PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2019-07-18, 18:31:27
News: Forum TIP:
The SHOUT BOX deletes messages after 3 hours. It is NOT meant to have lengthy conversations in. Use the Chat feature instead.

Pages: 1 ... 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 [18] 19 20 21 22 23 24
Author Topic: Investigating "anomalies" in Bifilar coils  (Read 131937 times)
Group: Guest
Something to consider  ;)

http://www.eevblog.com/forum/chat/rigol-ds1052e-inaccurate-voltage-readings/

0. It's always good to read EEVblog, thanks, it is one of my daily reads.
1. My scope is a DS1054z, unlocked to 100 MHz bandwidth and deeper memory.
2. As I posted yesterday, I have tested my scope's accuracy at reporting voltages in the 1-4 MHz range, by comparing its display and measurements of signal voltages, to a known reference DC voltage monitored by the Fluke 87-III DMM. Frequency accuracy is compared to the crystal oven PM6676 frequency counter, and phase measurements from cursors agree with the automatic measurement of phase angle to within a few tenths of a degree, usually.
3. A scope isn't really designed to be a super accurate voltmeter anyway; after all most scopes only have 8-bit vertical resolution. Or at least most that cost less than a house do. They are a lot more precise in the time domain.
   

Group: Administrator
Hero Member
*****

Posts: 3705


Buy me some coffee
Thanks ION thats the way to go  O0

TK It was fun while it lasted, been better than watching TV LOL
Will be interesting to see what Poynt says
   

Hero Member
*****

Posts: 1472
Here's the setup that gave those scopeshot measurements.
Nooooooo!

Do you think Pin is wrong or Pout is ?
Also, I saw on the scopeshot that Ch2 is offset vertically down and Ch2 is phase shifted wrt Ch3.  This is obviously wrong.
   
Hero Member
*****

Posts: 822
TK,

Attached is a sim of your test setup per Poynt with the values you gave me.  I assumed that the dcr of each coil in your iso transformer is .2 ohm and your signal generator was set to 50 ohm internal source.  I matched your measured levels at 1.69MHz by adjusting the K factor of the iso transformer L1/L2.

The measurements are as follows, CH1 = 3.7413v rms, CH2 = 2.0023/25 = 80.1ma rms (CH2 is 25x on the plot for easy visual comparison), CH3 = 1.5058v rms, I(R1) = 80.1ma rms, and the Math channel Pin = 120.56mw mean as you measured it.  These calculate to a Pout = 1.5058^2/18.8 = 120.6mw rms.  The COP = 120.6/120.56 = 1.00033.   

What is missing in these measurements is the power across R2 and the adjustment of the CH1 voltage due to the voltage across R2 and if taken into account (not shown), results in a COP<1.

The phase angle of CH2 to CH1 = 66.5 degrees and is not shown on the plot but it measured as such.

In addition to the points made by Verpies, I'll add that the most glaring discrepancy is your measurement of CH2 = 37.1ma rms across R2.  This should equal the current thru R1 which on your test is 79.8ma.  This is then used to calculate the Pin with a low result which produces the high COP.  Perhaps you could double check this.

I would try to replicate this circuit on my bench but I don't posses a coil equivalent to yours plus I don't think anyone will believe what I post at this moment in time anyway.

Pm
 
   
Sr. Member
****

Posts: 403
....
 plus I don't think anyone will believe what I post at this moment in time anyway.

Pm

Hi Partzman,

I do not think anyone here would have such opinion about you.  Please do not take it like that.

Thanks so much for sharing your findings on such interesting setups and please do continue this work.

Gyula
   

Hero Member
*****

Posts: 2099


Buy me a cigar
I " Second that emotion " !!

                O0


---------------------------
Nanny state ? Left at the gate !! :)
   

Hero Member
*****

Posts: 1472
I do not think anyone here would have such opinion about you.  Please do not take it like that.
Yes, Partzman proved himself as an honest experimenter, just like TK and Itsu.

@Partzman
I welcome your further musings about the input current measurement error modes.
What could case 2x lower input current readings ? - CSR inductance tends to cause errors on the high side.
   

Hero Member
*****

Posts: 3510


Buy me some coffee
Is it sinking in yet?

Honestly, I feel like I've been kicked in the stomach. I feel really badly about this.

Poynt99 asked me what the capacitance of my TBF coil was, by the method of disconnecting the series connection and measuring across the two half-coils with the capacitance meter. The capacitance measured that way is 2.19 nF.

