Odd behavior of LTSpice for NMOS and PMOS in series with zero gate voltage

[P E Schoen]

I have been working on an analog multiplexer using discrete NMOS and PMOS
devices. I have the gates connected to their respective sources through
resistors and their drains are connected together. Positive and negative
voltages are applied from source to source through a series load resistor.
The NMOS works as expected - the body diode conducts when a negative voltage
is applied and it blocks for positive voltage. But the PMOS device conducts
when it should block. Here is the ASCII file for a simple test that shows
this odd behavior.

http://enginuitysystems.com/pix/electronics/MOSFET_test.asc

A screen shot of the simulation:

http://enginuitysystems.com/pix/electronics/MOSFET_test.png

I have tried this with several models of the PMOS device. I also closed and
re-opened LTSpice. Am I not understanding PMOS devices or is something
strange going on?

Thanks,

Paul

[P E Schoen]

"P E Schoen" wrote in message news:ou5u9n$9b9$1...@dont-email.me...

Doh! I shouldn't try to design and simulate circuits while watching TV or
being otherwise distracted. It would have been obvious if the symbol
included the body resistor. Now to figure out how to do this.

Thanks,

Paul

[John Larkin]

On Fri, 10 Nov 2017 23:28:33 -0500, "P E Schoen" <pa...@pstech-inc.com>
wrote:

Makes perfect sense. The substrate diodes are in series and
conducting.

It would help if you labeled the nodes.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

[P E Schoen]

"P E Schoen" wrote in message news:ou61tv$rbs$1...@dont-email.me...

> Doh! I shouldn't try to design and simulate circuits while watching TV or
> being otherwise distracted. It would have been obvious if the symbol
> included the body resistor. Now to figure out how to do this.

I used two PMOS back to back, and that seems to work well for input voltages
above about 2.4 volts. Here are simulations:

http://enginuitysystems.com/pix/electronics/Analog_Mux_4.asc
http://enginuitysystems.com/pix/electronics/Analog_Mux_4.png

http://enginuitysystems.com/pix/electronics/Analog_Mux_4a.asc
http://enginuitysystems.com/pix/electronics/Analog_Mux_4a.png

Thanks,

Paul

[Tim Williams]

"P E Schoen" <pa...@pstech-inc.com> wrote in message
news:ou61tv$rbs$1...@dont-email.me...

> "P E Schoen" wrote in message news:ou5u9n$9b9$1...@dont-email.me...
>
> Doh! I shouldn't try to design and simulate circuits while watching TV or
> being otherwise distracted. It would have been obvious if the symbol
> included the body resistor.

Resistor?

It does show the diode!

I guess nobody talks about what a MOSFET symbol actually means, anymore?

Just as in any other semiconductor symbol: it's the triangle in/out of the
middle. That's the channel-substrate junction. What's the channel? There
are two doped regions (source and drain), and the channel is between them.
For an N-ch type, S and D are N doped, and the substrate is P doped. The
one arrow does double duty, representing both of those junctions.

Which, incidentally, means an N-ch MOSFET is also an N-P-N transistor, just
a crappy one, like the old lateral PNP hack. The wide base (equal to
channel length) gives low hFE, probably in the ballpark of 1 for most
designs.

Since source and substrate are strapped together, it's even a diode-strapped
transistor. But with the low hFE, there's not much point in calling it that
way, so a diode it is!

Indeed, any MOSFET symbol you see with the antiparallel diode is redundant*,
and if they draw it as a zener diode, then, redundant and silly! ;-)
(MOSFETs are usually rated for avalanche these days, so the device can
behave like a zener/avalanche diode. But that's just controlling the drain
junction to try to not destroy itself. It's not a separate diode.)

* "FETky" parts have an integrated schottky diode -- you can't make a FET
with schottky body junctions; heh, well, there are PHEMTs I guess, but no,
these are just the two parts, integrated together however they please.

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
Website: https://www.seventransistorlabs.com/

[John Larkin]

On Sat, 11 Nov 2017 02:16:02 -0500, "P E Schoen" <pa...@pstech-inc.com>
wrote:

This is one channel of my 32-channel VME SSR module.

https://www.dropbox.com/s/ijc5e6l9i9jws15/V270_SSR.JPG?raw=1

The protected mosfets turned out to have too much personality, so we
use regular mosfets with software overload protection.

The board is almost done with layout. I did one channel as a model for
my layout guy

https://www.dropbox.com/s/epcc6jpqip5vp5g/V270_Chans_1.jpg?raw=1

[John Larkin]

This is my preferred mosfet symbol

https://www.dropbox.com/s/hjxa2cutoch8bsa/Pdiss_Computer.JPG?raw=1

which visibly behaves like an NPN transistor. The usual symbol is too
fussy.

