PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
OverUnity Research > Benches > Grumage (Moderator: Grumage) > Akula0083 30 watt self running generator.
2024-11-26, 22:30:05
News: Check out the Benches; a place for people to moderate their own thread and document their builds and data.
If you would like your own Bench, please PM an Admin.
Most Benches are visible only to members.

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 ... 48
Author Topic: Akula0083 30 watt self running generator.  (Read 975226 times)
Topic Options img
img  Peterae
Re: Akula0083 30 watt self running generator. « Reply #575 on: 2014-04-11, 21:39:15 »

Group: Administrator
Hero Member
*****

Posts: 3961


Buy me some coffee
Grum is this with GroundLoops PCB, it's best to mention this guys,because what ever the solution is maybe applicable to a lot of people  O0

The next question is going to be who is first to get the single transistor fet driver, or how we can get it going, because at some point we will need to answer the question of whether we need to mod it for push/pull drive of the fet.

We need to develop a test procedure, a step by step list of things to do, to see the fet is being driven properly without the transformer & load drawing loads of current.

Maybe something like using a bulb in place of the primary and maybe using a DC supply to provide feedback, someone more acquainted with the theory of operation should be able to work out the best practise.
   
img  Grumage
Re: Akula0083 30 watt self running generator. « Reply #576 on: 2014-04-11, 21:57:04 »

Group: Moderator
Hero Member
*****

Posts: 2765


Buy me a cigar
Quote from: Peterae on 2014-04-11, 21:39:15
Grum is this with GroundLoops PCB, it's best to mention this guys,because what ever the solution is maybe applicable to a lot of people  O0

The next question is going to be who is first to get the single transistor fet driver, or how we can get it going, because at some point we will need to answer the question of whether we need to mod it for push/pull drive of the fet.

We need to develop a test procedure, a step by step list of things to do, to see the fet is being driven properly without the transformer & load drawing loads of current.

Maybe something like using a bulb in place of the primary and maybe using a DC supply to provide feedback, someone more acquainted with the theory of operation should be able to work out the best practise.

Dear Peter.

Yes it is the NON breadboard type. I have been in chat with Groundloop he is suggesting I may have a dry joint or two !!  :)

The single transistor is driving the Mosfet OK but there is some instability (probably dry joint)  The LED's are the brightest I have ever seen them !!   8) A must !!

Your suggestion of a dummy load in place of the transformer seems to be a good idea.  O0

Cheers Grum.


---------------------------
Nanny state ? Left at the gate !! :)
   
img  Peterae
Re: Akula0083 30 watt self running generator. « Reply #577 on: 2014-04-11, 22:13:43 »

Group: Administrator
Hero Member
*****

Posts: 3961


Buy me some coffee
Well if you get stuck take some high res pictures under a light and we remotely check things for you  O0
   
img  Hoppy
Re: Akula0083 30 watt self running generator. « Reply #578 on: 2014-04-11, 22:22:41 »
Group: Guest
Quote from: lost_bro on 2014-04-11, 20:31:16
Good day Grum

PLease see the attachment.....
The RED line will bias the MOSFET GATE directly without signal from TL494 or driver....

take care, peace
lost_bro

This is a circuit path that I do not see in the original Akula circuit schematic.
   
img  Groundloop
Re: Akula0083 30 watt self running generator. « Reply #579 on: 2014-04-12, 00:07:10 »
Sr. Member
****

Posts: 336
@All,

The attached drawing is what I used when I designed the board. This drawing was included in the zip archive that I posted.

The TL494 built in transistor is used as a pull down device. So then you actually need a current path up to the positive rail.

Let us start with the TL494 built in transistor being ON, conducting current. The MOSFET will be off because the gate is
drained to ground. The driving transistor (2SC945) will be OFF because the base is low voltage. Now we switch OFF
the TL494 built in transistor. The voltage of the gate of the MOSFET will go high and the MOSFET will start to
switch ON current. But the base of the driving transistor (2SC945) will also go high voltage, so this transistor will also
switch on carrying current. This will make the MOSFET switch ON much faster because of higher gate loading current.
So the sole purpose of the driving transistor (2SC945) is to get a faster MOSFET switch ON time.

So, yes, lost_bro is correct, a faulty TL494 or a removed TL494 and the MOSFET will stay ON conducting current to the max.

Grumage asked what the PHZ pot-meter is doing in MY design of the circuit. This pot-meter will allow the user to
tune the time constant of the Akyla R2/C4. I do not know WHY Akyla did include those components, but IF
this is important then it is better to use some tunable components here. That was my thought. In MY design,
the variable capacitor is for fine tuning and the variable resistor is for large tuning.

