PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2021-12-07, 21:29:47
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 2 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 157934 times)
Group: Experimentalist
Hero Member
*****

Posts: 1380
Yes, it was! I get those large angles at lower frequencies.

OK, I see the frequency was 1.449MHz.  I assume the load was ~10 ohms and the power output was rather low during this measurement.  You could try the same test when you have time with the output shorted and continue to lower the input frequency for the maximum phase shift possible and note the resultant negative Pin even if it's a small number.

Pm
   
Group: Tech Wizard
Hero Member
*****

Posts: 929

Gyula's approach looks promising, I like the idea of the tuned secondary, however I question the need for the linear regulator LM317, unless it is just for runaway protection. O0 Also a push pull Mazili type oscillator with resonant secondary may prove to be more efficient.
...

Hi ION,

Well, the linear regulator can be made "efficient" if we can establish a low voltage difference across its input and output, by making the oscillator's optimal supply voltage level to be just half to one volt less than what we receive at the output of the full wave rectifier (The use of a voltage doubler rectifier may or may not be needed and a tap on the winding of L2 could also help achieving the needed output DC level).

This way the total loss (including the 1 mA idle current) in the regulator could be kept at less than 10-15 mW (based on 8-9 VDC running supply voltage level for the oscillator). Should this loss prove still too high in this stage, then a dedicated DC-DC converter is to be used with at least around 90% efficiency.

I agree with a push-pull type oscillator like Mazili, or a dedicated Class-E type oscillator, this latter may also have 90+ % efficiency in the 1-2 MHz range. So if the rectifier diodes are Germanium types and could have really negligible losses, then the resulting 80-81 % efficiency of these two stages (0.9x0.9x100) would "demand" at least a COP of 1.3 or so from the bifilar coils (to be in the safe sustainable loop range). Perhaps two of such bifilar coils could be cascaded to increase their COP. 

I like your proposal of using reactances to do the looping, this method certainly involves the smallest loss possible. This method inherently involves that the operational frequency would always be at the bifilar coil's (self) resonant frequency and this then would "demand" that the highest COP should occur at their resonant frequency and not elsewhere in the frequency range. Logically, the highest COP may happen indeed at resonance, tests can confirm this. 

Thanks,
Gyula
   

Group: Renaissance Man
Hero Member
*****

Posts: 2443


Buy me a cigar


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

I toke the liberty to make a video of my coil and measurements, perhaps someone can spot what i am doing wrong here.
Sorry for the rather long video:  https://www.youtube.com/watch?v=OsvJNlYhNUM

As i have made measurements of 2 different coils with a set of 2 different resistors with similar (negative) outcome, i only can conclude for myself that
it behaves different as the other (already 3 i understand) replicators which all 3 report COP>1 results.

If no-one can point me to an obvious problem or error i make, i think it would be best that i step out of this thread so you guys can concentrate on exploiting
this effect by trying to further enhance it and even loop it.

Good luck.

By the way, i swapped over what TK calls his "red" and "blue" windings, and did not note a major change there.

Regards Itsu
   
Group: Tech Wizard
Hero Member
*****

Posts: 929
Is this any use?

http://www.ebay.co.uk/itm/10m-Professional-Quality-Pure-Copper-OFC-4-x-1mm-Blue-Bi-Wire-Loud-Speaker-Cable-/332043613736?hash=item4d4f577628:g:73IAAOSwImRYOCxf

Hi Graham,

I think the goal would be to have as high self capacitance between the tightly spaced windings as possible and thin wires with rectangular cross sections are probably among the best candidates for this.

Here are such wires in enamelled copper strip shape, in their open list the thinnest strip seems to be 1.12 mm while the width is 3.15 mm:
http://www.wires.co.uk/cgi-bin/sh000001.pl?WD=1%2e12mm%20x%203%2e15mm&PN=rt_ec_wire%2ehtml#art_2d3150x1120_2d1kg

Maybe they do have thinner than 1.15 mm while the width is still at least 2.5 mm or preferably wider like 3 mm, (you would need to ask them), to have big facing areas when you wind them bifilarly.

Maybe there are other sellers that have rectangular and enamelled copper stripes.

