a question about resistors in an arc experiment

I never looked specifically at JF's schematic, only one of the original sketches... which definitely had a negative resistance region.

...Jim Thompson

JF's updated proposal was meant to address my objection that the previous one requires a power supply. It attempts to use the input to power the Op-Amp, while purporting to draw a negative current, and thus have a negative total resistance. It doesn't though, which is hardly surprising, because if it did, one could construct a real life version that would provide free energy.

Sylvia.

Quit weaseling and look at the circuit. Can't admit that a "girl" might be right?

John

Damn, you *can't* admit that Sylvia is right.

Funny.

John

you've got R4 installed backwards (which effects the orientation of the current probe) that's why you see a negative trace,

what really happens is, as the voltage decreases so does the current flowing from the supply through the resistor.

To get it to behave as a negative resistor you need to have more voltage on the OP-amp's supply than on the input.

I think I saw this one in AOE.

O | | +------. | | |--' | | | | | .-->|--. | | | | | | | | | | | `--|-->' | | | | | .--| | | `------+ | | O

Phil is almost always right (when on topic) There are other regular participants here with a worse signal to noise ratio.

if there is any capacitance parallel to the negative resistor while it is disconnected, it quickly becomes unstable.

you can enter a negative value for a resistor in ltspice, it makes a handy comparison for the circuits everyone has been posting.

That's pure obvuscation.

For establishing the resistance of the circuit, the only relevant current is that going into, or out of, the circuit. The current through R4 is not that current. The offset of which you speak could change the sign of the current into the circuit, which in this context would be hugely significant. You can't right it off as being something that's of no concern.

Sylvia.

Yes, I have met some of the stubby fingered folk. The old COBOL code is trustworthy because it was tested to death. What it did was simple so it could be tested. The work is generally done in fixed point so the errors can be predicted.

It wouldn't matter if the opmp was powered from a separate supply. It would still be railed low, and the input to R4 would still look like a positive, about 1K, resistor to ground, and the circuit would still be useless.

You measured the current backwards.

John

I know a guy who used to program inter-bank money transfer systems. He had one VAX moving $17 billion a day at the BofA, back when $17 billion was a lot of money. He said that the code to do a transaction was horrible, and the code to back out a transaction was ten times worse. Imagine un-doing a chain of things with compound interest and transaction fees all the way. Makes me glad I'm a simple engineer.

John

On Sun, 18 Jan 2009 17:41:02 -0800, John Larkin wrote:

--- That seems to make sense, and I think the problem is the way LTspice handles currents.

Referring to the last schematic I posted:

Version 4 SHEET 1 880 680 WIRE 176 -32 128 -32 WIRE 320 -32 256 -32 WIRE 192 80 -16 80 WIRE 192 128 192 80 WIRE -512 144 -672 144 WIRE -16 144 -16 80 WIRE -16 144 -288 144 WIRE 0 144 -16 144 WIRE 128 144 128 -32 WIRE 128 144 80 144 WIRE 160 144 128 144 WIRE 320 160 320 -32 WIRE 320 160 224 160 WIRE 160 176 128 176 WIRE 192 208 192 192 WIRE 192 208 16 208 WIRE -672 240 -672 144 WIRE -512 240 -512 144 WIRE -288 240 -288 144 WIRE 128 256 128 176 WIRE 176 256 128 256 WIRE 320 256 320 160 WIRE 320 256 256 256 WIRE 128 304 128 256 WIRE -672 416 -672 320 WIRE -512 416 -512 320 WIRE -512 416 -672 416 WIRE -288 416 -288 320 WIRE -288 416 -512 416 WIRE 16 416 16 208 WIRE 16 416 -288 416 WIRE 128 416 128 384 WIRE 128 416 16 416 WIRE 128 464 128 416 FLAG 128 464 0 SYMBOL res 112 288 R0 SYMATTR InstName R1 SYMATTR Value 100k SYMBOL res 272 240 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R2 SYMATTR Value 100k SYMBOL res 272 -48 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R3 SYMATTR Value 490 SYMBOL voltage -288 224 R0 WINDOW 3 24 104 Invisible 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR Value PULSE(0 12 0 .01) SYMATTR InstName V3 SYMBOL res 96 128 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R4 SYMATTR Value 490 SYMBOL Opamps\\\\LT1797 192 96 R0 SYMATTR InstName U1 SYMBOL voltage -672 224 R0 WINDOW 3 24 104 Invisible 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR Value PULSE(0 12 0 .01) SYMATTR InstName V1 SYMBOL res -528 224 R0 SYMATTR InstName R5 SYMATTR Value 1000 TEXT -8 440 Left 0 !.tran .02 uic

If you lay the cursor over V1, even though current is going out of V1 it's shown as going to ground. Then, if you lay it over R5 it'll show it as going to ground also, the scheme seeming to be that charge flows in the direction of the highest voltage to ground.

Plot the currents and they're shown as going in different directions with charge flowing out of a device being designated as negative and into a device as positive.

More to the point though, plot the voltage across R5 and the current through it and they'll be in phase, just as expected for a normal resistor. Congruent, even, so if you want to see separate slopes, change the value of R5 to 750 ohms or so.

Now, if you plot the voltage out of V3 and the current through R4 you'll see that the relationship between current and voltage is definitely _not_ the same as in the V3 R5 case.

What's happening is that LTspice is plotting the current _into_ R4 as though E3 were a sink with charge flowing _out_ of R4.

Make sense?

JF

JF's circuit is weird, power tied to an input... not worthy of a look.

The Howland pump version, posted very early on WORKS!

...Jim Thompson

If I hover the probe over the V3 source, the "current meter" arrow points down, and the graphed current is ramping negative, which I suppose means the current is actually positive and ramping UP into R4.

If the arrow were reversed (pointing UP inside of V3) and the graph ramped positive and up, the situation would be more obvious.

Ditto R4; arrow points left, current graphs negative. One crosscheck as to the current direction is the voltage at the ends of R4: +12 on the left, +6 on the right, so the current is flowing to the right.

I like to check my circuits two or three (or as many as possible) different ways and make sure they're all consistent. And the next day, assume that an idiot designed the circuit and do it again. I can't make myself any smarter than I am, but I can make myself be very careful.

Is there any way to get a netlist from a newsgroup post directly into LT Spice? I've been highlighting, copying, opening an editor, pasting, saving as .ASC somewhere, finding the file, and dragging onto the LT Spice shortcut. All that is a nuisance.

John