PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
OverUnity Research > Electrical / Electronic Theory and Learning Center > Joule Thief Circuits > Lawrence Tseung sent a Prototype to test... any comments?
2024-11-26, 17:38:47
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 6 7 [8] 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Author Topic: Lawrence Tseung sent a Prototype to test... any comments?  (Read 342811 times)
Topic Options img
img  Groundloop
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #175 on: 2011-01-15, 00:43:41 »
Sr. Member
****

Posts: 336
Here is my way of testing if ANY boost converter is over unity or not.




You just connect a AA battery, an boost converter (JT) and a battery capacitor to the circuit and let it run.
The automated feed back circuit will take its power to run from the AA battery. The micro controller needs +5 Volt to run so we use an IC to boost the 1, 2 Volt up to 5 volt. The micro controller runs slowly on a 32 KHz crystal and uses approx. 60uA from the IC to run.  So the total current usage of the circuit is approx. 0,25mA from the AA battery. The circuit will monitor both the battery capacitor voltage and the AA battery voltage. If the AA battery voltage is too low, then the circuit will attempt to charge the AA battery from the battery capacitor. If the battery capacitor gets too high in voltage then the circuit will dump energy to the output load, thus keeping the battery capacitor below the recommended 2, 7 volt.

The project (design files, software and documentation) can be downloaded here: http://home.no/ufoufoufoufo/feedback.rar

Alex.
   
img  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #176 on: 2011-01-15, 00:51:57 »
Group: Guest
Quote from: poynt99 on 2011-01-14, 22:36:21
Professor,

Regarding the use of the DPO3032 Tek scope, go here to download the manual, then go to page 108 and follow the steps to create a multiply math function. It even shows an example of "power" by multiplying the voltage and current.

Then go to the "Measurement" menu and select "Mean" (not "Cycle Mean") as shown on page 99 of the manual. This of course is applied to the math wave form. Now you will have a running average of the REAL power.

As I have stated a couple times now, integration is not necessary to obtain a COP value.

.99



Oh God,

If you had commented on my numerous scope postings since July 2010 at the overunity.com forum, I could have quoted the correct COP value!   Even the China-made scopes had that functionality.

It is not too late.  I shall pass this extremely important piece of information to all the teams in Hong Kong and China.  We can claim that we have the correct COP now and ignore the Tseung FLEET comparison index.  There are dozens if not hundreds of COP > 1 devices in our possession.

Thanks so much Poynt99.

You are an angel. May God bless you.
   
img  poynt99
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #177 on: 2011-01-15, 01:04:13 »

Group: Administrator
Hero Member
*****

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

I am simply stating the facts. The most important factor regarding the scopes, is that they can perform both multiplication, and mean averaging.

Keep in mind however, that it's the old adage; "garbage in, garbage out", so ensure that your raw data is properly-acquired. I still have a great deal of doubt that good data can be obtained with passive probes, but you will have to take a chance I suppose, unless the Professor can borrow a current probe.

.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  aetherevarising
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #178 on: 2011-01-15, 01:10:12 »
Group: Guest
Quote from: poynt99 on 2011-01-14, 22:36:21
Professor,

Regarding the use of the DPO3032 Tek scope, go here to download the manual, then go to page 108 and follow the steps to create a multiply math function. It even shows an example of "power" by multiplying the voltage and current.

Then go to the "Measurement" menu and select "Mean" (not "Cycle Mean") as shown on page 99 of the manual. This of course is applied to the math wave form. Now you will have a running average of the REAL power.

As I have stated a couple times now, integration is not necessary to obtain a COP value.

.99



Poynty - I'm inclined to agree with you.  It is, perhaps, NOT NECESSARY to apply an integrated value to these results.  Unfortunately mainstream and academic experts consider it NECESSARY.  I'm afraid their argument is based on the fact that 'mean average' and the properly integrated values DO NOT CORRESPOND.  Since the latter is invariably the more modest value then I'm inclined to think this will become the required benchmark.  And as there really is no way within a simple averaging - to incorporate a proportional 'time' difference between when the energy is being returned and when it is being delivered - then - the chances are that they've got a strong argument.  If they corresponded - then your argument would carry. 

