Check Out This Month's "Microwaves & RF" Magazine

John Larkin · Jul 17, 2004

On 16 Jul 2004 17:19:56 -0700, Winfield Hill
<Winfield_member@newsguy.com> wrote:

John Larkin wrote:
Hey, Here's my latest gadget:

http://www.highlandtechnology.com/DSS/P400DS.html

Actually, three of us here worked on this for about three years as
sort of a background project, when we didn't have a paying customer
screaming for delivery on something. I never appreciated how much
hassle a benchtop instrument would really be until this got serious. A
VME or PCI board is blindingly simple compared to all the stuff you
have to put into a box like this. And by the time you finish it,
things have changed so much you're dying to redesign it again from
scratch. m.u.s.t..r.e.s.i.s.t..t.e.m.p.t.a.t.i.o.n.

I'm impressed. OK, how much does it cost? I'd like to see a
scaled-back version, with 0.1ns resolution, and 8 channels, etc.
Smaller perhaps and cheaper. Scalable to 16, 24, 32 channels.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

$3840, just a hair under brand S. I'd like to do a cheaper version
with a simpler technology, but my marketing people don't want us to
kill our own product just when it's getting up to speed. If you want a
lot of DDG channels, a rack full of VME modules is the usual way to
go. It's more common, actually, for people to want a lot of channels
of time measurement, as opposed to many channels of time delay.
Exceptions are big laser arrays like NIF, and implosion experiments.

Actually, a huge-number-of-channels delay gadget would be a fun thing
to do, but I'd guess you wouldn't sell many.

John

Kevin Aylward · Jul 20, 2004

dorinelu wrote:

Wich is the best sofware for electronic simulation?

No such thing as "best".

Download the demos from companies and see which one you like the most.

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Paul Burridge · Jul 20, 2004

On 20 Jul 2004 04:56:44 -0700, dorinel_ionascu@mymail.ro (dorinelu)
wrote:

Wich is the best sofware for electronic simulation?

LT Spice is very highly regarded among the cognoscenti here and it's
free from those great guys at Linear Technology: www.linear.com

Multisim, OTOH, has probably the worst cost/benefit ratio and is best
avoided.

--

"What is now proved was once only imagin'd." - William Blake, 1793.

Leon Heller · Jul 20, 2004

"dorinelu" <dorinel_ionascu@mymail.ro> wrote in message
news:5fd770a8.0407200356.7f7acbf9@posting.google.com...

Wich is the best sofware for electronic simulation?

SIMetrix SPICE has a very good reputation, it comes with the Pulsonix
software I use.

Leon

chrisv · Jul 20, 2004

tholen@antispam.ham wrote:

chrisv writes:

James Brown <Godfather@Of.Soul> writes:

Non sequitur.

PAPA'S GOT A BRAND NEW BAG !

Also non sequitur.

Amazing.

What is allegedly so amazing about Brown's non sequitur, chrisv?

You are erroneously presupposing that my "amazing" referred to Brown's
non sequitur, Tholen.

Illogical, chrisv;

Classic unsubstantiated and erroneous claim.

On your part, chrisv.

Illogical, Tolen.

what else could it be?

It could be that I think that you are amazing, Tholen.

Illogical, given that that would be non sequitur, chrisv.

Irrelevant, even if true, which it's not, Tholen.

Chaos Master · Jul 21, 2004

cat /etc/passwd | grep dorinelu returns:

Wich is the best sofware for electronic simulation?

LTSpice.

[]s
--
Chaos MasterŽ, posting from Brazil. REPLY TO GROUP!
"Bring me to life..."
visit http://marreka.no-ip.com. MSN: wizard_of_yendor@hotmail.com

Chaos Master · Jul 21, 2004

cat /etc/passwd | grep Leon Heller returns:

SIMetrix SPICE has a very good reputation, it comes with the Pulsonix
software I use.

