PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
OverUnity Research > Benches > PhysicsProf (Moderator: PhysicsProf) > Reliable Measurements and Simulations: Power In and Out and Efficiency n
2024-11-27, 00:31:21
News: If you have a suggestion or need for a new board title, please PM the Admins.
Please remember to keep topics and posts of the FE or casual nature. :)

Pages: 1 [2] 3 4 5
Author Topic: Reliable Measurements and Simulations: Power In and Out and Efficiency n  (Read 54623 times)
Topic Options img
img  EMdevices
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #25 on: 2011-03-22, 22:58:48 »
Hero Member
*****

Posts: 805
Quote
The question is, again:

How do we evaluate the output power when the power waveform has a significant AC component and the current and voltage are not in phase?  


I would say the basic fundamental equation for average powerflow doesn't "care" what the phase relation is between the voltage or current, it all works out just fine.

P_avg = (1/T)  * integral of [ v(t) * i(t) ] over one cycle, where 'T' is the period.  


But if you want to use a scope to average, like I sense most of you want to, then you must implement another CSR resistor, like I show in my previous two schematics,  then you can get away with just using the scope MEAN function, because now you have decoupled the two energy flows (or rather the output signal is well defined now).   Most importantly though,   use DC coupling for the probes.

P_in = MEAN [ V1 * V2 ] / 1ohm

P_out = MEAN [ V3 * V4 ] / 1ohm


so then efficiency can be calculated as:

n = MEAN[ V1 * V2 ]  /  MEAN[ V3 * V4]



Do you want me and Dumped to integrate that red signal in your picture?     I'm sorry, I think I failed in my explanation so far.      You see we do not have enough resolution in that spike, the answer would not be accurate if we did integrate.    I do understand that you have to go back to the university, so when you do get there and can take another zoomed in shot then I'll be more than happy to integrate.  I hope I didn't insinuate that you do not know how to integrate.  

EM
« Last Edit: 2011-03-23, 02:47:10 by EMdevices »
   
img  Itsu
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #26 on: 2011-03-23, 00:38:32 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159

I replicated physicsProf his "transistor resonator" as close as possible, now using a MPS2222a, 128Pf as C, the rest similar.

I used the by poynt99 proposed RC/DMM method for measuring the input, and trying to use the scope across a 1 Ohm output resistor for measuring the output

The input can be measured ok, but still i cannot accurately measure the current through the 1 ohm output resistor using the scopes MEAN setting (its all over the place).

Using the FFT function on the scope, i notice fairly strong signals around the 40MHz range, so possible much energy is going out as RF.

See here a short video of the above experiment: http://www.youtube.com/watch?v=6wG-p7TrXe0

Regards Itsu
   
img  feynman
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #27 on: 2011-03-23, 02:14:06 »
Group: Guest
@humbugger
Quote
Destroy some of the spikes, huh?  Using a low-pass filter made of physical parts will average the waveform, spikes and all in a very accurate way.  Why no objection to using scope math to get an average value?  Does that not "destroy" the spikes?
-humbugger

You're correct.   If the scope MEAN is doing any sort of MA (I mean, it is an average, so this is obvious) then I suppose my objection, if a valid one, should apply to both methods.  

The sampling rate is crucial and it would have to be twice the highest frequency harmonic , no?  Aka. the Nyquist rate to avoid aliasing?

Quote
Destroy and Average are different things, Feynman.  The professor is looking for the net DC equivalent average power values; that involves averaging when an irregular waveform is the subject.  You can average using an 8-bit scope with sampling errors or you can average using a resistor and capacitor with zero sampling errors.

If the analog RC network has theoretical zero sampling errors and effectively infinite Nyquist rate , then I withdraw my objections to the bandpass lowpass method.  

Quote
P.S.  Did you take that half hour of yours to whip up the Ainslie hardware and put an end to the ridiculous speculations yet?   Grin

Unfortunately, not yet.  I've been busy with my day job, as well as a promising device I learned about today with a potential COP=8.  I only have limited time/resources, so I have to focus my areas of research carefully.  I make my objections known to Rosemary (including my request for the raw data spreadsheets), so we'll see what happens.  I like to keep an open mind.  Hopefully, more data will be forthcoming when she finds somewhere to upload the information (spreadsheets, images, video if it's available).    

