Storing Cookies (See : http://ec.europa.eu/ipg/basics/legal/cookies/index_en.htm ) help us to bring you our services at overunity.com . If you use this website and our services you declare yourself okay with using cookies .More Infos here:
https://overunity.com/5553/privacy-policy/
If you do not agree with storing cookies, please LEAVE this website now. From the 25th of May 2018, every existing user has to accept the GDPR agreement at first login. If a user is unwilling to accept the GDPR, he should email us and request to erase his account. Many thanks for your understanding

User Menu

Custom Search

Author Topic: Testing the TK Tar Baby  (Read 1998270 times)

polln8r

  • Jr. Member
  • **
  • Posts: 81
Re: Testing the TK Tar Baby
« Reply #4845 on: September 16, 2012, 12:49:25 AM »
"Give a man an egg, he eats for a day. Teach a man to lay an egg..."

orbut 3000

  • Full Member
  • ***
  • Posts: 247
Re: Testing the TK Tar Baby
« Reply #4846 on: September 16, 2012, 01:06:06 AM »

By two different measurement methods, all mathematically correct. Right?


Kindest AND best and very well done FOR getting those omelettes.  I hope, soon, TO show you some OF our own.
Orbie

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Testing the TK Tar Baby
« Reply #4847 on: September 16, 2012, 09:43:32 AM »
Gaahh.... He still doesn't get it.

Gmeast, here is a very simple challenge for you. Please use your function generator to send a 62.5 kHz positive rectangular pulse of 25 percent duty cycle, at 12 volts peak, directly through a simple 5 Watt automotive light bulb, like a dome light. Apply the exact same power measurement techniques and calculations that you have done in the past three days with your mosfet circuit. Compare with straight DC at 3 volts -- or even 6 volts -- to the bulb. Measure and calculate and report. Use the exact same calculations, from a small shunt for current monitoring, that you have been using, just as if you were measuring your circuit.

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4848 on: September 16, 2012, 06:21:44 PM »
So, getting back to this discussion...

Uh... OK..... but the derivation in the Wiki article is shown to apply to complex waveforms as long as an rms current value can be obtained "somehow". But sure, if you are a bench tech measuring "these types of circuits" presumably meaning noisy PWMs or switching power supplies, you already have an integrating oscilloscope sitting in front of you and you know how to use it.... so you would probably not use your method directly.... even though the scope will be doing it internally.
Even with a good scope and passive probes, making a Pout measurement on an inductive-resistive switched "high-side" load is notoriously difficult, if not impossible to do accurately. Many have been down that road, including myself.

The only way to do it properly, is to use a differential voltage probe across the load resistor, and a current probe in series with it. The other option is to use an isolated front-end scope. These requirements eliminate 99.99% of all users on this forum, and at least 50% of everyone else.

Quote
I think the real contribution and novelty is in your insistence and demonstration that one can indeed make good measurements, reliable and accurate, with your method and without an expensive digital scope, as long as some basic knowledge is present.
Yes, but more importantly, this method allows the Pout measurement possible to make at all, fancy equipment or not. The only other way is to use a DC control as Rose and Greg have done, and we know this takes a fair bit of time to do, not to mention the added requirement for a variable power supply, which many don't have. My method works very well, and quickly, and without the need for a scope, nor variable power supply. AND, it is accurate.

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Testing the TK Tar Baby
« Reply #4849 on: September 16, 2012, 07:12:30 PM »
So, getting back to this discussion...
Even with a good scope and passive probes, making a Pout measurement on an inductive-resistive switched "high-side" load is notoriously difficult, if not impossible to do accurately. Many have been down that road, including myself.

The only way to do it properly, is to use a differential voltage probe across the load resistor, and a current probe in series with it. The other option is to use an isolated front-end scope. These requirements eliminate 99.99% of all users on this forum, and at least 50% of everyone else.
Yes, but more importantly, this method allows the Pout measurement possible to make at all, fancy equipment or not. The only other way is to use a DC control as Rose and Greg have done, and we know this takes a fair bit of time to do, not to mention the added requirement for a variable power supply, which many don't have. My method works very well, and quickly, and without the need for a scope, nor variable power supply. AND, it is accurate.
Yep, agreed..... I'm spoiled I guess because at the other location I have nice diff probes and that hugely expensive LeCroy DC-100MHz current probe to hand. Your method makes it possible for me to continue (if I want to for some reason) here with lesser equipment.

