12 Apr
2013
12 Apr
'13
5:22 a.m.
At the flip chip level it would include definitions for DEC3639B (aka 2N3639), D662 (aka
1N645) and D664 (aka 1N3606). Pretty standard stuff to read the datasheets and make SPICE
models.
A real design (or redesign) would pay extensive attention to margining just like the
original designers did. SPICE can help verify to an extent, but all the details you
actually need for design of RTL/DTL/flip-flop are in the 60's TI transistor design
books.
12 Apr
12 Apr
6:24 a.m.
On Apr 12, 2013, at 8:22 AM, "Shoppa, Tim" <tshoppa at wmata.com> wrote:
At the flip chip level it would include definitions
for DEC3639B (aka 2N3639), D662 (aka 1N645) and D664 (aka 1N3606). Pretty standard stuff
to read the datasheets and make SPICE models.
A real design (or redesign) would pay extensive attention to margining just like the
original designers did. SPICE can help verify to an extent, but all the details you
actually need for design of RTL/DTL/flip-flop are in the 60's TI transistor design
books.
Yeah, I can't stress enough that SPICE is NOT A DESIGN TOOL.
It is a verification tool. People coming out of college in
my generation fail to understand that because even a cheap
PC runs complex SPICE simulations in a fraction of the time
that the original machines did... that, and they're lazy.
The late, great Bob Pease was fond of pointing out that both
SPICE and spreadsheets are great at giving you the answer
you want, not the answer you need (the right one).
- Dave
7:09 a.m.
On 04/12/2013 09:24 AM, David Riley wrote:
Yeah, I can't stress enough that SPICE is NOT A
DESIGN TOOL.
It is a verification tool. People coming out of college in
my generation fail to understand that because even a cheap
PC runs complex SPICE simulations in a fraction of the time
that the original machines did... that, and they're lazy.
The late, great Bob Pease was fond of pointing out that both
SPICE and spreadsheets are great at giving you the answer
you want, not the answer you need (the right one).
And indeed, famously threw a computer off the roof of the parking garage,
stating that he then knew that THAT computer would never lie to him again.
I'm not certain that this was due to SPICE, but I'd bet it was, given his
overall feelings toward it.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
10:18 a.m.
On 04/12/2013 10:09 AM, Dave McGuire wrote:
On 04/12/2013 09:24 AM, David Riley wrote:
I use spice, XFTD, NEC2, and on and on...
They are all nice tools if you keep one thing in mind.
GIGO
For those that have forgotten of never heard that
"Garbage In produces Garbage Out". To avoid that
don't put garbage in. What is garbage? Simulations
with far to many unknown and parasitic items
unaccounted for.
If one takes the time and efffort to correlate and qualify the
results then the answers should be valid and often are.
That means if you use a mbt2222 transistor in spice
the first task is that spice realization a valid one!
Like anything computer simulation and test can be
good or bad and as a designer you can control that.
Allison
Yeah, I can't stress enough that SPICE is NOT
A DESIGN TOOL.
It is a verification tool. People coming out of college in
my generation fail to understand that because even a cheap
PC runs complex SPICE simulations in a fraction of the time
that the original machines did... that, and they're lazy.
The late, great Bob Pease was fond of pointing out that both
SPICE and spreadsheets are great at giving you the answer
you want, not the answer you need (the right one).
And indeed, famously threw a
computer off the roof of the parking garage,
stating that he then knew that THAT computer would never lie to him again.
I'm not certain that this was due to SPICE, but I'd bet it was, given his
overall feelings toward it.
-Dave
10:26 a.m.
On Apr 12, 2013, at 1:18 PM, Allison <ajp166 at verizon.net> wrote:
I use spice, XFTD, NEC2, and on and on...
They are all nice tools if you keep one thing in mind.
GIGO
For those that have forgotten of never heard that
"Garbage In produces Garbage Out". To avoid that
don't put garbage in. What is garbage? Simulations
with far to many unknown and parasitic items
unaccounted for.
If one takes the time and efffort to correlate and qualify the
results then the answers should be valid and often are.
That means if you use a mbt2222 transistor in spice
the first task is that spice realization a valid one!
Like anything computer simulation and test can be
good or bad and as a designer you can control that.