The other device , possible COP=8

http://www.overunity.com/index.php?topic=10518.0

Quote
And did you ever get any scope-shots from Rosie that were useful in studying the actual waveforms of the 1.5MHz oscillations?  Nope.  She never took any by her own accounting.  Just massively undersampled slow sweep traces showing the burst envelopes...completely useless for doing scope math based on sampling.

Well that's unfortunate if there were no scope-shots of the 1.5Mhz oscillation.  I haven't checked the thread much since after I made it clear we need to be scientifically rigorous in potential open-source OU devices.  I was hoping it would sort itself out without intervention or having to bench it out.

I was planning on waiting a couple of days to see where the thread went, saw what kind of photos and data ended up coming out , and so forth.   So I'm still reserving judgment.   I haven't seen the scope traces, but maybe that's because I haven't checked the thread.  Hopefully I'll have a few minutes tomorrow to do a look over new information on her circuit.

But this unrelated COP=8 thing looks promising.   I'll probably start a thread for it.

Cheers,
Feynman
« Last Edit: 2011-03-23, 04:01:34 by feynman »
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #28 on: 2011-03-23, 02:30:30 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Nice job Itsu. I'm not sure why your Pout measurement is not working out though.

Have you tried building and testing the standard JT circuit? maybe have a go at that first, as we have a pretty good foundation of test results already.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  humbugger
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #29 on: 2011-03-23, 03:02:57 »
Group: Guest
@ Feynman

I checked it out over there.  Man, you are really on it.  Lots done in just a couple of days as far as getting the build details pinned down and sources and communications and all.  Of course I think it's another crackpot worthless idea that will never work, but you know me  C.C

Anyway, it can't be any crazier than Rosie's thing and it looks like a fun build and not horribly expensive to try out so I'll keep an eye on your progress there.  Never know...one of these crazy gizmos might actually work someday!

Humbugger

P.S.  Why couldn't you just wind some thin steel "tape" or "ribbon" over the secondary?

P.P.S.  It's low pass not bandpass  O0
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #30 on: 2011-03-23, 03:47:43 »
Group: Moderator
Hero Member
*****

Posts: 3017
Quote from: humbugger on 2011-03-23, 03:02:57
@ Feynman

I checked it out over there.  Man, you are really on it. [snip]

Anyway, it can't be any crazier than Rosie's thing and it looks like a fun build and not horribly expensive to try out so I'll keep an eye on your progress there.  Never know...one of these crazy gizmos might actually work someday!

Humbugger


Well, I certainly agree with Humbugger here -- keep up the good work Feynman, and pls let us know how it turns out (over here).  I'm liking your attitude here also, Humbugger, hope you've had a good day.
Feynman, You asked David and he responded:
Quote
Oh lastly, is it okay if I publish your responses on overunity.com and/or
make a PDF?  It might save you some headache because if this actually works
you are going to be getting the same questions over and over.  :)

By all means push out all the info to everywhere under the sun and beyond.

That is great and corresponds with my "open source" philosophy as well.  I'm going to go out on a limb here -- if your replication gets to COP>3 and if it can be independently&reliably  measured, I've got $300 to make it worthwhile for you, no strings attached, just a gift.   $500 if you trust me to do the measuring (in either case, I need to be convinced by the data before the gift is offered, of course).  Wish I could offer more, but I would like to encourage you.    Also, I would encourage you to go after the prizes here and at OU, which together approach $20K...  :o

@Itsu --

I agree with .99 that you have done good work also, and thank you for replicating and doing this advancement of the study of this circuit AND of .99's suggestion for measuring Pin.
From you video, I get
Pin =  1.6 V (battery) * 0.016 A (V across the 1ohm resistor) = 25.6 mW -- which is in good agreement with what one would expect, cf 22 mW from my first post for this (or similar) circuit, for Pin calculated by the Tek 3032.

Good point about the high freq. also, which could result in RF losses.

