PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
OverUnity Research > Benches > AlienGrey > Dally, Shark & Ruslan workbench
2024-11-26, 06:33:44
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 ... 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 ... 101
Author Topic: Dally, Shark & Ruslan workbench  (Read 309317 times)
Topic Options img
img  NickZ
Re: Dally, Shark & Ruslan workbench « Reply #575 on: 2021-05-23, 14:02:53 »
Group: Guest
   Yet, it's the same thing that I've experienced with my very similar Stalker circuit. The push pull fets gets hot, burning hot, the grenade gets hot, my output cap(s) on the grenade output get hot, and my simple Kacher gets hot. And fans don't do enough. So, there is something wrong there.
   It does help to adjust the duty cycle, and frequency. But, the better the resonance, the higher the heat produced. So, it's a great heater, at least.
   Below is my circuit. Plus, due to the heat issues, I had to make another additional snubber circuit board to deal with the heat on the fets. But it doesn't really help that much. And I don't like to have to hear a bunch of fans going all the time.
   BTW: I don't use the big choke on the output, like is shown in your diagram. As I think that later schematic don't show it on there. Maybe it's not needed? But, I can make one, if really needed.  And no, both fets are not on at the same time. I know about that...
   Itsu, wait until you put a few hundred watts worth of bulbs on that output, plus the feed back circuit, on top of that. Toasty... 

   NickZ
« Last Edit: 2021-05-23, 15:31:09 by NickZ »
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #576 on: 2021-05-23, 16:15:57 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: NickZ on 2021-05-23, 14:02:53
   Yet, it's the same thing that I've experienced with my very similar Stalker circuit. The push pull fets gets hot, burning hot, the grenade gets hot, my output cap(s) on the grenade output get hot, and my simple Kacher gets hot. And fans don't do enough. So, there is something wrong there.
Yes, there is something wrong.  These components should not get that hot at kHz frequencies.

Most likely the ferrite has different characteristics compared to the author's ferrite. 
Low permeability or gapping being the first one that comes to mind and that is supported by the low inductance (34μH) measurement compared to author's inductance (110μH).

...badly driven transistors is another possibility, but that is less likely when using a PCB and predetermined components.
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #577 on: 2021-05-23, 16:25:56 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: Vasik041 on 2021-05-23, 10:42:23
It could be that you ferrite has significantly smaller permeability, so more turns and smaller gap needed.
I also think that his ferrite has a different permeability.

I support manipulating the air gap but I do not support adding more primary turns, because that just changes too many parameters, e.g. field geometry, LLEAK, CIT, saturation, which might be significant when replicating a device based on an unknown operating principle.
   
img  NickZ
Re: Dally, Shark & Ruslan workbench « Reply #578 on: 2021-05-23, 16:53:47 »
Group: Guest
Quote from: verpies on 2021-05-23, 16:25:56
I also think that his ferrite has a different permeability.

I support manipulating the air gap but I do not support adding more primary turns, because that just changes too many parameters, e.g. field geometry, LLEAK, CIT, saturation, which might be significant when replicating a device based on an unknown operating principle.


   All I know about the yoke is that it comes out of an old 32 Tv.  I've heard that they are made to run at around 15kHz. So, somewhat close to the inductor or grenade frequencies. But, the actual permeability is unknown. I've tried a couple of different yokes, both with same heat issue.
I use a different number of turns, 12t and 12t for yoke primaries, and 28t for the secondary, along with the 3 turns coil to the inductor circuit, as in the Ruslan schematic, TopRuslan 7.
 These are the big 3in plus size yokes. Without the fets and inductor connected up, the push pull runs cool, of course, and the scope signals looks like they should, normal. So, the push pull is working ok, by itself, and the fets stay cold, duty cycle, and frequency signals are normal. But, at resonance things can get hot, except when they don't, which also happens at times.

   I don't mean to distract from itsu's tests. But, the heat issue is a very important one to solve, I would think.

