PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2021-12-07, 21:36:30
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 2 3 4 5 [6] 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 ... 24
Author Topic: Investigating "anomalies" in Bifilar coils  (Read 157935 times)

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
"Ground" is a point. It cannot be in series with anything...anymore that it can be in parallel with anything.
Only multi-terminal components can be connected in series or parallel.
L1 forms a capacitance with much than L2.  This bypasses R2.
Designate the other side of R2 as the "ground" point and see for yourself

Your making no sense verpies.

You cannot make a length of rope tight by pulling on one end,while the other end is not fixed (grounded)

All the current flowing from the source(FG),must pass through R2-there is no other path for the current to flow through.

If you remove R2,and connect the ground to R1,then you have just altered the circuit.

If we add an R3 that has the same value as R2,at the point between V1 and L1,then R3 will show the same amount of current flowing through it as R2,meaning that all the current flowing into the system,must flow through R2.


Brad


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Brad,

OK, great. 

I am not understanding where you see a ground loop (not to say there isn't one) so could you perhaps be a little more specific.  Otherwise the equivalent circuit is as you show.

Pm

PM

I never mentioned a ground loop.
I did mention a current loop though.

Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
Another result from my PBF flat pancake coil.

FG set to 15 v p-p, tuned to 502 kHz for approx. -80 degree phase shift

IN: 4.45 x 0.0267 x cos -81.41 = 0.017747 W or 17.7 mW

OUT: 1.342/50.58 = 0.03550 W or 35.5 mW

COP 2.00


So if the fact that my 50 ohm resistor (actually 2ea. 100 ohm 5% in parallel, measuring 49.58 ohms) is not noninductve is the cause of my "success", what's wrong with that? Maybe "Realni OU" (tm TKLabs) requires a little bit of inductance there.

(where's that tongue in cheek smiley anyhow...)


   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Partzman, the solenoid coil is #33 green enamelled magnet wire, 376 turns (188+188) wound on a wooden dowel 12.67 mm diameter. I have not accounted for the DC resistance of the coil halves which is about 5.10 ohms per half.

I was waiting for this to come up.
So,our R1 is directly across L2-->so what is the true value of R1 that we are measuring the voltage drop across.

In my case,L2 has a very low resistance.


Brad


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Another result from my PBF flat pancake coil.

FG set to 15 v p-p, tuned to 502 kHz for approx. -80 degree phase shift

IN: 4.45 x 0.0267 x cos -81.41 = 0.017747 W or 17.7 mW

OUT: 1.342/50.58 = 0.03550

COP 2.00


So if the fact that my 50 ohm resistor (actually 2ea. 100 ohm 5% in parallel, measuring 49.58 ohms) is not noninductve is the cause of my "success", what's wrong with that? Maybe "Realni OU" (tm TKLabs) requires a little bit of inductance there.

(where's that tongue in cheek smiley anyhow...)

Simple-test your resistor for inductance,as i did with mine.
Not a sign of inductance up to 10 MHz--as far as i bothered to go.


Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
I was waiting for this to come up.
So,our R1 is directly across L2-->so what is the true value of R1 that we are measuring the voltage drop across.

In my case,L2 has a very low resistance.


Brad

Well, we have a 49.58 ohm resistor in parallel with a 5.1 ohm coil, so the resulting resistance is only  4.62 ohms. Plus the 1 ohm at R2 gives 5.62 ohms. Do we now use this value for the resistance in the output power equation P=Vch3rms2/R ?

So where before we had, say, an output value of 1.342/50.58 = 0.0355 W against an input of 17.7 mW for a COP of 2.0, now we have 1.342/5.62 = 0.3195 W or 319.5 mW, which raises the COP to 18.0 or some silly value like that.





Somehow I just don't think that's right.


   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
I think you don't have enough phase shift. Try tuning the frequency until you get around -75 to -85 degrees of phase shift between input voltage (yellow) and input current (green).

Ok, did that, had to go down from 1.4Mhz to 190Khz to get -80° phase shift, see screenshot

The input now calculates to be 22.5mW, see red math trace (current controller set to 10mA/Div., so similar as to the green vertical setting on the scope).
Output calculates to be 871mV² / 52 = 14.5mW

Be aware i measure the input current with my current probe at the red FG lead!

My 50 Ohm resistor is 51 Ohm 1%,  my 1 Ohm resistor is 1 Ohm 1%.

Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Hi Itsu,

Sorry to chime in, did you measure the load resistance to be 52 Ohm? 

(and in your formula you first wrote 51 Ohm and then calculated with 52 Ohm,  with 51 Ohm the output power would be 210.2 mW)

Gyula

Hi Gyula,

my load resistor is 51 Ohm 1%,  the csr is 1 Ohm 1%, totaling 52 Ohm when using the CH 3 (purple) voltage.
TK used 51 Ohm in his formula (50 +1), i adjusted that to 52 in my situation as i have 51 + 1.


Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Itsu,

There appears to be something wrong with the current to input voltage phase in your scope pix for some reason.  The current should be leading the input voltage by ~80 degrees or so and yours appears to be considerably less than that.  Compare your results with TK's post #97 for example.  What are the specs on your coil assembly like the inductance of each winding, inductance of the windings in series, and the capacitance between the windings?

Thanks for doing the test.

Pm

OK,  redid the test, now on 190Khz which makes the proper phase shift (-80°).

The specs of my TBP coil are:

L1= 132.8uH, 1.26 Ohm
L2= 135.3uH, 1.27 Ohm
L1 + L2 series= 533uH
C between windings= 2.2nF

Regards Itsu

 
   
Group: Guest
Itsu, what happens if you move your current probe to be right next to the 1 ohm resistor, preferably on the ground side of it?


Meanwhile, more results. Just kind of randomly twiddled knobs and measured.
FG set to 10 V p-p, 1.012 MHz.

IN: 2.22 x 0.0268 x cos -74.63 = 0.01577 W or 15.7 mW
Scope average of Math trace 19.8 mW

OUT: 1.052/50.58 = 0.021797 W or 21.8 mW,  COP 1.38

Or if I use the scope average of Math trace,  COP 1.10


The resistance of the half-winding of the pancake coil is only about 1.95 ohms. You don't even want to know what happens to the COP if I use that value in parallel with the 49.58 ohm load, as the resistance in the output power calculation. 1.88 ohms!
1.052/2.88 = 0.3828 W, COP around 20. 


 :D

   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Quote
Itsu, what happens if you move your current probe to be right next to the 1 ohm resistor, preferably on the ground side of it?

Not much, somewhat lower input calculated, 19.5mW instead of 22.5mW

I moved the current probe from the red lead to the black lead (close to the 1 Ohm resistor)

Itsu
   
Group: Guest
OK, all the above tests were done with the ElCheepo DDS FG (Ming-Ho MHS5200A) which runs on a wallwart 5VDC adapter without ground pin, so its outputs are isolated from ground, but as a direct digital synthesis unit its sine wave is made up of tiny steps and is only approximate especially at higher outputs and frequencies.

Now I've changed to the rusty trusty Interstate F43 High Voltage FG, which has an isolation switch that performs that function, isolating its outputs from ground. This old analog workhorse tops out at 3 MHz but performs great at lower frequencies and certainly has a better sine wave output than the Ming-Ho.

So with the F43 analog FG, here are some results at 951 kHz:

IN:  9.5 mW
OUT: 34.4 mW 
COP  3.82

I'll post the scopeshots here in a few minutes.



   
Group: Tech Wizard
Hero Member
*****

Posts: 929
Okay Itsu, I got it now, thanks for the explanations. 

Gyula

Hi Gyula,

my load resistor is 51 Ohm 1%,  the csr is 1 Ohm 1%, totaling 52 Ohm when using the CH 3 (purple) voltage.
TK used 51 Ohm in his formula (50 +1), i adjusted that to 52 in my situation as i have 51 + 1.


Itsu
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
"Ground" is a point. It cannot be in series with anything...anymore that it can be in parallel with anything.
Only multi-terminal components can be connected in series or parallel.
L1 forms a capacitance with much more than L2.  This bypasses R2.  Also, for input power calculations, the input voltage should not include the voltage drop of R2.
Designate the other side of R2 as the "ground" point and see for yourself

Verpies,

You have valid points here.  To address the first, I've attached scope pix of the probe placed above R2 as compared to being placed at the input.  There is a differential of 1.25 ma for an increase of 4.3% in this case.  This could be attributed to the capacitive coupling of the L1 primary to the common ground, a combination of this and probe placement, or simply probe placement as it will vary with position.  In any case we can use the differential as a correction factor in our calculations.

