PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2026-01-29, 08:39:48
News: The text input boxes (where you write your messages) are resizeable.  Just drag the bottom border of the text box to size it appropriately to your device.  The changes are persistent across your devices.

Pages: 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 [25] 26 27 28 29 30 31
Author Topic: partzmans board ATL  (Read 36157 times)
Topic Options img
img  Itsu
Re: partzmans board ATL « Reply #600 on: 2024-11-18, 09:49:46 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694
Quote from: partzman on 2024-11-17, 20:57:17
Itsu,

I have to assume you ran the above tests with C1 at zero bias as your schematic shows.  I also have to assume your primary turns is 20 making your V/t=2.4v.  Depending on the characteristics of the LED you used, it is very possible that the voltage across C1 would possibly provide a low power to the LED.

To really see if the energy in your replications is real, try substituting a high power LED in place of D1 as in your post #594.  High power meaning the LED having the ability to handle ~3.8 amps mean.  You could then measure the average amps times the average voltage across the LED to see the power across the LED which would be in addition to the power in Vload.

To help understand where the power is coming from, let's assume that your C1 which has a bulk capacitance (Cb) of 2680uf, has a charge separated capacitance (Ccs) of 134uf or 5% of Cb.  With your V/t=2.4v, then C1 increasing from zero volts to 2.4v, Ccs will reach an energy level of Ucs=2.4^2*134e-6/2=386uJ.  Now with a bias of 21.8v on C1 AND with an increase in C1 again being 2.4v, we see a peak voltage across C1 reaching 24.2v.  Now our gain in Ccs will be Ucs=(24.2^2-21.8^2)*134e-6/2=7.397mJ.  We take advantage of this gain by forcing energy in Vload.  This should not be the question.  The question should be, how much energy is lost in the bulk capacitance Cb due to this gain in Ccs?

IOW, is the voltage in C1 after a completed cycle really greater than the starting voltage in C1?  If so, why?  The answer is seen if one looks at the voltage at VL1a while measuring the ending voltage across C1.  It will be seen to be slightly negative.  This is equivalent to a slightly more positive voltage across the primary which will reflect in a slightly higher voltage across C1.  Now, what is the capacitance of C1 at this point in time?  Is it Cb or Ccs?

I have found that with film caps, there is very little difference if any, between Cb and Ccs.  With axial Lytics, Ccs will run from 5% to maybe 15% of Cb.  Radial Lytics with their axis vertically in the core will be ~10% while placing the axis horizontally in the core will result in ~30% of Cb.

Regards,
Pm

 




PM,


Quote
I have to assume you ran the above tests with C1 at zero bias as your schematic shows.  I also have to assume your primary turns is 20 making your V/t=2.4v.  Depending on the characteristics of the LED you used, it is very possible that the voltage across C1 would possibly provide a low power to the LED.


What you assume is correct, so the power in the led (load) depends on the characteristics of that led (load).



Quote
To really see if the energy in your replications is real, try substituting a high power LED in place of D1 as in your post #594.  High power meaning the LED having the ability to handle ~3.8 amps mean.  You could then measure the average amps times the average voltage across the LED to see the power across the LED which would be in addition to the power in Vload.


I could try that, allthough it involves using the Vload PS again which i wanted to avoid using.



Quote
To help understand where the power is coming from, let's assume that your C1 which has a bulk capacitance (Cb) of 2680uf, has a charge separated capacitance (Ccs) of 134uf or 5% of Cb.  With your V/t=2.4v, then C1 increasing from zero volts to 2.4v, Ccs will reach an energy level of Ucs=2.4^2*134e-6/2=386uJ.  Now with a bias of 21.8v on C1 AND with an increase in C1 again being 2.4v, we see a peak voltage across C1 reaching 24.2v.  Now our gain in Ccs will be Ucs=(24.2^2-21.8^2)*134e-6/2=7.397mJ.  We take advantage of this gain by forcing energy in Vload.  This should not be the question.  The question should be, how much energy is lost in the bulk capacitance Cb due to this gain in Ccs?

IOW, is the voltage in C1 after a completed cycle really greater than the starting voltage in C1?  If so, why?  The answer is seen if one looks at the voltage at VL1a while measuring the ending voltage across C1.  It will be seen to be slightly negative.  This is equivalent to a slightly more positive voltage across the primary which will reflect in a slightly higher voltage across C1.  Now, what is the capacitance of C1 at this point in time?  Is it Cb or Ccs?


