OT: United States of Feeding Tubes

[Jim Thompson]

On 28 Mar 2005 18:21:46 -0800, Winfield Hill
<hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:

Jim Thompson wrote...

Winfield Hill wrote:

What does the entire model expression look like for your plot?

See...

Newsgroups: alt.binaries.schematics.electronic
Subject: Re: 2N7000 Modeling Issues - Test2N7000-Level7.pdf -
Test2N7000-Level7-5V.pdf
Message-ID: <3a7h411ua171vsosvb3gl9t7011cctcq7a@4ax.com

That's not helpful. You've got an 87mV/decade slope for the MOSFET,
are you sure any fab MOSFET really has such a high transconductance?
It looks like a draftsman's artificially-too-good plot to me.

Well. We've now had several chip runs out of that process. Of
course, ALL the stuff I designed came out working EXACTLY as predicted
;-)

Keep in mind that this is a 5V process.

I'm bound to non-disclosure on the models (this device is simply
scaled from a _current_ project).

So I'll study the model and see what parameters move what, until
I get a DC match to 2N7000, then add in the capacitances from a
2N7000 and see how close we get.

That's good, the parameters you're using dramatically fail to match
the real 2n7000. But anyway it's good to see a subthreshold plot...

That was what I was after... not a match, but demonstrating
subthreshold behavior

Or bug the manufacturer to make a Level=7 model ;-)

Nah, won't happen.

Yep. I guess the powerFET manufacturers have enough business that
they don't care.

But I will pursue it as I learn the intricacies of Level=7... maybe I
can make a business of model-making ?

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[John Fields]

On Mon, 28 Mar 2005 17:17:02 -0700, Jim Thompson
<thegreatone@example.com> wrote:

On 28 Mar 2005 15:42:46 -0800, Winfield Hill
hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote:

Jim Thompson wrote...

Win, See...

Newsgroups: alt.binaries.schematics.electronic
Subject: 2N7000 Modeling Issues - Test2N7000-Level7.pdf
Message-ID: <703h41d0tcvctdfoqar06vpjd1525hpsq6@4ax.com

This is an XFAB, Level=7, model.

What does the entire model expression look like for your plot?

See...

Newsgroups: alt.binaries.schematics.electronic
Subject: Re: 2N7000 Modeling Issues - Test2N7000-Level7.pdf -
Test2N7000-Level7-5V.pdf
Message-ID: <3a7h411ua171vsosvb3gl9t7011cctcq7a@4ax.com

I can run Bsim3v2 level 7 and Bsim3v3.1 level 8, etc.

But even it seems to have an artificially flattened bottom end.

Perhaps it simply has an artificially-high 250pA leakage term.

The measured minimum current on my part was less than 2pA with
Vgs = 0 and Vds = 5V (I need to repeat the measurement, after
zeroing the instrument, so I how much was instrument leakage and
zero-offset error, vs actual 2n7000 component Id leakage at 5V).
For design purposes, maybe they feel one shouldn't count on such
low leakage. For myself, I'd prefer to add it back explicitly.

I'm bound to non-disclosure on the models (this device is simply
scaled from a _current_ project).

So I'll study the model and see what parameters move what, until I get
a DC match to 2N7000, then add in the capacitances from a 2N7000 and
see how close we get.

Or bug the manufacturer to make a Level=7 model ;-)

Ah, yes... instead of making the model satisfy the constraints of the
device, make a _real_ device to fit the desired model...

Nothing wrong with that...


--
John Fields

 

[Jim Thompson]

Win, Which Level=3 model matched a real 2N7000, ABOVE THRESHOLD, the
closest?

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Winfield Hill]

Jim Thompson wrote...

Win, Which Level=3 model matched a real 2N7000, ABOVE THRESHOLD,
the closest?

I don't have the answer, because I haven't taken the 2n7000 pulsed
measurements yet. But I can say they all have _substantially_
different curves, one from the other. But then, unlike othe 2nxxx
transistors, various company's datasheets have quite different high-
current specs. Perhaps I'll summarize this later today or tomorrow.