Ironically..... the Fluke 123 and 199 ScopeMeters that Ainslie used for the Quantum report, or claimed to use (I still don't know if it was really a 123 or 199 that she used) have isolated channel references. You can have 600 V between the "grounds" of the two probes; this allowed one (or Ainslie) to blithely make measurements all over the circuit without worrying.

Which reminds me... how is Gmeast measuring "across the load"? Is he considering Drain-to-Negative Rail to be "across the load"? Or does he have isolated probe references so he can actually measure across the load while also monitoring the main CVR on the neg rail? Or does he have enough channels available to use two probes in a differential manner for the "across the load" measurement?

It's hard for me to imagine anyone trying any kind of work like this without having a regulated, variable DC power supply or several of them. But I am constantly surprised whenever I dip into threads discussing electronics. I consider my own knowledge and bench setup to be rudimentary.... then I see people discussing resonance and coil tuning for example and they don't even have an oscilloscope and are using an old PC power supply (nothing wrong with that, I use one too for special purposes, but still.....)

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4850 on: September 16, 2012, 07:32:07 PM »
Which reminds me... how is Gmeast measuring "across the load"? Is he considering Drain-to-Negative Rail to be "across the load"? Or does he have isolated probe references so he can actually measure across the load while also monitoring the main CVR on the neg rail? Or does he have enough channels available to use two probes in a differential manner for the "across the load" measurement?
Greg thinks he is computing the input power (Pin) correctly, but he is not. For some odd unknown reason, he thinks that he should be multiplying the battery voltage by the duty cycle factor, when of course this is utterly incorrect.

As far as I know, he isn't attempting to measure the voltage across the load for any reason. Is he?

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4851 on: September 16, 2012, 11:15:58 PM »
Just a quick note to acknowledge PW and his excellent advice about using a 1 Ohm CSR vs. using a 0.25 or 0.1 Ohm CSR.

Two benefits:

1) It provides for a better signal to noise ratio and uses more bits of those 8-bit oscilloscopes, which is going to improve the measurement accuracy.

2) It reduces the Q of the CSR resistor, provided it is not a wire-wound CSR. Assuming the inductance is relatively unchanged among different values of CSR's this increases the resistance to inductance ratio and therefore the accuracy of the current being measured via the CSR.

One possible detriment is the introduction of 1 Ohm of resistance in the circuit, but most times this is innocuous.

fuzzytomcat

  • Hero Member
  • *****
  • Posts: 676
    • Open Source Research and Development
Re: Testing the TK Tar Baby
« Reply #4852 on: September 17, 2012, 01:22:46 AM »
@ gmeast,

Well I can see now you didn't research shit when it came to Rosemary's claimed Quantum COP>17 circuit ..... nothing ..... na da ...... and you call everyone else a idiot ?? Your a fool !!!!  :o

FTC
 :P

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4853 on: September 17, 2012, 01:51:52 AM »
Thanks Glen.  ;)

Actually, I posted the same pic over there in response to Greg's accusation.

ETA: Greg has now deleted my recent posts in his thread. I copied the example to my FG thread there.
« Last Edit: September 17, 2012, 04:39:33 AM by poynt99 »

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4854 on: September 17, 2012, 04:29:30 AM »
Greg has had another fit:
Hi Rosie,

Sorry, but I'm done here.  This poyn99t a_ _hole has ruined this for me.  Remove my moderator privileges and just ban me so I'm not tempted to post here again.  I've decided to NOT share any findings with anyone.

Thanks for the opportunity, but it's not worth putting in the hours and enthusiasm and then having to put up with a self righteous a _ _ wipe like poynt99.

