PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2021-10-22, 10:56:44
News: A feature is available which provides a place all members can chat, either publicly or privately.
There is also a "Shout" feature on each page. Only available to members.

Pages: 1 ... 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 155514 times)

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
I posted this some time back,but no one gave it a second thought.

As i stated,measuring the current through the input CVR to calculate the P/in,dose not seem to be right or the correct way to calculate the P/in.

Take another look at my scope shots below,and the description along with them.

The sense/pickup coil,is developing a voltage across it,that is in phase with the voltage(yes,voltage) across the BPC--not the current through the BPC,as we would expect.

And take note that the frequency is quite low as well.

So how can this be?


Brad

Brad,

depending on how far my pickup coil (TBP coil) is from the normal used TBP coil, i see a phase shift between the both coils voltages of 85° far away, to -54° back to back.
So there is no fixed phase shift between the both coils in my situation, it strongly depends on the distance between the 2 coils.


Itsu
   
Group: Experimentalist
Hero Member
*****

Posts: 1362
All,

Here are the results of re-tests of the pcb version of the MEI using the 1x 1Meg ohm probe on the CSR.  A mosfet driver is used for the generator with a 113uh inductor in series with the capacitance of the device resulting in a resonance frequency of 859kHz for maximum input.

The new test procedure was followed with the bottom line result being the device is not capable of COP>1.  I've attached scope pix of the comparison between the 10x and 1x probes used for measuring the CSR and both probes were connected directly across the CSR using spring clips for ground and the probe tip.  When ground leads were used, the measurements were undependable in particular with the 10x probe.  I'm not including the snapshots to conserve bandwidth but one thing they did prove and that is they can not be depended upon for accuracy.  The accuracy from the Math channel calculations can be depended upon IMO.

I apologize to everyone and especially TK, TM, Itsu, plus all others who chased this elusive rabbit down the hole.

Pm

   
Group: Guest
No apology is necessary Partzman, at least not to  me.  I've been learning a lot along this path.

But errors can go both ways, you know. (Except when someone gives me the wrong change for a dollar, it is always in their favour  ;)  )

So are you throwing in the towel now? I thought we were just getting started.

But I have to say that with my new setup it is much harder to obtain any COP>1 results, especially when using the Math average of full cycles between cursors as the Input Power (which I now also believe is correct.)  Even with high phase shifts.   :'(

Latest figures:

f=1.0037 MHz, tuned for max phase shift of -79.52 degrees
IN: Math average 325 mW, current through R2 CVR 124 mA
sine calculation 224.4 mW
OUT: 296 mW, current through R1 Load 125.5 mA

So... buttons.

Partzman, do I have your permission to make and post publicly another video on the topic?

   
Group: Experimentalist
Hero Member
*****

Posts: 1362
No apology is necessary Partzman, at least not to  me.  I've been learning a lot along this path.

But errors can go both ways, you know. (Except when someone gives me the wrong change for a dollar, it is always in their favour  ;)  )

So are you throwing in the towel now? I thought we were just getting started.

But I have to say that with my new setup it is much harder to obtain any COP>1 results, especially when using the Math average of full cycles between cursors as the Input Power (which I now also believe is correct.)  Even with high phase shifts.   :'(

Latest figures:

f=1.0037 MHz, tuned for max phase shift of -79.52 degrees
IN: Math average 325 mW, current through R2 CVR 124 mA
sine calculation 224.4 mW
OUT: 296 mW, current through R1 Load 125.5 mA

So... buttons.

Partzman, do I have your permission to make and post publicly another video on the topic?

TK,

Thanks!  I'm certainly discouraged but not giving up yet.  I have more variations to re-test and will probably try some variations in the test procedure along the way.

Yeah, go ahead and post whatever you like on the subject.

Pm
   
Sr. Member
****

Posts: 291
PW,

I have a BNC tee connector right at the FG bnc output plug, one side of the tee has a bnc to banana adapter where i measure the FG open circuit voltage 7.261V rms.
When connecting the TBP coil circuit to the other side of the tee, the voltage drops to 5.09V rms.

Not sure what you mean by your second sentence, but when i, instead of the TBP coil circuit, a 50 Ohm BNC dummy load connect to the other end of the tee, i get 3.63V rms


The FG is a Rigol DG4102.

Itsu


Itsu,

The 50R load you placed on the BNC Tee gave you exactly 1/2 the open circuit voltage just as would be expected from an FG with a 50R output impedance.