@EmD --  OK, I have to agree with you as I tried the "hand integration" for this curve that the spike is tough -- I really need to zoom in and get the detailed data for this rather strange (IMO) waveform.  Less strange though now that I see via the added LED with opposite polarity -- lights up along with the one I have been using.  (As I described in my last post.)  I immediately took a look at the Pout waveform using my ATTEN DSO and sure enough, now the Pout is mostly positive, so I have learned something valuable based on your suggestion and thank you for this.

Next, you wrote:

Quote
P_in = MEAN [ V1 * V2 ] / 1ohm

P_out = MEAN [ V3 * V4 ] / 1ohm


so then efficiency can be calculated as:

n = MEAN[ V1 * V2 ]  /  MEAN[ V3 * V4]

But Iout goes through the measuring resistor CSR1 also, so that power lost in CSR1 should be considered as part of Pout, right?  would you please augment your equation to include this?  Also, should one not add a CSR3 from the emitter-to-ground so as to get the power generated there? as part of Pout, you see.

Likewise, the power drop in CSR1 for Iin should be taken into account when calculating Pin.

I'd really like to see the math you come up with these "small" (but perhaps important) corrections.  In other words, the measuring resistors should not themselves skew the evaluation of n.  They are not needed for the circuit, and the power drops in the CSR's should be accounted for properly.

BTW,This is why we went to one resistor before, since it's resistance divides out when one take Pout/Pin (using the same resistor for both), but I see that your method also has merit -- and I very much like the idea of taking alternative-measurements (like the one you have proposed).

A good day!
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #31 on: 2011-03-23, 04:05:53 »
Group: Moderator
Hero Member
*****

Posts: 3017
PS - @Itsu, would you pls check  -- if you add a second LED, polarity reversed (and no diode of course), will both light up as my pair did?  
It might be a function of the toroid details, I suspect...
But if you get both LED's to light up as I did, would you check out the Pout then, using your DSO? 
« Last Edit: 2011-03-23, 04:45:13 by PhysicsProf »
   
img  EMdevices
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #32 on: 2011-03-23, 04:49:49 »
Hero Member
*****

Posts: 805
Quote
But Iout goes through the measuring resistor CSR1 also, so that power lost in CSR1 should be considered as part of Pout, right?  would you please augment your equation to include this?  Also, should one not add a CSR3 from the emitter-to-ground so as to get the power generated there? as part of Pout, you see.

Yes, in the boost configuration circuit Iout also passes through CSR1, and that's ok because the battery is also supplying power during the boost phase so this is Pin.

Also correct we are ignoring the small dissipations in the CSR1 resistor.

We can leave off the CSR3, because CSR1 accounts for the input current passing through the battery, our source of energy input.

EM

PS  power dissipation in the CSR resistors is orders of magnitude smaller then what comes out of the battery or what flows into the LED. 
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #33 on: 2011-03-23, 06:23:42 »
Group: Moderator
Hero Member
*****

Posts: 3017


Quote from: EMdevices on 2011-03-23, 04:49:49
Yes, in the boost configuration circuit Iout also passes through CSR1, and that's ok because the battery is also supplying power during the boost phase so this is Pin.

Also correct we are ignoring the small dissipations in the CSR1 resistor.

We can leave off the CSR3, because CSR1 accounts for the input current passing through the battery, our source of energy input.

EM

PS  power dissipation in the CSR resistors is orders of magnitude smaller then what comes out of the battery or what flows into the LED.  


Ah -- how do you know that power dissipation in the CSR's is " orders of magnitude smaller then what comes out of the battery or what flows into the LED"?  
.99 has promised measurements using a JT circuit to give values through CSR1 and CSR2 (and CSR3).   I'm still hoping for .99's measurements here!

 
« Last Edit: 2011-03-23, 08:37:28 by PhysicsProf »
   
img  EMdevices
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #34 on: 2011-03-23, 06:35:09 »
Hero Member
*****

Posts: 805
Here's why for these ball park values. ( I = 1 mA , Vb = 1V , R = 1 ohm)


Pr = I^2 R = 1e-6

While,

Pb = V * I = 1e-3


Got it?
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #35 on: 2011-03-23, 08:23:45 »
Group: Moderator
Hero Member
*****

Posts: 3017
Quote from: EMdevices on 2011-03-23, 06:35:09
Here's why for these ball park values. ( I = 1 mA , Vb = 1V , R = 1 ohm)


Pr = I^2 R = 1e-6

While,

Pb = V * I = 1e-3


Got it?

