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: Is joule thief circuit gets overunity?  (Read 600478 times)

ltseung888

  • Hero Member
  • *****
  • Posts: 4363
Re: Is joule thief circuit gets overunity?
« Reply #840 on: May 15, 2013, 12:55:04 AM »
Ok, that is interesting. However, something doesn't look right with your input current waveform. It is showing as mostly negative when it should be positive current pulses, at least for the most part. That could be where the problem is. Can you confirm that you had channel 2 set to DC coupling? It looks like the coupling was set to AC?
 
*** There was no problem with CH2.  It was set to DC.  Please read the poynt99 explanation related to the circuit diagram allowing common connection for all 4 probes.

I believe the RMS readings will not be useable if your waveform is not completely positive (pulsating DC) or if the waveform has both positive and negative swings and the positive and negative swings are not very symmetrical. In the scope shots you posted it looks like you didn't have your channel 2 probe set to DC coupling when you measured the input current. If the current waveform were showing as pulsating DC (all positive), I think the RMS reading would provide you with a quite different result.

I did a quick test with the input voltage set at 0.4V, and I do not get much noticeable negative swing on the input or output current waveforms. The output current is starting to get down fairly close to the noise level at this input voltage however. I also tried setting the input voltage to 0.36V, and although my LED still glows dimly at this input voltage, the output current is just barely detectable as very tiny blips on the scope trace, so the output current is unmeasurable on my scope at this input voltage. The output current is right in the scope's noise level at this input voltage.
 
*** I am not surprised that your results do not show the Zhou Board effects.  Two thirds of the Boards I built myself did not show that - even though they all lighted the LEDs.  The passing test of the Zhou Boards is the crossing of the 0 ref line.  I shall send you Boards 118 and 119 after the tests on them are done in here.  Take your time.  Be parepared to find a 4-CH Tektronics to do the final confirmation.
 


Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #841 on: May 15, 2013, 01:39:45 AM »
Please read the poynt99 explanation related to the circuit diagram allowing common connection for all 4 probes.
I am not surprised that your results do not show the Zhou Board effects. 
Two thirds of the Boards I built myself did not show that - even though they all lighted the LEDs. 
The passing test of the Zhou Boards is the crossing of the 0 ref line. 
I shall send you Boards 118 and 119 after the tests on them are done in here. 
Take your time.  Be prepared to find a 4-CH Tektronics to do the final confirmation.

Hi Lawrence. Ok, if the scope probe was set to DC then I believe that the problem with the input current waveform is that your input current waveform is inverted. Looking at the way you have your scope probes connected for the 4 channel scope measurement method (based on a diagram you posted earlier this year), it seems to me that this will invert your input current waveform. I believe this will throw off your instantaneous power calculations unless you set channel 2 for the input current measurement to the Inverted setting in your scope. The input current waveform should be pulsating DC, but I do see that it does go a bit negative in your waveform (when viewed inverted from what you showed in your scope screenshot). I can explain further if you like, but it has to do with the direction that the current is flowing and the way you have the channel 2 scope probe connected for the input current measurement in your 4 channel scope measurement method.


poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Is joule thief circuit gets overunity?
« Reply #842 on: May 15, 2013, 02:22:40 AM »
Void,

It does not really matter which way the current-measuring probe is placed, as long as there is not a grounding issue (i.e. placing the probe grounds at different points in the circuit).

The real issue is not whether the current trace is positive or negative (we know the battery loses energy), but if it is BOTH. I have shown, using two different scopes, a 4-CH Tek, and a 20MHz analog Hitachi, that the current trace does NOT cross the zero reference line while measuring board #33 that Lawrence sent me.

The ATTEN scopes seem to have a serious issue with DC offset when making this measurement, as there is a relatively significant 12mV or so causing the apparent "bipolar" current trace.

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Is joule thief circuit gets overunity?
« Reply #843 on: May 15, 2013, 02:40:09 AM »
@void: I think we've been over the meaning of the negative value for the current trace as detected by the probes in use according to the schematic posted earlier. Lawrence understands that a negative value for the current detected here means normal, conventional current flowing from positive to negative, moving power from the battery to be dissipated in the circuit. Not only that, the spreadsheet recorded values do not respect the "channel invert" function, so as long as the probes are positioned as they are, the spreadsheet will have negative values when the current is "normal", no matter whether the _displayed_ trace is inverted or not.

As .99 says, the important issue is the zero-crossing behaviour. I thought briefly that I had detected zero-crossing on my system but I had accidentally selected AC coupling for that measurement and didn't notice it... there's no indication on my analog screen of the coupling setting, like there is on the Atten screen. It's easy to see that Lawrence is using DC coupling for all his recent measurements, although he didn't appreciate the difference until we pointed out that his earlier use of AC coupling was an error. All my more careful measurements do not show this zero-crossing with my boards after all; the current is always "negative" meaning normal conventional current flowing out of the battery to dissipate power in the circuit.

