Zoltan's Simulation

[poynt99]

Correction - Some Revised Results

I've re-discovered that there is no such thing as RMS power. I had known this some time ago, but forgot and began using it again. My bad.

Turns out that getting the RMS power of any device in SPICE is not correct. RMS values are only valid for current or voltage. And once again, there is no such thing as RMS power! Only Average power.

How do you arrive at Average Power? You use an RMS voltage (V^2/R) or RMS current (I^2*R) , or both (V x I).

So the new results don't show overunity. The COP is now 0.867:1 (using R1 power/source power). See the attached scope shot of all the dissipative powers, which almost add up perfectly btw.

I have found a strange thing in how PSpice does an Average power calculation though. I plotted the V1 power using two methods: 1) by using Vrms x Irms for V1, and 2) by just using the built-in Average power math function for V1. The two values are not quite the same as you can see by the scope shot.

Any ideas why the difference?

.99

[poynt99]

Similar Problem with V2

Similar to the power discrepancy with V1, I have now shown a similar problem with V2.

There is definitely RMS current being used from V2 yet the average power is almost nil.

Again I used the two methods to calculate Average power in V2, and have obtained two very different results.

I wonder if the problem is reactive power vs. real power? V2 is driving mainly a capacitive load in the MOSFET parasitics, so is there any real power drawn from V2 then? not much, but we still have the two methods giving different values. Is one showing reactive and the other real power? Perhaps.

.99

[poynt99]

RMS-Power Article

Here's the short but excellent (and quite basic) paper that got me back on track regarding Average and RMS power.

.99

[MileHigh]

Bubble burst

.99: My heart is broken! lol

>
I have found a strange thing in how PSpice does an Average power calculation though. I plotted the V1 power using two methods: 1) by using Vrms x Irms for V1, and 2) by just using the built-in Average power math function for V1. The two values are not quite the same as you can see by the scope shot.

Any ideas why the difference?
>

Yes. When you use an RMS current or voltage you have to multiply/divide it by the resistor value only, a constant. That's because a single RMS value has allready factored in the root-mean square relationship of the energy waveform to normalize it. The RMS value "favours" higher voltages or currents and they "count for more" to make the power calculation correct. Therefore it does not make sense to multiply RMS current by RMS voltage, you are factoring in the "power correction" twice and that's why you get a larger wattage value.

My parents went to Florida and I couldn't do a drag-and-drop frown, all I got was a lousy URL.

[MileHigh]

.99:

> I wonder if the problem is reactive power vs. real power?

Absolutely. It also applies to my previous posting. The MOSFET input consumes no current, and the gate capacitor is microscopic. So the power is essentaially all reactive power =(roughly) no power.

But if you do the discredited(!) RMS x RMS multiplication, you are now compeltely ignoring all reactive effects, and are getting a big time fake power calculation, which you don't want! Did I say it also applies to my previous posting? lol

Incidently, there is a pretty much linear relationship between switching frequency and power consumption in cases like this. i.e.; your Core2Duo or 4004 consumes more power the faster you clock it.

MileHigh

[poynt99]

.99: My heart is broken! lol

Yes. When you use an RMS current or voltage you have to multiply/divide it by the resistor value only, a constant. That's because a single RMS value has allready factored in the root-mean square relationship of the energy waveform to normalize it. The RMS value "favours" higher voltages or currents and they "count for more" to make the power calculation correct. Therefore it does not make sense to multiply RMS current by RMS voltage, you are factoring in the "power correction" twice and that's why you get a larger wattage value.

Actually MileHigh,

This turns out NOT to be the case (much to my relief!). According to the article I posted, AND a simple test using only resistive elements, the two power calculations are equal. That is I put a 50% pulse wave (no neg excursion) into a resistor.

One calculation was done by multiplying VR(rms) x IR(rms), and the other using the AVG(P) math function built in. Again the results were the same and are correct.

.99

[MileHigh]

.99: The pulse wave has a flat top, like a square wave? That is not a good waveform to use because it's average value and it's RMS value are the same.

You should get different results if you use a sine wave.

MileHigh

[poynt99]

.99: The pulse wave has a flat top, like a square wave? That is not a good waveform to use because it's average value and it's RMS value are the same.

You should get different results if you use a sine wave.

MileHigh

You sure?

A 10V 50% pulse (again no negative excursion) has an:

average voltage of 5V
RMS voltage of 7.07V

.99

[MileHigh]

Where are you getting the 7.07 volts from?

[poynt99]

Where are you getting the 7.07 volts from?

Vrms = Root[V^2 (PW/T)]

In the example I gave above:
Root[10 x 10 (25u/50u)]
Root[100 (0.5)]
Root[50]
= 7.07V

.99

[poynt99]

MH,

I think you may be getting confused with a square wave vs. a pulse train (i.e pulsed-DC).

The Vrms of a 50% duty square wave (i.e. +/-5V) is 5V.

BUT, the Vave of the same wave is 0V.

.99

[MileHigh]

.99: Thanks for the 7.07 volts info and crunched the numbers and realize that I was just making a dumb mistake. I got thrown off because it is the same thing you multiply the peak of a sine wave to give you your RMS voltage also.

MileHigh

[RAMSET]

325% OU thanks Daniel at OU

99
325% OU transformer PDF

Got a hot one here

Witts Flux Switched Transformer 10a

EssexMagnetWireUltraShieldPlus

Chet

[RAMSET]

links

Links work in Rosemaries thread post # 878
Chet

[EMCSQ]

Sorry for the resurrection of this thread.
But can someone give me advice how to configure this 'mutual inductance coil with nonlinearity' in PSpice, because I want to figure out it myself.

My 50c measuring power:
AVG(Power) is proportional to RMS(Voltage) and RMS(Amps)
But when you have a time graph , as PSpice provides, you don't need to calculate AVG nor RMS. You only need to compare the Volts * Amps that are going in and that are going out. When finishing your simulations, with many many cycles you will have an AVG(Power) for IN and OUT for the whole time duration.
For this circuit (Time-Diagram)
The IN - Power is proportional to the IN-current because IN voltage is a constant. The OUT-Power we need both Volts and Amps as values multiplied.
The instant values of IN and OUT - Power should be drawn in the graph.