Maker Pro
Maker Pro

Rise and fall time of TI 'F' family

Hi,
I am making a high speed digital system board which will demonstrate
the pitfalls of high speed design, as part of a final project.
I am using a TI's F family to give me small rise and fall times to
demonstrate transmission line effects. However, I cant find the rise
and fall times of their F family anywhere. Does anyone know what it is
or better send me the datasheet where it is specified.

Kunal
 
J

John Fields

Jan 1, 1970
0
Hi,
I am making a high speed digital system board which will demonstrate
the pitfalls of high speed design, as part of a final project.
I am using a TI's F family to give me small rise and fall times to
demonstrate transmission line effects. However, I cant find the rise
and fall times of their F family anywhere. Does anyone know what it is
or better send me the datasheet where it is specified.

---
Go to:

http://www.ti.com

Search for the part number you have in mind, navigate to the data
sheet, and read it.
 
K

Kunal

Jan 1, 1970
0
Goto TI.com yourself.
Pick the 74F175 datasheet or any other F chip and see that only the
propagation delay is specified.

Kunal
 
J

John Larkin

Jan 1, 1970
0
Hi,
I am making a high speed digital system board which will demonstrate
the pitfalls of high speed design, as part of a final project.
I am using a TI's F family to give me small rise and fall times to
demonstrate transmission line effects. However, I cant find the rise
and fall times of their F family anywhere. Does anyone know what it is
or better send me the datasheet where it is specified.

Kunal


I have a National databook that cites their FAST as having Tr of
1.8-2.8 ns, and Tf of 1.6-2.6 ns.

Some of the TinyLogic cmos parts are well under 1 ns.

John
 
J

John Larkin

Jan 1, 1970
0
Of course, if you want to see where those pitfalls can **really** hurt
you, go to
http://www.infinibandta.org/

(I'm one of the contributing authors of the spec :)

Cheers

PeteS


How much of channel-based i/o, do you think, is based on the bloat of
Windows? People are putting stuff like tcp/ip stacks in hardware these
days; what's next? Word could sure use some hardware assist!

IBM's Cell thing is sort of the ultimate architecture... a processor
per process. That would result in interesting OS architectures.

Funny: I went to the Semi show on Tuesday. Microsoft had a big display
featuring a Microsoft-automated Hummer3 (which looks like a
horizontally-stretched SUV... really dumb looking) and I joked to a
guy about where were the ctrl/alt/delete buttons. He replied that
they'd been trying to boot it up all morning, no luck so far. And
they're pushing XpEmbedded for wafer fabs!

John
 
K

Kunal

Jan 1, 1970
0
Thanks John. Ill use the 2.8ns seconds for my calculations. That would
be the 'worst' case for me.
I am making a SI demo board with 5 experiments
1) Transmission line effects : See ringing/reflections on long traces
and how terminations reduce that
2) Crosstalk
3) Power Integrity : heavy loads connected to a DIP package with a thin
long trace to the power rails
4) Metastability : Straight out of Howard Johnson's book
5) Multi-drop bus : SHow how heavily loading an output can slow rise
time due to caps, thus reduce system freq.

My main problem is that I have to be able to see all these effects on a
100MHz scope (Tr = 3.5ns). I used howard johnson's 21:1 probe
configuration (in his book), lets hope it works out.
 
J

John Larkin

Jan 1, 1970
0
4) Metastability : Straight out of Howard Johnson's book

Maybe include an LS flipflop. They're notorious for metastability and
are slow enough that your scope could catch them. I've heard of an
LS74 oscillating for a microsecond if properly teased, and making
audible clocks on an FM radio every time is went metastable.

John
 
J

Joerg

Jan 1, 1970
0
Hello John,
Funny: I went to the Semi show on Tuesday. Microsoft had a big display
featuring a Microsoft-automated Hummer3 (which looks like a
horizontally-stretched SUV... really dumb looking) and I joked to a
guy about where were the ctrl/alt/delete buttons. He replied that
they'd been trying to boot it up all morning, no luck so far. And
they're pushing XpEmbedded for wafer fabs!

At least there is a clear quality indicator for cars after they have
been in production for a while (sometimes even if they haven't): The
quality and reliability is inversely proportional to the cost of the
extended warranty. That rule hasn't failed me since my college days.

