vcc must rise from 2.0 V to Vcc in less than
	25 ms.  If this is not possible, configuration can be delayed
	by holding RESET low until Vcc has reached 4.0 V (2.5 V for
	the XC3000L).  A very long Vcc rise time of >100 ms, or a
	non-monotonically rising Vcc may require >6-us high level on
	RESET, followed by a >6-us low level on RESET and D/P after
	Vcc has reached 4.0 V (2.5 V for the XC3000L)."

Our solution will probably be to hold RESET low during power-up.  Has
anyone else had similar problems?  If so, what was your solution?


------------------------------------------------------------------
-- 
   James C. LaLone              Email: lalone@mc.com
   Mercury Computer Systems
   199 Riverneck Rd.            Tel: (508) 256-0052 x305
   Chelmsford, MA 01824         Fax: (508) 256-3599


Article: 295
Subject: Re: FLEX8000 architecture / routability with fixed pin-out
From: sjsmith@netcom.com (Stephen J Smith)
Date: Fri, 14 Oct 1994 18:13:37 GMT
Links: << >>  << T >>  << A >>
this is in reply to "querying the altera flex 8000 architecture" ...

information about this, and other, details can be obtained from
the following sources:

application engineering: tel (800) 800-epld
application engineering: fax (408) 954-0348
marketing/training     : tel (408) 894-7000
marketing/product      : tel (408) 894-7104
customer service       : tel (800) sos-epld
literature             : tel (408) 894-7144
bbs (<= 14k4, 8,1,n)   : tel (408) 954-0104

i hope this is of some help.

Stephen Smith                 home: sjsmith@netcom.com
Product Planning Manager      work: sjsmith@altera.com


Article: 296
Subject: ANNOUNCE: WWW Page of FPGA-based Computing machines
From: guccione@foghorn.cc.utexas.edu (Steve Guccione)
Date: 14 Oct 1994 13:25:03 -0500
Links: << >>  << T >>  << A >>

I've (finally) gotten around to putting the list of FPGA-based
Computing Machines on the Web.  The URL is:

     List of FPGA-based Computing Machines
     http://bongo.cc.utexas.edu/~guccione/HW_list.html

It currently lists over 40 research and commercial systems, with a few
links to other appropriate Web pages.  As always, new systems and
corrections to existing entries are welcome.

For those without World Wide Web access, I will continue to post a
text version of this list to the comp.arch.fpga newsgroup
semi-periodically.

-- Steve Guccione
-- guccione@ccwf.cc.utexas.edu
-- 10/14/94



Article: 297
Subject: FCCM'95 Call for Papers
From: jma@descartes.super.org (Jeffrey M. Arnold)
Date: Fri, 14 Oct 1994 21:04:54 GMT
Links: << >>  << T >>  << A >>

     #######  #####   #####  #     #           ###    #####  #######
     #       #     # #     # ##   ##           ###   #     # #
     #       #       #       # # # #            #    #     # #
     #####   #       #       #  #  #           #      ###### ######
     #       #       #       #     #                       #       #
     #       #     # #     # #     #                 #     # #     #
     #        #####   #####  #     #                  #####   #####

                     C A L L    F O R    P A P E R S

                            THE THIRD ANNUAL
         IEEE SYMPOSIUM ON FPGAs FOR CUSTOM COMPUTING MACHINES
                            Napa, California
                               April 1995

            For more information, refer to the WWW URL page:  
             http://www.super.org:8000/FPGA/comp.arch.fpga


PURPOSE:   To bring together researchers to present  recent work 
in the use of Field Programmable Gate Arrays or other means for 
obtaining reconfigurable computing elements.  This symposium will 
focus primarily on the current opportunities and problems in this 
new and evolving technology for computing.

SOLICITATIONS:  Papers are solicited on all aspects of the use or 
applications of FPGAs or other means for obtaining reconfigurable 
computing elements in attached or special-purpose processors or 
co-processors, especially including but not limited to:
) Coprocessor boards for augmenting the instruction set of general-
  purpose computers.
) Attached processors for specific purposes (e.g. signal processing).
) Languages, compilation techniques, tools, and environments for 
  programming.
) Application domains.
) Architecture prototyping for emulation and instruction.
A special session will be organized in which venders of hardware and 
software can present new or upcoming products involving FPGAs for 
computing.