And regarding Professor's input here.  He has advised us REPEATEDLY that he is showing us his dirty washing.  We're all - I think - rather glad of this.  It will be ample opportunity to show us all where measurements have been erroneous and where they'll correspond to what it correct.  I see this as his 'teaching exercise' and frankly, I think it's salutory.  Just look at all the erroneous nonsense that's used in the analysis of motorised technologies.  As I see it - we'll be given an opporunity to forge an acceptable method of analysis.  Long overdue in my book.  So.  Thank you Professor.  There are many of us who are most grateful for your input here.

Rosemary
   
img  aetherevarising
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #179 on: 2011-01-15, 01:12:12 »
Group: Guest
Quote from: poynt99 on 2011-01-15, 01:04:13
Lawrence,

I am simply stating the facts. The most important factor regarding the scopes, is that they can perform both multiplication, and mean averaging.

Keep in mind however, that it's the old adage; "garbage in, garbage out", so ensure that your raw data is properly-acquired. I still have a great deal of doubt that good data can be obtained with passive probes, but you will have to take a chance I suppose, unless the Professor can borrow a current probe.

.99

Yes.  And if Professor can find a current probe that is able to measure at the frequencies that you're switching at.  To the best of my knowledge there is no such animal.  But I'd like to be proved wrong.  I would LOVE to find that probe. 
   
img  vonwolf
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #180 on: 2011-01-15, 01:14:58 »
Group: Guest
  I really didn't think it was getting anything by you poynt99 it's just throwing a made up term to muddy the waters. I'm not nearly as advanced in these things as most of you but I have been following LT's work for years and this is the 1st time in a long time that its got the attention a real expert's with real measurement's and math.
   It will be interesting to see the results but I'm having a hard enough time see how pulsed peak to peak power has anything to do with constant power output and throwing in new terms just makes it more difficult for a armature such as myself.
 Pete
   
img  aetherevarising
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #181 on: 2011-01-15, 01:17:47 »
Group: Guest
Sorry Lawrence - I missed your point.  I agree with Poynty.  Your results are more than enough proof provided only that COP is significantly greater than 1.  And in your case it is well in excess.

And well done.  I agree that your device is elegant evidence of a really challenging fact.

Kindest regards,
Rosemary
   
img  aetherevarising
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #182 on: 2011-01-15, 01:22:40 »
Group: Guest
Quote from: Groundloop on 2011-01-15, 00:43:41
Here is my way of testing if ANY boost converter is over unity or not.

You just connect a AA battery, an boost converter (JT) and a battery capacitor to the circuit and let it run.
The automated feed back circuit will take its power to run from the AA battery. The micro controller needs +5 Volt to run so we use an IC to boost the 1, 2 Volt up to 5 volt. The micro controller runs slowly on a 32 KHz crystal and uses approx. 60uA from the IC to run.  So the total current usage of the circuit is approx. 0,25mA from the AA battery. The circuit will monitor both the battery capacitor voltage and the AA battery voltage. If the AA battery voltage is too low, then the circuit will attempt to charge the AA battery from the battery capacitor. If the battery capacitor gets too high in voltage then the circuit will dump energy to the output load, thus keeping the battery capacitor below the recommended 2, 7 volt.

The project (design files, software and documentation) can be downloaded here: http://home.no/ufoufoufoufo/feedback.rar

That is a really elegant solution Groundloop.  I was not aware of it and have been looking for something like this.  Please check your pm's.

Rosemary

 :)  LOL  Sorry about the editing. 
   
img  poynt99
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #183 on: 2011-01-15, 01:43:08 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Quote from: aetherevarising on 2011-01-15, 01:10:12
Poynty - I'm inclined to agree with you.  It is, perhaps, NOT NECESSARY to apply an integrated value to these results.  Unfortunately mainstream and academic experts consider it NECESSARY
I very much doubt that, and, it absolutely depends on the characteristics of the device's power usage and delivery over time (see below).

Quote
I'm afraid their argument is based on the fact that 'mean average' and the properly integrated values DO NOT CORRESPOND.  Since the latter is invariably the more modest value then I'm inclined to think this will become the required benchmark.  And as there really is no way within a simple averaging - to incorporate a proportional 'time' difference between when the energy is being returned and when it is being delivered - then - the chances are that they've got a strong argument.  If they corresponded - then your argument would carry.
Power and Energy measurements in their pure form are not supposed to correspond. It is their respective ratios that does.  