Another issue, the Akyla board has ground planes. This is a lot of copper and very good heat sink. So when you
solder the GND connections, then use more heat and make sure you really DID solder the pads. For most of the
solder pads, a 18 Watt solder Iron is enough. But for the GND solder pads, I recommend a larger solder Iron of 40 Watt.

Also, if you use the recommended heat sink on the BD139 transistor, then make sure that the heat sink is lifted a little
up from the board before you solder the heat sink to the board. This because of small distance from the heat sink to
the transistor pads. The same problem are on the boards from Peter, because he did not download the newest version
of the files.

I also recommend using a variable voltage, current limited power supply, when you start testing on this board. Start low
voltage low current.

Added: I have received one board from Peter, thanks Peter,  and his board are identical to mine, except that on my board
there is gold plating on all pads. On Peters board there is standard tin plating on the pads.

GL.
« Last Edit: 2014-04-13, 17:14:32 by Groundloop »
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #580 on: 2014-04-12, 08:11:27 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Grumage on 2014-04-11, 21:57:04
Your suggestion of a dummy load in place of the transformer seems to be a good idea.  O0
That is a different idea.
e.g. When the MOSFET turns on for 100µs, the current through the transformer will gradually increase to 300mA, and with a dummy load (a light bulb) the current will immediately increase to 1A (or 10A ...depending on the light bulb).
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #581 on: 2014-04-12, 08:18:51 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Grumage on 2014-04-11, 19:34:19
Oh and BTW does anyone know what the FHZ pot is supposed to do ?? Id does not seem to work !!
On the original Akula diagram, the PHZ pot did not exist. It was the fixed 420kΩ resistor R2.
Normally, this pot/resistor is a part of a high pass filter or L2 spike scrubber.

Turning this pot close to 0Ω would be a bad idea as it would short the diode D2 associated with L2.  That's why R20 was added in series to prevent that.
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #582 on: 2014-04-12, 08:36:41 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Hoppy on 2014-04-11, 22:22:41
This is a circuit path that I do not see in the original Akula circuit schematic.
It's there, just drawn differently.

Without TL494 present, this path will indeed pull the gate of the MOSFET high, but when the TL494 is installed, pin 11 will pull it down through R7.
From the point of view of pin 11, the transistor VT2 and R18 act as an active pullup
In other words, TL494 pulls it down and VT2 pulls it up, achieving kind of a push-pull operation.
   
img  Hoppy
Re: Akula0083 30 watt self running generator. « Reply #583 on: 2014-04-12, 08:58:22 »
Group: Guest
Quote from: verpies on 2014-04-12, 08:36:41
It's there, just drawn differently.


Yes, I can now see it with my good glasses on!  :-[
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #584 on: 2014-04-12, 09:06:07 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Grumage on 2014-04-11, 19:34:19
The thing seems to have a mind of it's own !!  
I guess that is good news.
An anomalous circuit can be expected to behave unpredictably.

Quote from: Grumage on 2014-04-11, 19:34:19
I started with Hoppy's suggestion using the current limit on my bench PSU. Bringing the voltage up quietly. At around 7.8 V   I am getting a pulsating effect. The LED's shine like the sun pulling nearly 1.7 amps and then suddenly the current drops back to around 57 mA.  Strangely moving my hand around the circuit can trigger this response.
That is strange unless you have some floating HighZ inputs.
Do you have any activity at pin 3 ?

Quote from: Grumage on 2014-04-11, 19:34:19
I could really do with some advice now !! With the transformer fully disconnected the device has a standing current of 29/30 mA at 12 V. I placed my scope probe on the Mosfet gate resistor, with ref to ground. I am seeing no signal here. ??
...and you should.
Do you have pulses at pin 11 when the transformer is fully disconnected?
If not, does removing the Q1 (VT2) and the MOSFET, significantly increase the amplitude of pulses at pin 11, when the transformer is fully disconnected?

Quote from: Grumage on 2014-04-11, 19:34:19
With my transformer connected, the maximum voltage I can apply without serious overload is 7.8 V . I can see the pulses being applied to L1 with ref to ground.
In inductors, current pulses are much more interesting than voltage pulses.
The 0.22Ω resistor can be used as a Current Sensing Resistor (CSR) if it is not an inductive/wirewound resistor and when the nearby 100nF capacitor is removed.

