Trouble understanding MOSFETs

I'm having trouble working with MOSFETs. I've never actually designed them into any circuits before, and trying to do so based on what I think I've learned about 'em isn't working.

I'm attaching a schematic, a SPICE list, and a graph of the SPICE run. I would expect the graph to show V(100) dropping virtually to zero shortly after V(2) goes positive, but it ain't happening.

Perhaps the 2N7000 model I have isn't right, though I think I got it from the manufacturer. That's also hinted at by the fact that SPICE (ngspice, actually) thinks the gate is enough of a current sink to drop a comparator's high output from 5V to about 1V unless I introduce a 10MΩ gate resistor. However, the graph doesn't look much different using a generic NMOS model.

What really obvious and simple thing am I not getting here?

Thanks!

(The end result is to use the MOSFET to act as close to a zero-resistance path for a meter, but shut off current when the voltage goes over the meter's rating.)

Your spice list is not very readable since all the lines are run together with no line returns.

There's a disconnect between your schematic and the simulation since the schematic shows a 12V supply but the transistor output is sitting at 40V. Your spice list is apparently set for a 40V supply.

The gate is a very high input impedance so you shouldn't need a gate resistor.

Your spice net appears to show sb on the transistor connected to gate node 200. If sb is the transistor substrate then it should be connected to the source, node 300. If the substrate is connected to the gate the transistor will not work.

Step forward

Thanks, Carl. The SPICE list is Unixish line endings, so your browser is probably calling up Notepad, eh? Since your notes let me fix things to get a reasonable plot, I won't bother pasting the netlist here.

You're right about the disconnect between the schematic and the SPICE list; an experiment that wasn't reflected in both places.

Yes, Sb was the substrate; I must have misread where to connect it. Connecting it to the source makes for a better result.

This test circuit shows a curve I'd expect. So aside from the substrate issue, I wasn't as far out in the weeds understanding-wise as I feared.

Now to figure out what's wrong with how the MOSFET is wired into my shunt circuit..

Thanks!

Open it in a browser such as Internet Explorerm, FireFox, Opera or a word processor - even Windows WordPad.

Hero999 said:

Open it in a browser such as Internet Explorerm, FireFox, Opera or a word processor - even Windows WordPad.

Click to expand...

Thanks for the tip. Didn't realize that Notepad treats line endings differently than Wordpad. The file opens fine in Wordpad.

They don't treat line endings differently, it's just that Wordpad supports the UNIX format and Notepad doesn't.

Hero999 said:

They don't treat line endings differently, it's just that Wordpad supports the UNIX format and Notepad doesn't.

Click to expand...

Isn't that the same thing?.

Windows requires LF/CR (Notepad) and UNIX requires just LF (Wordpad).

No because Wordpad can open both LF/CR and LF files and Notepad can only open LF/CR files.

Download Em Editor @ http://www.emeditor.com/
or may fav crimson @ http://www.crimsoneditor.com/

Or Emacs.

https://www.gnu.org/software/emacs/

Even edit.com can open UNIX text files, Notepad has got to be the worst text editor invented.