I am afraid i cannot help you here answering those questions, but perhaps other members can as it touches the core of the workings of this device IMO.
I do not like that you have to assume here things like about the "charge separated capacitance (Ccs) of 134uf", i would rather like it to be measured before using it in any calculations.

Itsu
   
img  Itsu
Re: partzmans board ATL « Reply #601 on: 2024-11-18, 09:53:22 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694
Quote from: verpies on 2024-11-17, 23:54:30
From calculating the O/I energy only during the pulse instead over the entire cycle.

Ok,  sounds sound  :) , so as we have a 100ms repetition frequency, i would need to redo those earlier measurements using the whole cycle of at least 100ms.
And / or reduce this repetition frequency to 5 or 10ms to get a better resolution for the scope to measure it as now i measure something like 14us every 100ms.

Itsu
   
img  partzman
Re: partzmans board ATL « Reply #602 on: 2024-11-18, 14:47:13 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: Itsu on 2024-11-18, 09:49:46



PM,



What you assume is correct, so the power in the led (load) depends on the characteristics of that led (load).




I could try that, allthough it involves using the Vload PS again which i wanted to avoid using.




I am afraid i cannot help you here answering those questions, but perhaps other members can as it touches the core of the workings of this device IMO.
I do not like that you have to assume here things like about the "charge separated capacitance (Ccs) of 134uf", i would rather like it to be measured before using it in any calculations.

Itsu

Itsu,

I have to assume the Ccs in this case because it is your cap in your hands, not mine!  You missed my point as It was meant to be an example for the working concept.  Pick any numbers you want for the Ccs value and there will still be a gain.

You can measure the approximate value of your cap's Ccs by following my post #516.  I say approximate because you have to factor in the energy consumed by the primary during the inductive loading of C1.  Be sure the primary is not entering saturation.

Pm
   
img  partzman
Re: partzmans board ATL « Reply #603 on: 2024-11-18, 14:51:02 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: verpies on 2024-11-17, 23:54:30
From calculating the O/I energy only during the pulse instead over the entire cycle.

Verpies,

If the same measurements are taken over the period of 100ms, the COPs will remain approximately the same, but the calculated energy values will be extremely small and difficult to ascertain.

Pm
   
img  Itsu
Re: partzmans board ATL « Reply #604 on: 2024-11-18, 15:42:11 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694
Quote from: partzman on 2024-11-18, 14:47:13
Itsu,

I have to assume the Ccs in this case because it is your cap in your hands, not mine!  You missed my point as It was meant to be an example for the working concept.  Pick any numbers you want for the Ccs value and there will still be a gain.

You can measure the approximate value of your cap's Ccs by following my post #516.  I say approximate because you have to factor in the energy consumed by the primary during the inductive loading of C1.  Be sure the primary is not entering saturation.

Pm

PM,

you are right for me missing the point as your assumptions were meant to be an example for the working concept, sorry about that.
Your post #516 was made before i jumped in at post #521, so i must have missed that.

I will try to use that post #516 to calculate my Cap's Ccs value.

Itsu
   
img  Itsu
Re: partzmans board ATL « Reply #605 on: 2024-11-18, 15:43:59 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694
When i run my COP tests over a (more than) 100ms cycle, i get the following results:

Input (48V) 560mW   (see 1st screenshot).
Output (21.8V) 297mW   (see 2nd screenshot).

So COP now is 0.53.

But as PM mentioned, the energy values are very small, so the error margin will be very high too.
Even as i reduce the pulse repetition frequency to 5ms the O/I relation stays about the same.

Itsu
   
img  Itsu
Re: partzmans board ATL « Reply #606 on: 2024-11-18, 21:01:53 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694
Quote from: Itsu on 2024-11-18, 15:42:11
PM,

you are right for me missing the point as your assumptions were meant to be an example for the working concept, sorry about that.
Your post #516 was made before i jumped in at post #521, so i must have missed that.

I will try to use that post #516 to calculate my Cap's Ccs value.

Itsu


PM,

Here i show a similar measurement as done in your post #516
I had to use some different component values like L1 is 20 turns measuring 32mH
Vin is 41.4V, while C2 is 350uH


SP1 shows the basic waveforms seen, but here we focus on the average voltage on C1 of 1.961v that is measured with the CH3(pnk) cursors near the start of the discharge cycle.

SP2 then shows us the average voltage on C1 to be 1.956v near the end of the discharge cycle again measured with the CH3 cursors.