Not done yet -- pls calculate for V3 and V4, that is, involving the output voltage and current (typical).  Pout is the more difficult to measure...
 Really, I'd rather see experimental values.
   
img  Itsu
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #36 on: 2011-03-23, 14:08:30 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159

poynt99,

yes, i build a JT recently with good results, n=98.5%, see the "Possible breakthrough with the JouleThief (JT) circuit" thread, reply #24.


PhysicsProf,

i put in a second (polarity reversed) led (diode 1n4148 removed), and confirm both are lid.
The polarity reversed one is much brighter then the non-polarity reversed one.

But i understand from your opening post in this thread that you ARE using a 1n4148 diode on point 5 right?:

1.46V  AA battery (measured by DMM)
CSR1 = 1.1 ohms
Rb = 979 ohms
MPS2222A transistor
Capacitor (C2) = 151 pF  (note:  corrected to pF; measured)
Diode at point 5 = 1N4148    <<<<<<<<<<<<<<!
LED = red diode

So you have removed that in the mean time?

I redid the power measurements in the above situation (no 1n4148 diode, 2 x led).

Pin (RC/DMM method)                          1.487            x 0.0426            = 0.0633462 W     (63mW)
Po  (scope on emitter / across 1 Ohm)   0.09 (mean) x 0.009 (mean)  = 0.00081 W    (0.81 mW).

N = 0.81/63x100% = 1.28%

See the attachments for the circuit used, and the power results.
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #37 on: 2011-03-23, 15:55:48 »
Group: Moderator
Hero Member
*****

Posts: 3017
Itsu -- thank you for verifying that both LED's are lit, and yes of course, I did remove the diode in order to have both polarities of LED's to light up.

You're right that the JT circuit has given the best results so far -- that is true for me also.  Yet this circuit continues to intrigue and challenge...

I do recall the progress you made with the JT circuit (post 24 in the other thread), the one for which you achieved n of nearly unity. Have you tried the RC method for determining Pin for that circuit? that result could be compared with your earlier result (and perhaps posted on both threads, for continuity).
   
img  feynman
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #38 on: 2011-03-23, 20:17:07 »
Group: Guest
@PhysicsProf

Thank you PhysicsProf, that is a generous offer.  I will be more than happy to send you a device for independent testing -- if I ever get one working lol .  

Luckily, I have a job that affords me the ability to fund some degree of research.  Hopefully, this continues.  I bought two Nanoperm cores today... an M-416 and M-412 with relative magnetic permeability of u/u0 = 17,000  (very high, 2x - 100x higher than ferrite) , with moderate resistivity in the range of 1x10^-6 Ohm x Meters.  This is for a Gabriel device replication and for general experimentation.


http://www.magnetec.de/eng/universal-kerne1.htm

I'm also going to snag a couple of conventional ferrite toroidal cores (low resistivity, high permeability) ,as well as a couple of Micrometals powdered iron toroidal cores (moderate permeability, high resistivity).   This way I can do power experiments / replications with toroids with varying properties.

If all else fails, I'll make a giant 100amp Joule Thief with an IGBT and 60W bulb  , haha.  


@Itsu

Comments:

Well done on the earlier 98.5% efficiency result.  

Quote
i put in a second (polarity reversed) led (diode 1n4148 removed), and confirm both are lid.
The polarity reversed one is much brighter then the non-polarity reversed one.
-Itsu

So this is sort of like an LED-rectifier, huh?  Both of the LEDs light up with one brighter than the other?  Doesn't that mean they've got to be 'seeing' an AC signal?  

Quote
I redid the power measurements in the above situation (no 1n4148 diode, 2 x led).

Pin (RC/DMM method)                          1.487            x 0.0426            = 0.0633462 W     (63mW)
Po  (scope on emitter / across 1 Ohm)   0.09 (mean) x 0.009 (mean)  = 0.00081 W    (0.81 mW).