The Atten scopes would appear to be not suitable for these measurements, and should be put away. They simply are not precise and accurate enough to support claims of overunity or unusually high COP in electrical circuits of such low absolute power.


Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #844 on: May 15, 2013, 02:43:46 AM »
Void,

It does not really matter which way the current-measuring probe is placed, as long as there is not a grounding issue (i.e. placing the probe grounds at different points in the circuit).

The real issue is not whether the current trace is positive or negative (we know the battery loses energy), but if it is BOTH. I have shown, using two different scopes, a 4-CH Tek, and a 20MHz analog Hitachi, that the current trace does NOT cross the zero reference line while measuring board #33 that Lawrence sent me.

The ATTEN scopes seem to have a serious issue with DC offset when making this measurement, as there is a relatively significant 12mV or so causing the apparent "bipolar" current trace.

Hi Poynt99. Ok on that.  I wonder if Lawrence has ever run the self-calibration routines on his Atten scopes?
However, I would think that if the current waveform is inverted it will give errors in the instantaneous power calculations, as many of the input current waveform data points will be recorded as negative when they are actually positive, and vice versa if there is any actual negative component to the current waveform.  This would make the instantaneous power calculations and average power calculation invalid, no?


poynt99

  • TPU-Elite
  • Hero Member
  • *******
  • Posts: 3582
Re: Is joule thief circuit gets overunity?
« Reply #845 on: May 15, 2013, 03:12:17 AM »
Hi Poynt99. Ok on that.  I wonder if Lawrence has ever run the self-calibration routines on his Atten scopes?
That's an unknown.

Quote

However, I would think that if the current waveform is inverted it will give errors in the instantaneous power calculations, as many of the input current waveform data points will be recorded as negative when they are actually positive. This would make the instantaneous power calculations and average power calculation invalid, no?
It's not a problem. The only affect would be a reverse in polarity. The magnitude is most important, and it would remain intact.

Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #846 on: May 15, 2013, 03:20:30 AM »
Not a problem. The only affect would be a reverse in polarity. The magnitude is most important, and it would remain intact.

OK, thanks for the clarification. Lawrence is planning on sending me a couple of his boards for testing so we can compare my measurement results to his measurement results on those boards and see if there is a difference.

@Lawrence, based on the offset issue that Poynt99 suspects with your scope, I think that it would be a good idea to run the self-calibration routines on your Atten scopes if you have not done this previously. This could potentially correct any offset error in your scopes if there is some.

Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #847 on: May 15, 2013, 03:37:08 AM »
@Tinselkoala, Ok, on that. If Lawrence has not run the self calibration function on his scopes recently, then I will see if Lawrence is willing to do this and then re-do his measurements on his two boards before he sends them to me. The self-calibration may help correct any offset error his scopes have now.

ltseung888

  • Hero Member
  • *****
  • Posts: 4363
Re: Is joule thief circuit gets overunity?
« Reply #848 on: May 15, 2013, 06:08:58 AM »
@Tinselkoala, Ok, on that. If Lawrence has not run the self calibration function on his scopes recently, then I will see if Lawrence is willing to do this and then re-do his measurements on his two boards before he sends them to me. The self-calibration may help correct any offset error his scopes have now.
No problem.  I have run the Atten self califbration procedure many times now.  I shall redo it before any serious measurements.
@Poynt99
 
The Atten is not only showing "cross 0 reference line" that can be explained by DC Offset.  It is also showing the voltage spikes crossing the 0 reference line.  When that happens, the LED can be ON for days (Board 116 lasted for 3 days at 0.352V).  I shall try to repeat it for Board 118 and 119.

Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #849 on: May 15, 2013, 04:32:33 PM »
No problem.  I have run the Atten self califbration procedure many times now.  I shall redo it before any serious measurements.
@Poynt99
 
The Atten is not only showing "cross 0 reference line" that can be explained by DC Offset.  It is also showing the voltage spikes crossing the 0 reference line.  When that happens, the LED can be ON for days (Board 116 lasted for 3 days at 0.352V).  I shall try to repeat it for Board 118 and 119.

Thanks Lawrence. I actually have two DSO scopes. A 50MHz Hantek USB peripheral plug in type scope, which is what I  used on my previous tests on my joule thief circuit here, and a 100MHz Siglent 1102CNL scope, which is what I am going to use to run tests on your boards, since the Siglent should be a better scope overall, and it can be set down to 2mV/div, whereas my Hantek can only be set down to 10mV/div.