I second this. I have used SIMetrix Intro, enough for small projects. Very
good. Just a few problems (it sometimes gave me access violations, if the AVG
anti-virus was running in the background), *but I think they were caused by my
PC*.

[]s
--
Chaos MasterŽ, posting from Brazil. REPLY TO GROUP!
"Bring me to life..."
visit http://marreka.no-ip.com. MSN: wizard_of_yendor@hotmail.com

Jim Thompson · Jul 21, 2004

On 20 Jul 2004 10:11:32 -0700, s.fg@wanadoo.es (Sergio Gascón) wrote:

I am designing a circuit in Cadence using spectre but I want to
simulate an extermal power mos which model is not included in the
foundry kit (obviously). The mosfet is the irf3315 and the only model
I have found is a very simple model to be used in pspice. This model
could be enough for these first step simulations but I am not familiar
with model conversions and there are some lines I do not know how to
translate:

FI2 79 VFI2 -1

Since I see a similar device below listed correctly I presume a typo
here and it should be

FI2 7 9 VFI2 -1

This is a current-controlled current source, output is from node 7 to
node 9, "current measuring" voltage source is VFI2, and gain is -1

VFI2 4 0 0

Voltage

EV16 10 0 9 7 1

Voltage-controlled voltage source, output at nodes 10,0, input at
nodes 9,7, gain is 1

FI1 7 9 VFI1 -1

Same device type as FI2

VFI1 11 6 0

Voltage source

If anyone could give me a hand with the meaning or translation of this
lines it would be fantastic. A model suitable for spectre would be
even better but I think it is impossible.

480)460-2350 | |
| E-mail Address at Website Fax

Kevin Aylward · Jul 23, 2004

Paul Burridge wrote:

On Thu, 22 Jul 2004 15:40:56 -0700, Roy Lewallen <w7el@eznec.com
wrote:

John Woodgate wrote:

If you have the s-parameters (preferably at more than one
frequency) you can make an equivalent circuit and use that in
SPICE. OK for r.r.a.h., I expect, but maybe not for s.e.d.. (;-)

A SPICE model constructed from s-parameter data can be useful, but
only under a very limited set of circumstances. First, it's valid
only if the signal level is small. This is often not the case with
an oscillator, unless external amplitude limiting is used (as it
should be). Also, a circuit very often can oscillate (as well as
amplify) at more than one frequency, and a SPICE model that's valid
only over a small frequency range won't predict this behavior. (A
single frequency SPICE model is simple to construct from s-parameter
data, but one covering a wide range of frequencies is just about
impossible.)

What's the problem in making up your own (accurate) wideband Spice
model from measurements you make yourself on a VNA?
I'm also under the impression that a parameter such as (for example)
the base-collector capacitance lifted from a manufacturer's data sheet
and inserted into Spice won't work accurately. Why is this?

It should work ok if you use the right value. Go and have a look at the
basic spice capacitance equations. You can read I take it?

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Paul Burridge · Jul 23, 2004

On Fri, 23 Jul 2004 13:11:10 GMT, "Kevin Aylward"
<salesEXTRACT@anasoft.co.uk> wrote:

It should work ok if you use the right value. Go and have a look at the
basic spice capacitance equations. You can read I take it?

Hehe! You're such a card, Kev. I notice that you've avoided responding
to the original question in the thread, though! Bit beyond you was it?
;->
--

"What is now proved was once only imagin'd." - William Blake, 1793.

Kevin Aylward · Jul 23, 2004

Paul Burridge wrote:

On Fri, 23 Jul 2004 13:11:10 GMT, "Kevin Aylward"
salesEXTRACT@anasoft.co.uk> wrote:

It should work ok if you use the right value. Go and have a look at
the basic spice capacitance equations. You can read I take it?

Hehe! You're such a card, Kev. I notice that you've avoided responding
to the original question in the thread, though! Bit beyond you was it?