EDIT The Chinese produce it  http://www.windingwire.net/products/enameled-copper-wire/flat-copper-wire.html

Gyula
   
Group: Elite
Hero Member
******

Posts: 3541
It's turtles all the way down
Attached is another version using a current fed push pull oscillator similar to Mazzilli and some others. It uses Gyula's idea of the oscillator feed and the doubler.

Could possibly also use a bridge in place of doubler, if the output voltage is a bit higher than input.

I have included an LED surge load that becomes active should the voltage begin to soar much above the PP3 9 volt level. Also has a momentary push to start switch.

The next variation will have fail proof limiter and a crowbar shutdown circuit.

partzman and I were discussing the possibility of building the primary coil as part of the oscillator circuit by adding some taps, This would eliminate external transformer losses.

Values of components will be defined when other parameters are nailed down.

Comments / critiques welcome


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Experimentalist
Hero Member
*****

Posts: 1380
I toke the liberty to make a video of my coil and measurements, perhaps someone can spot what i am doing wrong here.
Sorry for the rather long video:  https://www.youtube.com/watch?v=OsvJNlYhNUM

As i have made measurements of 2 different coils with a set of 2 different resistors with similar (negative) outcome, i only can conclude for myself that
it behaves different as the other (already 3 i understand) replicators which all 3 report COP>1 results.

If no-one can point me to an obvious problem or error i make, i think it would be best that i step out of this thread so you guys can concentrate on exploiting
this effect by trying to further enhance it and even loop it.

Good luck.

By the way, i swapped over what TK calls his "red" and "blue" windings, and did not note a major change there.

Regards Itsu

Itsu,

First, your connections look correct and I think I see what you basic problem is but let's cover the inductance of those resistors first. 

I wouldn't necessarily depend on your meter to qualify the inductance due to the frequency it is using for the measurement,  but what I would do is set up a quick test with your generator, a low esr cap (mica if possible), and your resistors.  I think your 1 ohm used for R2 is OK because it appeared to have a resonance at 20-25MHz where you could see the voltage suddenly increase.  So, take your 51 ohm R1 and connect it in series with say a 1nfd cap, connect between your sig gen and ground and manually sweep the generator while looking for resonance with your scope.  You may have to change values of the cap to find resonance if there is any and from there you can calculate the inductance as you know.

OK, the basic problem you are experiencing is with your current probe/amp combo.  The deskew adjustment range of your scope at +- 10ns is not sufficient to compensate for your amplifier and the A6302 probe.  You actually need 25-35ns or possibly more and this is evident in your scope shots.  For example, in your video at the 8:07 time mark at 2.2MHz, CH4 is in phase with CH1 but compare to the phase of CH2 by holding a paper straight edge vertical to the display.  Likewise at time 8:35 or 3MHz, CH4 lags CH1 but CH2 leads CH1 and also the same is evident at time 8:54 or 5MHz.   IOW, you are basing your phase shift measurements on CH1-CH4 and your Math power measurement on CH1 x CH4.  These results are incorrect because you can't apply enough negative deskew to correct for the lag or delay in the probe system and at the frequencies where you should begin to see COP>1, this slight difference will matter.

The solution would be to try basing the phase and power measurement on CH1-CH2 and CH1 x CH2 and see what the results are.  I expect your phase angle will increase and your input power will decrease accordingly.

Pm   
   
Group: Guest
Well.... I have now added my last two 4.7 ohm precision NI resistors in series with the others on very short leads, to make a total of 18.8 ohms for R1, and I went back to the 1.00 ohm NI for R2.

Once again it seems "easy" to get the COP > 1 result.

Using the F43 FG set to 1.005Mhz (just a random setting to get a phase shift in the right range) I get this (scopeshot below):

Data:
Phase shift CH1>CH2  -76.06 deg (scope's auto measurement) or -75.9 deg (using cursors)
CH1 Vp-p 26.0 V
CH2 Ap-p 0.166 A
CH3 Vp-p 5.44 V

IN:
Scope Math CH1xCH2 average input = 0.148 W
Manual calculation using Vp-p and cursor phase values = 0.1314 W

OUT:
Manual calculation using CH3 Vp-p = 0.18677 W

COP 1.42 (manual) or 1.13 (using scope math average input)



Itsu: No no no! Please don't "step out", we need your input and we would really like for you to keep experimenting and trying to track down what variables are important to see why you aren't getting similar results to mine, since your coil is evidently pretty close in electrical parameters to mine.