So he came back with the suggestion to substitute a capacitor for the PBT coil in my test setup. I asked him to make it explicit by posting a schematic, so we all know we are on the same page. So he did that, and specified a capacitor value of 2.2 nF, a standard value. I found one in my parts stash, a good quality poly film unit.

I turned on the scope and let it and the Interstate F43 FG warm up. Both instruments were still set up from last night, so I didn't even change any settings except FG output amplitude (which I always turn down low when it's not in use.)  I connected up the parts and hooked them up, with the new FG coupling toroid I described above. And Bingo. All I did was turn up the FG amplitude to get a good vertical height for the signals, and tweak the frequency a bit to get full waveforms on the screen.

Saved the screenshot and the data dump and posted them here. And I wanted someone else to run the calculations before I told the story. Partzman kindly scrutinized the scopeshots and ran the numbers, and got the same result I got.



So I think talk of self-looping is still very premature. It's good to have suggestions as to how it might be done, but until a COP>1 is actually confirmed by something other than scope measurements, there isn't much to work with. How do you test a self-looping system for an OU device if the test device isn't actually OU?

 
Quote
What's the moral of this little story? Well, either I've got an overunity 35 cent capacitor on my bench.... or it makes it absolutely imperative for Partzman and the rest of us to try to find some _other_ indication of power levels, independently of oscilloscope measurements and 'traditional' power calculations, if this project is to survive. Of course, the scope is still necessary, obviously for tuning purposes, so that the _measurements_ indicate COP >1, but the scope traces and measurements alone cannot be used as evidence for OU any more.  Unless, of course, you are prepared to accept that my tiny poly film cap is also an OU device when connected in series with a couple of resistors.

So,are we now saying that scopes can no longer be used to calculate power  :o

So now-what is the point in even owning a scope ?
Why even bother learning how to use all the fancy functions of a scope,if it cannot even accurately calculate power in a simple circuit like this ?.

I think the gun is being jumped here.

1-What about your incandescent light bulb test TK ?
All that is needed for that,is an accurate P/in to the DUT,and the rest is done with your light meter, a DC power supply,and a couple of good DMMs.

2-I think poynt should now try and tell us as to why the scope is showing an OU capacitor--why this expensive bit of equipment is being fooled by a 20 cent capacitor and two resistors worth half as much at 10 cents each.

I have many of those PCU power supply 1:1 iso's,and no doubt a cap to boot--resistors by the hundreds.
So ,i will try the same test,and see if i can get the same result.


I think you need to just concentrate on the P/in of the DUT,and complete the !grain of wheat! bulb test-and not throw in the towel just yet.


Brad




---------------------------
Never let your schooling get in the way of your education.
   
Group: Elite
Hero Member
******

Posts: 3565
It's turtles all the way down
To all the guys that fought the good fight in this, my hat goes off to you  O0

The dedication and focus by the good men in these pages and the work done by them is uplifting.

Do not be discouraged, there is yet much to learn.

Some of the best minds and hearts are here and have contributed unselfishly, it will not be forgotten, it does not go unnoticed. The universe is listening! And it will be delivered.

Lets hope this effort can indeed solidify the bond of  brotherhood in the "quest".





---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Hero Member
*****

Posts: 3510


Buy me some coffee
Ok,below is my results from Poynt's cap test.
You will see that the cap value is 3.3nF,and not 2.2nF as specified--reason being,i dont have a 2.2nF cap.
This results in a lower frequency to achieve maximum amplitude across the 100R

While we are at it,i have a couple of questions about the math trace,which i included in each test.

1-Dose the math trace give us an indication of the power factor?-in that,the more of the math trace that is above the 0 volt line,the better-or closer the power factor is to 1,and the more the math trace becomes half above the 0 volt line,and half below the 0 volt line,the worse the power factor-->closer to 0.

This would seem to be the case,when looking at the V/I phase relationship in each test,where test two(the second scope shot and schematic)shows V and I to be nearly in phase with each other,while scope shot and schematic 1,show the V and I phases to be out by about 65*.

2- Using the value of the math trace,where it is all above the 0 volt line(as in scope shot 2),and converting that to an RMS value,dose this give you a P/out value?

Anyway,the tests are as shown below--who here can calculate the result's?.



So next i will change out the 100 ohm resistor for a grain of wheat bulb,and also place a bulb of the same between the inductor and cap.
These two bulbs should both glow with the same amount of brightness--correct?


Brad


---------------------------
Never let your schooling get in the way of your education.
   