Given the Vdrop you measured when connected to the circuit, it appears that the current flowing thru the FG output is 43.4ma.

PW

 

 
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
PW,

yes, thats right, thats what i also wrote in that first post to you, but the input power i was measuring was 147mW, so therefor
i mentioned that your method did not match with my calculations:   
Quote
40.9mW while the measured input power was 147mW

But probably that mismatch was caused by the now known mistake in measuring the input, allthough that 40 or 147 is still a big difference.


Thanks,  regards Itsu
« Last Edit: 2017-05-08, 21:50:49 by Itsu »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
All,

Here are the results of re-tests of the pcb version of the MEI using the 1x 1Meg ohm probe on the CSR.  A mosfet driver is used for the generator with a 113uh inductor in series with the capacitance of the device resulting in a resonance frequency of 859kHz for maximum input.

The new test procedure was followed with the bottom line result being the device is not capable of COP>1.  I've attached scope pix of the comparison between the 10x and 1x probes used for measuring the CSR and both probes were connected directly across the CSR using spring clips for ground and the probe tip.  When ground leads were used, the measurements were undependable in particular with the 10x probe.  I'm not including the snapshots to conserve bandwidth but one thing they did prove and that is they can not be depended upon for accuracy.  The accuracy from the Math channel calculations can be depended upon IMO.

I apologize to everyone and especially TK, TM, Itsu, plus all others who chased this elusive rabbit down the hole.

Pm

PM, 

ditto here,  no apologize needed,  i hope you will continue to pursue your thoughts on this and keep on posting your results.

Regards Itsu
   
Sr. Member
****

Posts: 291
PW,

yes, thats right, thats what i also wrote in that first post to you, but the input power i was measuring was 147mW, so therefor
i mentioned that your methode did not match with my calculations:   
But probably that mismatch was caused by the now known mistake in measuring the input, allthough that 40 or 147 is still a big difference.


Thanks,  regards Itsu

Itsu,

Using the FG's output Vdrop measurement, the current flowing to the circuit is the FG's Vdrop divided by 50 or in the stated case, 43.4ma.

Pin to the circuit is the measured current times the FG's loaded voltage or (43.4ma)X (5.094V)= 220.9mw

PW
« Last Edit: 2017-05-08, 20:42:39 by picowatt »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194

ooops,      ok, you are right of course, i was wrongly thinking of 40mW of Pin (power) as you mentioned then:

Quote
The difference between those two numbers divided by the 50R output Z will give you another Pin number

But the voltage difference divided by 50 Ohm is current, not power.     Got it   :-[


Itsu
   
Sr. Member
****

Posts: 291
ooops,      ok, you are right of course, i was wrongly thinking of 40mW of Pin (power) as you mentioned then:

But the voltage difference divided by 50 Ohm is current, not power.     Got it   :-[


Itsu

Itsu,

Sorry, I could have been a bit more clear...

The Vdrop divided by 50R gives you the current, the FG's loaded Vout times that current gives you the power.

This should be a fairly accurate method for measuring Pin.  The FG's 50R is being used as a rather large value CVR and even if that 50R or its internal connections are slightly inductive, the percentage of error produced due to any added reactance will typically be quite small.

It is important to make the measurements right at the FG's BNC (i.e., using a BNC Tee as you did) to eliminate cable inductance/reactance.

If you have the time, try comparing this method of determining Pin to others being discussed.

PW


   

Group: Professor
Hero Member
*****

Posts: 2382
I posted this some time back,but no one gave it a second thought.
As i stated,measuring the current through the input CVR to calculate the P/in,dose not seem to be right or the correct way to calculate the P/in.
What would be the correct way?
Would you like to forego measuring the input current altogether or measure it in another manner ?

The sense/pickup coil,is developing a voltage across it,that is in phase with the voltage(yes,voltage) across the BPC--not the current through the BPC,as we would expect.
That is unusual in an RL circuit but in RLC the distance between windings and thier capacitance can shift the current in the coil to be out of phase with the voltage appearing at their terminals..
« Last Edit: 2017-05-08, 21:27:55 by verpies »
   

Group: Professor
Hero Member
*****

Posts: 2382
The Vdrop divided by 50R gives you the current, the FG's loaded Vout times that current gives you the power.
Do you mean RMS averages of VOUT and VDROP or their instantaneous values?
   