   NickZ
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #579 on: 2021-05-23, 16:58:22 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: NickZ on 2021-05-23, 14:02:53
..., I had to make another additional snubber circuit board to deal with the heat on the fets. But it doesn't really help that much. And I don't like to have to hear a bunch of fans going all the time
Well, if you were willing to depart from the original circuit that far, why not go one step further and wind the primaries with a double wire cable, like this:



This is a lot better than your snubber (so you can ditch it*), it keeps the transistors cool and recycles the spikes back into the power supply - see this video:
https://youtu.be/RU15rQEV-Rg


* ...but keep your gate driving circuitry.
« Last Edit: 2021-05-29, 01:38:16 by verpies »
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #580 on: 2021-05-23, 17:14:30 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159

I opened up the both secondaries, used 24V on the logic and a separate 24V on the MOSFETs (return leads together), see picture.

Measurements made are both gate signals (yellow and blue), drain signal (purple) and drain current (green), see screenshot.

Yellow, purple and green are on the same MOSFET.

Now the MOSFETs do not get hot at all, voltage on drains was 24V @ 680mA according to the PS, Duty cycle on MOSFETs 30%.

My current probe (green trace) was set at its lowest value (1mA/div), but this one has its DC current part defective, so be aware.

Itsu
   
img  NickZ
Re: Dally, Shark & Ruslan workbench « Reply #581 on: 2021-05-23, 17:14:58 »
Group: Guest
   Verpies:
   Thanks, but I would like to try to stay on the same page for now, and follow what itsu comes up with, while dealing with Stalker circuits, and with the heat issue.  If he gets it working right, (nice and cool) while lighting a bunch of high wattage bulbs, I'll be right behind him... As my push pull/inductor circuit, and grenade are not very different from his, up to this point.

   NickZ
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #582 on: 2021-05-23, 18:02:23 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: Itsu on 2021-05-23, 17:14:30
Measurements made are both gate signals (yellow and blue),
The main gate pulses are not overlapping, so the driver is not driving them at the same time, but there are 4V oscillations on gates, which could turn on the MOSFET when the driver is not driving it.

Quote from: Itsu on 2021-05-23, 17:14:30
...drain signal (purple) and drain current (green)
The drain voltage is ringing a lot after the falling edge of the gate pulse. Is that intended or are the clamping diodes not doing their job ?
Drain oscillations which start immediately after the falling edges of the yellow or blue gate pulses are understandable, but oscillations which I circled in blue, are not like that.
The drain current waveform is such low resolution that I cannot see much in it. Therefore, I can't see the signs of ferrite saturation.

I am concerned about the large loop area of the leads coming from the yoke primaries, which must be responsible for a lot of the leakage inductance, which causes oscillations/ringing.

Quote from: Itsu on 2021-05-23, 17:14:30
I opened up the both secondaries, used 24V on the logic and a separate 24V on the MOSFETs (return leads together), see picture.
I don't see the wire I marked in red on this picture. Is it under the white board?
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #583 on: 2021-05-23, 19:48:33 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159

Well,  i have 15V on the MOSFET drivers, so the 18V zeners should be adequate for clamping the gates to max. 18V.
And in this situation the MOSFETs are not heating up.

The red marked wire is indeed inside the yoke and going below the white board to the 24V filtering caps, see pictures.


Itsu
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #584 on: 2021-05-23, 19:59:35 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: Itsu on 2021-05-23, 19:48:33
Well, I have 15V on the MOSFET drivers, so the 18V zeners should be adequate for clamping the gates to max. 18V.
I was asking about the drain clamping.
The 18V Zeners are fine for preventing gate oxide damage, but they cannot prevent the MOSFET from being turned on by lower amplitude gate oscillations which are still greater than the VGS_th
   
img  Vasik041
Re: Dally, Shark & Ruslan workbench « Reply #585 on: 2021-05-23, 20:21:40 »
Group: Guest
Quote from: Itsu on 2021-05-23, 17:14:30
Now the MOSFETs do not get hot at all, voltage on drains was 24V @ 680mA according to the PS, Duty cycle on MOSFETs 30%.

Itsu,

I think you can try now connect secondary windings. May be tune one first, than another.
You can reduce duty cycle during tuning time to avoid overheating or power output stage with lower voltage like 10v for beginning. If you get too high voltage on inductor/gradient coil you will need unwind few turns from yoke secondaries.