Quote from: Grumage on 2014-04-11, 19:34:19
I can's seem to alter the duty cycle though !!
Because the DTY pot is mislabeled.
Its purpose is not to vary the duty cycle directly, but to adjust the negative slope sensitivity (from -6.6V/ms to -18V/ms) to pulses appearing at the junction of the 1Ω resistor and the bank of white load LEDs.
This pot will have no effect if you don't have quick falling edges at this junction.  See this video.

P.S.
The inductance of the 1Ω resistor R7 (R5 on Akula's diagram) is of PARAMOUNT IMPORTANCE to the feedback loop in this circuit.
So don't be surprised if this circuit works completely differently with an inductive/wirewound resistor than with a non-inductive one.
« Last Edit: 2014-04-12, 13:50:26 by verpies »
   
img  Grumage
Re: Akula0083 30 watt self running generator. « Reply #585 on: 2014-04-12, 18:40:49 »

Group: Moderator
Hero Member
*****

Posts: 2765


Buy me a cigar
Dear Verpies.

Thank you for your detailed analysis.  O0

After finding my old Solon soldering iron from my youth had finally " pegged out " !! Wonder whether the new Chinese one will last as long ?? ;D  I went over the board and re heated some of the ground connections. Behold, no more strange behaviour !!

PIN 3

There is activity on Pin 3 . A DC level that starts to rapidly rise and fall ( in synch ) with the core vibration. I can bring this about with adjustments to both frequency and DTY. However the device current is still through the roof. Pushing over 3 A @ 8.2 Vdc !!

BTW. I am using Carbon Film resistors throughout the circuit build.

Cheers Grum.


---------------------------
Nanny state ? Left at the gate !! :)
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #586 on: 2014-04-12, 19:14:19 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Grumage on 2014-04-12, 18:40:49
There is activity on Pin 3 . A DC level that starts to rapidly rise and fall ( in synch ) with the core vibration.
That means that your signal feedback loop is functioning.

Quote from: Grumage on 2014-04-12, 18:40:49
I can bring this about with adjustments to both frequency and DTY. However the device current is still through the roof. Pushing over 3 A @ 8.2 Vdc !!
The wider the pulses on pin 3, the narrower the pulses on pin 11 (and at the MOSFET).
Narrower pulses mean less average current draw from the power supply.

Quote from: Grumage on 2014-04-12, 18:40:49
Pushing over 3 A @ 8.2 Vdc !!
Show me the scopeshot of current flowing through L1.  Maybe its inductance is too low due to excessive core gapping.

Quote from: Grumage on 2014-04-12, 18:40:49
BTW. I am using Carbon Film resistors throughout the circuit build.
Is the Akula persona using all non-inductive resistors, too?
   
img  Peterae
Re: Akula0083 30 watt self running generator. « Reply #587 on: 2014-04-12, 20:55:03 »

Group: Administrator
Hero Member
*****

Posts: 3961


Buy me some coffee
Quote
Is the Akula persona using all non-inductive resistors, too?
I'm on the wrong computer for screen shots, but having just looked at the video, they are blue resistors, so it could be carbon or wirewound, but they look like a dull blue coating which would tend to tell me they are wirewound, from my experience carbon tend to be shiny, and the 2 watt resistors in the video are dull.

Best guess is they are indeed inductive.

Also worth noting was that the vero board version has 2 pre driver transistors without heat sinks driving the fet and he is using a single high watt LED security light for the load, the video is on the first post first page.
   
img  Grumpy
Re: Akula0083 30 watt self running generator. « Reply #588 on: 2014-04-12, 22:14:48 »

Group: Tinkerer
Hero Member
*****

Posts: 3948
tExB=qr
I have several blue resistors that are all metal oxide.
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #589 on: 2014-04-12, 22:23:01 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Peterae on 2014-04-12, 20:55:03
...they look like a dull blue coating which would tend to tell me they are wirewound, from my experience carbon tend to be shiny, and the 2 watt resistors in the video are dull.
That might not hold in Akula's part of the world.

Quote from: Peterae on 2014-04-12, 20:55:03
Best guess is they are indeed inductive.
That would be a significant discovery
   
img  Peterae
Re: Akula0083 30 watt self running generator. « Reply #590 on: 2014-04-12, 22:34:20 »

Group: Administrator
Hero Member
*****

Posts: 3961


Buy me some coffee
Quote
That might not hold in Akula's part of the world.
True, although i did search google images for blue carbon and blue wirewound to see if any existed and how they looked.

How common are 2 watt carbon resistors in Russia, here they would be carbon film these days, not that that would change the situation i suppose.

