PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2019-11-14, 21:00:02
News: If you have a suggestion or need for a new board title, please PM the Admins.

Pages: 1 ... 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 [21] 22 23
Author Topic: Graham Gunderson Energy conference High COP demonstration  (Read 154090 times)
Group: Experimentalist
Hero Member
*****

Posts: 935
Here is the result of a first attempt to place alternately switched mosfets in the secondaries of the previous PM3 circuit. In essence, there was little output (~100mv) across the load with the mosfets actively switched but there appeared to be a negative input power of ~ -.4 watts!

So, the mosfets and the output cap were removed, the secondaries connected in parallel with the load and the following are the results.

The schematic again shows the probe connections.

Scope pix PM3P_A shows the measurement results which are most interesting. Power out is .59 watts and power in is -.405 watts.

I did suspect that the secondaries would not perform in a buck mode as well as a U core configuration due to the larger leakage inductance between them owing to the additional center leg flux path of the E core, but I did not expect these results!

pm

EDIT: I apologize for posting the above results as I have discovered the current sense probe was reversed giving the negative input power. However, the COP has improved to ~ 1.45 with the paralleled secondaries.
   
Group: Guest
Where does the .59 W output power measurement come from?
   
Group: Experimentalist
Hero Member
*****

Posts: 935
Where does the .59 W output power measurement come from?

CH3(pnk) is the voltage across RL so pout = 3.843^2/25 = .59w.

pm
   
Group: Experimentalist
Hero Member
*****

Posts: 935
The following is a scope pix of the same pm biased transformer layout with series connected secondaries as previously posted with one exception, there is no clamp phase as the input is a periodic sine wave at 30kHz.

The reactive pin is .189w and the resistive output is .279w for an apparent real/reactive COP = 1.48.

pm
   
Group: Guest
CH3(pnk) is the voltage across RL so pout = 3.843^2/25 = .59w.

pm

I see. So the input power you are citing is the MEAN (or average) of the Math trace which is the multiplication of CH1 and CH2 (with offset). And the output power is the RMS voltage value across the load resistance, squared and divided by the load resistance.  Am I understanding correctly? Pavg=V2rms/R

Could you try it another way? That is, set up the experiment just as you have it above and get the Mean Input power from the Math trace like you have done. Then set up a different Math trace, with the pout equation done by the scope Math: Math trace is (CH3)2 / 25, then let the scope compute the Mean of that trace as the Mean Output Power.

This is kind of a check on the scope's math and the logic behind the measurement and calculation.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 3651


Buy me some coffee
The following is a scope pix of the same pm biased transformer layout with series connected secondaries as previously posted with one exception, there is no clamp phase as the input is a periodic sine wave at 30kHz.

The reactive pin is .189w and the resistive output is .279w for an apparent real/reactive COP = 1.48.

pm

I also had some interesting results with a home made transformer from two shaded pole motors-that was made to be used as a generator. Just thought i would stick a PM in the rotor hole to see what would happen.
The output went up to 6 times the value than without the PM in there.

No P/in measurements at this time,just looking at the effect of having the PM in there.
Video is in the -can magnets do work -thread.

Brad


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

Posts: 935
I see. So the input power you are citing is the MEAN (or average) of the Math trace which is the multiplication of CH1 and CH2 (with offset). And the output power is the RMS voltage value across the load resistance, squared and divided by the load resistance.  Am I understanding correctly? Pavg=V2rms/R

Could you try it another way? That is, set up the experiment just as you have it above and get the Mean Input power from the Math trace like you have done. Then set up a different Math trace, with the pout equation done by the scope Math: Math trace is (CH3)2 / 25, then let the scope compute the Mean of that trace as the Mean Output Power.

This is kind of a check on the scope's math and the logic behind the measurement and calculation.

The attached scope shot below was taken of my current test set and has a different output voltage level than the previous post but is sufficient to answer your questions.  This shot uses the math channel for the output power calculation displayed in both mean and rms for comparison to a manual calculation. IOW, pout = 2.607^2/25 = .27185w which equals the scope's mean calculation.  If one were deceptive, he could quote the rms math calc and improve the COP.

pm
   
Group: Experimentalist
Hero Member
*****

Posts: 935
I also had some interesting results with a home made transformer from two shaded pole motors-that was made to be used as a generator. Just thought i would stick a PM in the rotor hole to see what would happen.
The output went up to 6 times the value than without the PM in there.