SUBMISSIONS: Authors should send submissions (4 copies, 10 pages
double-spaced maximum) before January 16, 1995, to Peter Athanas.  A 
Proceedings will be published by the IEEE Computer Society.  Specific 
questions about the conference should be directed to Kenneth Pocek.

SPONSORSHIP: The IEEE Computer Society and the TC on Computer Architecture.

CO-CHAIRS:
Kenneth L. Pocek					
Intel							
Mail Stop RN6-18					
2200 Mission College Boulevard 					
Santa Clara, California  95052				
(408)765-6705 voice (408)765-5165 fax			
kpocek@sc.intel.com					

Peter M. Athanas
Virginia Polytechnic Institute and State University	
Bradley Department of Electrical Engineering	
340 Whittemore Hall
Blacksburg, Virginia 24061-0111	
(703)231-7010 voice (703)231-3362 fax
athanas@vt.edu

ORGANIZING COMMITTEE:
Jeffrey Arnold, Supercomputing Research Center		
Brad Hutchings, Brigham Young Univ.
Duncan Buell, Supercomputing Research Center		
Tom Kean, Xilinx, Inc. (U.K). 
Pak Chan, Univ. California, Santa Cruz			
Wayne Luk, Oxford Univ.
Apostolos Dollas,  Technical Univ. of Crete		



Article: 298
Subject: CFP (final) -- Great Lakes Symposium on VLSI
From: shambhu@acsu.buffalo.edu (Shambhu J. Upadhyaya)
Date: Fri, 14 Oct 1994 22:45:38 GMT
Links: << >>  << T >>  << A >>
                          FINAL CALL FOR PAPERS
                    FIFTH GREAT LAKES SYMPOSIUM ON VLSI
              March 16 - 18, 1995  *  Buffalo, New York, USA

SPONSORS:
--------
IEEE Computer Society, IEEE Circuits and Systems Society,
ACM SIGDA and The State University of New York at  Buffalo.

LOCATION:
--------
Buffalo Marriott. The Buffalo/Niagara Falls metropolitan area is the second 
largest in the state of New York and offers many closeby tourist attractions. 

SCOPE:
-----
The Great Lakes Symposium on VLSI is an IEEE and ACM sponsored annual event 
and has now become a truly international conference, attracting participants 
from all over the world. Original and unpublished articles on all aspects of 
VLSI are invited.  Each article should clearly state its contributions which 
could be either theoretical or experimental. Major topics include  but are 
not limited to:
   1. Physical Design                  2. Analog and Mixed Signal ICs
   3. Memory Design                    4. Testing, Design for Testability
   5. Synthesis and Verification       6. Field Programmable Gate  Arrays
   7. Multichip Modules                8. Low Power Design
   9. VLSI Technology                 10. Specialized VLSI Architectures
  11. Hardware/Software Codesign      12. High Level Synthesis

SUBMISSION GUIDELINES:
---------------------
Authors should submit FOUR COPIES of a review package which
should include:
 * A TITLE PAGE containing title, most relevant topic number(s), name
   and affiliations of all authors(s), telephone number, Fax number and
   e-mail address  of the author to whom correspondence should be sent. 
 * A 50 word ABSTRACT.
 * An EXTENDED SUMMARY, not exceeding 2,000 words, describing 
   the problem(s) addressed and its significance, techniques used
   and specific results.

TUTORIALS:
---------
Tutorial sessions are planned for March 18, 1995. Two page proposals
on topics of current interest in VLSI are solicited.

KEYNOTE ADDRESS:
---------------
The keynote speaker will be James Meindl, Joseph M. Pettit Chair Professor
of Microelectronics, Georgia Institute of Technology. 

IMPORTANT DEADLINES:
-------------------
 * Submissions must be received by Nov. 4, 1994.
 * Notification of Acceptance by Dec 19, 1994.
 * Camera-Ready Paper due by Jan. 18, 1995 (Proceedings published 
   by IEEE Computer Society).