Secondly, you are correct that the voltage drop across the sense resistor R2 should be subtracted from the input voltage.  I have attached a pix that gives us this result with the scope Math channel with the input voltage at 6.48v rms and the corrected input at 6.475v rms.

There is yet a third correction factor that should be discussed which was pointed out by Smudge previously and that is the actual power output.  Referring to the "Sense Voltage Comparison" shot, I normally use the measured output voltage which in this case is 1.502v rms to calculate the output power across a total load of 51 ohms but sometimes depending on the power levels this actually understates the output.  For example 1.502^2/51 = 44.24mw.  Since the voltage across the sense resistor is nearly in phase with the output voltage, we need to subtract this from the overall output voltage.  The resultant voltage is now across R1 only so with this correction we have, (1.502-.02984)^2/50 = 43.34mw (power across R1) and (.02984)^2/1 = 890uw (power across R2) for a total Pout = 44.23mw.

When using these correction factors in my original example, the corrected input voltage of 6.492v rms = 6.487 and the corrected input current of 30.53ma = 31.84ma for a corrected Pin = .207*cosine(-81.8 degrees) = 29.46mw.  The original uncorrected and calculated Pin = (6.492*.03053)*cosine(-81.8 degrees) = 28.27mw for comparison.  The corrected COP = 43.7mw/29.46mw = 1.48.

Pm
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
PM

I never mentioned a ground loop.
I did mention a current loop though.

Brad

Brad,

Oops, my bad!  My brain saw "ground loop", sorry about that!

Pm
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
Ok, did that, had to go down from 1.4Mhz to 190Khz to get -80° phase shift, see screenshot

The input now calculates to be 22.5mW, see red math trace (current controller set to 10mA/Div., so similar as to the green vertical setting on the scope).
Output calculates to be 871mV² / 52 = 14.5mW

Be aware i measure the input current with my current probe at the red FG lead!

My 50 Ohm resistor is 51 Ohm 1%,  my 1 Ohm resistor is 1 Ohm 1%.

Itsu

Itsu,

Let's start with the deskew of your current probe which your scope will allow but may be limited in range.  A good check of this since you have a non-inductive 1 ohm sense resistor and you also should have the HF spring contact to use on your scope probe for a direct contact across the sense resistor, place your current probe on a short lead between the load resistor R1 and the top of sense resistor R2 and compare the current probe trace with the scope probe trace and adjust the deskew to align the traces if possible.  If you are using a current probe amplifier, set it to the highest bandwidth possible.

If the traces do not align, simply use the voltage across the sense resistor R2 in your Math calculation.

I just re-checked your scope pix on this post and I see that your measurements are taken across the entire screen which shows and incomplete number of cycles which will affect your overall measurements.  You can either adjust the frequency for a precise number of cycles or turn on your vertical cursors and activate measurements between cursors.

Pm

Edit: I'm curious if you can achieve a COP>1 at this frequency!
   
Group: Guest
The mighty Interstate F43 function generator is capable of producing 40 v p-p into 50 ohms, so I cranked it up to near the top of its output range. 

510  kHz results:

IN:
11.9 x 0.0712 x cos -83.08 = 0.10208 W or 102 mW

OUT:
3.652/50.58 = 0.2634 W or 263 mW.

COP 2.58

But I'm using 1/4 watt resistors for the 49.58 ohm "load". Are they getting warm or not? Yes, they are perceptibly warming very slightly by touch. I have no way of measuring this though.



   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

PM,

ok, i will follow these procedures.

I am calibrating my scope at the moment which will take some time, then i will show you the results of the current probe versus the 1 Ohm resistor.

Concerning your "measurements are taken across the entire screen which shows and incomplete number of cycles" i don't think it will matter, but will double check on that.
 
By the way, i did also calculate the input power by using the R2 current like done by TK and TM, and it was within a few mW's (20mW input versus 22.5mW input
using the current probe).

Itsu

 
   
Group: Tech Wizard
Hero Member
*****

Posts: 929
Hi Partzman and all,

I wonder whether you have already considered to test your bifilar coil circuit to embed into a looped circuit?

IF not yet, then I would propose to build a simple one transistor oscillator that would have a normal output coil inductively coupled to the main oscillator LC tank circuit and would feed your bifilar coil circuit as the FG does now. 
And at the output of your bifilar coil circuit a full wave bridge rectifier (made from say Germanium diodes) would drive a DC voltage stabilizer, this latter should have very low idling current of course. Then the output of this DC stabilizer would drive the one transistor oscillator. Such simple looping would eliminate any measuring issues... if there are any, that is. 