No P/in measurements at this time,just looking at the effect of having the PM in there.
Video is in the -can magnets do work -thread.

Brad

Brad,

Thanks for the heads up but I am following that thread and have watched your video of the voltage gain with the PM insertion. I'm anxious to see your results on the power measurements.

pm
   
Group: Experimentalist
Hero Member
*****

Posts: 935
I am surprised that no one has asked for the power measurements off the PM3 core arrangement without any pm bias! So, below are scope shots of measurements without the pms and supporting ferrite parts. As can be seen, the pin = .361w and the pout = .353w for a COP = .98.

The PM3 device is very sensitive to the pm placement on the ferrite supports and any gaps. The next two pairs of scope pix A and B demonstrate this quite well and the measurements are taken from the same circuit as above only with the pms added and adjusted. I'll leave it to the reader to calculate the COPs.

Please note the CH1(yel) to CH2(blu) (input voltage and current) phase differences between each of the test pairs.

pm

 
   
Group: Elite Experimentalist
Hero Member
*****

Posts: 705
WOW partzman... PM3x B is excellent.

I would like to replicate your test device if you don't mind.
If that's okay with you, I would need some closer pics of your cores so I can see how you have it all put together. If you can provide sizes and core material details that would be great as well.

Thanks for sharing your excellent work

Luc
   
Group: Experimentalist
Hero Member
*****

Posts: 935
WOW partzman... PM3x B is excellent.

I would like to replicate your test device if you don't mind.
If that's okay with you, I would need some closer pics of your cores so I can see how you have it all put together. If you can provide sizes and core material details that would be great as well.

Thanks for sharing your excellent work

Luc

Luc,

Thanks. I will pm the info to you.

pm
   
Group: Professor
Hero Member
*****

Posts: 1275
You will be pleased to know that there is a good chance that Graham Gunderson will post some videos that he took before the conference as a form of taking notes and trying to describe the device.  As soon as he sorts out his password problem with our administrators.

Smudge
   
Group: Experimentalist
Hero Member
*****

Posts: 935
You will be pleased to know that there is a good chance that Graham Gunderson will post some videos that he took before the conference as a form of taking notes and trying to describe the device.  As soon as he sorts out his password problem with our administrators.

Smudge

I look forward to that!

pm
   
Group: Experimentalist
Hero Member
*****

Posts: 935
Here is a report on my latest findings on my PM3 device that was previously posted.

In an effort to determine the source of apparent COP sensitivity, I discovered that the positioning of the CH1 probe lead (measuring the input voltage) in relation to the transformer assembly would affect the input power calculation by changing the amplitude and phase of the voltage across the sense resistor. The PMs are removed from the transformer assembly for these tests.  Also, a shorted loop of stranded wire would have the same effect when positioned close to the core. This is with the CH2 probe setup with the coil ground connection and tip placed directly on the sense resistor leads which should provide the shortest, most direct and accurate connection possible. With the regular pull clip and ground lead used on CH2, the effect is still there but not as pronounced.

What is interesting is that a separate current probe/amp combo connected to CH4 of the same scope did not reflect the current changes seen in CH2 as described above.

The scope used for the above tests is a Tektronix MDO3034 with TPP0500B probes.

For comparison, I used a Tektronix TDS3034 with P6139A probes and the readings remained rock solid with no affect on the sense current in CH2 with the CH1 input lead placed on and near the transformer assembly.

I plan to contact Tektronix and discuss what possibly may be the problem. The moral of all this is that one can not assume the equipment being used is performing as expected!

I do have a PM3000 power analyzer I will test the device with as well for additional comparison.

pm 
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846
I was playing with a Full H-bridge in trying to copy the primary voltage signal as being presented in GG presentation, see picture 1 (yellow trace)

After some Full H-bridge setup attempts, i finaly got something working (thanks ION), see diagram below (using IRFP260N MOSFETs).

The input comes from my (opto-coupler isolated) FG which drives the verpies designed pulse motor drive system and so generate the needed 2 inputs (A and B).

The output of the full H-bridge can be seen in the screenshot below as the blue trace.
Via a "swinging" choke (7H) it drives the primary of a 1:6 transformer (220V : 1450V) with a parallel 5.6nF capacitor to get the primary into resonance.
The yellow trace shows this resonance signal and we can see that the flat line at the 30% zero part is not there.
So i think somehow Graham did flatten out (shorted) this somehow as in a resonance system it will be hard to get this flat line due to decaying signals.