Most if not all boost converter circuits and their derivatives will use and deliver a fixed average value of power after the unit has had sufficient time to stabilize, such as 20 or 30 minutes. This is assuming of course that the voltage supply and it's impedance do not change, ambient temperature does not change, and the load impedance does not change after stabilization has been reached.

The other important fact to realize, is that the device's cyclic wave form does not change over time. So in all aspects, the device, after stabilization, uses and delivers a steady continuous value of power. As such, it matters not if the output and input powers are divided as a ratio, or the output and input Joules are divided as a ratio, the results will be equal. In the latter case however, an extra unnecessary step was performed.

.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: Lawrence Tseung sent a Prototype to test... any comments? « Reply #184 on: 2011-01-15, 01:51:13 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Quote from: aetherevarising on 2011-01-15, 01:12:12
Yes.  And if Professor can find a current probe that is able to measure at the frequencies that you're switching at.  To the best of my knowledge there is no such animal.  But I'd like to be proved wrong.  I would LOVE to find that probe. 

This is simply not true.

The TCP202, DC-50MHz current probe is a good example. A 50MHz probe is sufficient to capture 10ns rise and fall times. And there are probes capable of >100MHz as well.

.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  aetherevarising
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #185 on: 2011-01-15, 01:53:05 »
Group: Guest
Quote from: poynt99 on 2011-01-15, 01:51:13
This is simply not true.

The TCP202, DC-50MHz current probe is a good example. A 50MHz probe is sufficient to capture 10ns rise and fall times. And there are probes capable of >100MHz as well.

.99

WELL.  That's news indeed.  So.  Can you PLEASE tell me where I can get this probe?  I've been looking.  EVERYWHERE.  

Rosemary

added  I've just seen that DC- number.  Presumably it will manage our complex waveforms.  I know we've tried one that simply flattens the waveform and gives an approximate number.
   
img  poynt99
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #186 on: 2011-01-15, 02:03:45 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Quote from: aetherevarising on 2011-01-15, 01:53:05
WELL.  That's news indeed.  So.  Can you PLEASE tell me where I can get this probe?  I've been looking.  EVERYWHERE.  

Rosemary

added  I've just seen that DC- number.  Presumably it will manage our complex waveforms.  I know we've tried one that simply flattens the waveform and gives an approximate number.

To recap, the TCP202 is good to 50MHz, so it will manage quite fast transients. These and several other current probes are available from every major oscilloscope manufacturer out there. Go to their sites or speak to their sales people. Tell them you want as a minimum, an AC 50MHz current probe. They will either sell you direct, or provide a dealer list.

DC capability is nice, but more expensive and not really necessary for dynamic circuits such as these.

.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  MileHigh
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #187 on: 2011-01-15, 02:06:36 »
Group: Guest
Quote from: Groundloop on 2011-01-15, 00:43:41
Here is my way of testing if ANY boost converter is over unity or not.

You just connect a AA battery, an boost converter (JT) and a battery capacitor to the circuit and let it run.
The automated feed back circuit will take its power to run from the AA battery. The micro controller needs +5 Volt to run so we use an IC to boost the 1, 2 Volt up to 5 volt. The micro controller runs slowly on a 32 KHz crystal and uses approx. 60uA from the IC to run.  So the total current usage of the circuit is approx. 0,25mA from the AA battery. The circuit will monitor both the battery capacitor voltage and the AA battery voltage. If the AA battery voltage is too low, then the circuit will attempt to charge the AA battery from the battery capacitor. If the battery capacitor gets too high in voltage then the circuit will dump energy to the output load, thus keeping the battery capacitor below the recommended 2, 7 volt.

The project (design files, software and documentation) can be downloaded here: http://home.no/ufoufoufoufo/feedback.rar

Alex.

Alex, as a suggested exercise for you, you should try to further qualify your tester with respect to what approximate COP level from the JT would be required for the device to run continuously.  Without knowing that threshold specification, the information about the tester is incomplete.

Your tester discharges the battery to the point where it starts to drop in voltage as it charges the big capacitor.  Then it starts to redirect the energy in the big capacitor back to the source battery. The battery has an unknown charging efficiency, and you indicated there is the overhead required to run the microcontroller also.  These two effects reduce the net amount of energy that the big cap puts back into the battery.  Therefore the JT has to have a COP sufficiently high to overcome these losses to self-sustain.