There are loads of postings I don't respond to. Usuallly its the trivial
ones.

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

ldg · Jul 23, 2004

On 20 Jul 2004 04:56:44 -0700, dorinel_ionascu@mymail.ro (dorinelu)
wrote:

Which is the best sofware for electronic simulation?

This is really a large question. There are many analog and digital
simulators, each optimized for a specific market segment.

For instance, IC designers have used Hspice or Smartspice for smaller
circuits where accuracy is really important. This class of simulator
is analog-only. (There are many others.)
http://www.silvaco.com/products/circuit_simulation/smartspice.html
http://www.synopsys.com/products/mixedsignal/hspice/hspice.html

Full chip simulators are now available that use various schemes to
reduce simulation time. From what I've seen, these simulators are
really expensive and beyond my needs. Nanosim is an example:
http://www.synopsys.com/products/mixedsignal/nanosim/

There are mixed mode simulators that allow both analog and
intermediate sized logic circuits to be simulated simultaneously.
Pspice AD is a good example of this. Since Pspice was bought by Orcad
and Orcad was bought by Cadence, this simulator has stayed mostly
targeted at board work, though I've used it in the past for IC design,
as have others.
There are many others in this category as well, including SIMetrix
SPICE. http://www.catena.uk.com/index.html

Many of these simulators have been created by combining Berkeley
spice and a digital simulator called Xspice that was created by
Georgia Tech http://users.ece.gatech.edu/~mrichard/Xspice/

A high end mixed simulator example might be Cadence's Spectre
http://www.cadence.com/products/custom_ic/spectre/index.aspx

Yet another interesting mixed mode simulator is the Smash simulator by
Dolphin: http://www.dolphin.fr/medal/smash/smash_overview.html . This
one does vhdl and verilog while simultaneously doing detailed analog
simulations.

Saber is sort of in a class by itself in that it can easily simulate
physical as well as electrical/electronic devices.
http://www.synopsys.com/products/mixedsignal/saber/saber_ds.html
It can, for instance, simulate entire car systems where a mechanical
device operates a sensor, which then is processed by electronics, then
drives a mechanical actuator. It would also be great for modeling
MEMs devices since both the physical device and interface circuits can
be modeled at once. It also does mixed mode. New models are easily
made using its MAST programming language. Since both Saber and Hspice
were owned by Avant!, the actual hspice models have been ported to it,
making it even more acceptable for low level IC design. I used this
program years ago to design IR focal planes. This is a great
simulator, but expensive.

I'm not a user of pure digital simulators, but I'm sure others in this
group are knowledgeable in this area.

Many here use the free LTspice simulator. It has a great user
interface and the support is unusually good. It's primary purpose is
for board level work, allowing Linear Technology customers to easily
simulate circuits with Linear Technology devices, though the author's
intention clearly is beyond this. The user interface is so good I've
often wished it would front-end for other simulators I use in my work
(hspice, smartspice). To me, however, the simulator itself doesn't
quite meet my needs for doing IC design. I'd use it for small
circuits and board level work.

Hope any of this helps. Much of this is subjective, and everyone has
their own opinion.

Regards,
Larry

Roy Lewallen · Jul 23, 2004

Paul Burridge wrote:

What's the problem in making up your own (accurate) wideband Spice
model from measurements you make yourself on a VNA?

The problem is in figuring out what components to use, and how to
connect them, to imitate the (usually complex) way the s-parameter data
change with frequency. If the circuit is very simple, with one or two
components dominating the frequency response, it's quite possible. But
an MMIC isn't likely to be in that category. Grab some s-parameters from
an MMIC data sheet and give it a try, and you'll see what I mean. Then,
for a real eye-opener, dust off your VNA, make your own measurements,
and see how close they come to the manufacturer's.

I'm also under the impression that a parameter such as (for example)
the base-collector capacitance lifted from a manufacturer's data sheet
and inserted into Spice won't work accurately. Why is this?

