Claimed OU circuit of Rosemary Ainslie

[PaulLowrance]

Paul, the problem is more than just the shunt resistor inductance. There is inductance in the probe ground lead, and it can be quite appreciable, such as 0.15uH for a 4" lead. Double that for an 8" lead, which some folks may be using in their testing.

In my PSpice simulations, as little as 0.3uH of inductance will significantly skew the voltage and hence current measurement, as the reactance becomes quite appreciable (with 50ns rise times) relative to the DC shunt resistance.

Regarding the load resistor, yes it is a ceramic air-core as shown in the test plan and photos previously posted here.

.99,

That's true for the inductance of a single wire where the ground wire is separated, but that's not true when the two wires are next to each other. You're testing a small shunt resistor, right? So the ground wire should hug against the probe as much as possible.

Anyhow, I thought we got past this inductance issue. My advice is to forget about that type of measurement, and just do the temperature measurements on the mosfet. If you think the shunt or whatever else consumes much power, then go ahead and test them as well.

Furthermore, you can do the same temperature measurements on the battery shunt as well. Just get a low power resistor, maybe 1/4 or 1/8 watt, 0.25 ohms, that will heat up enough, and measure the temperature. Then do the control experiment to see how much current was flowing through the battery shunt. That way you don't need to place any probes or anything on the Ainslie electrical circuit. The gang is happy. We're happy. Everyone's happy! 

Paul

[poynt99]

Paul,

SPICE does lossy and ideal transmission lines, and it does them well.

What do transmission lines have to do with modeling energy going back to the source anyway?

I've proven in SPICE that energy goes back to the source, and I've not used any TL's! That's a done deal. Why are you arguing that fact? Energy can go back into an ideal (or non-ideal) DC source, AND it can go back into an ideal (or non-ideal) capacitor. If it can do this in SPICE, it surely can and will do it in real life.

The writeup in Rose's article states that grid-powered sources were used as well, and I highly doubt that it makes any difference.

Furthermore, it is pointless to argue about using a battery vs. a lab supply until/unless the measurement technique can be worked out and agreed upon by all as valid. Up to now, no one imo has used a workable technique to acquire solid measurements, and it is looking as though the thermal/CONTROL method may be the best choice. However, the Ainslie team has not commented to weigh in on this issue AFAIK.

.99

[PaulLowrance]

Paul,

SPICE does lossy and ideal transmission lines, and it does them well.

No fundamental spice does not. Spice does not even consider propagation delays or transmission line reflectivity. You have add the models yourself, and it's complex.

What do transmission lines have to do with modeling energy going back to the source anyway?

I already mentioned one example, Barkhausen effect. It produces high frequency noise found in magnetic components.

I've proven in SPICE that energy goes back to the source, and I've not used any TL's! That's a done deal.

I just proved you wrong. Show your spice model here that includes propagation delays or transmission line reflectivity. I've spent way to much time doing spice models to know that there's no way you've included such complex models in your simulations. Sorry, .99, but in terms on any high frequency effect that Ainslie might have discovered, your spice simulations are worthless.

Why are you arguing that fact? Energy can go back into an ideal (or non-ideal) DC source, AND it can go back into an ideal (or non-ideal) capacitor. If it can do this in SPICE, it surely can and will do it in real life.

No argument. What I'm telling you is a fact. If you think otherwise, then show your spice models. Show a spike as it traverses down the line. Show the reflectivity. That's antenna theory that you will not find it spice. Again, such high frequencies exist in all magnetic components.  Are you actually arguing against this? It's conventional physics, .99. I have no idea what heated discussions occurred with you and others here, and maybe you convinced them that your spice sims are accurate for anything that Ainslie may have discovered, but I know better.



Paul

[PaulLowrance]

.99,

BTW, here's a quick intro in doing transmission lines in pspice that I recommend,

http://w2.cadence.com/appnotes/TransmissionLineApplicationsInPSpice.pdf

As stated, spice can do it, but it's complex. Enjoy! 

Whatever you do, don't think that pspice can easily simulate real world transmission lines because it's not that simple. Pspice relies on the user knowing about the particular task at hand, and correctly modeling it into pspice.


Paul

[poynt99]

No spice does not. Spice does not even consider propagation delays or transmission line reflectivity.

PSpice does indeed model these effects. I know because I have used them.

I just proved you wrong. Show your spice model here that includes propagation delays or transmission line reflectivity. I've spent way to much time doing spice models to know that there's no way you've included such complex models in your simulations. Sorry, .99, but in terms on any high frequency effect that Ainslie might have discovered, your spice simulations are worthless.

No, I have not included TL effects in my modeling. I see no need at this point. The modeling of the circuit was not to obtain OU results, as those are not possible in SPICE. It has been used as a tool, and it has been quite useful here I might add.

No argument. It's a fact. Spice does not include propagation delays or transmission line reflectivity, etc. You have add your own complex models for a specific design if you want it. If you don't believe me, then show us a spice simulation that includes such transmission delays. Why would you even argue against that?

Again, SPICE is being used as a tool for exploring some aspects of the circuit's behaviour. No OU effects were ever expected. The theory of shunt inductance was tested and proven to adversely affect the results by using SPICE. Energy flow back to the source was proven using SPICE.

All this without having to model propagation delays and reflections using TL's. It never was my intention to model the circuit's every nuance. I've already explained the reasons for using it, and it has been a useful resource, for which I believe even Rose would agree with to a certain extent.

Anyhow Paul, you seem to be hell-bent on going off on a tangent away from the real subjects at hand, and I'm really beginning to tire of trying to bring you back into focus and up to speed on the project. I really don't think I owe you the trouble.

Cheers,
.99