MileHigh
   
img  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #188 on: 2011-01-15, 02:59:17 »
Group: Guest
Quote
Dear xxx,

Please read the post from poynt99 at:
http://www.overunityresearch.com/index.php?topic=538.msg9174#msg9174
very carefully.

We have been looking for the integration function for months to get a true COP value.  That is not necessary.  Once we have the Channel 1 and Channel 2 multiplied to get the Instantaneous Power curve, we could have used the “mean” function to display the running Power!

Our displayed graphs need not be as complicated as we have been showing for months.  All we need is to display the Instantaneous Power Curve and the mean running Power value.  There is NO need to do any calculations at all.  All the work on calculating the COP peak-to-peak and COP rms are NOT necessary and confusing.

Please redo the calibration of all the FLEET prototypes according to the directions from poynt99.  Drop all the COP peak-to-peak calculations and do the correct COP comparisons.  Please send me some sample screen shots and photos as soon as possible.

Skype me if you need more information or direction.  Thank you.

Best regards and Happy Chinese New Year

Lawrence

The Hong Kong and China teams will be happy with this information.  Thanks to poynt99 once again.
   
img  PhysicsProf
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #189 on: 2011-01-15, 06:16:44 »
Group: Professor
Hero Member
*****

Posts: 3017
Quote from: poynt99 on 2011-01-14, 22:36:21
Professor,

Regarding the use of the DPO3032 Tek scope, go here to download the manual, then go to page 108 and follow the steps to create a multiply math function. It even shows an example of "power" by multiplying the voltage and current.

Then go to the "Measurement" menu and select "Mean" (not "Cycle Mean") as shown on page 99 of the manual. This of course is applied to the math wave form. Now you will have a running average of the REAL power.

As I have stated a couple times now, integration is not necessary to obtain a COP value.

.99



"Now you will have a running average of the REAL power."  This is very helpful indeed. I have downloaded the manual and started reading. Thank you.  

If we can get the running average for the real INPUT power and OUTPUT power separately, I would think the next step will be to take the "running ratio" of these, Pout/Pin.

Will this not provide a reasonable COP value?  Perhaps there is a better way...
 Seems this would require 4 channels on the Tek 3032... not sure how to do all this, but sounds feasible.

"As I have stated a couple times now, integration is not necessary to obtain a COP value"  -- OK, but can you specify what IS needed then to obtain a reasonable COP value using the Tek 3032?   Does the method I suggest above do the job?

This is a very important discussion -- thanks again.
   
img  MileHigh
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #190 on: 2011-01-15, 06:43:15 »
Group: Guest
PhysicsProf:

You don't need two more DSO channels to measure the power consumption of the overall Joule Thief circuit.  All that you need is some form of low-pass filtering system between your power source and the the Joule Thief circuit load.  I am copy/pasting one version of this for your consideration below.  I am assuming that you have or can get a variable power supply and a large capacitor or two?  That's all that you really need.

Ion or Poynt probably have similar or better versions of what I explain below so you have some choices.

You should be able to make a very accurate power consumption measurement of the Joule Thief circuit using pure analog components and a very accurate power output measurement on the Joule Thief secondary output using the two channels of the DSO and the built-in averaging function.  You can run both measurement systems live at the same time.  So you are almost there.

For your consideration below with a few comments to tweak things for your low-power application.  Instead of a 30,000 uF capacitor you could probably get away with a  5,000 or 10,000 uF cap.  You should also put a 1 uF and a 0.1 uF capacitor in parallel with the big capacitor.  You may need to experiment with the value of the resistor below suggested as 100 ohms.  A 500-ohm resistor might be a better choice for your low power application.  (It could possibly be even higher since you want to simulate a 1.5-volt battery, not a 12.6-volt battery.  Something like running your bench power supply at a nominal 5 volts and having the resistor cause a 3.5 volt drop to simulate the 1.5-volt battery.  That gives you reasonable 'breathing room' with the variable power supply.)

Best of luck,  MileHigh

>>>

So let's discuss a very simple technique to get a very accurate power consumption measurement for your Bedini motor.  There are many ways of doing this but let's start with one simple example.

Here is what you need:

1) A large electrolytic capacitor.  Perhaps something like 30,000 uF at 25 volts.  That might be the size of a 300 ml Coke can or larger.
2) A 1% precision resistor, let's say 100 ohms @ 1/2 watt.  Alternatively you can take a regular 100 ohm resistor and measure the resistance of it with your multimeter.
3) A bench power supply that has a variable output voltage from 0 to 25 volts.