Depends on the accuracy you're looking for and the range of operation. A
simple transistor model, such as one made from data sheet parameters,
will work surprisingly well for many purposes. But it's a linear model,
good only for a very limited range of operating conditions. In a real
transistor, Ccb (for example) varies strongly with Vcb. A complete SPICE
model accounts for this, along with several tens of other factors
ignored by simpler models. If you're operating the transistor with a
small signal and at voltages, currents, and frequency near those on the
data sheet, a simple data sheet model can often be good enough.

One of my most treasured texts is _Modeling The Bipolar Transistor_, by
Ian Getreu, published by Tektronix in 1976 (Tek P/N 062-2841-00). It's
long out of print, but it might show up on eBay or elsewhere from time
to time and I highly recommend it to anyone seriously interested in
transistor modeling.

Roy Lewallen, W7EL

Paul Burridge · Jul 23, 2004

On Fri, 23 Jul 2004 11:06:28 -0700, ldg <ldg@ix.netcom.com> wrote:

Many here use the free LTspice simulator. It has a great user
interface and the support is unusually good. It's primary purpose is
for board level work, allowing Linear Technology customers to easily
simulate circuits with Linear Technology devices, though the author's
intention clearly is beyond this.

I agree entirely. And the free and generous support of the head
developer on this group - thought it can't of course be taken for
granted - is another bloody good reason to look no further.
--

"What is now proved was once only imagin'd." - William Blake, 1793.

Mike Engelhardt · Jul 23, 2004

Deji,

Thanks a lot Mike. That FIXED the problem!!!! I had
thought that M and G prefixes meant Mega and Giga in
spice. i may have to double-check that.

G does mean Giga, as M should mean Mega, but SPICE is
case insensitive(and predates the convention that
metric multipliers greater than 1 should all start
with a capital letter so KHz is correct instead of kHz)
so Meg is used to mean Mega and m(or M) is milli.

--Mike

Paul Burridge · Jul 23, 2004

On Fri, 23 Jul 2004 22:36:02 GMT, "Mike Engelhardt" <nospam@spam.org>
wrote:

Deji,

Thanks a lot Mike. That FIXED the problem!!!! I had
thought that M and G prefixes meant Mega and Giga in
spice. i may have to double-check that.

G does mean Giga, as M should mean Mega, but SPICE is
case insensitive(and predates the convention that
metric multipliers greater than 1 should all start
with a capital letter so KHz is correct instead of kHz)
so Meg is used to mean Mega and m(or M) is milli.

Happily, (as you will know!) One can simply use exponents instead, for
the avoidance of doubt.
--

"What is now proved was once only imagin'd." - William Blake, 1793.

Jim Thompson · Jul 23, 2004

On 23 Jul 2004 10:32:38 -0700, s.fg@wanadoo.es (Sergio Gasc?n) wrote:

I usually work with Cadence and this is the first time I use spice for
years. I have translated and spice model into spectre to simulate it
with cadence and I would like to compare the results with the original
model.

The file is a *.spi file. What should I do in order to place this
transistor into a schematic?

I know it is a very basic question but I tried doing it like I did
years ago and something goes wrong. The .slb file is missing.

I am using Orcad 9.2.

Thanks in advance.

480)460-2350 | |
| E-mail Address at Website Fax

Roy Lewallen · Jul 24, 2004

Paul Burridge wrote:

. . .
Your reply throws up another interesting question: which model
parameters are essential for adequate modeling of the BJT up into UHF,
and which, by their omission, call the model's reliability at beyond
such high frequencies into question?
Thanks,

There is no answer to that question. It's sort of like asking which
words are essential in order to communicate.

Roy Lewallen, W7EL

Kevin Aylward · Jul 24, 2004

ldg wrote:

On 20 Jul 2004 04:56:44 -0700, dorinel_ionascu@mymail.ro (dorinelu)
wrote:

Which is the best sofware for electronic simulation?