   
Group: Guest
Another: F43 FG at 1.255 Mhz,  Scope auto phase angle -73.8, cursor phase angle -74.17, R1 = 18.8 ohm NI, R2 = 1.00 ohm NI

IN:
Math average 187 mW
Manual average 147.7 mW

OUT:
using (7.4Vp-p x 0.3535)2/19.8 = 0.3456 W

COP 2.35 (manual average In) or 1.84 (scope math average In)

   
Group: Guest
Another: F43 FG at 1.255 Mhz,  Scope auto phase angle -73.8, cursor phase angle -74.17, R1 = 18.8 ohm NI, R2 = 1.00 ohm NI

IN:
Math average 187 mW
Manual average 147.7 mW

OUT:
using (7.4Vp-p x 0.3535)2/19.8 = 0.3456 W

COP 2.35 (manual average In) or 1.84 (scope math average In)

I have discovered one thing at least: My CH2 (current through R2) probe measurements are influenced by the RF emissions from the coil and are probably not reliable for that reason.  I don't know if I have enough hands to eliminate the spring tip clip and gator ground lead on the CH2 probe and use the spring ground pin instead but I'll try it. I also think I have a "Steve Weir" kelvin-type probe around here somewhere that I might be able to find and use.
   
Sr. Member
****

Posts: 340
I'm not experienced in electronics like you are  TinselKoala, though it still surprise me why you all dismiss the thermal measurements on resistors (placed outside the RF from pancake coils) as the simplest way to calculate average power dissipated without not reliable scope measurements in RF field.Any reasons ?
   
Group: Guest
Still using old probing system.... Please check my math. R1= 18.8 ohms, 1%, non-inductive; R2 = 1.00 ohm, 1%,  non-inductive.

   
Group: Guest
I'm not experienced in electronics like you are  TinselKoala, though it still surprise me why you all dismiss the thermal measurements on resistors (placed outside the RF from pancake coils) as the simplest way to calculate average power dissipated without not reliable scope measurements in RF field.Any reasons ?

Yes, at such low power levels it is extremely difficult to obtain accurate and precise enough temperature measurements without doing some pretty elaborate calorimetry using fixed-loss-to-ambient techniques and very sensitive and precise thermometers.  In addition, this circuit is very sensitive to stray and parasitic inductance, so the lead wires to external resistors in a calorimeter will add significant inductance and alter the performance of the circuit in unpredictable and perhaps unquantifiable ways.

The output measurements here are probably fairly reliable anyway; where the difficulty lies is in the _input_ power measurements.  Since the input power is provided by a function generator it is more difficult to use, say, a capacitor charged with a known amount of energy to provide the input. It could be done though, for example with a preamp stage clocked by the FG and providing power to the circuit, through a bipolar transistor or a mosfet, from a charged supercap. 

Perhaps more importantly, we still have not resolved the issue of whether it is legitimate and proper to use the current through R2 as the correct input current measurement. We have several opinions that it is not proper to do so. Alternate means of measuring the input current should be tried and qualified.
   
Sr. Member
****

Posts: 340
Yes, at such low power levels it is extremely difficult to obtain accurate and precise enough temperature measurements without doing some pretty elaborate calorimetry using fixed-loss-to-ambient techniques and very sensitive and precise thermometers.  In addition, this circuit is very sensitive to stray and parasitic inductance, so the lead wires to external resistors in a calorimeter will add significant inductance and alter the performance of the circuit in unpredictable and perhaps unquantifiable ways.

The output measurements here are probably fairly reliable anyway; where the difficulty lies is in the _input_ power measurements.  Since the input power is provided by a function generator it is more difficult to use, say, a capacitor charged with a known amount of energy to provide the input. It could be done though, for example with a preamp stage clocked by the FG and providing power to the circuit, through a bipolar transistor or a mosfet, from a charged supercap. 