Regards, Joerg
 
B

Bob Stephens

Jan 1, 1970
0
The
quality and reliability is inversely proportional to the cost of the
extended warranty. That rule hasn't failed me since my college days.

That's a pretty good one.


Bob
 
J

Joerg

Jan 1, 1970
0
Hello Bob,
That's a pretty good one.

When I bought my first "real" car I had one more: I picked a few models
that fit my needs. Then I went to the dealers of each brand and,
pretending I already had one, asked for the price of a spare part. The
left mirror, since that gets clipped a lot. BIG differences. Then I went
ahead and bought the car where this part had the most reasonable price.

Oh, and the company that made the mirrors was close to where I lived. So
I knew what these mirrors cost. They were all in the same ball park.

Regards, Joerg
 
K

Ken Smith

Jan 1, 1970
0
John Larkin said:
IBM's Cell thing is sort of the ultimate architecture... a processor
per process. That would result in interesting OS architectures.

Remember the "transputer" with many ALUs for a single task. Now If IBM
would just make a multiple version of a chip like that so that each thread
in XP could have thousands of ALUs to work on it, a 3GHz machine might be
able to keep up with RS232 data coming from an 8051. That would be nice
wouldn't it.
 
B

Bob Monsen

Jan 1, 1970
0
Ken said:
Remember the "transputer" with many ALUs for a single task. Now If IBM
would just make a multiple version of a chip like that so that each thread
in XP could have thousands of ALUs to work on it, a 3GHz machine might be
able to keep up with RS232 data coming from an 8051. That would be nice
wouldn't it.

You sound bitter... ;)

I just upgraded my sister in law to 512 MB of memory, so she can surf
the web. She was running 128 MB, and XP SP2 was SO SLOW that she simply
could not do anything. After upgrading, the system was zippy.

Same situation for a neighbor of mine, who, in addition to having over
600 bits and pieces of spyware and a rootkit trojan on the system, was
running McAfee Internet Security. After I reformatted her disk,
reinstalled everything, and upgraded to SP2, her 128 MB system was quite
slow. Turns out McAfee sucks up about 50% of the cpu just sitting there,
according to Task Manager. McAfee is actually worse than having the
trojan; at least the trojan was free.

Now, my first home computer was a 128K (not M, K) mac. At work, I had
used minicomputers with a maximum address space of 2 MB, that supported
compiles for 20 users.

My suspicion is that microsoft made a deal with compact, dell, etc, at
some point in the past (probably when they were killing off OS/2) that
guaranteed that if the mfgrs sold dos/windows, and made it hard to get
anything else, then MS would make the OS would bloat so badly, and
perform so miserably, that people would be required to constantly
upgrade their computers to get the same performance they used to get.

(whatever happened to the connection machine? The BBN butterfly
architecture? All of that massively parallel stuff has fallen by the
wayside, probably in favor of massively parallel stuff built on racks of
hundreds of cheap PCs, running Linux.)

--
Regards,
Bob Monsen

If a little knowledge is dangerous, where is the man who has
so much as to be out of danger?
Thomas Henry Huxley, 1877
 
J

Joerg

Jan 1, 1970
0
Hello Bob,
Now, my first home computer was a 128K (not M, K) mac. At work, I had
used minicomputers with a maximum address space of 2 MB, that supported
compiles for 20 users.

My father worked on the design of a system that had 2K of memory and
controlled a complete cold-rolled steel production line. IIRC it never
crashed. Probably around the time Bill Gates was in first grade.
My suspicion is that microsoft made a deal with compact, dell, etc, at
some point in the past (probably when they were killing off OS/2) that
guaranteed that if the mfgrs sold dos/windows, and made it hard to get
anything else, then MS would make the OS would bloat so badly, and
perform so miserably, that people would be required to constantly
upgrade their computers to get the same performance they used to get.

(whatever happened to the connection machine? The BBN butterfly
architecture? All of that massively parallel stuff has fallen by the
wayside, probably in favor of massively parallel stuff built on racks of
hundreds of cheap PCs, running Linux.)