Hero Member
*****

Posts: 3510


Buy me some coffee
To all the guys that fought the good fight in this, my hat goes off to you  O0

The dedication and focus by the good men in these pages and the work done by them is uplifting.

Do not be discouraged, there is yet much to learn.

Some of the best minds and hearts are here and have contributed unselfishly, it will not be forgotten, it does not go unnoticed. The universe is listening! And it will be delivered.

Lets hope this effort can indeed solidify the bond of  brotherhood in the "quest".

Excellent and true words ION.  O0


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
I think the simulation needs some parasitic inductances inserted.

Here are a couple more results from the Poynt99 Capacitor setup.

I carefully re-soldered joints, made sure that CH2 probe tip and reference were tight against the body of the 1.00 ohm NI CVR, and did some retuning of frequency. There is still about 8 cm of lead-length wiring, and of course I'm still using ordinary probes with their approx. 10 cm reference clipleads.

The CH2 offset and most of the phase shift wrt CH3 went away with these slight changes.

So I proceeded to test with, and without, the FG coupling transformer. Without the transformer,  I had to adjust the frequency a bit.  The COPs went down to a bit over 1.5.  I'm pretty sure that similar changes to the actual PBT setup would also reduce the measured COP there.

# 72 is with the coupler in place, #73 and #74 are without it.
   
Group: Guest

So,are we now saying that scopes can no longer be used to calculate power  :o

So now-what is the point in even owning a scope ?
Why even bother learning how to use all the fancy functions of a scope,if it cannot even accurately calculate power in a simple circuit like this ?.

I think the gun is being jumped here.

1-What about your incandescent light bulb test TK ?
All that is needed for that,is an accurate P/in to the DUT,and the rest is done with your light meter, a DC power supply,and a couple of good DMMs.

2-I think poynt should now try and tell us as to why the scope is showing an OU capacitor--why this expensive bit of equipment is being fooled by a 20 cent capacitor and two resistors worth half as much at 10 cents each.

I have many of those PCU power supply 1:1 iso's,and no doubt a cap to boot--resistors by the hundreds.
So ,i will try the same test,and see if i can get the same result.


I think you need to just concentrate on the P/in of the DUT,and complete the !grain of wheat! bulb test-and not throw in the towel just yet.


Brad

No, it doesn't mean a scope is useless or unreliable. It does mean that extra care must be taken in making measurements and interpreting results. It also means that if the scope is the _only_ evidence for COP>1, it really can't be used to make a solid claim, since it is possible -- even easy -- to obtain COP>1 readings from systems that we know are not OU. The scope readings must be shown to have what we call "concurrent validity", that is, they must be backed up with other kinds of measurements that also show the same degree of OU (or not). And of course they must be shown to be "correct" within themselves (internal validity).

Yes, the problem here is with the measurement of the input power. Since the value is so dependent on the phase shift between CH1 and CH2, and the "desired" shift is so close to 90 degrees, tiny inductances become significant in producing the COP >1 values. This is true for the Capacitor test and also for the full PBT systems.
   

Hero Member
*****

Posts: 3510


Buy me some coffee
Below,pic 1 showing the DUT with resistors.
Pic 2 showing the results of the light bulb test.

It looks like both bulbs are the same brightness,and both measure the same voltage across them with the scope.


Brad


---------------------------
Never let your schooling get in the way of your education.
   

Hero Member
*****

Posts: 3510


Buy me some coffee
No, it doesn't mean a scope is useless or unreliable. It does mean that extra care must be taken in making measurements and interpreting results. It also means that if the scope is the _only_ evidence for COP>1, it really can't be used to make a solid claim, since it is possible -- even easy -- to obtain COP>1 readings from systems that we know are not OU. The scope readings must be shown to have what we call "concurrent validity", that is, they must be backed up with other kinds of measurements that also show the same degree of OU (or not). And of course they must be shown to be "correct" within themselves (internal validity).

Yes, the problem here is with the measurement of the input power. Since the value is so dependent on the phase shift between CH1 and CH2, and the "desired" shift is so close to 90 degrees, tiny inductances become significant in producing the COP >1 values. This is true for the Capacitor test and also for the full PBT systems.

OK,so now my question is-how else are you going to measure the power in,in such a system-other than using the scope ?.


Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
Ok,below is my results from Poynt's cap test.
You will see that the cap value is 3.3nF,and not 2.2nF as specified--reason being,i dont have a 2.2nF cap.
This results in a lower frequency to achieve maximum amplitude across the 100R

While we are at it,i have a couple of questions about the math trace,which i included in each test.