This is really a large question. There are many analog and digital
simulators, each optimized for a specific market segment.

Many of these simulators have been created by combining Berkeley
spice and a digital simulator called Xspice that was created by
Georgia Tech http://users.ece.gatech.edu/~mrichard/Xspice/

This is not really an accurate description. XSpice *is* already the
combined Berkeley spice with Xtensions that add digital capabilities.
That is, there is not a separate Georgia Tech digital engine that was
combined with a separate Berkeley spice. XSpice is not a "digital
simulator". It is a mixed mode simulater.

Saber is sort of in a class by itself in that it can easily simulate
physical as well as electrical/electronic devices.

As can any generic spice, so its not.

http://www.synopsys.com/products/mixedsignal/saber/saber_ds.html
It can, for instance, simulate entire car systems where a mechanical
device operates a sensor, which then is processed by electronics, then
drives a mechanical actuator. It would also be great for modeling
MEMs devices since both the physical device and interface circuits can
be modeled at once. It also does mixed mode. New models are easily
made using its MAST programming language. Since both Saber and Hspice
were owned by Avant!, the actual hspice models have been ported to it,
making it even more acceptable for low level IC design. I used this
program years ago to design IR focal planes. This is a great
simulator, but expensive.

This is not specific to Sabre at all. *Any* and all spices have behavial
modelling, e.g. B sources that allows pretty much any function to be
implemented. In fact, spices like isspice have full scripting with
constructs such as if else.

I'm not a user of pure digital simulators, but I'm sure others in this
group are knowledgeable in this area.

Many here use the free LTspice simulator. It has a great user
interface

?

and the support is unusually good.
It's primary purpose is
for board level work, allowing Linear Technology customers to easily
simulate circuits with Linear Technology devices, though the author's
intention clearly is beyond this. The user interface is so good I've
often wished it would front-end for other simulators I use in my work
(hspice, smartspice).
To me, however, the simulator itself doesn't
quite meet my needs for doing IC design.

What aspects are you refering to here? Obviously, I will take this
opertunity to point out that SuperSpice has quite a few i.c. specific
hooks directly built in, such as mosfet binning, and automatic worst
case reruns. I am not aware of any in the "cheap" simulaters that do
this.

SS is also set up that it can automatically drive any engine if it can
run in standard spice batch mode.

Much of this is subjective, and everyone has their own opinion.

Indeed. Mine is that SuperSpice has the best GUI in the known 3
universes.

Kevin Aylward
salesEXTRACT@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Paul Burridge · Jul 24, 2004

On Sat, 24 Jul 2004 07:37:36 GMT,

As yet unrecognized quantum electrodynamicist, astrophysicist,
mathematician, logician, neo-Darwinist, talented musican, programmer,
electronic designer, polymath and hierophant, Kevin Aylward wrote:

History....

[the usual self-aggrandizement snipped]

If you're as smart as you make yourself out to be, you might address
the real questions here, rather than make continual attempts to gain
acceptance and recognition by blowing your own trumpet in an
embarrassingly obvious fashion.
HTH.
--

"What is now proved was once only imagin'd." - William Blake, 1793.

Check Out This Month's "Microwaves & RF" Magazine

John Larkin

Guest

Kevin Aylward

Guest

Paul Burridge

Guest

Leon Heller

Guest

chrisv

Guest

Chaos Master

Guest

Chaos Master

Guest

Jim Thompson

Guest

Kevin Aylward

Guest

Paul Burridge

Guest

Kevin Aylward

Guest

ldg

Guest

Roy Lewallen

Guest

Paul Burridge

Guest

Mike Engelhardt

Guest

Paul Burridge

Guest

Jim Thompson

Guest

Roy Lewallen

Guest

Kevin Aylward

Guest

Paul Burridge

Guest

Log in

Welcome to EDABoard.com

Sponsor