MAILING ADDRESS:
---------------
Program Chair, GLSVLSI
201  Bell Hall
Dept. of Electrical and Computer Engineering
The State University of New York at Buffalo
Buffalo, NY 14260

FOR FURTHER INFORMATION:
-----------------------
e-mail: glsvlsi@eng.buffalo.edu
FAX:(716) 645-3656;  Phone: (716) 645-2422

General Chair                  Program Chairs           Local Arrangements
-------------                  --------------           ------------------
S. Chakravarty          R. Sridhar & S. Upadhyaya         V. Demjanenko
(SUNY, Buffalo)              (SUNY, Buffalo)              (SUNY, Buffalo)

Program Committee
-----------------
J. Abraham                  C. Agagnostopoulos        V. D. Agrawal
(Univ. of Texas, Austin)    (Kodak, Rochester)        (AT&T Bell Labs)

J. Cong                     W. Debany                 M. I. Elmasry
(UCLA)                      (USAF, Rome Labs)         (Univ. of Waterloo, Canada)

E. Friedman                 D. Hill                   R. Jain 
(Univ. of Rochester)        (Synopsys)                (Univ. of Wisconsin) 

B. Kaminska                 S. Kang                   J. Oldfield            
(Ecole Polytechnic, Canada) (Univ. of Illinois)       (Syracuse Univ.)       

C. Papachristou             C. A. T. Salama           E. Sha                 
(Case Western Univ.)        (Univ. of Toronto, Canada)(Univ. of Notre Dame)  

N. Sherwani                 S. Verdonckt-Vandenbroek
(Western Michigan Univ.)    (Xerox, Rochester)

Steering Committee
-----------------
N. Sherwani
(Western Michigan Univ.)



Article: 299
Subject: Re: PALASM versions?
From: ian@myhost.subdomain.domain ()
Date: 15 Oct 1994 09:48:48 GMT
Links: << >>  << T >>  << A >>
Graham Seaman (seamang@westminster.ac.uk) wrote:
: Does anyone know what the latest version of PALASM is and if it
: is public domain? I'm unsure of the status of this and related
: software (eg. ABEL/CUPL) - are they p.d., shareware, proprietary
: or what? Any general info or ftp sites would be much appreciated!

I use Palasm ver 1.5a and MachXL version 1.3 
I have always found AMD Technical Support very good, frequent well organised
upgrades (I hope after saying that I am up to date 8-})


In the UK AMD have a person who liases with Universities and AMD have always
helped education. I got an OHP set, sets of manuals etc when I ran a short 
course for Industry on PLd design.

For a definitive answer to the "Public Domain" question contact Stuart
Tindell of AMD. I think his number is 01925 830380. Basically if you buy one
copy for the university to get the manual you can use the software in the
University, but do talk to AMD about it.

re  PLD software;

There was a copy of LOG/IC offered for use on the NET, see COMP.LSI.CAD faq,
I got it, tried it and didn't like it. I have been using PALASM for years
and for teaching purposes it is great, it has JK flipflop support with AMD's
version of the PALCE610 and I use it to augment first and final year
classes. The latest AMD MACHs are very useful CPLDs and although MACHXL is
not the best CPLD fitter about it is easy to use. 

I have put a DRAM refresher and a UART for a 68000 into MACHs and have a guy
working to put a big design into 5 or 6 MACHs connected together( Simulated
sparately). This is unusual since bigger FPGAs would normally be used but my
premise is that the very constraints Machs place on the designer produce
"better" designs and that MACHs are reprogrammable and cheap. The latest
MACH does not need a programmer, merely a lead to the parallel port and it
programs itself using JTAG signals; very neat trick!

I have used (read tried) abel, a very old version of Cupl and Orcad/PLD
For my money ORcad/PLD was the best for simple PLDs. It uses a Dongle and
costs money, hence my use of Palasm. I am away for a fortnight, email me
then for further discussion.
--
Regards from  Ian McCrum, Lecturer in Digital Systems and ECAD   
Email:  <IJ.MCCRUM@ulst.ac.uk>   WEB:  http://www.eej.ulst.ac.uk/index.html
POST-MAIL: Department of Electrial & Mechanical Engineering 
University of Ulster at Jordanstown,   Northern Ireland, BT37 0QB
Tel: +44 232 365131 ext 2364 or at home +44 247 882889
Fax: +44 232 362804 or preferably       +44 247 882894
------------------------------ ends ------------------------------------ 8-}




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