Sr. Member
****

Posts: 291
Do you mean RMS averages of VOUT and VDROP ?

Verpies,

Yes... 

This is a fairly measurement if a BNC tee is placed on the FG's BNC,  You need to know the FG's output impedance, most are 50R, but some can be selected to be various values.  If you are uncertain if your FG has a 50R output Z, measuring the difference between the FG's open circuit voltage and the FG's voltage when loaded with a known resistance will allow you to calculate/confirm the FG's output Z.  As Itsu did, a 50R resistor connected to his FG output caused the FG output to drop by exactly half, confirming his FG has a 50R output Z.

So, if your FG has a 50R output, disconnect the circuit from the FG, measure the FG's open circuit output voltage.  Reconnect the circuit and measure the FG's loaded output voltage.  The difference between those two measurements is Vdrop.  Vdrop divided by the FG's 50R gives you the input current.  The input current times the FG's loaded output voltage gives you Pin.

Measurements should be made right at the FG's BNC.

PW

   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
Itsu,

Sorry, I could have been a bit more clear...

The Vdrop divided by 50R gives you the current, the FG's loaded Vout times that current gives you the power.

This should be a fairly accurate method for measuring Pin.  The FG's 50R is being used as a rather large value CVR and even if that 50R or its internal connections are slightly inductive, the percentage of error produced due to any added reactance will typically be quite small.

It is important to make the measurements right at the FG's BNC (i.e., using a BNC Tee as you did) to eliminate cable inductance/reactance.

If you have the time, try comparing this method of determining Pin to others being discussed.

PW

PW,

when using your Pin method i get:

7.211V rms - 5.052V rms = 2.159 V
2.159 / 50 = 0.04318A
Pin = 0.04318 x  5.052 = 0.218A   =  218mA


When using the scope math function i get a Pin of 166mW
Vin yellow = 5.095V rms (across the coil)
IR2 blue   = 57.5mA (voltage across the 1 Ohm csr R2)
Phase shift -57°

When calculating the above values manually i get
Pin = Vin x Ir2 x Cos -57°
Pin = 5.095 x 57.5mA x  0.544  =  159.5mW



Pout measured by scope is 167mW
Pout calculated by Vr1² / R1+R2 =   2.91V rms / 51 = 166mW


These new input / output values are now showing around COP=1 (@ 1MHz) because i have, like TK, cleaned up the circuit and minimized the stray inductance
So your method still shows a much higher input compared to the other 2 methods and compared to the output.
No 1:1 toroid was used.


Itsu
   

Group: Professor
Hero Member
*****

Posts: 2382
The input current times the FG's loaded output voltage gives you Pin.
...but only for purely resistive DUTs (loads).
   
Sr. Member
****

Posts: 291
PW,

when using your Pin methode i get:

7.211V rms - 5.052V rms = 2.159 V
2.159 / 50 = 0.04318A
Pin = 0.04318 x  5.052 = 0.218A   =  218mA


When using the scope math function i get a Pin of 166mW
Vin yellow = 5.095V rms (across the coil)
IR2 blue   = 57.5mA (voltage across the 1 Ohm csr R2)
Phase shift -57°

When calculating the above values manually i get
Pin = Vin x Ir2 x Cos (-57°)
Pin = 5.095 x 57.5mA x  0.544  =  159.5mW



Pout measured by scope is 167mW
Pout calculated by Vr1² / R1+R2 =   2.91V rms / 51 = 166mW


These new input / output values are now showing around COP=1 (@ 1MHz) because i have, like TK, cleaned up the circuit and minimized the stray inductance
So your methode still shows a much higher input compared to the other 2 methodes and compared to the output.

Itsu

Itsu,

Personally, I would be inclined to believe the input power as calculated using the Vdrop across the FG's 50R.

Although it is possible that there is some phase shift or other strangeness throwing off the FG derived measurement, it is also quite possible that there is signal leakage elsewhere with regard to the circuit.  Perhaps, for example, the coils are cap coupling to the bench surface and current is flowing via that path, and/or equipment/signal grounds believed to be "floating" are only cap coupled at best.

You might try a new FG Vdrop based measurement after disconnecting all other probes from the circuit and elevating or suspending the coil via  dielectric standoffs or fishing line to reduce any possible leakage paths.

