PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2021-10-22, 10:23:14
News: If you have a suggestion or need for a new board title, please PM the Admins.
Please remember to keep topics and posts of the FE or casual nature. :)

Pages: 1 ... 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 155510 times)

Group: Administrator
Hero Member
*****

Posts: 3174
It's not as complicated as it may seem...
If you think it is really worthwhile, I'm happy to do it.

I'll PM you my address.

I think it is worthwhile to have alternate methods of determining Pin, and I think I can say with confidence this isn't the last circuit you are ever going to measure Pin.
   

Group: Administrator
Hero Member
*****

Posts: 3174
It's not as complicated as it may seem...
OK, here you go. I found a 1uF 10 percent poly film cap in my stash and measured it to be 0.997 uF.

Without coupling transformer, at 3.531 kHz (The F43 is hard to set precisely, being analog with a big knob), the phase shift was -71.49 degrees.
IN: Math average 371 mW, sine calculation 310 mW
OUT: 385 mW
You would think that Pin and Pout would agree closely in this case, so something is up with this measurement. Possibilities (not saying these are the case, just possibilities):

1) Equipment (scope or probes).
2) Measurement protocol.
3) CVR or Load inductance (should have negligible effect at 3.5kHz)
4) Erroneous assumptions.
5) Incorrect calculations.
6) Earth Power Vortex spot.
7) Telekinetic OU  :o
   
Group: Guest
You would think that Pin and Pout would agree closely in this case, so something is up with this measurement. Possibilities (not saying these are the case, just possibilities):

1) Equipment (scope or probes).
2) Measurement protocol.
3) CVR or Load inductance (should have negligible effect at 3.5kHz)
4) Erroneous assumptions.
5) Incorrect calculations.
6) Earth Power Vortex spot.
7) Telekinetic OU  :o

Well, 385 mW out using a sine assumption (Vrms2/R) pretty much equals 371 mW from the scope math , doesn't it ?  If one took a bunch of test samples and then averaged them together to get rid of random errors it might get even closer.

   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
sanity check here to,

1uF poly film cap (992nF) @ 3.5KHz, without 1:1 toroid

Input 221mW, 
Output 220mW

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

Posts: 928
Hi TK and All,

Perhaps doing a differential voltage measurement across a NI resistor connected (as a CSR) in series with the hot (red) output lead of the FG would be another method not yet tried?

I mean this differential voltage measurement across a series say 1 Ohm would reveal only the current taken out from the FG by the DUT.

The 'normal' measurements applied so far should be done after the measurement of the voltage drop across CSR i.e. the CSR could be left in its place and Pin and Pout may be obtained as before in the measurement setup shown so far.  The dissipation in the CSR would be known when CH1 shows the voltage level directly across the output of the FG during the 'normal' measurement setup.

I do not think the two tips of the probes would introduce a noticable change in the operation of the DUT during the input current measurement. All the ground clips of all the scope probes should be left floating during this measurement, we would be interested in the voltage drop across the series CSR resistor.

Introducing a series (say) 1 Ohm) resistor at the hot (red) output of the FG would not change power levels in the DUT too much.

What do you think?

Gyula
   

Group: Professor
Hero Member
*****

Posts: 2382
3) Compute Pin via Pin = Ptot - Pidle
The trouble is that with some amplifier topologies Pidle is not constant and decreases when of Pout increases, keeping Ptot almost constant :(
   

Group: Professor
Hero Member
*****

Posts: 2382
Perhaps there are a number of individuals here that might be willing and eager to design something up; Verpies?
We did something like that with Itsu recently using Elantec's EL2009C ?
There is even a video somewhere about it.
   

Group: Professor
Hero Member
*****

Posts: 2382
I have a 1.00 ohm NI resistor that I mounted to a scope probe cable with the probe cut off. I used this long ago at Steve Weir's urging when dealing with the Ainslie affair. SO I dug it out and installed it on the Capacitor Test circuit,
So the frequency response of the scope's input is not flat anymore :(
Did you see how a square current waveform looks like now ?

...and eliminated the other probe ground clips.
I did not suggest eliminating probe grounds - only the long clipleads.
   
Group: Guest
Hi TK and All,

Perhaps doing a differential voltage measurement across a NI resistor connected (as a CSR) in series with the hot (red) output lead of the FG would be another method not yet tried?

I mean this differential voltage measurement across a series say 1 Ohm would reveal only the current taken out from the FG by the DUT.

The 'normal' measurements applied so far should be done after the measurement of the voltage drop across CSR i.e. the CSR could be left in its place and Pin and Pout may be obtained as before in the measurement setup shown so far.  The dissipation in the CSR would be known when CH1 shows the voltage level directly across the output of the FG during the 'normal' measurement setup.