With these components you are going to simulate a 12-volt power supply with the large capacitor.

The circuit is very simple:  You connect the output of the power supply to the 100-ohm resistor, and that is connected to the positive of your big capacitor.  The positive of the big capacitor is also connected to the power input of the Bedini motor.

You set your power supply to 12.6 volts and you charge up the big capacitor to 12.6 volts.  Then you connect your Bedini motor and get it running.

When you do this you are monitoring the voltage across the big capacitor.   You notice the voltage across the capacitor starts to drop.  So you compensate by increasing the voltage from your variable power supply.

Let's assume that after a few minutes, the Bedini motor is running normally at it's stabilized speed.  You have also stabilized the voltage in the big capacitor back up to 12.6 volts to simulate a normal battery.   To do this let's assume that the voltage from the power supply has to be set to 13.84 volts.

You have to check the voltage across the capacitor with your oscilloscope to see if it is relatively stable.  If the ripple voltage is less than +/- 0.1 volts, then let's say that the voltage is stable enough.  If there is too much voltage ripple, then simply add a second big capacitor in parallel to the first capacitor.

So how do you calculate the power consumption of the motor?  It is very easy, you know the voltage supplied to the motor, it is 12.6 volts.  And you know that the average current that you are feeding into the capacitor from the power supply is identical to the average pulsing current that is going into the Bedini motor.  The current in must be equal to the current out if the capacitor voltage is stable.

So the current is (13.84 - 12.6)/100  = 12.4 milliamperes.

Therefore the power consumption of the Bedini motor is (0.0124 x 12.6) = 0.156 watts.

Note that you don't actually measure 13.84 volts and 12.6 volts and do the subtraction.  You simply put your multimeter across the 100-ohm resistor and measure 1.24 volts.  You are taking advantage of the fact that the multimeter is more accurate when measuring lower voltages.

Let's check the power dissipation in the resistor = (0.0124 x 0.0124 x 100 ) = 0.015 watts.  So you are fine using a 1/2 watt resistor in this example and you are sure that it will not burn up.

Here is the big advantage with this setup:  It will be able to measure the power consumption of a Bedini motor or any other device no matter how crazy the current waveform is.  As long as your big capacitor is big enough to smooth out the voltage then it will work.  It will allow you to make very accurate power consumption measurements without worrying about the limitations of digital and analog meters.

The disadvantage is that you have to work with a variable voltage power supply and you have to wait a few minutes to let everything stabilize before you make your final measurements.  If you are doing research this should not be a problem.

MileHigh
   
img  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #191 on: 2011-01-15, 08:01:48 »
Group: Guest
Quote from: PhysicsProf on 2011-01-15, 06:16:44
"Now you will have a running average of the REAL power."  This is very helpful indeed. I have downloaded the manual and started reading. Thank you.
*** Get someone familiar with the Oscilloscope to help is even better.  My experience in Hong Kong was that purely learning from manuals took over two weeks and many subtle features were not properly digested.  (Not knowing the “mean” function was the biggest mistake!)

If we can get the running average for the real INPUT power and OUTPUT power separately, ***please do this separate step first.   When you have only one 2 channel DSO, this is the best you can do.  This result will be important for any further experiments.  Please share the results with photos, screen shots etc.

***If you can borrow another identical scope, you can have one to display the Input Power waveform and the mean Input Power reading.  You can use the other to display the Output Power waveform the mean Output Power reading.  It will be easy to see if the COP is greater than 1.

***With the two oscilloscope set up, you can modify the physical set up of the prototype (even with the breadboard).  You can easily observe the change in frequency and waveform as you make the modifications.  That is the technique we use to get high COP prototypes.  (Notice that I now drop the Tseung Fleet Comparison Index.)

This is a very important discussion -- thanks again.


Please do one step at a time.  The result may surprise you.  With luck, you may be able to demonstrate a prototype with COP > 10 and average Output Power greater than 10 watts.  Such results will be so good that you do not need to worry about the “noise”.  Good luck with the tuning.
   
img  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #192 on: 2011-01-15, 08:14:18 »
Group: Guest
Quote from: MileHigh on 2011-01-15, 06:43:15
PhysicsProf:

You don't need two more DSO channels to measure the power consumption of the overall Joule Thief circuit.  All that you need is some form of low-pass filtering system between your power source and the the Joule Thief circuit load.  I am copy/pasting one version of this for your consideration below.  I am assuming that you have or can get a variable power supply and a large capacitor or two?  That's all that you really need......