N = 0.81/63x100% = 1.28%

-Itsu

Okay... interesting.  I keep hearing what Bolt / overunity.com told me,  "a carefully tuned Joule Thief can get COP>1."     How I have no idea, but now there are two people with COP>1 results, so the condition (whether a measurement error or true COP>1) has been replicated.

Itsu, I have a few ideas...

1) Can you conduct several more COP trials with the exact same circuit?  It's always good science to do multiple trials , and publish all the values.  Then you can calculate mean, median, std_dev, etc.

2) Can you try powering off the circuit (remove the battery), then connect it back up and test the COP several times?  Call these "Power Cycle Trials" or something.

3) Can you try different AA batteries for power, and measure the output COP vs battery voltage?  

4) Can you try removing the scope probe and the scope ground, and see if the input power is the same?   Maybe try the above tests #1 - #3 with the scope disconnected, so you can make sure your input power is in the correct range?

I mean , either
(A) There is something wrong with the measurement protocol (RC/DMM method for input, scope on emitter for output) or
(B) JT circuits are actually overunity under certain conditions which remain to be determined, perhaps related to voltage/current phase , or perhaps to something else .

Caveats:

Could current be entering the circuit via a ground loop on the scope?  Like I suggested in #4,  Itsu, Can you see if your input power consumption changes if you remove the scope from the circuit?   One thing to remember is that removing the scope might change the Out-of-phase condition (OOP), if some sort of phasing issue is the source of the extra power.  So you'll probably have to do multiple trials for all these tests.

Other questions:

Electrical Engineers: Is there a way I can do input/power measurements with just DMMs + 1 channel scope?  I have a variety of RC components, several DMMs, and one single-channel Tektronix analog scope.  I can calculate AUC of a periodic signal by taking a photograph, and integrating the pixels.  My scope has a grid that is backlit.


My oscilloscope; Tektronix 465 with one probe

Cheers,
Feynman
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #39 on: 2011-03-24, 00:56:42 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Itsu,

It appears to me that for your Pout measurement, you are taking the mean of the voltage and current, then multiplying these together to obtain your power. This is not correct.

You must have an oscilloscope that performs MATH functions to the traces, such that you can produce an instantaneous power trace by multiplying the voltage by the current traces. THEN you apply a mean function to that product.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #40 on: 2011-03-24, 04:52:18 »
Group: Moderator
Hero Member
*****

Posts: 3017
  @.99, as I recall, Itsu explained earlier that he was using this rough way of estimating and fully realizes this is not an accurate method of determining n (multiplying Vrms*Irms, etc.)
And note that he added the RC method you recommended for determining Pin.  
  NOW if we could just get an RC method like that for Pout -- that would be great!  ;)
I have actually been thinking about some kind of capacitor to effectively integrate the output power, along with a cap for the input power... Not sure if that would work.
I'm learning, but feel like my graduate work in E&M did not prepare me very well for circuit analysis like this!   ^-^

As for myself, I'm snowed under right now working on taxes.  Ugh...  looks like I'll be pretty much out for a day or perhaps a few days from the forum.

I would like to ask .99 --  I have replies from Itsu and a PM from Humbugger that their Sim packages really can't handle the coupled-inductor component of this circuit, as presented in Post #1 of this thread.    So I'm stuck, with regard to any simulation -- unless I can ask:

1.  can your Sim package handle the coupled-inductor component of this circuit?

2.  If so, would you be willing to take a crack at it, produce some (even preliminary) results from it?  
I sure would appreciate it.

I realize you are very busy... no rush, really, as I'm tied up with tax forms at this time myself.  But if you could take a crack at this circuit, as you did with the standard JT circuit, it would be much appreciated...  Indeed, is there some sort of tutorial for using a simulation package?    I'm an experimentalist and fairly content with testing the devices at my electronics bench; but a decent simulation would sure be helpful.

TTFN  (Ta-Ta for Now)
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #41 on: 2011-03-24, 05:21:45 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Professor,

Yes, I will try my hand at simulating that version. I'm confident it will work.