SP3 then shows us the mean current in L2 to be 72.1mA over 32.10us via the CH4 cursors plus, a peak current reached in L2 of 145.3mA .

Using these numbers we'll first solve for the charge separated capacitance.  Since dV=di*dt/C, C=di*dt/dv .  So, dV=1.961-1.956=0.005v .  So, C=(0.0721*32.1e-6)/0.005=463uF .

Next we'll calculate the energy in L2 at 32.30us when the peak current has reached 145.3mA .  Ul2=0.1453^2*350e-6/2=3.69uJ .

Finally we see what energy was lost in C1 to charge L2.  UC1=(1.961^2-1.956^2)*463e-6/2=4.53uJ . 

So my Ccs seems 463uF if those last 2 calculations (3.69uJ and 4.53uJ) are close enough.

Itsu
   
img  partzman
Re: partzmans board ATL « Reply #607 on: 2024-11-19, 14:06:33 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: Itsu on 2024-11-18, 21:01:53

PM,

Here i show a similar measurement as done in your post #516
I had to use some different component values like L1 is 20 turns measuring 32mH
Vin is 41.4V, while C2 is 350uH


SP1 shows the basic waveforms seen, but here we focus on the average voltage on C1 of 1.961v that is measured with the CH3(pnk) cursors near the start of the discharge cycle.

SP2 then shows us the average voltage on C1 to be 1.956v near the end of the discharge cycle again measured with the CH3 cursors.

SP3 then shows us the mean current in L2 to be 72.1mA over 32.10us via the CH4 cursors plus, a peak current reached in L2 of 145.3mA .

Using these numbers we'll first solve for the charge separated capacitance.  Since dV=di*dt/C, C=di*dt/dv .  So, dV=1.961-1.956=0.005v .  So, C=(0.0721*32.1e-6)/0.005=463uF .

Next we'll calculate the energy in L2 at 32.30us when the peak current has reached 145.3mA .  Ul2=0.1453^2*350e-6/2=3.69uJ .

Finally we see what energy was lost in C1 to charge L2.  UC1=(1.961^2-1.956^2)*463e-6/2=4.53uJ . 

So my Ccs seems 463uF if those last 2 calculations (3.69uJ and 4.53uJ) are close enough.

Itsu

Itsu,

That all looks correct.  Your Ccs is on the high side percentage wise which is good.

I have yet to determine what causes the variation of Ccs but I think there are many factors involved including the shield effect of the aluminum case.

Pm
   
img  Itsu
Re: partzmans board ATL « Reply #608 on: 2024-11-19, 15:27:17 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

Yes, i would expect that the aluminum case would almost completely block out the Electric field coming from the toroid, but apparently it does not do that completely.
   
img  Verpies
Re: partzmans board ATL « Reply #609 on: 2024-11-19, 17:57:15 »

Group: Administrator
Hero Member
*****

Posts: 4513
Quote from: Itsu on 2024-11-18, 15:43:59
When i run my COP tests over a (more than) 100ms cycle, i get the following results:
Calculating the O/I energy ratio merely over a longer time period is not the goal.
The goal is to calculate the O/I energy ratio over an integer number of cycles (including 1 cycle).
   
img  Itsu
Re: partzmans board ATL « Reply #610 on: 2024-11-19, 20:18:57 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

Well, that's what i wanted to say when saying in my post #605:

Quote
Even as i reduce the pulse repetition frequency to 5ms, the O/I relation stays about the same.

If i do that i have 20 cycles and the output versus input wattage is 730mW versus 973.5mW for a COP of 0.749

Itsu
   
img  partzman
Re: partzmans board ATL « Reply #611 on: 2024-11-21, 17:08:32 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: Itsu on 2024-11-19, 20:18:57
Well, that's what i wanted to say when saying in my post #605:

If i do that i have 20 cycles and the output versus input wattage is 730mW versus 973.5mW for a COP of 0.749

Itsu

Itsu,

I see Verpies or no one else has responded to this post so I will.  I know you are well aware (but perhaps there are those who are not) that with oscilloscopes we have two dimensions of resolution we have to be concerned with being the vertical and horizontal.  The vertical of course is the units we are measuring, and the horizontal is time.

With that being said and with all due respect, your trace captures above are poorly lacking in both!  So, the conclusion that the device is operating with a COP=.749 is highly suspect in accuracy and here's why.