Perhaps more importantly, we still have not resolved the issue of whether it is legitimate and proper to use the current through R2 as the correct input current measurement. We have several opinions that it is not proper to do so. Alternate means of measuring the input current should be tried and qualified.

so,maybe it's time to power it from standalone circuit created for this purpose ? Then you can power it from a capacitor
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3270

Thanks PM,   Thanks TK,   

I will take another thorough look at my resistors inductance and measurement methods (like TK mentioning using a RF probe tip, which i do not have on my Tek probes, but do on the Owon probes)
Mind you, concerning the use of my current probe/amp combo, i also did the same power calculations with same results using the R2 current, but only at 190Khz, so i will use that further up in frequency too


Itsu
   
Group: Guest
TK

Just out of curiosity ,can you re-do you P/in P/out measurements ,with  R2 in the position of the circuit  below.


Brad

Brad, in one test I have running right now, I put a 1.00 ohm NI resistor in your location and a CH4 probe on the coil side of this resistor. I still have the 1.00 ohm in the original position too, on CH2. Unfortunately I seem to have hit another Rigol bug because I can't get a good Math CH1-CH4 differential reading, but using cursors I see that CH1 p-p is at 22.6 Vp-p and CH4 is at 22.2 Vp-p for a difference of about 0.4 V which translates to about 0.4 A p-p through the new resistor. I can't get more precise than that because of the cursor readouts and the high voltage to begin with. The current through the "old" resistor using CH2 at the same time is almost exactly the same at 394 mA p-p using cursors or 388 mA p-p using the scope's auto measurements. As near as I can tell the phase angle of the CH1->(CH1-CH4 difference) is also very close to the phase angle of the CH1->Ch2 measurement.
   

Group: Professor
Hero Member
*****

Posts: 2602
If no-one can point me to an obvious problem or error i make, i think it would be best that i step out of this thread
You've got to be kidding!

In scientific research, negative data is as much valuable as positive.

You've got plenty of things to try.  Such as:
- Using your TG + spectrum analyzer to sweep your resistors (and capacitors - if you use any).
- Putting a 1:1 ferrite isolation transformer on the output of your FG and measuring the input power this way.
- Calibrating the skew and amplitude difference of you current probe vs. CSRs. at various frequencies.
- Shielding /putting the device in a metal box (not the CSR though).
« Last Edit: 2017-05-03, 11:34:04 by verpies »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Brad, in one test I have running right now, I put a 1.00 ohm NI resistor in your location and a CH4 probe on the coil side of this resistor. I still have the 1.00 ohm in the original position too, on CH2. Unfortunately I seem to have hit another Rigol bug because I can't get a good Math CH1-CH4 differential reading, but using cursors I see that CH1 p-p is at 22.6 Vp-p and CH4 is at 22.2 Vp-p for a difference of about 0.4 V which translates to about 0.4 A p-p through the new resistor. I can't get more precise than that because of the cursor readouts and the high voltage to begin with. The current through the "old" resistor using CH2 at the same time is almost exactly the same at 394 mA p-p using cursors or 388 mA p-p using the scope's auto measurements. As near as I can tell the phase angle of the CH1->(CH1-CH4 difference) is also very close to the phase angle of the CH1->Ch2 measurement.

So you are seeing a 400mA difference at each end?
Im not sure where the original 1.00 ohm resistor was in your circuit.


What i was looking for,is there a difference in input current value,between having the CVR at the ground side of the circuit-as PM has it,and having it at the input end of the circuit-as MH said it should be.


Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
OK... Now I've checked the Math CH1-CH4 subtraction values against the input traces, and the Math value IS giving me the correct figures for the difference between the traces, just at a coarse resolution. 

So I get a RMS value of about 880 mA rms on the Math trace for the differential CH1-CH4 (current through R2b at TinMan's position) , against a CH2 measurement of 138 mA rms (current through R2a at original position) . Phase angle is about the same.

So if this is correct, this brings down the COP to well below 1. 
   
Group: Guest
So you are seeing a 400mA difference at each end?
Im not sure where the original 1.00 ohm resistor was in your circuit.


What i was looking for,is there a difference in input current value,between having the CVR at the ground side of the circuit-as PM has it,and having it at the input end of the circuit-as MH said it should be.


