For Joerg: No more *PHUT*?

J

Joerg

Jan 1, 1970
0
J

Joerg

Jan 1, 1970
0
Tim said:
I tried to get a SPICE model of a switching regulator from National
recently, and got a very helpful apps engineer who ran several web bench
simulations for me, but didn't have a model to share.

_I_ can run National's web bench, and when I do I can actually plug in
the numbers that I'm looking for instead of the misunderstood crap that
he did.

Like NXP with you, National lost my business on that one. (chances were
high it was going to be a microprocessor -> gate drive -> FET anyway, but
they certainly lost out on whatever chance they had to start with).

I completely fail to understand why companies sometimes behave that way.
TI lost a design-in on a regulator as well, similar reason. Web Bench
has so far never been useful for me. It typically resulted in some
message like "The design cannot be completely with the values ... yada,
yada, yada". And they are all in full production now :)

While I don't do many IC designs (but some) I know that every time we
did extensive SPICE sims. So the models must be there. Now I can
understand why they want to keep a chip under the cloak of secrecy. But
an LDMOS device? Not releasing SPICE models for that is almost the
perfect definition of "leaving money on the table". Well, the
competition gladly took it.
 
D

Don Klipstein

Jan 1, 1970
0
In <[email protected]>,
guess it is a response to this:
and it is in stock at digikey, farnell etc.

that blf578xr is still in "development" and the datasheet isn't even
available

Come to think of it, I could supply a "XBP-BP13" xenon flashlamp, which
would resemble many used in "1-time-use" cameras and digital cameras.

I could give a guaranteed "safe area of operation" - anode voltage 850
to 1350 volts, flash energy .1 to 500 megajoules, with average power
dissipation over any period including more than one flash attempt not
exceeding 1 watt.
After the flashlamp goes through a flash attempt without flashing, it
can be warranty-returned (entire flashlamp must be returned) for
customer's choice of refund or a replacement flashlamp, and no more.
 
D

Don Klipstein

Jan 1, 1970
0
I tried to get a SPICE model of a switching regulator from National
recently, and got a very helpful apps engineer who ran several web bench
simulations for me, but didn't have a model to share.

_I_ can run National's web bench, and when I do I can actually plug in
the numbers that I'm looking for instead of the misunderstood crap that
he did.

Like NXP with you, National lost my business on that one. (chances were
high it was going to be a microprocessor -> gate drive -> FET anyway, but
they certainly lost out on whatever chance they had to start with).

I actually like National Semi's WebBench. Then again, I use it only
for stuff simpler than microprocessor-controlled.
 
J

Joerg

Jan 1, 1970
0
Tim said:
I know that most SPICE models you do get are behavioral, and that it's
work to get a model that's accurate enough for circuit design yet doesn't
shove hair crosswise up the butts of the Defenders of the Intellectual
Property.

So I understand the impulse not to do SPICE models at all.

But it still leaves me severely disinclined to design in products, and it
leaves me favorably inclined to folks that _do_ publish SPICE models,
even when I know their models are 'just' behavioral.

FET models are usually not behavioral. There isn't much IP value in them
either. If someone wanted to disect a FET in order to copy it then SPICE
won't help them much. They'd have to send it to a materials science lab
for slicing and tear-down, along with a big fat check. So why they don't
release those SPICE models is something I really don't understand. But
oh well, what's the old sales mantra? If you don't take care of your
customer, someone else will :)

And it's not that they don't give out info. They usually furnish models
for high falutin RF simulators that cost as much as a Cadillac. Many
folks don't have that, and it's often not all that useful for pulse apps
anyhow. S-parameters aren't all that helpful there.
 
J

Joerg

Jan 1, 1970
0
Jim said:
[snip]
I know that most SPICE models you do get are behavioral, and that it's
work to get a model that's accurate enough for circuit design yet doesn't
shove hair crosswise up the butts of the Defenders of the Intellectual
Property.

Most systems-on-a-chip are modeled behaviorally; and not very well. Most
companies that inquire what I charge for a good model, decline my services.
:-(

Yet they will likely not have a problem paying a lawyer 10x that amount
for setting up some contract. Been there, and that company went kaputt.
As expected.
 
D

DonMack

Jan 1, 1970
0
The real question is: "How well do these work for audio amplifiers"!!! ;)
 
J

Joerg

Jan 1, 1970
0
DonMack said:
The real question is: "How well do these work for audio amplifiers"!!! ;)

Oh, they should work well. The gold plating makes sure that electrons
have the correct traction but not too much, don't uncontrollably slither
around, and the music sounds crisp. The mounting bracket forms a beker,
to make sure none of the sound spills while in the trainsistah.

Of course, when uncle Leroy bangs against the wall and claims his teevee
is on the fritz ever since you fired it all up that means it oscillateth
and that pretty soon some of the neighbors will come visit, with
pitchforks and all that :)
 
F

Fred Bartoli

Jan 1, 1970
0
Nico Coesel a écrit :
Its from NXP. Trust me, it will do *PHUT* in a real fault situation.

From what we know about Joerg I take that PHUT would mean PHilips
Unobtainium Transistor.

Is that right?
 
J

Joerg

Jan 1, 1970
0
Fred said:
Nico Coesel a écrit :

From what we know about Joerg I take that PHUT would mean PHilips
Unobtainium Transistor.

Is that right?

NXP transistors are usually not unobtanium. What is often unobtanium for
their RF transistors are SPICE model. They don't always release them
which is IMHO a serious business mistake. Or at least a costly one
because they lose design-ins for that reason.
 
N

Nico Coesel

Jan 1, 1970
0
Michael A. Terrell said:
*PHUT* *PHUT* (in stereo)

No, nowadays its 5.1 surround:
*PHUT* *PHUT* *PHUT* *PHUT* *PHUT* . <- micro *phut*
 
Top