Best of luck,  MileHigh

MileHigh

Dear MileHigh,

The FLEET prototype is a pseudo resonance circuit.  Any small change or attachment is likely to shift the frequency and change the waveform.  This is particularly true when the prototype exhibit high COP.

My advice is for the PhysicsProf to get the Input and Output Power curves and mean values first.  He is not familiar with the oscilloscope yet.  Let him get help and/or study the manual first.  That is a tough task already.

Other tests can wait.
   
img  Groundloop
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #193 on: 2011-01-15, 08:24:26 »
Sr. Member
****

Posts: 336
Quote from: MileHigh on 2011-01-15, 02:06:36
Alex, as a suggested exercise for you, you should try to further qualify your tester with respect to what approximate COP level from the JT would be required for the device to run continuously.  Without knowing that threshold specification, the information about the tester is incomplete.

Your tester discharges the battery to the point where it starts to drop in voltage as it charges the big capacitor.  Then it starts to redirect the energy in the big capacitor back to the source battery. The battery has an unknown charging efficiency, and you indicated there is the overhead required to run the microcontroller also.  These two effects reduce the net amount of energy that the big cap puts back into the battery.  Therefore the JT has to have a COP sufficiently high to overcome these losses to self-sustain.

MileHigh

MileHigh,

I will say it the other way around, if the COP value is high enough in the JT to allow for the device to run continuously, then the JT is over unity.

Alex.
   
img  MileHigh
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #194 on: 2011-01-15, 13:02:19 »
Group: Guest
Quote from: ltseung888 on 2011-01-15, 08:14:18
Dear MileHigh,

The FLEET prototype is a pseudo resonance circuit.  Any small change or attachment is likely to shift the frequency and change the waveform.  This is particularly true when the prototype exhibit high COP.

My advice is for the PhysicsProf to get the Input and Output Power curves and mean values first.  He is not familiar with the oscilloscope yet.  Let him get help and/or study the manual first.  That is a tough task already.

Other tests can wait.

Lawrence:

Your first sentence ironically highlights the reason it is advisable and important to measure both the power supplied to the device and the power output from the device simultaneously.  If you measure the power output from the device only, and then change the test configuration and measure the power supplied to the device, you run the very risks that you stated.  That being the two different test steps may change the frequency or waveform of the operating Joule Thief circuit.

The tough task is to get the oscilloscope set up and measuring the two waveforms.  In contrast, it is very easy to set up the low pass filter for providing power to the Joule Thief.  A capacitor charged to 1.5 volts with a trickle charge through a resistor keeping it charged at 1.5 volts while it powers the Joule Thief circuit will emulate the 1.5 volt battery perfectly.  This can easily be verified.  All that you have to do is look at what the Joule Thief waveforms look like while it is powered by the 1.5 volt battery.  Then substitute the trickle-charged capacitor for the 1.5 volt battery and stabilize the capacitor voltage to 1.5 volts as described in my previous posting.  You will then observe that the Joule Thief waveforms and operating frequency will be identical to the case when the the Joule Thief is powered by the 1.5 volt battery.  This is a task that is very easy to accomplish.

Once this is done you are able to observe the power supplied to the Joule Thief and the power power output by the Joule Thief in real time.  You are free to make changes to Joule Thief configuration and reestablish the proper capacitor voltage and record a new set of power measurements.  In addition, you now have the flexibility to change the supply voltage to the Joule Thief if you want to and see how it performs at 2 volts or 3 volts, etc.  You are free from the restriction of supplying the circuit with 1.5 volts.

When you compare the 1.5 volt battery to a large capacitor that maintains a 1.5 volt potential all of the time, from the point of view of the Joule Thief circuit they will look like identical power sources.  As I previously stated, this can be easily proven by comparing the Joule Thief waveforms for the two different power sources.  The two sets of waveforms will be identical.

The most credible way to measure your COP values is to monitor the power supplied to the device and the power output by the device at the same time with the identical configuration.

MileHigh
   
img  MileHigh
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #195 on: 2011-01-15, 13:25:56 »
Group: Guest
There is a very very minor issue that might crop up that I will address just in case some people believe it is relevant.