1-Dose the math trace give us an indication of the power factor?-in that,the more of the math trace that is above the 0 volt line,the better-or closer the power factor is to 1,and the more the math trace becomes half above the 0 volt line,and half below the 0 volt line,the worse the power factor-->closer to 0.

This would seem to be the case,when looking at the V/I phase relationship in each test,where test two(the second scope shot and schematic)shows V and I to be nearly in phase with each other,while scope shot and schematic 1,show the V and I phases to be out by about 65*.

I think that is correct.

Quote

2- Using the value of the math trace,where it is all above the 0 volt line(as in scope shot 2),and converting that to an RMS value,dose this give you a P/out value?


That's a good question. Does the "average" equal to the "RMS" value in that case?  What we need is the "average" or "mean" of the Math trace. The Math gives you the instantaneous power at any point in time and should be correct no matter what the wave shape or phase shift between the two signals. Average power is what should be equivalent to the (sine wave) manual calculation of Vrms x Irms x cos (phase angle).

Quote

Anyway,the tests are as shown below--who here can calculate the result's?.


I'm not sure we can since we don't have actual readouts of the phase angle or the average value of the math trace.

Quote



So next i will change out the 100 ohm resistor for a grain of wheat bulb,and also place a bulb of the same between the inductor and cap.
These two bulbs should both glow with the same amount of brightness--correct?


Brad

I'm afraid to make any guesses at this point.  It might be important to use the _same_ bulb in both cases, since bulbs can be expected to vary a little.
   
Group: Guest
Brad, I've had a look at the Atten scope manual and according to what I read there you should be able to get an "automatic" measurement of the Math average and also of the Phase Shift in degrees.

Check Section 2.11.3 "Auto Measurement" on pages 71-74 for a description of how to do it. I know we've been through this thing before with the Math, but the manual says you can select Math as the Source for auto measurements and there is a "mean" listed there. Also in the "delay" measurements it is supposed to give you the phase angle in degrees between two channels.

   

Hero Member
*****

Posts: 3510


Buy me some coffee
 author=TinselKoala link=topic=3445.msg61787#msg61787 date=1494127680]





Quote
No, it doesn't mean a scope is useless or unreliable.
Yes, the problem here is with the measurement of the input power. Since the value is so dependent on the phase shift between CH1 and CH2, and the "desired" shift is so close to 90 degrees, tiny inductances become significant in producing the COP >1 values. This is true for the Capacitor test and also for the full PBT systems.

Here is what im saying.

Through this thread,there has been mention of removing the clip on part of the probe tip's,and using just the point of the probe to take measurements.

Now,your scope is a 100MHz scope-correct?,and you were only at a frequency of 1.5-2MHz--cant quite remember.
Anyway,you are well below the rated frequency of the scope,and yet the scope probes seem to be cause for consideration--even well below what the scope should be able to cope with.

I find it odd that the scope would come with leads/probes that would throw out power measurements well below it's rated frequency range.
So that was my point-why have a scope to make power measurements,when the very scope it self could be throwing those power measurements out-E.G,the scope probes them self.

This is like having a lawnmower that runs fine,as long as you dont cut grass with it. C.C

Speaking of lawn's,im off to mow mine  ;D


Brad


---------------------------
Never let your schooling get in the way of your education.
   

Hero Member
*****

Posts: 3510


Buy me some coffee
Brad, I've had a look at the Atten scope manual and according to what I read there you should be able to get an "automatic" measurement of the Math average and also of the Phase Shift in degrees.

Check Section 2.11.3 "Auto Measurement" on pages 71-74 for a description of how to do it. I know we've been through this thing before with the Math, but the manual says you can select Math as the Source for auto measurements and there is a "mean" listed there. Also in the "delay" measurements it is supposed to give you the phase angle in degrees between two channels.

Thanks TK,i will have another look.


Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
OK,so now my question is-how else are you going to measure the power in,in such a system-other than using the scope ?.


Brad

Yes, that is the challenge, isn't it. I don't know the answer at the moment. I'm not sure even a broadband power analyser would do. Some power analyzers use thermal means to do it, they might work.
   
Group: Guest
author=TinselKoala link=topic=3445.msg61787#msg61787 date=1494127680]





Here is what im saying.

Through this thread,there has been mention of removing the clip on part of the probe tip's,and using just the point of the probe to take measurements.