As I previously mentioned, confirming the FG measurement (i.e., looking for any phase error) requires measuring an internal point in the FG prior to its 50R.         

PW
   
Sr. Member
****

Posts: 291
...but only for purely resistive DUTs (loads).

Verpies,

Is not the FG's 50R the "purely resistive" load in this case?

PW
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194

Ok PW,

my TBP coil is placed vertically on a wooden stand, see picture, so probably has very little coupling to the bench, but i agree, there could be any leakage
when using the probes etc. which i cannot eliminate as i have no data then.


Thanks,   Itsu

   

Group: Professor
Hero Member
*****

Posts: 2382
Is not the FG's 50R the "purely resistive" load in this case?
Which case?
- If you short the FG's output then - yes.
- If you connect a resistive load to the FG's output then - the load is entirely resistive but the 50Ω internal resistance does not contribute to all of the load's resistance.
- If you connect a reactive load* to the FG's output then - no.


* By " reactive load" I mean a load that stores energy and feeds it back to the FG later in time.
   
Sr. Member
****

Posts: 291
Ok PW,

my TBP coil is placed vertically on a wooden stand, see picture, so probably has very little coupling to the bench, but i agree, there could be any leakage
when using the probes etc. which i cannot eliminate as i have no data then.


Thanks,   Itsu

Itsu,

Your coil's mount does look pretty good from a cap leakage stand point, so likely it is not related to that  But still, raising it up a bit with a plastic spacer (grab an empty parts drawer) and repeating your measurements might be worthwhile (around here even dry wood is considered a decent conductor).  I understand that disconnecting probes will preclude making any measurements, but by disconnecting all but the FG leads and repeating the FG Vdrop measurement, it might help determine if there are any unintended leakage paths.

PW   
   
Sr. Member
****

Posts: 291
Which case?
- If you short the FG's output then - yes.
- If you connect a resistive load to the FG's output then - the load is entirely resistive but the 50Ω internal resistance does not contribute to all of the load's resistance.
- If you connect a reactive load* to the FG's output then - no.


* By " reactive load" I mean a load that stores energy and feeds it back to the FG later in time.

Verpies,

Now you have me wondering...

We have the FG's low Z output amp (assumed to be of unchanging amplitude and phase) driving a 50R non-inductive resistor connected to an unknown load.

We are measuring across that 50R resistor.

PW
   

Group: Professor
Hero Member
*****

Posts: 2382
We have the FG's low Z output amp (assumed to be of unchanging amplitude and phase) driving a 50R non-inductive resistor connected to an unknown load.  We are measuring across that 50R resistor.
For the sake of analysis, we can assume that the output impedance of the FG's DAC+amp is 0Ω.

Now let's consider a situation in which the DUT has stored some energy in the past and now at this particular instance in time is presenting 10V to the FG.

Which way will the current flow through that internal 50Ω resistor ?
To know this, you'll need to know what voltage the FG's DAC+amp is outputting at that moment:
1) If the FG's DAC+amp is outputting >10V then the current will flow from the FG to the DUT.
2) If the FG's DAC+amp is outputting 10V then no current will flow between the FG and the DUT.
3) If the FG's DAC+amp is outputting <10V volts then the current will flow from the DUT to the FG.

...but, but, you don't know what the FG's DAC+amp is outputting at this moment so you can't tell which way the current is flowing nor the sign of the energy flow (power).

For your power measuring method to work, you could hack the FG and get at the voltage signal before the internal 50Ω resistor (which is not a bad idea) or with a FG like Itsu's you could mirror the Ch1 signal on Ch2 and get at the internal driving voltage signal that way, without opening the FG.
« Last Edit: 2017-05-09, 06:43:45 by verpies »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Verpies,

Now you have me wondering...

We have the FG's low Z output amp (assumed to be of unchanging amplitude and phase) driving a 50R non-inductive resistor connected to an unknown load.

We are measuring across that 50R resistor.

PW

What is the difference in measuring across the FGs t0 ohm resistor,and the next resistor in the series-being the 1 ohm NI precision  resistor,when an inductor is the load?


Brad


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
t0 Ohm?

Besides the typo t0 instead of 50,that is not what i mean.


Brad


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

Group: Professor
Hero Member
*****

Posts: 2382
Besides the typo t0 instead of 50, that is not what i mean.
What don't you mean besides the typo?
   
Pages: 1 ... 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 2021-10-22, 10:56:44