The issue is that a 1.5 volt alkaline AA battery might have an output impedance of perhaps 0.1 ohms. (It may be much less than that.)  This is in contrast to the capacitor emulation of the 1.5 volt battery where the output impedance in this case for all practical intents and purposes is nearly zero.  This might mean that you see a slight AC ripple on the battery voltage with your oscilloscope when the Joule Thief is running but you will not see the ripple on the capacitor voltage.

So If during the operation of the circuit you observe that the battery voltage has a slight AC ripple and the capacitor setup has no AC ripple it is highly unlikely that this will affect the overall operation of the Joule Thief circuit at all.  Observing anything more than a very slight AC component on the 1.5 volt battery source may be an an indication that the battery itself is getting old.  If you took any issue with the fact that the the battery shows a slight AC component it its output voltage and the capacitor emulation of the battery does not show an AC output voltage, then this "issue" can be fixed also.

The fix for this "issue" is to put a resistor in between the capacitor and the power input of the Joule Thief circuit so you can simulate the output impedance of the battery.  It would have to be a resistor of perhaps 0.1 ohms.  Then the voltage supplied to the the Joule Thief circuit by the capacitor setup would have a nearly identical slight ripple just like the battery.

The reality is that whether the power source for the Joule Thief has a 0.1 ohm output impedance or a near-zero output impedance, this will not affect the COP measurement for the Joule Thief in any significant way at all.  If anyone wants to double-check this they can simply take the COP measurements with and without the 0.1 ohm output impedance resistor and they should not see any real change in the COP measurements.

MileHigh
   
img  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #196 on: 2011-01-15, 13:47:05 »
Group: Guest
Quote from: MileHigh on 2011-01-15, 13:02:19
Lawrence:

Your first sentence ironically highlights the reason it is advisable and important to measure both the power supplied to the device and the power output from the device simultaneously.  If you measure the power output from the device only, and then change the test configuration and measure the power supplied to the device, you run the very risks that you stated.  That being the two different test steps may change the frequency or waveform of the operating Joule Thief circuit.

MileHigh

Dear MileHigh,

Thank you for reminding me.  The main reason why I use two oscilloscopes is that I do not need to disturb the set up.  With only one oscilloscope, I have to measure the Input Power and then disconnect to measure the Output Power.  The test configuration is likely to be changed.  The result will not be scientifically valid.

@PhysicsProf, please borrow another identical or similar oscilloscope with the multiply and mean functionality to do the experiment.  The additional good aspect of the two oscilloscopes is that you can tune for highest COP by looking at the two Power Waveforms and Mean values at the same time.  Even small changes of moving wires closer together etc. can be observed and tuned.  The highest COP value can be captured by camera and oscilloscopes.  The result will be scientifically valid.

I also received the following email from Hong Kong:

Quote
Let me digest that first. Thanks for the advancement.
Aaron

Mr. Aaron Quant brought 50 one-inch toroids at the same time.  I shall ask him to buy 100 for me to prepare for the coming expected requests for free prototypes.  Now I shall insist that the verifiers must have access to two good oscilloscopes with the multiply and mean (or the multiply and integrate) functionality.

Thank you to all for the excellent discussion and suggestions.  I am sure that our dear PhysicsProf will do many tests to confirm that the prototypes are really capable of COP > 1.   May God bless you all.  Amen.
   
img  poynt99
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #197 on: 2011-01-15, 14:07:28 »

Group: Administrator
Hero Member
*****

Posts: 3217
It's not as complicated as it may seem...
Quote from: PhysicsProf on 2011-01-15, 06:16:44
"Now you will have a running average of the REAL power."  This is very helpful indeed. I have downloaded the manual and started reading. Thank you.  

If we can get the running average for the real INPUT power and OUTPUT power separately, I would think the next step will be to take the "running ratio" of these, Pout/Pin.

Will this not provide a reasonable COP value?  Perhaps there is a better way...
 Seems this would require 4 channels on the Tek 3032... not sure how to do all this, but sounds feasible.
While having a running Po/Pi ratio (COP) would be ideal, it is not necessary, unless attaching the scope probes has a significant effect on the power levels. In such a case, then a 4-channel scope, or 2x 2-channel scopes would be preferable.