Vasik
   
img  Vasik041
Re: Dally, Shark & Ruslan workbench « Reply #586 on: 2021-05-23, 20:25:19 »
Group: Guest
Quote from: Itsu on 2021-05-23, 19:48:33
The red marked wire is indeed inside the yoke and going below the white board to the 24V filtering caps, see pictures.

That is probably not very significant now, but the proper place for these caps is close to the PCB socket,
so the length of wires (ground, +24v) would be minimal.

Vasik

PS there are already caps on PCB, most probably you don't need extra caps.
   
img  picowatt
Re: Dally, Shark & Ruslan workbench « Reply #587 on: 2021-05-23, 20:33:46 »
Sr. Member
****

Posts: 453
Quote from: Itsu on 2021-05-23, 17:14:30
I opened up the both secondaries, used 24V on the logic and a separate 24V on the MOSFETs (return leads together), see picture.

Measurements made are both gate signals (yellow and blue), drain signal (purple) and drain current (green), see screenshot.

Yellow, purple and green are on the same MOSFET.

Now the MOSFETs do not get hot at all, voltage on drains was 24V @ 680mA according to the PS, Duty cycle on MOSFETs 30%.

My current probe (green trace) was set at its lowest value (1mA/div), but this one has its DC current part defective, so be aware.

Itsu

Itsu,

I am confused.  The purple trace (a MOSFET drain) shows 48vdc when the MOSFET is off.  I thought you said you were running at 24vdc.

Also, the purple trace is pulled low (MOSFET apparently conducting) during a time period when neither gate is high (between the blue gate falling edge and the yellow gate rising edge).

What am I missing?

PW 
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #588 on: 2021-05-23, 20:34:39 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159
Quote from: verpies on 2021-05-23, 19:59:35
I was asking about the drain clamping.
The 18V Zeners are fine for preventing gate oxide damage, but they cannot prevent the MOSFET from being turned on by lower amplitude gate oscillations which are still greater than the VGS_th

Ok,  so you mean the MUR460 clamping diodes.

Indeed, the zeners will not prevent any gate turnon due to <18V but higher then the Gate Source threshold signals.


Itsu
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #589 on: 2021-05-23, 20:50:29 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159
Quote from: Vasik041 on 2021-05-23, 20:21:40
Itsu,

I think you can try now connect secondary windings. May be tune one first, than another.
You can reduce duty cycle during tuning time to avoid overheating or power output stage with lower voltage like 10v for beginning. If you get too high voltage on inductor/gradient coil you will need unwind few turns from yoke secondaries.

Vasik

Vasik,


i connected the 3 turn secondary and things kept stable (24V @ 700mA input now).
But i noticed 1 MOSFET getting hotter (twice) as the other one, and when swapping primary leads it follows that primary.
So i guess i need to more balans the primaries.


I then measured across the inductor coil (differential methode) and tuned for max. amplitude wich seems at 113KHz, see red trace below (TL494 frequency 22.5KHz).

Now i am pulling 24V @ 1A.

Concerning the extra caps at the yoke, i am using a different PS to supply 24V to the primaries, so they need this filtering.
Besides, i was following Stalker his yoke layout which also has these caps at that location.

Itsu
   
img  AlienGrey
Re: Dally, Shark & Ruslan workbench « Reply #590 on: 2021-05-23, 20:54:38 »
Group: Guest
Quote from: Vasik041 on 2021-05-23, 20:25:19
That is probably not very significant now, but the proper place for these caps is close to the PCB socket,
so the length of wires (ground, +24v) would be minimal.

Vasik

PS there are already caps on PCB, most probably you don't need extra caps.
Yoke primary windings if you bifilar wind them instead of the way you are
doing it will absorb most of that wast energy back into the system.

Regards
« Last Edit: 2021-05-24, 08:28:50 by AlienGrey »
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #591 on: 2021-05-23, 20:57:22 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159
Quote from: picowatt on 2021-05-23, 20:33:46
Itsu,

I am confused.  The purple trace (a MOSFET drain) shows 48vdc when the MOSFET is off.  I thought you said you were running at 24vdc.

