PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2019-11-14, 22:18:03
News: If you have a suggestion or need for a new board title, please PM the Admins.

Pages: 1 2 3 4 5 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 154097 times)
Jr. Member
**

Posts: 79
To All,

Here is a possible simplified version per TK's initial review.

Graham said that only 25% of the circuit was being used. Apparently that is a correct statement.

Spokane1
   
Jr. Member
**

Posts: 79
Dear TK,

Perhaps this is a good time to review Reiyuki's timing diagram.

The attached photo is from his note book. These are timing notes that Graham wrote into his book on Sunday July 10, 2016 during lunch.

Gotoluc's outstanding  work is done by using an oscillator and adjusting the frequency to get the desired input response.  Graham's system is a little more complex than that. You can see how two timing pulses are processed through the logic to get that backend control pulse coordinated. This is what I want to figure out at this point.

Dose the logic we have so far appear to function in this way? I haven't dug into it yet.

Spokane1
   
Group: Guest
Dear TK,

Perhaps this is a good time to review Reiyuki's timing diagram.

The attached photo is from his note book. These are timing notes that Graham wrote into his book on Sunday July 10, 2016 during lunch.

Gotoluc's outstanding  work is done by using an oscillator and adjusting the frequency to get the desired input response.  Graham's system is a little more complex than that. You can see how two timing pulses are processed through the logic to get that backend control pulse coordinated. This is what I want to figure out at this point.

Dose the logic we have so far appear to function in this way? I haven't dug into it yet.

Spokane1
Yes, I think so. The line going from U8 to U5 triggers one pulse generator in U5 which triggers the other one, and this is synchronized with the signal to the H-bridge from U8 which is "processing" the pulse signal from U7. What I still don't understand is how this system is synched to the 556 clock system which is also sent to the H-bridge.

It turns out that my local supplier allegedly has some 74HC123 (or maybe HCT versions) in stock so I'm off to try to pick up 3 of them to play around with. We shall see what we shall see.

   
Jr. Member
**

Posts: 79
It turns out that my local supplier allegedly has some 74HC123 (or maybe HCT versions) in stock so I'm off to try to pick up 3 of them to play around with. We shall see what we shall see.

Dear TK,

AT least you live some place where you have a supplier. such is not the case in Spokane with a population of about 400,000. I will have to put in an order to DigiKey or Mouser and then wait a week.

Keep in mind that all my capacitor values on those drawings are WAG's.

Again thanks for all the footwork you are doing on this.

Spokane1
   
Group: Guest
OK... after a rather frustrating all-night session I have finally gotten some results from the U5-U7-U8 circuit.

The 74HC123s do not oscillate on their own. Apparently there must be an input from an external clock for the system to work. I supplied a clock signal from one of my function generators to Pin 10 of U7. I could not get it to work with a signal from the 556 circuit that I tested earlier. The result is shown in the screenshot below.

The CH1 Yellow trace is the Pin 13 output of U5 which goes to the Synchronous Diode. The CH2 Light Blue and CH3 Purple (Cyan and Magenta?) traces are the complimentary outputs of the 7414 U8 which go to the H-bridge. The CH4 Dark Blue trace is the input from my FG to Pin 10 of U7.

The two pots on U5 control the Yellow trace pulse width and pulse delay. One of the pots on U7 controls the pulse width of the two complimentary outputs of U8 to the H-bridge. I couldn't see any effect of the other pot on U7 but that may be due to where I'm injecting the clock signal. Too burned out at the moment to track that one down.

I'm not completely sure if the U7 chip is working on both sides. I'll try replacing it with a fresh chip to see if anything changes.

There may be a mistake on Spokane1's schematic because 1) there apparently needs to be a clock input to the 74hc123 system for it to work at all,  and 2) otherwise there is no synchronization between the 556 signal to the H-bridge and the signals from the 74HC123 system to the H-bridge and the synchronous diode.