In order to achieve even an 8 bit vertical resolution in the sample modes on both the TDS and MDO series scopes, the traces need to reach a peak to peak level as close to the maximum vertical deflection as possible.  If the vertical deflections reach lower peak to peak levels, the resolution falls below 8 bits severely limiting the accuracy.

The same problem exists in the horizontal resolution.  In your MDO scope shot above, your resolution is 100kS/s or only 10k sample points over the full 100ms sweep.  This again limits the accuracy of the samples taken when compared to the periods of the actual events.  Compare this horizontal resolution to your SP1 pix in post #606 where we have 100MS/s or 10k samples over the 100us sweep.  This is an increase in sampling accuracy of 1000x over the scope pix in post #610.

There are other factors not taken into consideration here such as offsets plus rise and fall times and other accuracy limitations specified by Tek, but I think the idea is conveyed.

Regards,
Pm 
   
img  Itsu
Re: partzmans board ATL « Reply #612 on: 2024-11-21, 19:53:54 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

PM,

thanks for your clear explanation on the limitations of the used scopes.

I think it is obvious to anyone looking at my latest scope screenshots that they are lacking resolution, so that the results are very doubtful.

That's why i mentioned this in an earlier post (#605):
Quote
the error margin will be very high too.

I could further decrease the pulse repetition time, but i think that would interfere with what you are trying to show us with your present setup.


Itsu
   
img  Verpies
Re: partzmans board ATL « Reply #613 on: 2024-11-22, 10:08:23 »

Group: Administrator
Hero Member
*****

Posts: 4513
The optimum scope setup for O/I calculations is:
1) The horizontal time base set so that slightly more than 1 cycle is captured by the scope (set the measurement cursors to span exactly 1 cycle and do the calculation between these cursors only).
2) The vertical gain/amplitude set all the way up ...but without clipping.
   
img  Itsu
Re: partzmans board ATL « Reply #614 on: 2024-11-22, 13:04:03 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

Here i further reduced the repetition frequency to 100us (this might compromise Partzman his setup) to get a better horizontal resolution.

I measure over 1 cycle (100us) between cursors

Input shows 321.8mW (in between the green cursors)
Output shows 49.05mW (in between the red cursors).


The shoot-through current pulses prevent me from further increase the vertical resolution.

Itsu
   
img  partzman
Re: partzmans board ATL « Reply #615 on: 2024-11-22, 20:49:21 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: Itsu on 2024-11-22, 13:04:03
Here i further reduced the repetition frequency to 100us (this might compromise Partzman his setup) to get a better horizontal resolution.

I measure over 1 cycle (100us) between cursors

Input shows 321.8mW (in between the green cursors)
Output shows 49.05mW (in between the red cursors).


The shoot-through current pulses prevent me from further increase the vertical resolution.

Itsu

Well, there you have it!

Pm
   
img  partzman
Re: partzmans board ATL « Reply #616 on: 2024-12-02, 21:07:54 »
Group: Moderator
Hero Member
*****

Posts: 2232
Below is a block diagram of a concept that utilizes my previous disclosures of RLE, dielectric induction/charge separation, and displacement current as seen in post#379 on page 16 of this thread.

An initial working bench device exhibits a COP=1.6 with a theoretical COP=3 .

I will post no more info on this device until I am able to demonstrate a stand alone generator.  If one does their homework, the operating principle will become apparent.

Regards,
Pm
   
img  Itsu
Re: partzmans board ATL « Reply #617 on: 2024-12-03, 10:18:42 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

PM,

Quote
An initial working bench device exhibits a COP=1.6 with a theoretical COP=3 .

That sounds very interesting, i hope you can demonstrate your working device anytime soon  O0

Itsu
   
img  PhysicsProf
Re: partzmans board ATL « Reply #618 on: 2024-12-04, 20:33:39 »
Group: Professor
Hero Member
*****

Posts: 3230
Quote from: partzman on 2024-12-02, 21:07:54
Below is a block diagram of a concept that utilizes my previous disclosures of RLE, dielectric induction/charge separation, and displacement current as seen in post#379 on page 16 of this thread.

An initial working bench device exhibits a COP=1.6 with a theoretical COP=3 .

I will post no more info on this device until I am able to demonstrate a stand alone generator.  If one does their homework, the operating principle will become apparent.

Regards,
Pm

So glad to hear that you are continuing on, and seeking to demonstrate a stand-alone generator!  More power to you!