I do not think the two tips of the probes would introduce a noticable change in the operation of the DUT during the input current measurement. All the ground clips of all the scope probes should be left floating during this measurement, we would be interested in the voltage drop across the series CSR resistor.

Introducing a series (say) 1 Ohm) resistor at the hot (red) output of the FG would not change power levels in the DUT too much.

What do you think?

Gyula
Yes, I did try that, when we were trying to see if the current in that leg was truly the same as the current sensed by the low-side CVR. The problem is that the p-p voltage across that resistor is high enough so that the differential voltage obtained by subtracting one channel from the other is hard to resolve. In other words, it is hard to see a hundred milliamps difference between two signals that are in the 10-15 volt or even higher range. At least on my scope, the resulting trace was pretty much useless.
   
Group: Tech Wizard
Hero Member
*****

Posts: 928
Yes, I did try that, when we were trying to see if the current in that leg was truly the same as the current sensed by the low-side CVR. The problem is that the p-p voltage across that resistor is high enough so that the differential voltage obtained by subtracting one channel from the other is hard to resolve. In other words, it is hard to see a hundred milliamps difference between two signals that are in the 10-15 volt or even higher range. At least on my scope, the resulting trace was pretty much useless.

Okay, thanks. 

What if (as a sanity check) you would use your Fluke 87 true rms DMM to check the voltage drop directly across such CVR at 3.5 kHz?  (Such DMM may work up to 100 kHz, right?)

EDIT:  returning to the differential voltage measurement, suppose the value of the CVR is increased so that the voltage drop increases across it, thus the difference also increases to be able to evaluate it much better?  A higher CVR value surely reduces input power to the DUT but it would not matter much.

Gyula
« Last Edit: 2017-05-07, 21:45:18 by gyula »
   
Group: Guest
So the frequency response of the scope's input is not flat anymore :(
Did you see how a square current waveform looks like now ?
I did not suggest eliminating probe grounds - only the long clipleads.

Well, maybe.
   

Group: Administrator
Hero Member
*****

Posts: 3174
It's not as complicated as it may seem...
It is looking to me that the problem is with the CVR or its wiring back to common.

At 1.6882 MHz and using the 2.2n cap, you are reporting a phase difference of 82.7°. The theoretical phase difference is about 65°, and that makes all the difference in the world; cos(82.7)=0.127, cos(65)=0.422.

How could the phase be off that much? It turns out very easy. In my sim all that was needed for parasitic inductance between the CVR probe and common was 28nH, and the phase changed to 82.7°. Wire has an inductance of about 20nH per inch, so 28nH represents about 1.4 inches of connecting wire.

Interestingly, a parasitic inductance of 100nH in any other leg of the circuit has little effect on this phase angle. Why is that you ask? Well, at 1.6882MHz, the reactance of 28nH is a significant result (30%) relative to the value of the CVR;  XL=0.3R compared to the 1R CVR.

How can we overcome this problem? I would suggest using the CVR low-side as the common point for the circuit; that is, right as close to the CVR body as possible. If this is already being done, then it may be the CVR itself that is exhibiting the 28nH inductance. That is where I would be starting to look for the problem.
   

Group: Administrator
Hero Member
*****

Posts: 3174
It's not as complicated as it may seem...
We did something like that with Itsu recently using Elantec's EL2009C ?
There is even a video somewhere about it.
Excellent.

Is its idle bias current a concern as you pointed out?

Where did you guys find them? Nothing at Newark nor Digikey. There are a couple on ebay.

I'll see if I can find the thread. These would work great for the idea, and there is a macro-model I can play with to try it out.
   

Group: Professor
Hero Member
*****

Posts: 2382
Is its idle bias current a concern as you pointed out?
Yes, it drifts from an A-class to an AB-class and the bias current changes.

Where did you guys find them? Nothing at Newark nor Digikey. There are a couple on ebay.
Me thinks AliExpress  8)
We use  them to buffer the 50Ω outputs of our Signal Generators and get them down to 1Ω ...and several Watts.
   
Sr. Member
****

Posts: 291
Yes, I did try that, when we were trying to see if the current in that leg was truly the same as the current sensed by the low-side CVR. The problem is that the p-p voltage across that resistor is high enough so that the differential voltage obtained by subtracting one channel from the other is hard to resolve. In other words, it is hard to see a hundred milliamps difference between two signals that are in the 10-15 volt or even higher range. At least on my scope, the resulting trace was pretty much useless.

TK,

Have you tried using the FG's 50R as a CVR as I previously suggested?  Set and measure the FG's output while open circuit, then note the FG output's Vdrop when you connect the circuit.  That difference divided by the FG's 50R should give you a decent FG output current measurement.