Regarding Itsu's test; I have no problem with his input power measurement. The issue is with the Pout measurement. I am not certain if Itsu is aware, but the averaging can not be performed before the multiplication when looking at the output power. It is not even worth doing to get a rough idea, as we explored when the RMS values were also an issue of contention a while back. This is no different actually.

If I misunderstood what he was doing or saying, then I apologize in advance.

Unfortunately, there is no way to use this averaging method for determining output power. I will be presenting a method in the near future that will allow for an easy and accurate Pout measurement. ;)

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  Itsu
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #42 on: 2011-03-24, 09:44:39 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159

Hi Poynt99,

i indeed used the average (mean) V and i to calculate the power (mean) as this seems to me the logical thing to do having no DSO with a MATH function.
(my DSO has a math function, but only as a wave form, no values   :(  ).
But i stand corrected, and will reread your post about this as obviously i did not fully grasped it.


PhysicsProf,

I will retest my JT setup with the MPS2222a transistor to be more "in line" with your setup, and see if i can get similar results as with the bc547.
But i obviously have a "problem" in getting the correct Pout values, so have to wait for a "rock solid" Pout measurement method.


Feynman,

thanks for your comments, indeed, on the emitter i have a 2.38Mhz rf signal, see IMG_2144.jpg above.
Concerning the mentioned COP>1, my result (1.28%) can hardly considered to be that, so i think you overlooked the . (period).
But thanks for the ideas, i will continue to experiment with these both circuits.

Regards Itsu.
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #43 on: 2011-03-24, 14:04:14 »
Group: Moderator
Hero Member
*****

Posts: 3017
(Taking a break to reply briefly:  )
Quote from: poynt99 on 2011-03-24, 05:21:45
Professor,

Yes, I will try my hand at simulating that version. I'm confident it will work.

[snip] I will be presenting a method in the near future that will allow for an easy and accurate Pout measurement. ;)

.99

Great news!  I look forward, .99, to both the simulation and your proposed method for " an easy and accurate Pout measurement" !   O0

@Feynman:  The n ~ 1.06 result was obtained with the reverse-JT circuit described in post #1 of this thread, with one addition:  a 221 pF capacitor ACROSS (parallel to) inductor L1 (call the cap C1 that goes with L1, and C2 that goes with L2).  See circuit schematic in Post #1.  (Yes, just one LED in the circuit at the time; as noted above, adding a second LED with reversed polarity will allow current to flow both ways and should allow for larger Pout).   C2 was 151 pF . 
Again, the measurements were made with the Tek 3032, Mean Pin and Mean Pout evaluated by the scope MATH function, as described earlier. 

I was just experimenting with the circuit, but it makes some sense, perhaps...  A simulator would be nice for this, but I've used the following to calculate the resonance frequencies of the two TANK CIRCUITS set up with the L1-C1 pair, and the L2-C2 pair, using:

http://chemandy.com/calculators/tank-resonance-calculator.htm

Results of the tank-circuit calculations:
L2 = 81 uH (OK, approximate at this high frequency; measured at 10KHz; see post #1)
C2 = 151 pF
Resonant f = 1.44 MHz.

Observed self-resonant frequency of this reverse-JT oscillator:  1.75 MHz (data given in post #1; f = 1/Period)
So that is close and suggests that the the L2-C2 tank-circuit may dominate the freq that the whole circuit settles into...   Oh, yes, I would like to know  how the circuit "chooses" its self-resonant frequency.

Next, with the added cap over L1, that gave the rather anomalous 1.06 result for n:

L1 = 67 uH (again, approximate at this high frequency; measured at 10KHz; see post #1)
C1= 221 pF
Resonant f for this tank circuit = 1.31 MHz.

Now, when both tank circuits are tuned to the same freq, AND two LED's are present in the output circuit (having reversed polarities), we should, I think, realize a maximum "push" of the current through the LED's.  (sorry, little time right now to explain my reasoning -- but briefly, the L1-C1 tank circuit acts as 'infinite' resistance at its resonant f, forcing current to go through the LED's.)