Also the data sheet for the 74HC123 chip says that the pins 6 and 14 of the chip should be hard-wired to ground (in addition to being connected to the timing components) for noise rejection. This does not appear to have been done on the Gunderson board, as far as I can tell, but I did it on the U7 chip on my breadboard (but not the U5 chip.) Go figure.
   
Newbie
*

Posts: 29
OK... after a rather frustrating all-night session I have finally gotten some results from the U5-U7-U8 circuit.

The 74HC123s do not oscillate on their own. Apparently there must be an input from an external clock for the system to work. I supplied a clock signal from one of my function generators to Pin 10 of U7. I could not get it to work with a signal from the 556 circuit that I tested earlier. The result is shown in the screenshot below.

The CH1 Yellow trace is the Pin 13 output of U5 which goes to the Synchronous Diode. The CH2 Light Blue and CH3 Purple (Cyan and Magenta?) traces are the complimentary outputs of the 7414 U8 which go to the H-bridge. The CH4 Dark Blue trace is the input from my FG to Pin 10 of U7.

The two pots on U5 control the Yellow trace pulse width and pulse delay. One of the pots on U7 controls the pulse width of the two complimentary outputs of U8 to the H-bridge. I couldn't see any effect of the other pot on U7 but that may be due to where I'm injecting the clock signal. Too burned out at the moment to track that one down.

I'm not completely sure if the U7 chip is working on both sides. I'll try replacing it with a fresh chip to see if anything changes.

There may be a mistake on Spokane1's schematic because 1) there apparently needs to be a clock input to the 74hc123 system for it to work at all,  and 2) otherwise there is no synchronization between the 556 signal to the H-bridge and the signals from the 74HC123 system to the H-bridge and the synchronous diode.

Also the data sheet for the 74HC123 chip says that the pins 6 and 14 of the chip should be hard-wired to ground (in addition to being connected to the timing components) for noise rejection. This does not appear to have been done on the Gunderson board, as far as I can tell, but I did it on the U7 chip on my breadboard (but not the U5 chip.) Go figure.

Morning TK,

Try this circuit:

   
Group: Guest
Thanks! That actually works!

Now I don't need the external FG clock, and the previously  non-functional pot now controls the frequency. Now we just need to play with the capacitor and resistor values in the various timers to get the right duty cycle and frequency range. I already changed the capacitor in the frequency timer section to slow it down to around 70 kHz minimum but should probably add even more capacitance to get down to the 30 kHz range.

The Dark Blue trace in the shot below is now the Frequency output at Pin 12 of the rewired U7 chip.  But does it actually correspond to what's visible on Gunderson's board?

And what about the synchronization problem? How to synch with the signal from the 556 that goes to both the H-bridge and the synchronous diode?

Anyhow, time for me to take the rest of the day off and get some sleep.

Thanks again, nice work K4zep !!    O0
   
Jr. Member
**

Posts: 79
Dear TK,

I don't know how you can carry on with this kind of work with out some sleep. Maybe things will be better for you this weekend. My wife also suffers from a sleep disorder. She lost her job at the VA because of  it.

Here is my first pass at a logic diagram.

I was wondering if the "System Clock" U3 has anything to do with the actual timing of the power MOSFETS?  It might be the master clock for the switch mode power supplies on the H-Bridge and the Backend Diode. This would allow that frequency to be adjusted to avoid harmonic interference between the main logic frequency and/or the instrumentation.

It appears that U7, U8 and U5 alone can supply the basic timing signals.

Have a good weekend.

Spokane1
   
Group: Guest
Some further adjustment of cap values allows me to tune to the 50Khz range with considerable margin on either side, and to get exactly 50 percent duty cycle, again with a nice range. And the short Yellow pulse can be adjusted from very narrow to about half the width of the main pulse and can be positioned anywhere on a nice range from before to after rising edge of the main pulse.
   