Incidentally, I have characterized an enhancement-mode PHEMT as a
diode. A very fast, 1 pF, 1 amp diode!

[P E Schoen]

"John Larkin" wrote in message
news:emce0dd8jc4l75ted...@4ax.com...

> This is my preferred mosfet symbol

> https://www.dropbox.com/s/hjxa2cutoch8bsa/Pdiss_Computer.JPG?raw=1

> which visibly behaves like an NPN transistor. The usual symbol is too
> fussy.

> Incidentally, I have characterized an enhancement-mode PHEMT as a diode. A
> very fast, 1 pF, 1 amp diode!

I made symbols for NMOS and PMOS, with the body *diode*.

http://enginuitysystems.com/pix/electronics/NMOS_PMOS.png

Having that as part of the symbol makes it more obvious which way it will
conduct current.

Paul

[Tim Williams]

"John Larkin" <jjla...@highlandtechnology.com> wrote in message
news:emce0dd8jc4l75ted...@4ax.com...

> This is my preferred mosfet symbol
>
> https://www.dropbox.com/s/hjxa2cutoch8bsa/Pdiss_Computer.JPG?raw=1
>
> which visibly behaves like an NPN transistor. The usual symbol is too
> fussy.

NPN transistor points in the direction of conventional current flow.

So, uh, you've shorted out the supply?

[John Larkin]

On Sat, 11 Nov 2017 23:00:35 -0600, "Tim Williams"
<tmor...@gmail.com> wrote:

>"John Larkin" <jjla...@highlandtechnology.com> wrote in message
>news:emce0dd8jc4l75ted...@4ax.com...
>> This is my preferred mosfet symbol
>>
>> https://www.dropbox.com/s/hjxa2cutoch8bsa/Pdiss_Computer.JPG?raw=1
>>
>> which visibly behaves like an NPN transistor. The usual symbol is too
>> fussy.
>
>NPN transistor points in the direction of conventional current flow.

Ditto in my mosfet symbol. The mosfet is just an NPN with an insulated
base. Very intuitive.

>
>So, uh, you've shorted out the supply?

What would make you think that?

[Tim Williams]

"John Larkin" <jjla...@highlandtechnology.com> wrote in message

news:3fmf0dl6h9ib1s1gi...@4ax.com...

> Ditto in my mosfet symbol.

MOSFETs are majority not minority carrier devices. Minority carriers flow
through the body diode, as indicated by the triangle in all semiconductor
symbols......

> The mosfet is just an NPN with an insulated
> base. Very intuitive.

No, that's an IGBT. Very slow for your things.

> What would make you think that?

But anyway, we've trolled this before.

[John Larkin]

On Sun, 12 Nov 2017 04:30:54 -0600, "Tim Williams"
<tmor...@gmail.com> wrote:

>"John Larkin" <jjla...@highlandtechnology.com> wrote in message
>news:3fmf0dl6h9ib1s1gi...@4ax.com...
>> Ditto in my mosfet symbol.
>
>MOSFETs are majority not minority carrier devices. Minority carriers flow
>through the body diode, as indicated by the triangle in all semiconductor
>symbols......
>
>> The mosfet is just an NPN with an insulated
>> base. Very intuitive.
>
>No, that's an IGBT. Very slow for your things.
>
>
>> What would make you think that?
>
>But anyway, we've trolled this before.
>
>Tim

Nobody seems impressed with my analog multiplier thing.

[John S]

I can count on one finger the number of people here who wish to massage
your ego. Besides, you forgot to add the duct tape. -1.

[Tom Del Rosso]

Tim Williams wrote:
> "P E Schoen" <pa...@pstech-inc.com> wrote in message
> news:ou61tv$rbs$1...@dont-email.me...
>> "P E Schoen" wrote in message news:ou5u9n$9b9$1...@dont-email.me...
>>
>> Doh! I shouldn't try to design and simulate circuits while watching
>> TV or being otherwise distracted. It would have been obvious if the
>> symbol included the body resistor.
>
> Resistor?
>
> It does show the diode!
>
> I guess nobody talks about what a MOSFET symbol actually means,
> anymore?

When I was a little kid the arrow pointing in meant "iN" and the arrow
pointing out meant "Pee"

[John Larkin]

On Sun, 12 Nov 2017 12:14:39 -0600, John S <Sop...@invalid.org>
wrote:

It's a cool circuit. The transistors operate with bases at ground and
collectors at ground. It works well with a cheap non-monolithic dual
transistor. It lets you really push a mosfet safely, far better than a
primitive current limiter, even a foldback limiter. Can save a ton of
money on mosfets and heat sinks and fans.