The DC voltage level for the one transistor oscillator could easily be chosen to be very close to the output DC level of the stabilizer (chosen strategically) so that a linear, micropower regulator could be used without too much loss.

With the COPs around 1.5 to 3 or so as the present measurements indicate, such looping should be possible and even if the power levels involved are in the some ten to some hundred mW range, that would be enough to maintain the operation of the looped circuit, perhaps even a dimly lit LED could be run from it.

Gyula
   
Group: Guest
Partzman, I am assuming that the corrections you calculated up above being small, will also be small in my setup. Do you have enough information to calculate the corrections for one of my datasets?

Also, what effect do you think my casual probing and my non-noninductive resistor might be having on my results? I would like to have some confidence that I am seeing the same effect you have seen but I'm worried that my results may be artifacts, whereas yours may not be, since you are using tighter probing and you have your proper resistor.

I'm happy that the effect can be reproduced even if it's not due to the same cause... but I think it probably is. However this brings up some other issues. Why did you not get similar results when you tried round coils yourself, but only got them with the Golden Mean proportions? Can Itsu reproduce the results using his current probe, wherever placed?  And of course we have the issue of whether or not our probing scheme is actually valid for this circuit. At present I am just looking at data, I have not started trying explain it or to find errors if they are there.

I've got to say though, that I haven't had this much fun in the lab in quite some time!   >:-)
   

Group: Renaissance Man
Hero Member
*****

Posts: 2443


Buy me a cigar
But I'm using 1/4 watt resistors for the 49.58 ohm "load". Are they getting warm or not? Yes, they are perceptibly warming very slightly by touch. I have no way of measuring this though.

Dear TinselKoala.

If those resistors aren't of too great importance you could attach the end of a cheap digital thermometer with a dab of Epoxy resin and monitor the temperature rise.

I had thoughts of this some time ago as an aid to measurements by passing a known current and voltage through and calibrating wattage against temperature.

Cheers Graham. 


---------------------------
Nanny state ? Left at the gate !! :)
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
Hi Partzman and all,

I wonder whether you have already considered to test your bifilar coil circuit to embed into a looped circuit?

IF not yet, then I would propose to build a simple one transistor oscillator that would have a normal output coil inductively coupled to the main oscillator LC tank circuit and would feed your bifilar coil circuit as the FG does now. 
And at the output of your bifilar coil circuit a full wave bridge rectifier (made from say Germanium diodes) would drive a DC voltage stabilizer, this latter should have very low idling current of course. Then the output of this DC stabilizer would drive the one transistor oscillator. Such simple looping would eliminate any measuring issues... if there are any, that is. 

The DC voltage level for the one transistor oscillator could easily be chosen to be very close to the output DC level of the stabilizer (chosen strategically) so that a linear, micropower regulator could be used without too much loss.

With the COPs around 1.5 to 3 or so as the present measurements indicate, such looping should be possible and even if the power levels involved are in the some ten to some hundred mW range, that would be enough to maintain the operation of the looped circuit, perhaps even a dimly lit LED could be run from it.

Gyula

Gyula,

I have considered looping the device in the past but have not done so however, your idea certainly has merit and should be given a try.

Pm
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
Partzman, I am assuming that the corrections you calculated up above being small, will also be small in my setup. Do you have enough information to calculate the corrections for one of my datasets?

Also, what effect do you think my casual probing and my non-noninductive resistor might be having on my results? I would like to have some confidence that I am seeing the same effect you have seen but I'm worried that my results may be artifacts, whereas yours may not be, since you are using tighter probing and you have your proper resistor.

I'm happy that the effect can be reproduced even if it's not due to the same cause... but I think it probably is. However this brings up some other issues. Why did you not get similar results when you tried round coils yourself, but only got them with the Golden Mean proportions? Can Itsu reproduce the results using his current probe, wherever placed?  And of course we have the issue of whether or not our probing scheme is actually valid for this circuit. At present I am just looking at data, I have not started trying explain it or to find errors if they are there.