Group: Guest
Dear TK,

I don't know how you can carry on with this kind of work with out some sleep. Maybe things will be better for you this weekend. My wife also suffers from a sleep disorder. She lost her job at the VA because of  it.

Here is my first pass at a logic diagram.

I was wondering if the "System Clock" U3 has anything to do with the actual timing of the power MOSFETS?  It might be the master clock for the switch mode power supplies on the H-Bridge and the Backend Diode. This would allow that frequency to be adjusted to avoid harmonic interference between the main logic frequency and/or the instrumentation.

It appears that U7, U8 and U5 alone can supply the basic timing signals.

Have a good weekend.

Spokane1

Looks great, except that I've rewired the U7 according to the information from K4zep so that it does not need an external clock any more.  I don't know if this corresponds to Gunderson's board version, I'll leave that part up to you!

Don't worry about my sleep cycle, it's partially an adaptation to the heat down here in Northeast Aztlan. Unseasonably cool this morning at 10 am, it's "only" 84 degrees F. with relative humidity about 80 percent. So I tend to sleep during the hottest part of the day and am more active evening-night.
   
Group: Guest
Visualizing the timing delay introduced by running the basic clock signal through multiple gates in the 74C14 hex Schmitt trigger inverter:

Again, the Blue trace is the raw pulse signal generated by the oscillating 74HC123. The Magenta and Cyan traces are the result of running this signal through 2 and 3 gates in the 74C14.

   
Jr. Member
**

Posts: 79
Dear TK,

You need to take a break and go have a beer, soda, or protein smoothie.

I really appreciate all the work you are doing on this circuit. With your assistance we are already at least couple of weeks ahead in this investigation.

I have a couple of old 74XX123's that I shall attempt to breadboard this weekend, depending upon how many honey-do's I have on the weekly list.

Spokane1
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846

Finally i was able to replicate GotoLuc's scopeshot of his version of the GG scope signals using a very simple circuit, see screenshot 1

It uses 2 MOSFETs (IRF530) and an isolated FG (to avoid ground lead shorts) set at BELOW the LC resonance frequency (LC resonance is at 24.5KHz, to get this signal i had to go to 20.7KHz)
L=10mH, C=3.5nF, csr=50 Ohm  (using a 1 Ohm csr and/or my current probe results in very fuzze (weak) signals).
See picture of diagram.

Input signal from FG is 7Vpp square wave at 33% duty cycle, see purple trace in screenshot 2.

Regards Itsu
   
Newbie
*

Posts: 29
Visualizing the timing delay introduced by running the basic clock signal through multiple gates in the 74C14 hex Schmitt trigger inverter:

Again, the Blue trace is the raw pulse signal generated by the oscillating 74HC123. The Magenta and Cyan traces are the result of running this signal through 2 and 3 gates in the 74C14.

Scope traces says it all, shows gate delays, etc!  Damn good work TK!

Ben K4ZEP
   
Group: Guest
@Itsu: Nice work! You probably already realize this but if you move the common Scope Probe ground point to the top of the 50R resistor you can avoid including it in the Vdrop across the coil-capacitor tank. This works if your PSU has isolated negative lead (as usual)  and your FG is also isolated. Of course you will have to invert the CH2 signal in the scope to get the correct current direction.     O0

Two questions: What happens if the mosfets are reversed so that the Drains are connected together, instead of the Sources? And what happens if you wind a secondary on top of the primary windings of the 10mH coil, can you light up a light bulb from the secondary?
 

@K4zep: Thanks! Gate propagation delays are often neglected or forgotten about but when driving high-power H-bridge they can cause problems -- or sometimes even solve problems.
It's important to note that I'm using a 74C14N because that's all I have on hand, but the Gunderson breadboard appears to use a 74AC14PC which may be faster. I'll have to see if I can find one at my supplier to see if there is any significant difference in gate speed.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846


I was kind of struggling with the current signal as it looks too ragged to me compared to GG and Gotoluc's current signals, so it kept it as straight forward as possible.
But i can try to change the common scope probe ground point to where you showed.