Quote
"As I have stated a couple times now, integration is not necessary to obtain a COP value"  -- OK, but can you specify what IS needed then to obtain a reasonable COP value using the Tek 3032?   Does the method I suggest above do the job?

This is a very important discussion -- thanks again.
As mentioned, if the DPO3032 is a 2-channel scope, you will have to obtain another scope that performs multiplication, preferably the same model, or get a 4-channel scope.

The DPO3032 (being a 2-channel scope) will only allow you to make one power measurement at a time.

The absolutely ideal scenario is to have a 4-channel scope, 2 current probes, and 2 differential voltage probes.

Here is what I suggest:

Place the entire JT in a metal enclosure, and attach BNC (or something suitable) connectors to the enclosure so you can measure the required points in the circuit. This should minimize influence on the JT operation and allow for using a single 2-channel scope to make your measurements.

Also, see if you can borrow a current and differential probe. I wouldn't be surprised if the electrical engineering department had them.

.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  ltseung888
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #198 on: 2011-01-15, 14:24:36 »
Group: Guest
Quote from: MileHigh on 2011-01-15, 13:25:56
There is a very very minor issue that might crop up that I will address just in case some people believe it is relevant.

The issue is that a 1.5 volt alkaline AA battery might have an output impedance of perhaps 0.1 ohms. (It may be much less than that.)  This is in contrast to the capacitor emulation of the 1.5 volt battery where the output impedance in this case for all practical intents and purposes is nearly zero.  This might mean that you see a slight AC ripple on the battery voltage with your oscilloscope when the Joule Thief is running but you will not see the ripple on the capacitor voltage.

 
MileHigh

Dear MileHigh,

If you look at the sample screen shots in reply 106, you would find that the ripple on the FLEET is more than 1 volt!

That is why I do not trust any simulation for pseudo resonance set ups.

Lawrence
   
img  MileHigh
Re: Lawrence Tseung sent a Prototype to test... any comments? « Reply #199 on: 2011-01-15, 15:03:44 »
Group: Guest
Quote from: ltseung888 on 2011-01-15, 14:24:36
Dear MileHigh,

If you look at the sample screen shots in reply 106, you would find that the ripple on the FLEET is more than 1 volt!

That is why I do not trust any simulation for pseudo resonance set ups.

Lawrence


The ripple you see on the battery output voltage in reply 106 is actually the JT output capacitively or inductively coupling back to the input.  That's what it looks like by examining the JT output waveform.  You should see approximately the same waveform if you use the proper output impedance resistor for the capacitor emulation of the battery setup.

Anyway I gave you my thoughts on how to accomplish the testing without having to use a second digital storage oscilloscope.  It's up to PhysicsProf or any other testers to decide if they want to try them out or not.  As Humbugger stated, if you believe that your COP values are greater than 2, you should not have to worry too much about the fine nuances in the test setup.  If the gains are there they should be readily apparent as long as the test setup is reasonably well done.

Personally if I was to do the testing I would try the capacitor emulation of the supply battery and the thermal profiling of the output.  For the very low power this system is running at I am assuming that you can get thermometers or thermocouples that would work fine in this application.  Since it is low power you can put the output load components in a small insulated box so that the increased thermal resistance to the outside world results in higher and easier to measure temperature differentials.  If the COP is greater than 2 then you should be able to generate conclusive data without the need for any fancy equipment.  All that you need is the right test and measurement strategy and a high quality multimeter and a high quality temperature measurement device.

It's worth mentioning that when you do thermal profiling of the output, the output component of the system has a thermal mass, a.k.a. thermal capacitance.  Therefore the instantaneous output power is being thermally integrated.  In other words, if you took the output components and put them in a small container of thermal paste, that container is performing integration on the power waveform in exactly the same way the DSO could calculate the integration and do the averaging.  So a thermal mass is just as good or better an integrator of instantaneous power as a DSO, provided you have a very accurate way to measure the temperature and you do proper thermal profiling.  So the thought is to let Mother Nature do the difficult and complicated task of instantaneous integration and averaging for "free."  A thermal mass is a low-pass filter that averages the high frequency thermal/electrical power waveform.

MileHigh
« Last Edit: 2011-01-15, 15:33:48 by MileHigh »
   
Pages: 1 2 3 4 5 6 7 [8] 9 10 11 12 13 14 15 16 17 18 19 20 21 22
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-26, 17:38:47