One technique is simply not to upgrade. I am writing this on what some
would consider a clunker. Pentium 2, 64MB of RAM, old NT. Mozilla seems
to have some memory leaks so once in a while I have to re-start and it
speeds up again. But not more often than once a day. You can kind of see
when it's time because it starts grinding on the hard disk page file too
much. This is the machine I use to keep online data sheets on screen
while designing on the other PC.

I just don't follow the bloat unless there is a very compelling reason.
The only reason I can think of is if some SW isn't backwards compatible.
However, I tend not to buy such SW so only a client requirement could
trigger that situation.

Regards, Joerg
 
G

Guy Macon

Jan 1, 1970
0
Joerg said:
One technique is simply not to upgrade. I am writing this on what some
would consider a clunker. Pentium 2, 64MB of RAM, old NT.

This can work well, but not if you are using a Microsoft OS and are
not behind a standalone firewall. The good news is that FREESCO
running on an old 486 makes a great firewall.
 
T

Tom MacIntyre

Jan 1, 1970
0
Hello Bob,


My father worked on the design of a system that had 2K of memory and
controlled a complete cold-rolled steel production line. IIRC it never
crashed. Probably around the time Bill Gates was in first grade.

I have no idea how much memory it had, but the TV event switching
computer we used when I first started in broadcasting in 1980 used
about a 15 or 20 foot length of ticker tape to load its operating
software.

It was replaced in 1987 or so with an XT with 256k of RAM. Only 256k
was enabled, I later discovered from examining the silkscreen on the
board, so I righted that wrong, and upped it to the full 640k that was
actually there. No difference in performance that I could tell. :)

Tom
 
Joerg said:
One technique is simply not to upgrade. I am writing this on what some
would consider a clunker. Pentium 2, 64MB of RAM, old NT.

Even better is to upgrade hardware but not OS. 98 is still ok, and is
quick on a modern PC. I mean why pay MS yet again for something thats
only going to cripple your work output?

Nother approach is to have 2 sets of each app, one full featured and
the other a quick version. Only use the full bloat version when its
needed, most of time use the quickie. Makes the whole business miles
faster.


NT
 
Bob said:
I just upgraded my sister in law to 512 MB of memory, so she can surf
the web. She was running 128 MB, and XP SP2 was SO SLOW that she simply
could not do anything.

I'd like to upgrade mine too... How do you open the head? Does the port
labelled 'mouth' take CD or DVD? what are those ear connectors for?

I've been trying the nose connectors, but there seem to be
compatibility issues, the system refuses to accept any of the nose
plugins I've tried, even though they fit correctly. What am I doing
wrong?

slow. Turns out McAfee sucks up about 50% of the cpu just sitting there,
according to Task Manager. McAfee is actually worse than having the
trojan; at least the trojan was free.

Once a day scan is a more sensible setting, unless in a hostile
situation. Scanning every file every time does cane performance.

My suspicion is that microsoft made a deal with compact, dell, etc, at
some point in the past (probably when they were killing off OS/2) that
guaranteed that if the mfgrs sold dos/windows, and made it hard to get
anything else, then MS would make the OS would bloat so badly, and
perform so miserably, that people would be required to constantly
upgrade their computers to get the same performance they used to get.

That certainly is the game, or part of it, though theres no deal
needed. MS builds in more features to impress buyers: these features
mean more code which means slower, and that means constant upgrading of
hardware. Sometimes its worth it, eg the move from 3.1 to 95. Sometimes
its not, eg xp.

And why do the hardware sellers push MS? Because its what will bring
the customer back in 1-2 years for another system. And another, and
another.

MS has tried making deals that hardware vendors only sell MS OSes or
else MS wont sell them Win, but IIRC it was ruled illegal.

If you dont want to play that game, stick with an older less bloated OS
and non bloated apps.


NT
 
J

Joerg

Jan 1, 1970
0
Hello NT,
Nother approach is to have 2 sets of each app, one full featured and
the other a quick version. Only use the full bloat version when its
needed, most of time use the quickie. Makes the whole business miles
faster.

Better yet try to live with the simple things. I run my whole business
accounting on MS-Works database. It works like a champ, is simple and
fast. Then I use Works spreadsheet for EE math and other fancy stuff. It
is amazing what you can do with a $100 software that often even comes
free with a new PC.

Regards, Joerg
 
Top