Absolutely. All I'm saying is that I work with (or rather
clean up after) a lot of people who use simulators, whether
SPICE or a logic simulator or whatever, as a DESIGN tool.
They end up with fragile code that looks like garbage
because e.g. they'll iteratively tweak their design by
simulating it and throwing in pipeline delays until it all
"works", at least under the specific conditions of the sim.
So yes, like any good tool, you have to be using it the
right way. Just because you CAN use a circular saw in
place of a pencil or a hammer doesn't mean you SHOULD.
It's important to do the math BEFOREHAND, ideally on paper
or paper equivalent, and simulate AFTERWARD. That is,
sadly, a lesson I'm still occasionally relearning.
- Dave
8:26 a.m.
The late, great Bob Pease was fond of pointing out
that both SPICE
and spreadsheets are great at giving you the answer you want, not the
answer you need (the right one).
There's a bit of doggerel which I saw presented as coming from a
graffito in a terminal room (back when there _were_ terminal rooms, of
course):
I really hate this damn machine
I wish that they would sell it
It never does quite what I want
But only what I tell it
Computers are really good at answering the question you asked, not the
question you meant to ask - or the question you thought you asked - or
the question you actually need the answer to....
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
13 Apr
13 Apr
10:15 a.m.
On 4/12/2013 5:22 AM, Shoppa, Tim wrote:
At the flip chip level it would include definitions
for DEC3639B (aka 2N3639), D662 (aka 1N645) and D664 (aka 1N3606). Pretty standard stuff
to read the datasheets and make SPICE models.
A real design (or redesign) would pay extensive attention to margining just like the
original designers did. SPICE can help verify to an extent, but all the details you
actually need for design of RTL/DTL/flip-flop are in the 60's TI transistor design
books.
Last night I tinkered with what was posted and I threw in some garbage
(actually for the diodes, I put in 1N4004 and for the transistor, a
2N918) just to get spice to shut up and it proved I had the right sort
of syntax to satisfy the basic parser.
Thanks to Tim, I'll find better models now that I know what I'm looking for.
This is just an exercise to bring up a corner of what is out there I've
not touched to see how it works. This is a fairly simple circuit, so
should be possible to spot garbage in - garbage out symptoms warned of
by Alison.
thanks
Jim
14 Apr
14 Apr
12:01 p.m.
On 04/13/2013 01:15 PM, Jim Stephens wrote:
1N4004, I assume thats for the power supply. That diode is so slow I
use them as
PIN attenuators at RF. The 1n914 is more suitable. Another good choice
is the
1n5711. Those are leaded parts there are SMT parts worth looking at as
well.
On 4/12/2013 5:22 AM, Shoppa, Tim wrote:
> At the flip chip level it would include definitions for DEC3639B (aka
> 2N3639), D662 (aka 1N645) and D664 (aka 1N3606). Pretty standard
> stuff to read the datasheets and make SPICE models.
>
Replace DEC with 2n as is 2n3639. The D662 and friends, there are
better part out there.
generally its possible to do what they did and do it much better.
A real design
(or redesign) would pay extensive attention to
margining just like the original designers did. SPICE can help verify
to an extent, but all the details you actually need for design of
RTL/DTL/flip-flop are in the 60's TI transistor design books.
Last night I
tinkered with what was posted and I threw in some garbage
(actually for the diodes, I put in 1N4004 and for the transistor, a
2N918) just to get spice to shut up and it proved I had the right sort
of syntax to satisfy the basic parser.
Thanks to Tim, I'll find better models now that I
know what I'm
looking for.
This is just an exercise to bring up a corner of what is out there
I've not touched to see how it works. This is a fairly simple
circuit, so should be possible to spot garbage in - garbage out
symptoms warned of by Alison.
Transistors out there use the most common types either leaded or SMT
that have switching
fast enough. Minimally 2n3904, 4401, 4124 or their complements SMT parts
will have
MBT prefix.
Allison
4271
days inactive
4273
days old
7 comments
6 participants
participants (6)
-
ajp166＠verizon.net -
fraveydank＠gmail.com -
jws＠jwsss.com -
mcguire＠neurotica.com -
mouse＠Rodents-Montreal.ORG -
tshoppa＠wmata.com