I've got to say though, that I haven't had this much fun in the lab in quite some time!   >:-)

TK,

OK, looking at your post #141, the voltage across your sense resistor appears to be ~71.2v rms and with a phase angle of ~83 degrees so, you would subtract .0712 *cos(83) = 8.7mv rms from your 11.9v rms input leaving a net input of 11.89v rms.  Assuming you would have approximately the same outside capacitive coupling as I appear to have, increase the sensed current by 5% to 75ma rms.  Your corrected Pin is now 11.89*.075*cos(83) = 109mw for a new COP = 2.41 vs 2.58.

I prefer to use the recommended RF style measurement technique with the probe tip and spring ground right at the sense resistor with all other grounds connected directly at the bottom of R2.  Moving your probes around while taking measurements should give you an indication of any circuit to probe interaction.

The inductive load resistor could lead to erroneously high COPs depending on how inductive it is at the operating frequency.  You might try to series resonate the load resistor with a known capacitor to profile it's inductance.  I used this simple test to qualify the Caddock non-inductive film resistors I use and basically found insignificant inductance up to ~10MHz.

The round coil tests I ran that never provided a COP>1 were using 3-5 stacked coil configurations with variations.  I have never tried 2 coils in a round configuration.  I did have success with triangular, square and rectangular shapes plus coils wound on toroidal forms but all in 3 coils or more.

I will be pleasantly surprised if Itsu can get a COP>1 at his low operating frequency.  The lowest frequency I have achieved COP>1 is at ~650kHz using 3 coil devices.  The goal is to reach a COP~1.5 to 2.0 at 100kHz or below.  If this is reached, a self runner can be realized IMO.

I really glad to hear you are having fun with this O0 !

Pm

Edit: I should comment that I have never tried using coils in the configuration that you TK, TM, and Itsu, are using in the Tesla wound configuration- Hmmmm!   
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Itsu,

Let's start with the deskew of your current probe which your scope will allow but may be limited in range.  A good check of this since you have a non-inductive 1 ohm sense resistor and you also should have the HF spring contact to use on your scope probe for a direct contact across the sense resistor, place your current probe on a short lead between the load resistor R1 and the top of sense resistor R2 and compare the current probe trace with the scope probe trace and adjust the deskew to align the traces if possible.  If you are using a current probe amplifier, set it to the highest bandwidth possible.

If the traces do not align, simply use the voltage across the sense resistor R2 in your Math calculation.

I just re-checked your scope pix on this post and I see that your measurements are taken across the entire screen which shows and incomplete number of cycles which will affect your overall measurements.  You can either adjust the frequency for a precise number of cycles or turn on your vertical cursors and activate measurements between cursors.

Pm

Edit: I'm curious if you can achieve a COP>1 at this frequency!

After a fresh calibration, these are the signals from my current probe (green) placed between the resistors R1 and R2 (17.36mA rms) and the yellow
probe placed across the R2 resistor (19.36mA rms).

This is with the same setting from my FG (20Vpp sine wave @ 190KHz) as used before.

The yellow signal is kind of weak, so the between vertical cursor test is not very stable.

Itsu

EDIT    i added a screenshot of my FG with csr only at 26MHz frequency using the current probe (50mA/Div.) and yellow probe across the csr, see screenshot 2
The 3th screenshot is as screenshot 2 but now at 190KHz, and with square wave at 25% duty cycle (current probe at 100mA/Div.)

EDIT 2   i added a 4th screenshot similar as screenshot 2, but now deskewed (max. + / - 10ns) so the green current probe trace matches the yellow voltage probe trace.
 
« Last Edit: 2017-05-02, 18:16:06 by Itsu »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270
Using a different TBP coil (with similar properties) i redid my measurements using the same 51 Ohm and 1 Ohm resistors.

Input calculated by the scope/current probe is 22.7mW.
Input calculated by using the R2 current is: 6.81V rms * 0.0189mA rms * cos -80° =  22.3mW
See screenshot 1

Output calculated by scope is 15.4mW
Output calculated by Ch3 (purple)² / 52 Ohm = 0.8682² / 52 =  14.49mW
See screenshot 2

So this different coil also shows via 2 different methods a similar cop<1 as the first TBP coil.

 
This 2e coil specs:

L1= 128.9uH,  1.2 Ohm
L2= 128.8uH,  1.2 Ohm
L1 + L2 series = 528.6uH
C between windings = 2.37nF

Itsu

   
Pages: 1 2 3 4 5 [6] 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 ... 24
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2021-12-07, 21:36:30