Make the measurements right at the FG's output BNC.  Use of a BNC Tee might prove handy.

There are a few possible pitfalls with this method (assumes FG's output amp is Low Z/stiff, uses a NI output 50R, etc.), but this measurement should get you close. 

If necessary, opening the FG and grabbing a signal right at the FG's output amp prior to the 50R might be helpful.  I have an FG in which I repurposed a rarely used BNC and connected it directly to the output amp as both a measurement point and aux low Z output (there are some caveats associated with using it as an output).  Some FG's use a switched attenuator instead of a single 50R.  For those a schematic may be needed to locate the amp out.  Your low cost DDS probably just has a 50R at the output.

Calibrated GOW's?  Really?  DC versus AC with inductance, non-linearity, thermal response, etc.

PW 
   
Group: Guest
It is looking to me that the problem is with the CVR or its wiring back to common.

At 1.6882 MHz and using the 2.2n cap, you are reporting a phase difference of 82.7°. The theoretical phase difference is about 65°, and that makes all the difference in the world; cos(82.7)=0.127, cos(65)=0.422.

How could the phase be off that much? It turns out very easy. In my sim all that was needed for parasitic inductance between the CVR probe and common was 28nH, and the phase changed to 82.7°. Wire has an inductance of about 20nH per inch, so 28nH represents about 1.4 inches of connecting wire.

Interestingly, a parasitic inductance of 100nH in any other leg of the circuit has little effect on this phase angle. Why is that you ask? Well, at 1.6882MHz, the reactance of 28nH is a significant result (30%) relative to the value of the CVR;  XL=0.3R compared to the 1R CVR.

How can we overcome this problem? I would suggest using the CVR low-side as the common point for the circuit; that is, right as close to the CVR body as possible. If this is already being done, then it may be the CVR itself that is exhibiting the 28nH inductance. That is where I would be starting to look for the problem.

My CVRs are Ohmite WND1R0FE.

I presume the "FE" stands for..... well, you know.

Why is it a "problem"? Maybe OU requires a little bit of inductance. 

Or... just maybe.... all the reported COP>1 results are due to stray inductances that are impossible to eliminate entirely from a real construction.  I think we are accumulating some good evidence for that. Since the Cap Test circuit consistently produces COP>1 but that gets smaller and smaller as inductances are reduced and probing is made better, and the PBT appears to behave the same way... well, it is definitely true that similar effects (COP>1 in both Cap Test and PBT) don't necessarily have the same causes (stray inductance vs. "realni Overunity"), but often they do.

   
Group: Guest
TK,

Have you tried using the FG's 50R as a CVR as I previously suggested?  Set and measure the FG's output while open circuit, then note the FG output's Vdrop when you connect the circuit.  That difference divided by the FG's 50R should give you a decent FG output current measurement.

Make the measurements right at the FG's output BNC.  Use of a BNC Tee might prove handy.

There are a few possible pitfalls with this method (assumes FG's output amp is Low Z/stiff, uses a NI output 50R, etc.), but this measurement should get you close. 

If necessary, opening the FG and grabbing a signal right at the FG's output amp prior to the 50R might be helpful.  I have an FG in which I repurposed a rarely used BNC and connected it directly to the output amp as both a measurement point and aux low Z output (there are some caveats associated with using it as an output).  Some FG's use a switched attenuator instead of a single 50R.  For those a schematic may be needed to locate the amp out.  Your low cost DDS probably just has a 50R at the output.

The F43 has a stack of big carbon composition resistors. It does have an attenuation switch as well. I don't think I'm willing to open it up and mess with its innards, although I have repaired it before (blown output transistors). The "low cost" DDS has who knows what inside, it's all surface mount stuff and I doubt if its impedance is actually 50 ohms, it is just a toy really.

Quote

Calibrated GOW's?  Really?  DC versus AC with inductance, non-linearity, thermal response, etc.

PW

Well, how do you account for the fact that my nomograph works to give values that are in fairly close agreement with scope and DMM measurements? Coincidence?

I think that I have adequately reproduced the _apparent_ COP>1 effect in the original PBT circuit for which Partzman provided the schematic. I have also shown that Poynt99's capacitor test circuit provides the same sort of COP>1 measurements when tested in the same manner. And so have others working here in this forum and elsewhere.  I think it is no longer up to _ME_ to perform variation after variation in an attempt to track down the reason for this effect -- I believe this has also been adequately shown to be a result of high phase shift due to unresolvable inductance in the input power parameters in both the Cap Test and in the PBT. It requires a real stretch of the imagination to pretend that the COP>1 result in the Cap Test could be due to mere inductance but the COP>1 result in the PBT could be due to real OU.  Should someone want to verify that the effect in the PBT is due to some actual OU process or excess energy, they are going to have to demonstrate it in some other way than by using pretty coloured lines on an oscilloscope, or even its numbers in boxes. Using the bulb is an attempt to do this, but even though its leads have plenty of inductance the OU is very shy, it goes away and hides when the light is shining on it.
   