Now,your scope is a 100MHz scope-correct?,and you were only at a frequency of 1.5-2MHz--cant quite remember.
Anyway,you are well below the rated frequency of the scope,and yet the scope probes seem to be cause for consideration--even well below what the scope should be able to cope with.

I find it odd that the scope would come with leads/probes that would throw out power measurements well below it's rated frequency range.
So that was my point-why have a scope to make power measurements,when the very scope it self could be throwing those power measurements out-E.G,the scope probes them self.

This is like having a lawnmower that runs fine,as long as you dont cut grass with it. C.C

Speaking of lawn's,im off to mow mine  ;D


Brad

Most probe kits include the little "spring thing" which makes the ground connection when you have the spring cliptips off. Using these presents its own problems though, you have to hold the probe by hand or with a fixture, and you have to have your active and ground points close together in your DUT.

The 100 MHz bandwidth of a scope doesn't mean it can't measure signals at higher frequencies. It means that this is the point where the amplitude of the measurement drops 3dB below the actual signal value. It also means that higher harmonics aren't measured correctly. The general rule is that, for example, for an ideal square wave to show up accurately you need to have at least up to the fifth harmonic included in the measurement, so a "100MHz" scope should show a 20 MHz square wave fairly accurately, but over that frequency the higher harmonics drop off so the square wave will look more and more like a sine wave as you go up in frequency.
 
People have measured the actual bandwidth of the DS1054z and its "3dB" dropoff point is actually more like 130 MHz.

But we are in a situation here where there _may_ be significant power in higher harmonics, although it doesn't seem to be the case looking at the signal using my scope's FFT. But you really need a Spectrum Analyzer to answer this question properly.
   
Group: Guest
Here you can see that the scope's calculation of "Average" for the Math trace (all positive) is not the same as the "RMS" calculation for the trace.

(I got this to happen simply by reversing the direct FG connection to the Cap Test circuit, swapping Red and Black leads. I don't understand why this should happen but there it is. It does _not_ happen when using the toroid isolation transformer between FG and circuit, reversing the FG Red and Black has no effect then. )

   

Group: Administrator
Hero Member
*****

Posts: 3705


Buy me some coffee
Hi Partzman,

I do not think anyone here would have such opinion about you.  Please do not take it like that.

Thanks so much for sharing your findings on such interesting setups and please do continue this work.

Gyula
Indeed we all have the utmost respect for Partzman and TK and many others here working on this.
Do we yet know that the capacitor is not COP >1  8)
   

Hero Member
*****

Posts: 3510


Buy me some coffee
Most probe kits include the little "spring thing" which makes the ground connection when you have the spring cliptips off. Using these presents its own problems though, you have to hold the probe by hand or with a fixture, and you have to have your active and ground points close together in your DUT.

The 100 MHz bandwidth of a scope doesn't mean it can't measure signals at higher frequencies. It means that this is the point where the amplitude of the measurement drops 3dB below the actual signal value. It also means that higher harmonics aren't measured correctly. The general rule is that, for example, for an ideal square wave to show up accurately you need to have at least up to the fifth harmonic included in the measurement, so a "100MHz" scope should show a 20 MHz square wave fairly accurately, but over that frequency the higher harmonics drop off so the square wave will look more and more like a sine wave as you go up in frequency.
 
People have measured the actual bandwidth of the DS1054z and its "3dB" dropoff point is actually more like 130 MHz.

But we are in a situation here where there _may_ be significant power in higher harmonics, although it doesn't seem to be the case looking at the signal using my scope's FFT. But you really need a Spectrum Analyzer to answer this question properly.

So,this may explain Graham Gunderson's power measurements then,also showing a COP>1
Equipment measurement error  :)

Perhaps they too worked this out,and decided to remain silent on the subject--rather than have the balls to show the error.
It is a great thing that that dose not happen here on OUR,where error's are posted -if and when found.

That is to say,if there is an actual error in the measurements that have been taken.


I think i will press on with this a little while longer  O0


Brad


---------------------------
Never let your schooling get in the way of your education.
   

Hero Member
*****

Posts: 1472
This is like having a lawnmower that runs fine,as long as you dont cut grass with it. C.C
Almost, but it is like pissing in its gas tank and expecting it to run well.

In this case, the signal coming from the Current Sensing Resistor was erroneous ( 215% too small ).
The scope was doing its job properly with this garbage signal ...and Garbage In = Garbage Out.
« Last Edit: 2017-05-07, 11:54:21 by verpies »
   
Pages: 1 ... 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 [18] 19 20 21 22 23 24
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2019-07-18, 18:31:27