Xilinx Spartan3 Timing Problems - Whats about the chips

[itsme]

Hi all,
I am actually trying to switch a virtex2 project to spartan3. While
implementing the design I got several timing errors. One reason is that the
timing parameter "CLK to valid Data on Pad" (Tiockon). The data sheet of
virtex2 gives a value of 3.51ns. The value for spartan3 is 5.865ns - which
is much higher!!! . I got this value for spartan3 from Timing Analyzer
(Vers. 6.1.03i). In the newest datasheet of the spartan3 the chapter
"switching characteristics" is still quite empty.
So I really wonder what the XILINX guys do. Selling a chip without timing
specs? Is the spartan3 really much slower than the Virtex2?
Any comments? Any Spartan3 Users?

 

[Steven K. Knapp]

One quick thing ...

The Spartan-3 speeds files were recently updated to better reflect actual
silicon measurements. The previous speeds files, v1.27, were considered
PREVIEW level and were based primarily on simulation data. The recent
speeds files, v1.29, have now graduated to ADVANCE level, indicating that
the data is based on actual device characterization data measured on
Spartan-3 devices.

The particular parameter that you mentioned, Tiockon, ranged from 1.099 ns
to 6.197 ns using the -4, v1.27 speeds file. The value reported by the
timing analyzer, 5.865 ns, is specific for the way that your application was
implemented in the FPGA.

The Tiockon parameter now ranges from 1.127 to 4.000 ns using the -4, v1.29
speeds file. Re-running the design using the speeds files will improve the
performance. Similarly, setting some timing constraints on the particular
path may improve the results further still.

Optionally, the v1.29 speeds files now also include a -5 speed grade, which
is about 13% to 15% faster than the new revision of the -4 speed grade.

Now for the next logical question: Where can I obtain the new speeds file?

Please point your browser to the following web link.
http://www.xilinx.com/xlnx/xil_ans_display.jsp?getPagePath=18557

Then please download and extract the referenced ZIP file. Then, please
re-run the Xilinx implementation software to take advantage of the extra
performance.
---------------------------------
Steven K. Knapp
Applications Manager, Xilinx Inc.
General Products Division
Spartan-3/II/IIE FPGAs
http://www.xilinx.com/spartan3
---------------------------------
Spartan-3: Make it Your ASIC


Please check which version
"itsme" <itsme@gmx.de> wrote in message
news:buo43e$ji4$06$1@news.t-online.com...

Hi all,
I am actually trying to switch a virtex2 project to spartan3. While
implementing the design I got several timing errors. One reason is that
the
timing parameter "CLK to valid Data on Pad" (Tiockon). The data sheet of
virtex2 gives a value of 3.51ns. The value for spartan3 is 5.865ns - which
is much higher!!! . I got this value for spartan3 from Timing Analyzer
(Vers. 6.1.03i). In the newest datasheet of the spartan3 the chapter
"switching characteristics" is still quite empty.
So I really wonder what the XILINX guys do. Selling a chip without timing
specs? Is the spartan3 really much slower than the Virtex2?
Any comments? Any Spartan3 Users?

 

[Anjan]

Hi Steve,
I have a question about spartan 3. We are xilinx customers and have
ordered spartan 3 device long back. Also we have ordered engineering
samples. But the distributor can't answer when the order comes to us.
Can you please tell me whether spartan 3 1.5m devices are being
shipped(engineering samples)?
Anjan

 

[Morten Leikvoll]

"Anjan" <anjanr@yahoo.com> skrev i melding
news:5a5faf7b.0401222006.2ad19b2b@posting.google.com...

Hi Steve,
I have a question about spartan 3. We are xilinx customers and have
ordered spartan 3 device long back. Also we have ordered engineering
samples. But the distributor can't answer when the order comes to us.
Can you please tell me whether spartan 3 1.5m devices are being
shipped(engineering samples)?
Anjan

We have the same problem here.. Prototyping boards are ready and we have
been promised devices for a quite long time. I hear xilinx has problems in
production, wich is boring for us having quite a few boards ready :/ We were
also promised a few samples with 2.5v max io, but they showed up either.
From now on , we will be VERY careful to go for future devices. However, I
hope they will appear in the near future.

 

[Ray Andraka]

This sounds like a spartan2 deja-vu, no?

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

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