So Rose will no doubt ban me; this time on purpose as opposed to by mistake. Greg's threatening to leave, and once Rose calms him down, he'll agree to stay and Rose will ban me...again. LOL  :-X

Greg specifically asked to see what I have built and tested in regards to Rose's circuit, so I obliged him and posted that pic of my setup from 2009. He deleted that too. ???

Evidently, Greg is afraid of something.  :o

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Testing the TK Tar Baby
« Reply #4855 on: September 17, 2012, 04:44:25 AM »
Amazing.
The guy just doesn't want to do any of his own homework, it seems. And when he's proven wrong he deletes the proof and continues to make the same accusatory claim. That is truly amazing. He's graduated from NERD Novice to full-blown NERD Collaborator with that move, right out of the Red Queen Ainslie's playbook.

It's too bad that he never actually admitted that last fundamental error. Perhaps he saw it, realised how inefficient his circuit actually is when analyzed properly that way,  and that was what prompted him to want to discard power analysis and start looking at energy instead. A fresh new fertile field to err in.   

And when the deletions and edits that change meaning start happening, you know the end is near.

poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Testing the TK Tar Baby
« Reply #4856 on: September 17, 2012, 05:06:30 AM »
Ah well, it's par for the course in these circles.  ;)

Life goes on.  ;D

polln8r

  • Jr. Member
  • **
  • Posts: 81
Re: Testing the TK Tar Baby
« Reply #4857 on: September 17, 2012, 07:43:01 AM »
Poof! Whoosh! And--in the blink of an eye--it's gone.

fuzzytomcat

  • Hero Member
  • *****
  • Posts: 676
    • Open Source Research and Development
Re: Testing the TK Tar Baby
« Reply #4858 on: September 17, 2012, 09:10:13 AM »
Poof! Whoosh! And--in the blink of an eye--it's gone.

 :-* :o :-X


FTC
 ???

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Testing the TK Tar Baby
« Reply #4859 on: September 17, 2012, 09:14:18 AM »
Well.... let's review.

Gmeast began by looking at the NERD circuit and rejecting it out of hand as unworkable, but he was curious about the other COP>17 claim from the Quantum article.

So he started building that one. Immediately he posted the "Has anyone actually built this thing" post with the "It does not work"... when he discovered that it made an exactly inverted duty cycle and could not be used to make the dutycycle claimed by Ainslie in the Quantum article.

So he modified the circuit to use his own timer or pulse generator or oscillator. And he decided to use gate driver chips since he knows mosfet circuit design and PWM circuitry and measurements so well. And he had trouble making the "Ainslie oscillations" and never really did duplicate her "random aperiodic Hartley oscillations"... since he used a scope that didn't have the aliasing and Moire patterns of the Fluke 199's low- resolution display.

But he found that a true duty cycle of 3.7 percent ON didn't produce any appreciable load heating, so he went to a duty cycle of 25 percent ON and a higher frequency than Ainslie used in order to have anything to measure at the load that could be called "heating".

But he still couldn't see anything like a recharge going back to the battery so he had to incorporate a recirculation diode that didn't appear in the Quantum circuit at all. Nevertheless, he finally got measurable load heating and was able to make DC power comparisons.... and found massive OverUnity in this "replication" of a circuit Ainslie couldn't even contemplate much less produce and test. Not 17:1, but still some respectable numbers. Clearly OU.... and clearly wrong.

He repeatedly "made the same mistake twice" by applying the duty cycle twice to his input power calculations, in several different ways, giving him a spuriously low calculated input power value. An example of this incorrect calculation method is shown below. Calculated correctly his measurements yield perfectly ordinary efficiency numbers, which at least speaks to his ability to write down numbers from boxes.

He clung to his "overunity" result in the face of several different clear... or more or less clear... explanations and examples of his major error, until finally -- apparently with egg on his face, 0.75 egg -- he melted down in a squall of insulting foulmouthed vituperation.... and a short while later he caused to be removed the entire record of his work from public view. The whole thread is simply gone.

Fortunately I have many images of his posts from that thread, so it's not completely gone from the ken of man.