Also, the purple trace is pulled low (MOSFET apparently conducting) during a time period when neither gate is high (between the blue gate falling edge and the yellow gate rising edge).

What am I missing?

PW

PW,

well,  yes, i am running on 24V, but thats how the signals appear in a PushPull environment as i understand it.
Verpies made a hell of a good explaination a few years back on OU.com when we also stumbled across that.

Also he explained why we cannot rely on the voltage trace to see what a MOSFET is doing, the current trace tells it all, but for now i have no good current traces.
 
I will see if a can dig up a link.

Itsu 
   
img  Vasik041
Re: Dally, Shark & Ruslan workbench « Reply #592 on: 2021-05-23, 21:08:37 »
Group: Guest
Quote from: Itsu on 2021-05-23, 20:50:29
Vasik,
i connected the 3 turn secondary and things kept stable (24V @ 700mA input now).
But i noticed 1 MOSFET getting hotter (twice) as the other one, and when swapping primary leads it follows that primary.
So i guess i need to more balans the primaries.
I then measured across the inductor coil (differential methode) and tuned for max. amplitude wich seems at 113KHz, see red trace below (TL494 frequency 22.5KHz).
Now i am pulling 24V @ 1A.
Concerning the extra caps at the yoke, i am using a different PS to supply 24V to the primaries, so they need this filtering.
So it seems that primary windings are different ? How you made them ?
I would suggest simultaneous winding with two wires of equal length.
This way they will be very close to each other ever if core have some irregularities.

I think you can ground one of inductor ends for tuning (just remember disconnect it later), so you will not need differential signal.
Inductor circuit need to be tuned to 122Khz (24.4 x 5)

Quote
Besides, i was following Stalker his yoke layout which also has these caps at that location.
From power supply design rules it is clearly wrong. But you can change it later if it creates trouble.
Just make sure that there is short thick ground wire from PCB to the caps to avoid any loops.

BTW also please pay attention to kacher power wire, you really need to place a capacitor close to the PCB
(forgot to mention it before)

Vasik
   
img  picowatt
Re: Dally, Shark & Ruslan workbench « Reply #593 on: 2021-05-23, 21:15:47 »
Sr. Member
****

Posts: 453
Quote from: Itsu on 2021-05-23, 20:57:22
PW,

well,  yes, i am running on 24V, but thats how the signals appear in a PushPull environment as i understand it.
Verpies made a hell of a good explaination a few years back on OU.com when we also stumbled across that.

Also he explained why we cannot rely on the voltage trace to see what a MOSFET is doing, the current trace tells it all, but for now i have no good current traces.
 
I will see if a can dig up a link.

Itsu

Itsu,

Even with the primaries inductively coupled to one another, I cannot understand why the purple drain is apparently on around 60% of the time.  Look at the time period between the blue gate falling edge and the yellow gate rising edge, neither MOSFET should be on at that time, yet the purple drain is being pulled low (apparently the MOSFET is on). 

The 48vdc seen at the drain still makes no sense to me.  Are you certain your connections are correct?  Is there a good ground path between your two 24vdc supplies?

PW




   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #594 on: 2021-05-23, 21:37:55 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159
Quote from: Vasik041 on 2021-05-23, 21:08:37
So it seems that primary windings are different ? How you made them ?
I would suggest simultaneous winding with two wires of equal length.
This way they will be very close to each other ever if core have some irregularities.

I think you can ground one of inductor ends for tuning (just remember disconnect it later), so you will not need differential signal.
Inductor circuit need to be tuned to 122Khz (24.4 x 5)
From power supply design rules it is clearly wrong. But you can change it later if it creates trouble.
Just make sure that there is short thick ground wire from PCB to the caps to avoid any loops.

BTW also please pay attention to kacher power wire, you really need to place a capacitor close to the PCB
(forgot to mention it before)

Vasik

Vasik,

i did just that, "simultaneous winding with two wires of equal length."

I will have to see if i can make them equal induction wise.