Brad

Yes, I think so, quite a bit if I did it right. (Edit: I'm not so sure now) See the previous post for the test scopeshots and Math for the differential CH1-CH4 voltage drop across the new R2 at the "MH" position. The test schematic is attached here:

« Last Edit: 2017-05-03, 12:55:12 by TinselKoala »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
OK... Now I've checked the Math CH1-CH4 subtraction values against the input traces, and the Math value IS giving me the correct figures for the difference between the traces, just at a coarse resolution. 

So I get a RMS value of about 880 mA rms on the Math trace for the differential CH1-CH4 (current through R2b at TinMan's position) , against a CH2 measurement of 138 mA rms (current through R2a at original position) . Phase angle is about the same.

So if this is correct, this brings down the COP to well below 1.

So now we have a situation,where the current flowing into the circuit,is more than the current flowing out of it !apparently! ???

How can this be ?,as there is no way in hell that the capacitive coupling between the open end of L1 and ground,could pass that much current--just no way  :o
Something funny going on here-some where.

There are other things being missed in my original circuit as well-video to come shortly.



---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
Actually, I think the problem is with the differential measurement across R2b. As you can see from my scopeshot above there is a tiny phase difference between the CH1 signal and the CH4 signal. This is what produces the phase shift and the voltage amplitude in the Math trace. I think the differential measurement across the CVR to give a voltage drop that can then be turned into a current... is only valid if there is _no_ phase shift between the signals at the two probes on either side of the resistor. I think.

SO here I've twiddled my scope to delay CH4 a bit so that the tiny phase shift between CH1 and CH4 is removed, see below. This brought the Math differential measurement CH1-CH4 down to reasonable values that are in the ballpark of the CH2 R2a measurement.  It also removed the phase shift from the Math trace as well. The 1.20 Vp-p reported is affected by the coarse resolution and should probably be more like 600 mV p-p or even less. Which brings it almost equal to the CH2 current from R2a. Is this cheating, or valid, or not? Let's hear from some more experienced scoposcopists on this one.

How to interpret all of this? Too far past my bedtime to be sure.    :D



   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
Actually, I think the problem is with the differential measurement across R2b. As you can see from my scopeshot above there is a tiny phase difference between the CH1 signal and the CH4 signal. This is what produces the phase shift and the voltage amplitude in the Math trace. I think the differential measurement across the CVR to give a voltage drop that can then be turned into a current... is only valid if there is _no_ phase shift between the signals at the two probes on either side of the resistor. I think.

SO here I've twiddled my scope to delay CH4 a bit so that the tiny phase shift between CH1 and CH4 is removed, see below. This brought the Math differential measurement CH1-CH4 down to reasonable values that are in the ballpark of the CH2 R2a measurement.  It also removed the phase shift from the Math trace as well. The 1.20 Vp-p reported is affected by the coarse resolution and should probably be more like 600 mV p-p or even less. Which brings it almost equal to the CH2 current from R2a. Is this cheating, or valid, or not? Let's hear from some more experienced scoposcopists on this one.

How to interpret all of this? Too far past my bedtime to be sure.    :D

The way i checked,was simply to swap the polarities of the FG leads across the circuit.
I also seen no difference at all.

Brad


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
OK,back to my original setup-circuit.

Can we actually count on the current value shown on R2 to be correct ?

https://www.youtube.com/watch?v=OyG3F_4-TLo


Brad


---------------------------
Never let your schooling get in the way of your education.
   
Group: Guest
This is a really good illustration of how tiny inductances can really screw up measurements. I had my two probes CH1 and CH4 on either side of the R2b resistor, but the CH1 probe was separated from the resistor by about 2 inches of wiring and connectors. SO I moved both probes to be right up against the body of the resistor where the lead wires go in, and I made sure the probe ground clips were symmetrical and to the exact same ground point, and I got a lot better result. Only needed -4 ns on CH4 to align phase, and the Math trace went way down. So I'm saying at this point that I see actually no significant difference in the current between the values read at the two resistors.

That is, if I am correct about the effect of phase of the differential measurement across R2b.
   
Pages: 1 2 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 2021-12-07, 21:29:47