(PS - I don't understand why a ramp generator is needed on the right-hand side, in your diagram...)
   
img  Verpies
Re: partzmans board ATL « Reply #619 on: 2024-12-04, 21:05:01 »

Group: Administrator
Hero Member
*****

Posts: 4513
Quote from: partzman on 2024-12-02, 21:07:54
I will post no more info on this device until I am able to demonstrate a stand alone generator.  If one does their homework, the operating principle will become apparent.
I hope that does not include questions about the block diagram itself.

I have a conceptual one. Which statement is true about this diagram you posted?:
1) The the L1 and Cs constitute the E-field source.
2) An external E-field source is superposed on the fields of L1 and Cs.
   
img  partzman
Re: partzmans board ATL « Reply #620 on: 2024-12-05, 14:55:32 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: PhysicsProf on 2024-12-04, 20:33:39
So glad to hear that you are continuing on, and seeking to demonstrate a stand-alone generator!  More power to you!

(PS - I don't understand why a ramp generator is needed on the right-hand side, in your diagram...)

The purpose of the voltage ramp connected to Cs is to create an instant constant displacement current in Cs that is transferred to Vload.  This occurs in sync with the instantaneous voltage rise in Cs that is generated by the charge separation in Cs via the E-Field.  This constant current is available to Vload until the voltage ramp reaches a predetermined peak voltage level that is held at which time the current in Cs falls to zero.  This is one source of output in the device.

Pm
   
img  partzman
Re: partzmans board ATL « Reply #621 on: 2024-12-05, 15:02:03 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: verpies on 2024-12-04, 21:05:01
I hope that does not include questions about the block diagram itself.

I have a conceptual one. Which statement is true about this diagram you posted?:
1) The the L1 and Cs constitute the E-field source.
2) An external E-field source is superposed on the fields of L1 and Cs.

The most true statement is #1 but not totally correct technically.  The E-Field source is L1 (along with any included core) with the recipient of said field being Cs.

Pm
   
img  Itsu
Re: partzmans board ATL « Reply #622 on: 2024-12-05, 20:42:06 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4694

Concerning this E-Field source (L1 plus core) and the recipient being Cs, are you sure it's the E-Field that is transferring the energy?

When i isolate Cs with a copper tape like shown in the below picture and i ground this copper tape, it (Cs) should be shielded significantly from any E-Field IMO.

But when i do so, there is no difference to be seen on the received pulse, like can be seen in the below two screenshots.

I use the diagram as shown below.

Green is the current into Vload, and pink is the voltage of Vload

Itsu
   
img  partzman
Re: partzmans board ATL « Reply #623 on: 2024-12-06, 14:50:53 »
Group: Moderator
Hero Member
*****

Posts: 2232
Quote from: Itsu on 2024-12-05, 20:42:06
Concerning this E-Field source (L1 plus core) and the recipient being Cs, are you sure it's the E-Field that is transferring the energy?

When i isolate Cs with a copper tape like shown in the below picture and i ground this copper tape, it (Cs) should be shielded significantly from any E-Field IMO.

But when i do so, there is no difference to be seen on the received pulse, like can be seen in the below two screenshots.

I use the diagram as shown below.

Green is the current into Vload, and pink is the voltage of Vload

Itsu

Itsu,

I really appreciate your efforts in replication and I especially appreciate this particular experiment! 

My analysis of your test is that you have charge separation in the foil shield you've placed around Cs which will appear nearly if not truly identical to the charge separation in Cs.  Please review my post #500 on page 21 of this thread.  There are other posts as well that compare the equality of the voltages of a single piece of wire verses Cs.

What I have wanted to try but haven't thus far, is to shield the open end of the 'Lytic used for Cs to the point that the positive wire has just enough clearance so it doesn't short to the shielded case.  This should reduce the measured charge separated capacitance Ccs slightly from the unshielded Ccs.

Perhaps an easier way to test for this would be to use a relatively large film cap (1-4uf) for Cs, then wrap shielding around Cs again preventing the positive lead from shorting on the shield with the grounded lead shorted to the shield.  This should show a reduction in the Ccs value.

The question is, can aether energy be shielded?  I do not have that answer! 

Regards,
Pm

 
   
img  partzman
Re: partzmans board ATL « Reply #624 on: 2024-12-06, 16:24:32 »
Group: Moderator
Hero Member
*****

Posts: 2232
Itsu,

As I reviewed your test, I'm not sure if you had the copper shield connected to ground or not.  It would be interesting if you connected the shield to ground and re-tested!

Regards,
Pm
   
Pages: 1 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 [25] 26 27 28 29 30 31
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2026-01-29, 08:39:48