a question about resistors in an arc experiment

I'm about to conclude that Sylvia is just a female rendition of AlwaysWrong... either obstinate to the point of ignorance or a troll.

If she continues to exhibit a closed mind I'll write her off.

...Jim Thompson

A lot of our VME modules (say, half of our production) goes into test stands for FADECS, "full authority" jet engine control computers. They test the hell out of them, simulating all the sensors and actuators, normal and failed, running scripts that take days to execute, and generate feet of paper reports. If they change a tiny bit of code, they run it all again. That's cool, because they can never take down a test stand, so we're always supplying new modules.

"Full Authority" means no mechanical or hydraulic backups. Code fails, plane falls.

John

John L & Sylvia... sounds like a '60's rock group ;-)

...Jim Thompson

If you "can't be bothered" why are you persisting in this thread?

Too ignorant, yet too obstinate to listen to your peers?

...Jim Thompson

is

Well, to me a proper negative resistor is one that has zero current when shorted. Its I vs E curve is just the same as a real resistor of the same absolute value, but the current sign is flipped. It obeys Ohm's law in all quadrants, in all situations.

The nice thing about this resistor is that you can plug its negative value into all the classic equations... series and paralleled resistors, voltage dividers, capacitor rc exponentials, rlc resonators, whatever... and the equations still describe what you'll simulate or measure, small-signal or large.

This is also what LT Spice will do if you assign a negative value to a resistor.

What's the mathematical definition of the other type?

John

"Sylvia and John" sounds better. But I am about as un-musical as Horatio Hornblower.

But don't you agree about JFs circuit? If it is a negative resistor, it violates conservation of energy.

John

e:

not the

iour

he

age.

result is

. Since

the side

the

es even

al

r,

One type does:

I =3D V/R

the other does

V =3D I * R

in both cases without doing math on the situation. In one case, the voltage serves as the input and in the other the current does.

This circuit: R GND ---!!---))))---/\\/\\--- GND

wants the second type to make an oscillator. With the first type. The small disturbance will send the thing rocketing to the rails.

A parallel resonant system needs the first type.

Reeling? What strange ideas you harbor.

Both the LTC opamp model, and LT Spice, are working fine in your simulation. The opamp output is railed low, there's no negative resistance, current is flowing OUT of V3, Conservation of Energy is not being violated, the whole mess looks about like a 1K regular resistor (plus the opamp Icc), and you were probably reading the current direction backwards.

Build one, use a 12 volt battery for V3, and see if it charges the battery.

John

Clearly a 2-terminal negative resistor with no internal power source, which his circuit claims to be, violates conservation of energy.

Conservation of Energy is one of the few laws I'm not willing to mess with.

You can make an unpowered 2-terminal circuit with a region of incremental negative resistance, but his circuit doesn't do that, either.

John

So what _is_ your dispute? Hot and bothered that a negative resistance region obtained using an active circuit (in a way a form of gyrator) does not fit your definition of a negative resistor?

...Jim Thompson

Jim, have you run JFs circuit? Do you think it behaves like a negative resistor?

Here. I'll do it for you. I loaded it exactly as he posted it, and ran the transient sweep as he set it up. I changed nothing.

ftp://jjlarkin.lmi.net/JF.jpg

The nodes plotted are the V3 supply, the R3-R4 junction, and the low end of R3, the opamp output.

Well, who is right, JF or Sylvia? Boy or girl?

John

OK, now meet some of the COBOL programmers who are handling all your money.

John

Change that "emaciated" to "dried up" ;-)

...Jim Thompson

On Sun, 18 Jan 2009 11:08:04 -0800, John Larkin

This is a very neat scope. We have one at work. It has a couple of deficiencies, though. I wish it had more than 2500 capture points.

And, you can't look at all captured data previous to the trigger point (with the sweep stopped) if you expand the sweep speed too much.

And, finally, on the Tek forum there is a description of a measurement problem:

More like the atonality of Berg, Dvorak, Schoenberg, Webern, Bartok and Stravinsky, all scrambled together ;-)

...Jim Thompson

Heck, I have a huge gap, 24 kilobytes, between my last HELP message text and the start of the Xilinx config stuff. It only assembles to about 40 kilobytes so far.

A 68332 can do .W (16 bit addressing) on the bottom 32K of code space, and sign extends that to include the top 32K of the 32-bit space, which is i/o and CPU ram. I use .L addressing above 32K, 0x00008000 and up, which is just lookup tables and help text and Xilinx config stuff and some chunks of flash. Easy.

John