Changing the MOSFETs to a common drain configuration, is a problem as the FG should be connected to the both gates and both sources (i think), but i can try.

The present 10mH coil is a small ferrit U-cores choke with little room to add a secondary (its just what came close to Luc's 10mH coil), so i have to wind/use another
transformer like toroid to check.

Itsu
   
Newbie
*

Posts: 29
@Itsu: Nice work! You probably already realize this but if you move the common Scope Probe ground point to the top of the 50R resistor you can avoid including it in the Vdrop across the coil-capacitor tank. This works if your PSU has isolated negative lead (as usual)  and your FG is also isolated. Of course you will have to invert the CH2 signal in the scope to get the correct current direction.     O0

Two questions: What happens if the mosfets are reversed so that the Drains are connected together, instead of the Sources? And what happens if you wind a secondary on top of the primary windings of the 10mH coil, can you light up a light bulb from the secondary?
 

@K4zep: Thanks! Gate propagation delays are often neglected or forgotten about but when driving high-power H-bridge they can cause problems -- or sometimes even solve problems.
It's important to note that I'm using a 74C14N because that's all I have on hand, but the Gunderson breadboard appears to use a 74AC14PC which may be faster. I'll have to see if I can find one at my supplier to see if there is any significant difference in gate speed.

Morning TK,

The 74AC is about 10 times faster than the 74C......At 5VDC, Into a 50pf load, the AC has a min/mas rise time of 3-10ns with typical 8ns.  The C is much slower.  At 5VDC, it is an average of 220ns.  74AC is dam better chip if speed is the need!

Ben K4ZEP

   
Group: Guest

I was kind of struggling with the current signal as it looks too ragged to me compared to GG and Gotoluc's current signals, so it kept it as straight forward as possible.
But i can try to change the common scope probe ground point to where you showed.

Changing the MOSFETs to a common drain configuration, is a problem as the FG should be connected to the both gates and both sources (i think), but i can try.

The present 10mH coil is a small ferrit U-cores choke with little room to add a secondary (its just what came close to Luc's 10mH coil), so i have to wind/use another
transformer like toroid to check.

Itsu

Oh, yes, I forgot about the FG's connection to the sources. So it probably won't be possible to use a common Drain configuration. I was thinking how cool it would be to just strap the mosfets together back-to-back with a bolt. Oh well, nothing wrong with common-Source.

I used a 10 mH choke that I pulled from an old CRT TV chassis, with an extra winding over the outside of it for my favourite Joule Thief:
   
Group: Guest
Morning TK,

The 74AC is about 10 times faster than the 74C......At 5VDC, Into a 50pf load, the AC has a min/mas rise time of 3-10ns with typical 8ns.  The C is much slower.  At 5VDC, it is an average of 220ns.  74AC is dam better chip if speed is the need!

Ben K4ZEP

Well, thanks for looking that up! I need to try to get my hands on a few 74AC14 chips to play with then. But it does appear that I'm getting a bit better risetime from my C version, perhaps slightly less than 100 ns as shown on the scopetrace above. Of course the only load is the scope probe and input, so somewhat less than 50 pF.
But I'm actually also interested in the gate delay rather than just the risetime.
Unfortunately my supplier isn't open on weekends so I won't even be able to find out if he has them in stock until Monday. I guess I should have asked him when I was there on Thursday. Oh well.
The risetime of the pulses from the 74HC123 is impressive. I haven't measured it yet but even at 200 ns/div the rising edges are almost vertical. (Dark Blue and Yellow traces)


By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.
   