I did some testing with my Siglent scope yesterday. I warmed up my scope for 20 minutes and then configured each probe/channel to DC, x1, and 2mV/div, and then grounded the probe tips to their ground clips. There was some DC offset showing on each channel, so I ran the self-calibration function on my scope. (My Hantek scope instructs connecting the scope probe tips to the ground lug on my scope when running the self-calibration, but my Siglent scope instructs to disconnect the probes from the scope when running the self-calibration). With the scope probes again connected to the scope, I then double checked that each probe/channel was still configured for DC, x1, and 2mV/div, and then grounded the probe tips to their ground clips. Channel 1 was showing no offset, but Channel 2 was still showing about + 0.6mV of offset on the 2mV/div setting with the scope tip grounded to its ground clip. I then left my scope turned on overnight and then re-ran the self-calibration function on my scope in the morning, and then rechecked the Channel 1 and Channel 2 DC offsets with the probe tips grounded. Both channels are now showing very close to zero offset (with settings: DC, x1, and 2mV/div). Just wanted to mention this in case it helps with your scopes. I am not sure if leaving the scope to warm up a lot longer helped, or if you just sometimes have to run the self-calibration a few times and keep checking for the least amount of DC offset.

By the way Lawrence, I read that Siglent is actually the manufacturer of Atten scopes, so our scopes should be very similar. The self-calibration method I describe above should be very similar if not exactly the same for your Atten scopes. What model of Atten scopes do you have?
« Last Edit: May 15, 2013, 07:27:33 PM by Void »

ltseung888

  • Hero Member
  • *****
  • Posts: 4363
Re: Is joule thief circuit gets overunity?
« Reply #850 on: May 15, 2013, 09:22:22 PM »
Thanks Lawrence. I actually have two DSO scopes. A 50MHz Hantek USB peripheral plug in type scope, which is what I  used on my previous tests on my joule thief circuit here, and a 100MHz Siglent 1102CNL scope, which is what I am going to use to run tests on your boards, since the Siglent should be a better scope overall, and it can be set down to 2mV/div, whereas my Hantek can only be set down to 10mV/div.

I did some testing with my Siglent scope yesterday. I warmed up my scope for 20 minutes and then configured each probe/channel to DC, x1, and 2mV/div, and then grounded the probe tips to their ground clips. There was some DC offset showing on each channel, so I ran the self-calibration function on my scope. (My Hantek scope instructs connecting the scope probe tips to the ground lug on my scope when running the self-calibration, but my Siglent scope instructs to disconnect the probes from the scope when running the self-calibration). With the scope probes again connected to the scope, I then double checked that each probe/channel was still configured for DC, x1, and 2mV/div, and then grounded the probe tips to their ground clips. Channel 1 was showing no offset, but Channel 2 was still showing about + 0.6mV of offset on the 2mV/div setting with the scope tip grounded to its ground clip. I then left my scope turned on overnight and then re-ran the self-calibration function on my scope in the morning, and then rechecked the Channel 1 and Channel 2 DC offsets with the probe tips grounded. Both channels are now showing very close to zero offset (with settings: DC, x1, and 2mV/div). Just wanted to mention this in case it helps with your scopes. I am not sure if leaving the scope to warm up a lot longer helped, or if you just sometimes have to run the self-calibration a few times and keep checking for the least amount of DC offset.

By the way Lawrence, I read that Siglent is actually the manufacturer of Atten scopes, so our scopes should be very similar. The self-calibration method I describe above should be very similar if not exactly the same for your Atten scopes. What model of Atten scopes do you have?
I have Atten ADS1062CA and Atten ADS1062CAL.  Previously, I always thought that the warm up time is to get the Atten to normal room temperature of around 20 degrees C.  In cold countries, the warm up time may take longer.  But I can easily turn it on overnight.
 
Did you detect any spikes at all with your tests?  I can easily detect them.  Example - see reply 835.

Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #851 on: May 15, 2013, 09:58:14 PM »
I have Atten ADS1062CA and Atten ADS1062CAL.  Previously, I always thought that the warm up time is to get the Atten to normal room temperature of around 20 degrees C.  In cold countries, the warm up time may take longer.  But I can easily turn it on overnight.
 
Did you detect any spikes at all with your tests?  I can easily detect them.  Example - see reply 835.

Hi Lawrence. Warm up time is to allow time for all sensitive components in the circuitry to reach their steady state operating temperature. 30 mins. to 1 hour should be long enough. After sufficient warm up time, if you test the DC offset for each scope channel as I described in my previous post, and then try the self-calibration, and then retest the DC offset, I would be interested to hear what kind of DC offsets you are measuring on each channel.