Itsu
   
img  Itsu
Re: Dally, Shark & Ruslan workbench « Reply #595 on: 2021-05-23, 21:40:25 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4159
Quote from: picowatt on 2021-05-23, 21:15:47
Itsu,

Even with the primaries inductively coupled to one another, I cannot understand why the purple drain is apparently on around 60% of the time.  Look at the time period between the blue gate falling edge and the yellow gate rising edge, neither MOSFET should be on at that time, yet the purple drain is being pulled low (apparently the MOSFET is on). 

The 48vdc seen at the drain still makes no sense to me.  Are you certain your connections are correct?  Is there a good ground path between your two 24vdc supplies?

PW


Here i have eliminated the 2nd PS, now running on one 24V PS only (750mA) , same signals as above.


Link to verpies explaination 48V push Pull:   https://overunity.com/12736/kapanadze-cousin-dally-free-energy/msg474877/#msg474877


Itsu

 
   
img  picowatt
Re: Dally, Shark & Ruslan workbench « Reply #596 on: 2021-05-23, 21:51:15 »
Sr. Member
****

Posts: 453
Quote from: Itsu on 2021-05-23, 21:40:25

Here i have eliminated the 2nd PS, now running on one 24V PS only (750mA) , same signals as above.

Itsu

Itsu,

Still continues to make no sense to me.

Before tearing into the coil, I would replace the yoke windings with a pair of resistors again and confirm all is as it should be. 

The 48vdc at the drain makes no sense to me, nor does the purple drain being pulled low when both the blue and yellow gate are low. 

What happened to your current probe?

PW
   
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #597 on: 2021-05-23, 21:56:10 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: Vasik041 on 2021-05-23, 21:08:37
Just make sure that there is short thick ground wire from PCB to the caps to avoid any loops.
Yes, it is very important to minimize the area of the loops formed by the 3 wires connecting the two primary windings.
I just twist them together all the way to the mosfets...

Quote from: Vasik041 on 2021-05-23, 21:08:37
So it seems that primary windings are different ? How you made them ?
Yes, he wrote that they have different inductances, see below:

Quote from: Itsu on 2021-05-23, 10:14:38
Yoke primary inductances 34.7uH / 6.4uH    and    32.5uH / 5 uH    all at 100KHz  (the 2nd values are with both sec's shorted).

Maybe the ferrite's gaps under these windings are located asymmetrically.
   
img  verpies
Re: Dally, Shark & Ruslan workbench « Reply #598 on: 2021-05-23, 22:00:47 »

Group: Professor
Hero Member
*****

Posts: 3498
Quote from: picowatt on 2021-05-23, 21:15:47
Even with the primaries inductively coupled to one another, I cannot understand why the purple drain is apparently on around 60% of the time. 
I think you are not accounting for the time needed to disspate the magnetic flux energy stored in the core when both MOSFETs turn off.
Also, it is important to notice that what appears to be 0V when both MOSFETs are off, is in fact the -1.2V voltage drop of the diode, which at low vertical magnification gives a false impression that the MOSFET is on (i.e. pulling the drain down to ground).

The analysis of this problem started at the link Itsu has posted and concluded in the following post on the OU forum:
https://overunity.com/12736/kapanadze-cousin-dally-free-energy/msg475096/#msg475096
The posts between the first and last message of this discussion are interesting, too.
   
img  picowatt
Re: Dally, Shark & Ruslan workbench « Reply #599 on: 2021-05-23, 22:22:11 »
Sr. Member
****

Posts: 453
Quote from: verpies on 2021-05-23, 22:00:47
I think you are not accounting for the time needed to disspate the magnetic flux energy stored in the core when both MOSFETs are off.
Also, it is important to notice that what appears as 0V when both MOSFETs are off, is really -1.2V voltage drop of the diode.

The analysis of this problem started at the link Itsu has posted and concluded in the following post on the OU forum:
https://overunity.com/12736/kapanadze-cousin-dally-free-energy/msg475096/#msg475096
The posts between the first and last message of this discussion are interesting, too.

Verpies,

Thanks, I'll give it a look.

Is the shorter period of oscillations that occur during both gate off times the beginning of clamp diode turn off?  If not, what is causing those oscillations?

PW
   
Pages: 1 ... 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 ... 101
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-26, 06:33:44