Loading the secondary will kill the resonance and manipulating with ceramic magnets has no visual influence.

Video here:  https://www.youtube.com/watch?v=HlRo3ILTGko&feature=youtu.be   

Regards Itsu
« Last Edit: 2016-09-11, 00:36:33 by Itsu »
   

Group: Professor
Hero Member
*****

Posts: 1479
I was playing with a Full H-bridge in trying to copy the primary voltage signal as being presented in GG presentation, see picture 1 (yellow trace)
A worthy goal

After some Full H-bridge setup attempts, i finaly got something working (thanks ION), see diagram below (using IRFP260N MOSFETs).
The input comes from my (opto-coupler isolated) FG which drives the verpies designed pulse motor drive system and so generate the needed 2 inputs (A and B).
A full H-bridge is a tricky beast on the MOSFET/iso driver level, but to drive it you don't need the 4047 sequencing circuit that I designed for Tinman's motor, because you have a wonderful 2-ch FG that can do all that without any help.

The yellow trace shows this resonance signal and we can see that the flat line at the 30% zero part is not there.
That is because you are entering the "dead period" (the period when all MOSFETs are off) when there is still current flowing in the primary winding . 
If you enter this dead period when there is zero current flowing through the primary, then you will get a flat current line during this period.  As you already know, a freely oscillating LC circuit, crosses the zero-current line twice per cycle.
So, this is just a matter of matching the ½ period of the LC self-oscillation to the H-bridge's ON-period.
   
Group: Experimentalist
Hero Member
*****

Posts: 935
Itsu,

If I may suggest, try placing the "choke" outside the H bridge on the DC input side and then the H bridge provides the current polarity reversal to the parallel resonant primary and capacitor. Ideally, the choke should be 5-10 times the inductance of the transformer primary which would really be the leakage inductance of the primary with the secondary loaded.

During the "dead" time of the waveform, all four mosfets in the H bridge are turned on which clamps or freezes the current in the primary and then your voltage and current waveform on the input should more closely resemble Graham's scope pix.

pm
   

Group: Professor
Hero Member
*****

Posts: 1479
Ideally, the choke should be 5-10 times the inductance of the transformer primary which would really be the leakage inductance of the primary with the secondary loaded.
Yes, and if you'd like to let us know the value of this leakage inductance then measure the inductance of your primary with an LCR meter (or a scope) while your secondary winding is shorted.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846

Thanks verpies and partzman,

i did had the choke outside the LC, but then there is no resonance at all at the primary, just the square wave like signal as in the blue trace.

The primary measures 2.3H @ 1KHz and has a DC resistance of 3.2 Ohm.
Shortening the secondary drops the primary inductance to 2.3mH.

I will see if i can get a more flat line at the "dead periode" by "matching the ½ period of the LC self-oscillation to the H-bridge's ON-period".


Itsu
   
Group: Experimentalist
Hero Member
*****

Posts: 935
Thanks verpies and partzman,

i did had the choke outside the LC, but then there is no resonance at all at the primary, just the square wave like signal as in the blue trace.

The primary measures 2.3H @ 1KHz and has a DC resistance of 3.2 Ohm.
Shortening the secondary drops the primary inductance to 2.3mH.

I will see if i can get a more flat line at the "dead periode" by "matching the ½ period of the LC self-oscillation to the H-bridge's ON-period".


Itsu

Itsu,

OK, you can choose an operating frequency you desire and resonate with that 2.3mh primary leakage inductance to calculate the value of "C" needed. You will now need to use a relatively low value of load resistance say 20-100 ohms and connect a choke with a minimum inductance of ~10mh to the DC supply and the top of the H bridge. A larger choke value is fine as it will supply a more constant current with less ripple and less will also work but will have a larger current ripple as the value is decreased.

For the timing of your bridge mosfets, look at my post #458 which shows the schematic used and the phasing required to generate one complete sine wave at the resonant frequency of your LC plus the clamp phase which freezes the primary current and core flux. Using your dual output Rigol generator as Verpies suggested would allow you to generate the required signals and also vary the clamp phase timing while maintaining the required timing for the resonant period.

pm   
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
Hi Itsu

You might recall that GG claimed to use a swinging choke in his device. I see that you are using a wall power supply transformer for the inductor, which may not be the same. (see attachment).