I am not detecting current spikes like you show in those scope shots in my circuit.

I also tried another couple of informal tests where I changed the collector winding turns count to 18 turns and also to 39 turns, and I did not see any AC current spikes (or any zero crossing at all to speak of on the current waveform), and the efficiency of the circuit also did not change more than about 1% with these different collector winding turns count changes as well. Efficiency stayed within +/- 1% or so of about 70% with collector winding turns counts of 18, 28, and 39 turns. Changes to the collector winding turns count does noticeably change the frequency of oscillation however.
18 turns - 71.53 kHz,   efficiency = 70.93%
28 turns  - 57.14 kHz,  efficiency = 68.93%
39 turns - 43.48 kHz,   efficiency = 71.29%
I was surprised to see that the circuit efficiency seems to have remained about the same for these different collector winding counts.

Note: I was using the RMS measurement feature on my scope to make these measurements as it was just an informal test to see what effect changing the collector windings turn count might have on circuit performance. I will use data logging and instantaneous and Average power calculations when doing any tests on your boards. I made the above measurements using my Siglent scope, which may be somewhat more accurate than my Hantek scope. That may account for the difference of measuring around 70% efficiency this time and measuring around 75% efficiency previously with my Hantek scope. It might also be that my ferrite core which is in two halves and is just taped together, has separated a little bit more than before. I really should bolt the two halves together tight so they can't move, but these tests I am doing right now are just informal to get a general idea.





ltseung888

  • Hero Member
  • *****
  • Posts: 4363
Re: Is joule thief circuit gets overunity?
« Reply #852 on: May 16, 2013, 08:25:55 AM »
I turned on the Atten for over an hour before doing anything.  The first thing is to press the default setup button and do the calibration.  I then set the reading to the ones normally used.
 
The five boards were tested:
Board 89, 124, 125, 118 and 119.
 
The Input Power supply was adjusted so that the LED was just turned on.  Here are the results.  All five boards showed current crossing 0 ref line.....

Void

  • Hero Member
  • *****
  • Posts: 2333
Re: Is joule thief circuit gets overunity?
« Reply #853 on: May 16, 2013, 08:39:29 AM »
I turned on the Atten for over an hour before doing anything.  The first thing is to press the default setup button and do the calibration.  I then set the reading to the ones normally used.
 
The five boards were tested:
Board 89, 124, 125, 118 and 119.
 
The Input Power supply was adjusted so that the LED was just turned on.  Here are the results.  All five boards showed current crossing 0 ref line.....

That is very interesting Lawrence. Thanks for checking the calibration on your scope. I don't really understand what is causing those AC current spikes in your scope shots. Your input current traces are very different from what I see in my circuit. I look forward to running tests with your boards to see if I see the same sort of results.

TinselKoala

  • Hero Member
  • *****
  • Posts: 13958
Re: Is joule thief circuit gets overunity?
« Reply #854 on: May 16, 2013, 11:18:53 AM »
This is getting rather ridiculous, don't you think?

Are you really looking for, and trying to reproduce, some kind of OU evidence that only shows up with the cheapest possible DSOs used incorrectly? With absolutely NO concurrent validation (a technical term meaning evidence of OU from some OTHER experimental or measurement technique running alongside the Atten measurements)?

Note again, that after all this time, the definitive experiment is still not being done. Charge a known capacitor to a known voltage. Then you know precisely the energy available for whatever. Use this capacitor to run the JT circuit while monitoring its voltage and the time accurately. Monitor the output power in your favourite manner. When the capacitor has dropped by a known voltage increment, you will KNOW just how much energy you have put into your circuit, and you will KNOW for how long.... thus you will KNOW, not guess or interpret, just how much input ENERGY you are dealing with during that time. This figure can be directly compared to the output power, or more precisely to the integrated output power over the time interval concerned (the output ENERGY).

It will do no one any good if I do this measurement. It must be done on several boards that Lawrence claims do show OU behaviour, preferably by Lawrence himself, but anyone who has an "approved" board, an oscilloscope, a stopwatch, an accurate voltmeter, and a bunch of capacitors can do the work in a day or two with enough runs for statistical significance to result IF there is anything to see.

When you see that every capacitor you try results in efficiencies of under 100 percent, then you can start resorting to your special pleading that "the robust and confirmed OU effect only happens with batteries. No, not those batteries, or those, not just any batteries, but only rechargeable batteries of specific chemistry and manufacturer, and only when they are between 0.341 and 0.351 volts".... and only when they are pointed northeast, on Tuesday mornings, when the weather is clear......

Whatever, dudes.