I've also been gathering a small assortment of manufacturer's parts
to see how the measurements tend to cary. For now, I think they're
all pretty similar in the 1mA to 50mA region, which I imagine is
where you'll be most interested in using them. Linear operation at
currents above 50mA or so would push one toward overheating, right?


--
Thanks,
- Win

 

[Winfield Hill]

Winfield Hill wrote...

Jim Thompson wrote...

Win, Which Level=3 model matched a real 2N7000, ABOVE THRESHOLD,
the closest?

I don't have the answer, because I haven't taken the 2n7000 pulsed
measurements yet. But I can say they all have _substantially_
different curves, one from the other.

Consider the 2n7000 spice drain current plots in the figure posted
on a.b.s.e. These are all with Vds = 5V, and are summarized below.
First, note they're all pretty similar at Vgs=3.5V, with about 100mA
drain current, except the old Motorola model is 8% lower at 92mA.

At Vgs=5V they're still together, except now the new Motorola curve
is lower than the rest. At 7V the old Motorola model is much higher,
while the new one stays lower. At 9V the Philips and new Motorola
and models are in agreement at high currents, and the Zetex model
is in the middle. At 10V the Philips model shows the lowest current.
See what I mean, all substantially different from each other?

2n7000 spice, Drain current vs Vgs, with Vds = 5V, various models:

type Vgs = 3.5V 5V 7V 9V 10V
-------- ------ ----- ------ ------ ------
Mot - new 101mA 330mA 769mA 1250mA 1452mA
Philips 113 460 1007 1270 1360
Zetex 106 457 1084 1680 1880
Mot - old 92 485 1360 2210 2530

But then, unlike other (JEDEC) 2nxxx transistors, various
company's datasheets have quite different high-current specs.

I'll try to post some datasheet comparisons tomorrow morning.


--
Thanks,
- Win

 

[Winfield Hill]

Winfield Hill wrote...

Jim Thompson wrote...

Win, WHAT is important to you in power-FET modeling, subthreshold?
Maybe we can mutually determine the "fix" for the models?

OK, here's what I wrote early in the discussion, last December 31st,

Winfield Hill wrote...

Well, then surely you'll want viable Spice MOSFET models, because
you'll be using the FETs throughout their useful linear range, and
the output-stage crossover region is critical.

Sadly most standard Spice library VMOS models simply don't do the
subthreshold linear region. For example, see the 10-decade plots
on page 123 of our book. A jellybean 2n7000 is rather similar to
the VN01 that we show in figure 3.14, and certainly a proper Spice
model should be able to make that plot. But I'd be surprised if
your standard Spice libraries work properly below say 5 to 20mA,
which is not that far below the FET's maximum current. Keep in
mind that linear power FET circuitry always operates well below
the maximum rated FET switching current, to keep power dissipation
junction heating under control.

With respect to the Id-vs-Vgs curves on page 123 of our book, and
the g_m plots on page 132, the spice models should be able to show
this performance. At subthreshold currents a FET acts very much
like a transistor with respect to transconductance, etc., and when
a power FET is used in linear audio amplifiers, e.g. in class AB,
it may go through this region during each cycle. So in using Spice
to determine distortion and evaluate various design configurations,
surely it's necessary for the FET model to smoothly simulate the
subthreshold region, and properly progress to the current-saturated
regions that are normally accurately modeled. I'm going to go in
to the lab and take some detailed 2n7000 measurements later today.

I've completed a set of measurements over the current range of 1pA
to 10mA, and posted the results on a.b.s.e. (Above 10mA I'll have
to take pulsed measurements, which requires a different setup.) As
anticipated, the data looks similar to the VN01 data, AoE page 123.

As expected, the 2n7000 has an exponential gate-voltage-programmed
drain current in the under-1mA subthreshold region, similar to a BJT.
It follows the formula Id = Is exp(Vgs/Vt) and for my tested 2n7000
Is = 0.1pA and Vt = 78mV. The transconductance is g_m = Id / 0.078
(by comparison BJT transconductance, g_m = Id / 0.025, is 3x higher).