You may be able to simulate a swinging choke with a properly step-gapped regular inductor. By properly, I mean the gap must be distributed and progressive i.e. a stepped gap or (see attachment) some type of controlled spacing between portions of the laminations rather than just a single slot at one location.

The attached paper from "Magnetics" is the best but here other links also:

http://electronics.stackexchange.com/questions/204107/swinging-chokes

http://www.vias.org/crowhurstba/crowhurst_basic_audio_vol3_025.html

https://library.e.abb.com/public/56a1ca38920d46ddc125746b0027f201/SwingingChoke-US-10.pdf

On another note:

In the last couple of days, I was  consulted on a custom design which used a 200 Volt 50 amp H bridge using the IR2184 driver chip. I became  fairly familiar with the part having worked with ITSU only a few days earlier and thus decided to model it in LTSpice. The customer was now and again blowing the FET's. At first we thought it was a gate capacitance problem causing "shoot through" or inadvertent short time conduction of the totem pole FETs. The FET's used had 28000pF of gate capacitance so we at first thought the drive resistors to the gates might be too high (22 ohms) causing the shoot through current, but after modelling the circuit and running some bench tests we found there was adequate delay between FET's. (fixed at 500ns)

The problem turned out to be a race condition between the chip power supply and the 250 volt power supply.
The chip power supply must be energized first or the outputs of the chip go high impedance and the gates can float. The solution was to insure the chip power supply is energized first (before the 250 volt supply) and to add some gate to source resistors for safety.
 
I have the LTSpice model for the IR2184 full H bridge drive if anyone needs it. (.asc file)

Regards, ION


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

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846


Partzman,

thanks for the info, i will load the secondary and then try to get the primary into resonance and move the choke to the DC side.
The problem with the Rigol FG as driver for the H-bridge is the fact that it is grounded, so my scope ground lead probes will cause problems.
Verpies his pulse motor driver acts like an isolated buffer.

With isolating the FG (or the scope) from ground i have bad experiences, so i try to avoid that.



ION,

thanks for the info on the swinging choke, i thought it was just a made up name by Graham or Spokane1, but it really exists.
I use the 220V side of a 220V/24V 50Hz transformer as choke.

I have LTSpice on my PC, so please could you send me the .asc file for the ir2184 H bridge, perhaps it will clear up some things in my mind.


Regards Itsu
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
Hi Itsu and partzman

I forgot to add that the IR2184 does not allow for all for transistors to be turned on at the same time as seems to be needed by the GG device for the "clamp phase". I don't know how you can get around that as it is a forbidden state.

Regarding the attached simulation, maybe partzman can take a look at it, because there is some kind of a problem that I can't figure out. I need to hit the escape key a few seconds after I energize the simulation, then it will escape the hangup and run ok.

 The program is crunching away doing something for a long time until the escape key is hit, then it runs fine. I added small series resistors in all the supplies to help with convergence but no luck. Maybe I need to add some small resistance in the capacitors also?

Regards

ION



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

Posts: 935
Hi Itsu and partzman

I forgot to add that the IR2184 does not allow for all for transistors to be turned on at the same time as seems to be needed by the GG device for the "clamp phase". I don't know how you can get around that as it is a forbidden state.

Regarding the attached simulation, maybe partzman can take a look at it, because there is some kind of a problem that I can't figure out. I need to hit the escape key a few seconds after I energize the simulation, then it will escape the hangup and run ok.

 The program is crunching away doing something for a long time until the escape key is hit, then it runs fine. I added small series resistors in all the supplies to help with convergence but no luck. Maybe I need to add some small resistance in the capacitors also?

Regards

ION

ION,

Would you post the IR2184 sub circuit.

Thanks,

pm
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
ION,

Would you post the IR2184 sub circuit.

Thanks,

pm

Hi pm

Besides the IR2184.sub, I have included a couple of their test models and the IR2184.asy.

Also there is a half bridge I built upon one of the circuits.

I'm new at this and am sure I have made many errors.
-----------------------------------------------------------------------------
For Itsu: I doubt that 220VAC wall power supply can mimick a swinging choke. You would have to test the inductance at several DC current levels, and I'm not yet sure how to do that. I'd bet that the wall transformer saturates sharply rather than softly.

Regards, ION

Edit: .asy not accepted by forum so changed it to .doc


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Pages: 1 ... 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 [21] 22 23
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2019-11-14, 21:00:02