Sr. Member
****

Posts: 291
The F43 has a stack of big carbon composition resistors. It does have an attenuation switch as well. I don't think I'm willing to open it up and mess with its innards, although I have repaired it before (blown output transistors). The "low cost" DDS has who knows what inside, it's all surface mount stuff and I doubt if its impedance is actually 50 ohms, it is just a toy really.

Have you tried measuring the difference between the FG's open circuit voltage versus its connected to the circuit voltage right at the FG's BNC? (no need to go into the FG)

Quote
Well, how do you account for the fact that my nomograph works to give values that are in fairly close agreement with scope and DMM measurements? Coincidence?

I'd have a bit more confidence if you were using an AC freq similar to the test frequency..  But as non-linear versus temp as the GOW's are, well, you already know what my arguments are...

PW
   
Group: Experimentalist
Hero Member
*****

Posts: 1362
FWIW, here is a test protocol that works for my bench with my particular equipment but it does have room for improvement.  In an attempt to find the cause of the current phase and magnitude problem, I determined that a 10Meg 10x probe is susceptible to induction from the DUT therefore it's measurements are dependent on position in reference to the DUT.  So, to improve the induced signal to noise and lower the probe impedance, I used a Tek P6129B in the 1x mode for the CH2 measurements on R2.  A .1x 100K probe would be better but would have to be built as I do not see any available.  The 1x 1Meg probe seems to be unaffected by movement and position and gives consistent measurements over time in the low MHz range. 

Attached is the data from testing the circuit as shown and comparisons between measuring across the CSR or using a current probe.

These are the basic measurement conditions-

1)   A compensated 1x 1Meg probe is used for the measurement across the CSR R2.  The ground for this probe is the only connected ground from the scope to the DUT.

2)   The 1x probe is de-skewed with equal magnitude in reference to CH1 at the frequency of interest.

3)   A tightly coupled isolation transformer is to be used.

4)   Set all scope channels so the displayed waveforms are at the max value of the input dacs.  The digital Tek scopes are 25 bits digitization levels/div so 10.24 divisions of deflection can be utilized with the fine scaling feature.

I plan to use this protocol in re-testing my various MEI devices.
 
Pm

Edit:

 
« Last Edit: 2017-05-08, 04:13:24 by partzman »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194
Have you tried measuring the difference between the FG's open circuit voltage versus its connected to the circuit voltage right at the FG's BNC? (no need to go into the FG)

I'd have a bit more confidence if you were using an AC freq similar to the test frequency..  But as non-linear versus temp as the GOW's are, well, you already know what my arguments are...

PW


I did,  see here:    http://www.overunityresearch.com/index.php?topic=3445.msg61578#msg61578


Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4409


Buy me some coffee
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


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

Posts: 291

I did,  see here:    http://www.overunityresearch.com/index.php?topic=3445.msg61578#msg61578


Itsu

Itsu,

Sorry I missed that post.  Did you measure right at the FG?  What model is your FG?

Repeating the open/closed circuit measurement using just a resistor (50R, etc) right at the FG will confirm your FG's output Z (Vdrop versus loading).

PW
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3194

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
   
Group: Guest
OK, I worked on my PBT setup, making even shorter connections to the Load resistor stack, cleaning up the CH2 Rigol probe so it works well at 1x attenuation, verified that it does not pick up RF and is not sensitive to orientation, suspended probes so they come onto the circuit vertically from above, removed the ground clips except for the CH2 clip, ditched the coupling transformer, adjusted traces to use most of the vertical screen space with fine V/div adjustment, set the scope to Measure All between cursors positioned as in Partzman's shots above exactly across 4 full cycles of the Math waveform (my scope does not show the cursor positions with lines when doing this but they are set properly.)

And at a frequency that formerly gave me COP>1 I now get very close agreement with Input (math trace) and Output power. Also very close agreement with Current measured by Vdrop across R1 and Vdrop across R2. Phase shift between CH2 and CH3 is essentially zero.

f = 1.4076 MHz  phase angle = -73.82
IN: Math Average 353 mW
sine wave calculation 285 mW
OUT: 357 mW
Current through Load = 2.59/18.8 = 0.138 A
Current through CVR = 0.137 A

Next back to the Cap Test setup to see if things have "improved" there too.
   
Group: Guest
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

Not no one.
Is it possible that your pickup coil has a phase shift of its own?
   
Pages: 1 ... 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 2021-10-22, 10:23:14