Leaving C2 the same, one finds that C1 ~ 182 pF will bring the L1-C1 tank circuit to 1.44 MHz, the same as the L2-C2 tank circuit.  By the equations.  In practice, I will add a variable cap to the circuit and see if I can tune things in...  whew...   ^-^
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #44 on: 2011-03-25, 15:28:52 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
I have some time today (I'm at home burnt out from 3 busy weeks at work) so I am going to try this simulation. Hopefully I'll be posting some results in the next hour or so.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #45 on: 2011-03-25, 17:58:20 »
Group: Moderator
Hero Member
*****

Posts: 3017
I like forward to your results, .99.

I'm heading up to the University for testing of the reverse-JT with two LED's (in parallel, opposite polarities) in the output.  Just a short time today, but hope to have results to post either late tonight or tomorrow am.
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #46 on: 2011-03-25, 19:40:26 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
The simulation ran first time out of the box.  O0

Below are the schematic and wave forms.

Other than a lower duty cycle, I think I've faithfully reproduced your wave forms professor.

You will note that the computed MEAN power of the csr1 voltage times the collector voltage is -7mW. The power is negative because the csr1 voltage is measured backwards. So actually, this indicates a sourced power of 7mW to the transistor part of the circuit. You will also note in the next wave form shot, that the actual power in both the diode and LED are nearly zero watts, a mere 14uW to be exact.

This is a good example of when this method does not work as a power measurement. Making the assumption that there is significant current in the LED may be incorrect, even though it may be illuminating.

Your Pout measurement should have come out to a higher negative value, according to my simulation. Perhaps the undersampling is having a marked effect here? At any rate, there should be very little power in the diode/LED oriented A-K as per your diagram, because the positive voltage off the collector does not seem to go much more than +1.5V. The fact that the LED does light is still interesting though. The reversed LED should light quite well however, as the negative voltage off the collector is over -6V in my sim. It appears to be about -5V in your scope measurement.

The battery power (-20.5mW) is in line with what you measured.

As a side note, it would appear that this configuration is a poor one if the goal is to transfer power to the LED, as there is about 12mW of power wasted just in the base resistor R1.

Cheers,
.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  poynt99
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #47 on: 2011-03-25, 20:55:06 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Itsu,

If your scope has the ability to export it's raw data, then we can probably perform your Pout measurement that way. Let me know if you are interested in trying this.

.99


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  PhysicsProf
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #48 on: 2011-03-26, 17:31:00 »
Group: Moderator
Hero Member
*****

Posts: 3017
Thanks for doing the SIM, .99...

I saw your schematic after I had left for the university, and to compare with yesterday's data especially, I ask if you could make a few changes:

a) 2 Kohm resistor to the base R1.
(you're right:  "As a side note, it would appear that this configuration is a poor one if the goal is to transfer power to the LED, as there is about 12mW of power wasted just in the base resistor R1."-- that was with 1Kohm.  What will it be with 2kOhms?  and is there any way to remove this resistor once the circuit is oscillating??  I haven't found a way...)  This is the sort of thing that I think is great about a good simulation -- tells you where the power is being wasted and allows you to "tweak" the circuit to make improvements in efficiency.

b)  Remove the 1N4148 diode and put in TWO red Led's, in parallel, with opposing polarities.  
c) Voltage?  Both light up brightly for Vin of about 1.45 V (AA battery) -- is this the voltage you used?   Also, please repeat for 1Volt in, and 2Volts input.
d) Remove CSR-sim off the LED (I'm not sure what you mean by this actually; clarify?) and do all measurements across CSR1 (to compare with the method I used yesterday, method suggested by Laneal).

Now, I had a big puzzle when I started up yesterday, with the configuration described above -- and with AA battery for input.  n was coming out consistently > 140%...  I show the data below -- took me just a little while to spot there was a BIG problem... can anyone see the clues that there is a problem in the measuring?  
Took me a little longer to locate the root cause of the problem (which affected the measurements!).    For fun, and education.
Then I'll get back later with "real data" -- with the problem removed.
   
img  feynman
Re: Reliable Measurements and Simulations: Power In and Out and Efficiency n « Reply #49 on: 2011-03-26, 18:48:43 »
Group: Guest
Was your problem the sign (+/-) of the power curve?
   
Pages: 1 [2] 3 4 5
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-27, 00:31:21