This isn't about ego, it's about circuits. Really, my ego doesn't need
any outside approval.

[Jim Thompson]

On Sun, 12 Nov 2017 11:01:21 -0800, John Larkin
<jjla...@highlandtechnology.com> wrote:

>On Sun, 12 Nov 2017 12:14:39 -0600, John S <Sop...@invalid.org>
>wrote:
>
>>On 11/12/2017 9:40 AM, John Larkin wrote:

[snip]

>>>
>>> Nobody seems impressed with my analog multiplier thing.
>>
>>I can count on one finger the number of people here who wish to massage
>>your ego. Besides, you forgot to add the duct tape. -1.
>
>It's a cool circuit. The transistors operate with bases at ground and
>collectors at ground. It works well with a cheap non-monolithic dual
>transistor. It lets you really push a mosfet safely, far better than a
>primitive current limiter, even a foldback limiter. Can save a ton of
>money on mosfets and heat sinks and fans.
>
>This isn't about ego, it's about circuits. Really, my ego doesn't need
>any outside approval.

John L, Nobody commented on your circuit because 99.5% of the lurkers
here have no clue as to how it works. Most can't even properly bias a
transistor let alone tell you how it's working.

As for John S, he couldn't find his "S" even with the help of an
assistant holding the mirror and flashlight >:-}

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

It's what you learn, after you know it all, that counts.

[kevin93]

On Sunday, November 12, 2017 at 11:01:30 AM UTC-8, John Larkin wrote:
..

> >> Nobody seems impressed with my analog multiplier thing.
> >
> >I can count on one finger the number of people here who wish to massage
> >your ego. Besides, you forgot to add the duct tape. -1.
>
> It's a cool circuit. The transistors operate with bases at ground and
> collectors at ground. It works well with a cheap non-monolithic dual
> transistor. It lets you really push a mosfet safely, far better than a
> primitive current limiter, even a foldback limiter. Can save a ton of
> money on mosfets and heat sinks and fans.
>
> This isn't about ego, it's about circuits. Really, my ego doesn't need
> any outside approval.

...

It seems a good application of this circuit arrangement to use it to measure device dissipation in real-time.

I first saw the circuit about 30 years ago in the National Semi Linear Applications Handbook and Bob Pease described it in one his articles in 2002.

http://www.pa4tim.nl/wp-content/uploads/2012/01/figure_01.gif

http://www.electronicdesign.com/test-amp-measurement/what-s-all-wattmeter-stuff-anyhow

kevin

[P E Schoen]

"kevin93" wrote in message
news:ba919f9a-725a-4ebb...@googlegroups.com...

> On Sunday, November 12, 2017 at 11:01:30 AM UTC-8, John Larkin wrote:
..
>> Nobody seems impressed with my analog multiplier thing.

>> It's a cool circuit. The transistors operate with bases at ground and
>> collectors at ground. It works well with a cheap non-monolithic dual
>> transistor. It lets you really push a mosfet safely, far better than a
>> primitive current limiter, even a foldback limiter. Can save a ton of
>> money on mosfets and heat sinks and fans.

>It seems a good application of this circuit arrangement to use it to
>measure device dissipation in real-time.

> I first saw the circuit about 30 years ago in the National Semi Linear
> Applications Handbook and Bob Pease described it in one his articles in
> 2002.

> http://www.pa4tim.nl/wp-content/uploads/2012/01/figure_01.gif

> http://www.electronicdesign.com/test-amp-measurement/what-s-all-wattmeter-stuff-anyhow

When I worked for EIL Instruments in the 70s and 80s we sold wattmeters and
watt transducers that used a non-linear arrangement of resistors and diodes
in a bridge circuit to produce a DC signal proportional to watts. Here is
the patent showing how it works:

http://www.freepatentsonline.com/3662264.pdf

Paul

[k...@notreal.com]

PNP = Piss iN Pot

[John Larkin]

That's pretty horrible.

The Jones multiplier dates to 1963. Gilbert reinvented and improved it
in 1967.

>
>http://www.electronicdesign.com/test-amp-measurement/what-s-all-wattmeter-stuff-anyhow

The 2 uF calculation is obviously wrong.

[piglet]

On 13/11/2017 00:12, John Larkin wrote:
>
> The 2 uF calculation is obviously wrong.
>
>

Could be a typo, 2uF for 20uF or 90VA for 9VA

piglet

[John Larkin]

On Mon, 13 Nov 2017 07:11:28 +0000, piglet <erichp...@hotmail.com>
wrote:

Silly typo, that nobody caught. Everybody knows that 1 uF is 2600 ohms
at 60 Hz, and 1 pF is 160 ohms at 1 GHz.