My measurements show conduction starting at 200mV, 1nA at Vgs = 0.7V,
and 1.0mA at Vgs = 1.8V, with a 2.303 Vt = 178mV/decade slope (see
plot on a.b.s.e.). This compares to 58mV/decade for BJT transistors.

Above 0.1mA the 2n7000 drain current (and transconductance) begins to
fall below that predicted by the exponential formula. I tried improving
the model by reducing Vgs by a term Id*Rs. This worked over 0.1 to 5mA
with Rs = 15 ohms, but above 5mA the term overcorrects and the formula
predicted too little current.

Note, the spice model I tried earlier showed 0pA up to 2.4V, where it
suddenly and dramatically soared to about 100nA, and it continued with
excessive transconductance up to over 100uA. It required Vgs = 2.51V
for 1mA, so it was *very badly* off the mark on all counts, and nearly
useless for linear modeling.


--
Thanks,
- Win

 

[Winfield Hill]

Winfield Hill wrote...

Winfield Hill wrote...

Jim Thompson wrote...

Win, Which Level=3 model matched a real 2N7000, ABOVE THRESHOLD,
the closest?

I don't have the answer, because I haven't taken the 2n7000 pulsed
measurements yet. But I can say they all have _substantially_
different curves, one from the other.

Consider the 2n7000 spice drain current plots in the figure posted
on a.b.s.e. These are all with Vds = 5V, and are summarized below.
First, note they're all pretty similar at Vgs=3.5V, with about 100mA
drain current, except the old Motorola model is 8% lower at 92mA.

At Vgs=5V they're still together, except now the new Motorola curve
is lower than the rest. At 7V the old Motorola model is much higher,
while the new one stays lower. At 9V the Philips and new Motorola
and models are in agreement at high currents, and the Zetex model
is in the middle. At 10V the Philips model shows the lowest current.
See what I mean, all substantially different from each other?

2n7000 spice, Drain current vs Vgs, with Vds = 5V, various models:

type Vgs = 3.5V 5V 7V 9V 10V
-------- ------ ----- ------ ------ ------
Mot - new 101mA 330mA 769mA 1250mA 1452mA
Philips 113 460 1007 1270 1360
Zetex 106 457 1084 1680 1880
Mot - old 92 485 1360 2210 2530

But then, unlike other (JEDEC) 2nxxx transistors, different
company's datasheets have quite different high-current specs.

I won't be able to take my pulsed measurements until this weekend,
but here're some comparative numbers taken from 2n7000 datasheet
curves. Note the 4:1 typical Id range at Vgs = 3.5 volts.

type Vgs = 2V 3.5V 5V 7V 9V 10V
-------- ----- ----- ----- ----- ------ ------
Motorola 0mA 150mA 380mA 800mA 1320mA 1480mA
Vishay 1 90 420 1050 - -
Siliconix ?0 140 480 1200 - -
Philips 0 200 >600 - - >1000
Fairchild 0 250 800 1400 est 2500 - (NSC process)
Samsung 50 600 1300 1850 - -

All the datasheets have high-current curves that fail to properly
indicate the FET's currents at low gate voltages, where one would
use them for linear circuitry. For example, the Vgs = 2V drain
current should be about 10mA, not 0mA as the plots show.

Philips does have a transconductance vs Id curve that shows g_m
increasing proportional to current below 50mA, approx g_m/Id = 4 S/A,
which is 1/10 that of a BJT. You'll recall my 8-decade measurements
showed g_m/Id = 12.8 S/A below 1mA, dropping off somewhat above 1mA.

The problem with the manufacturer's datasheet plots is they're linear
rather than log-linear, so that no detail is visible at low currents.


--
Thanks,
- Win

 

[Clarence_A]

"Winfield Hill" <hill_a@t_rowland-dotties-harvard-dot.s-edu> wrote
in message news:d1t46u03n1@drn.newsguy.com...

Richard the Dreaded Libertarian wrote...

You can't reason with anti-gun nuts. They're used to abject
obedience
to whichever king or emperor happens to be in the catbird
seat.

Another helpful post.
--
Thanks,
- Win

I agree, the truth is always helpful.