Group: Ambassador
Hero Member
*****

Posts: 3256
You've wandered into the Deep end of the pool....over most heads...mine included .

respectfully

Chet K
   
Newbie
*

Posts: 29
Well, thanks for looking that up! I need to try to get my hands on a few 74AC14 chips to play with then. But it does appear that I'm getting a bit better risetime from my C version, perhaps slightly less than 100 ns as shown on the scopetrace above. Of course the only load is the scope probe and input, so somewhat less than 50 pF.
But I'm actually also interested in the gate delay rather than just the risetime.
Unfortunately my supplier isn't open on weekends so I won't even be able to find out if he has them in stock until Monday. I guess I should have asked him when I was there on Thursday. Oh well.
The risetime of the pulses from the 74HC123 is impressive. I haven't measured it yet but even at 200 ns/div the rising edges are almost vertical. (Dark Blue and Yellow traces)


By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.
Hi TK and all,
If you have time, go to:

http://teslaenergysolutionsllc.com/our-research/coil-shorting-experiments/

Using those H bridge switchers.  Switches 100Khz effortlessly!

Other good stuff there too! 

It looks like using GG's logic circuits and those output devices, home free on the input side.

Now if we can find some high speed super high efficiency Rectifiers........
And of course figure out the "core's"

Ben K4ZEP
   

Group: Professor
Hero Member
*****

Posts: 1479
By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.
I care ...but the electronic implementation is not so interesting to me.  I can think of several different ways to achieve such input waveforms.  I think the magic smoke is in the magnetics (a la Smudge's thoughts) or faulty measurements. In any case don't let it out.
If the output waveform contained high and narrow spikes, then I would suspect McFreey's Modus Operandi and if it did not - I would expect CARA type MO. 
Without accurate schematics, nice scopeshots, probe positions and proof of calibration, I suspect the ORBO MO ;)

P.S.
If immobilizing the core halves to prevent their movement/vibration, makes the device transition form OU operation to UU operation, then I would lean towards the CARA type of MO.
   
Group: Elite
Hero Member
******

Posts: 3637
It's turtles all the way down
Quote from TK

Quote
By the way, I'm starting to think that the "usual suspects" really don't care much for Gunderson's device. Nobody but us chickens are commenting much on the threads on OU and EF.

Doubt that I am a usual suspect, but for now I am reserving (and have been asked to reserve) my opinions.

I may break silence in the future, after the film debut.

BTW,
 We  used the 74xx123 dual monostable back in the early 80's as a watchdog timer for our first 6500 based uP product.

On that note, I might add that the use of TTL logic is oftentimes not subject to the lockup conditions that can occur  in a  uP based design when it loses it's stack pointer and winds up in a tight loop, no longer servicing other routines.

Hence the need to "feed or kick the dog".


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

Group: Administrator
Hero Member
*****

Posts: 2905
It's not as complicated as it may seem...
IMHO there isn't anything further to comment on if and until the exact measurement protocol is illustrated. And following that (if it indeed is shown in sufficient detail) no doubt there will be several questions that most likely won't be answered.

In the mean time I would suggest that folks study up on real power vs. imaginary power, and what constitutes an input power source vs. something that is part of the device. Getting this straightened out would be an important technicality to consider before such an exercise took place.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1846
I created a 100uH secondary over the 10mH (5 Ohm) now primary.
I made the common Scope Probe ground point to be the top of the 50R resistor and inverted the ch2 (current) channel.

Without a load the voltage / current signal looks like as in screenshot 1

With a 6V / 0.6W incandescent light bulb (glowing dimly) as load to the secondary these signals become as in screenshot 2 (vertical settings changed!)
I am unable to obtain similar signals as in screenshot 1 with this load.

Measuring the load shows the signals as in screenshot 3.
Yellow the voltage across the bulb, green the current throught the bulb, red the math trace yellow x green.
Be aware that the current probe controller was set to 100mA/div. so the reported current and power needs to be taken times 10

The diagram is the updated circuit


Itsu
   
Pages: 1 2 3 4 5 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, 22:18:03