He also had a huge resistor near the LED, i cant see them making a carbon or film resistor that size, although it looks unusual to me for a wire wound resistor, i cannot imagine it being anything else other than a wirewound.

Maybe someone else with a lot of practical experience would also take a look.

EDIT this is what they look like to me
   
img  Hoppy
Re: Akula0083 30 watt self running generator. « Reply #591 on: 2014-04-12, 22:36:13 »
Group: Guest
I have populated the PCB supplied by Peterae and have found that if the direct DC path to the mosfet (as shown by lost_bro in post 576), by cutting the track to the mosfet (see attached) just leaving the 750R resistor biasing the drive transistor, the transistor does not switch cleanly and with the track intact the mosfet fails to switch 'OFF', thereby passing a considerable continuous current through LI as Grum has reported.

Hoppy
   
img  Groundloop
Re: Akula0083 30 watt self running generator. « Reply #592 on: 2014-04-12, 22:42:20 »
Sr. Member
****

Posts: 336
Quote from: Hoppy on 2014-04-12, 22:36:13
I have populated the PCB supplied by Peterae and have found that if the direct DC path to the mosfet (as shown by lost_bro in post 576), by cutting the track to the mosfet (see attached) just leaving the 750R resistor biasing the drive transistor, the transistor does not switch cleanly and with the track intact the mosfet fails to switch 'OFF', thereby passing a considerable continuous current through LI as Grum has reported.

Hoppy

Hoppy,

If you cut the track then you need to add a 10K resistor from gate on MOSFET to ground.
This because with the track cut, there is no current path for the gate to discharge to ground.

Also, do you have a o-scope image (DC) of the gate voltage to the MOSFET?

GL.
   
img  Hoppy
Re: Akula0083 30 watt self running generator. « Reply #593 on: 2014-04-13, 00:14:54 »
Group: Guest
Quote from: Groundloop on 2014-04-12, 22:42:20
Hoppy,

If you cut the track then you need to add a 10K resistor from gate on MOSFET to ground.
This because with the track cut, there is no current path for the gate to discharge to ground.

Also, do you have a o-scope image (DC) of the gate voltage to the MOSFET?

GL.

Yes, I had a gate grounding resistor in place. No scope shot at the moment as TL494 output gone completely dead. I think the voltage reg got fried! Earlier on I was over-volting C10 as if the load had become OC but when checked it was not. I've changed the regulator and have the correct output voltage at the TL494 but no output. Changed the TL494 but still no output. Time for bed.

Hoppy
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #594 on: 2014-04-13, 01:54:49 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: Hoppy on 2014-04-12, 22:36:13
...thereby passing a considerable continuous current through LI as Grum has reported.
I talked to Grum and his high current seems to be caused by exceeding the reverse blocking voltage of his diodes.
   
img  lost_bro
Re: Akula0083 30 watt self running generator. « Reply #595 on: 2014-04-13, 02:02:01 »
Sr. Member
****

Posts: 281
Quote from: verpies on 2014-04-12, 19:14:19
That means that your signal feedback loop is functioning.
The wider the pulses on pin 3, the narrower the pulses on pin 11 (and at the MOSFET).
Narrower pulses mean less average current draw from the power supply.
Show me the scopeshot of current flowing through L1.  Maybe its inductance is too low due to excessive core gapping.
Is the Akula persona using all non-inductive resistors, too?

Helllo Verpies

As pointed out earlier, because the configuration is common emitter, sinking current ( with pull-up resistor), when the pulse is NARROWER @ PIN 11; then the inverse of the signal would correspond to a LONGER pulse at the MOSFET gate, ie; pulse = 10%DC @ pin 11 then 90%DC @ gate.  IF the circuit is functioning properly, of course....

take care, peace.
lost_bro

EDIT; needed to change the BOLDed out  part of quote 
« Last Edit: 2014-04-13, 16:30:25 by lost_bro »
   
img  4Tesla
Re: Akula0083 30 watt self running generator. « Reply #596 on: 2014-04-13, 05:47:25 »
Group: Guest
Here are screenshots of both Akula's and XName41's boards.  It looks like Akula has a normal beige color resistor, while XName41 has a blue resistor.

They look like 2 or 3 watt resistors.

Beige = Carbon Film
Blue = Metal Film

(in most cases)

The load or current limiting resistor on the LED in the video looks to be wirewound.
« Last Edit: 2014-04-13, 08:15:55 by 4Tesla »
   
img  Hoppy
Re: Akula0083 30 watt self running generator. « Reply #597 on: 2014-04-13, 10:04:55 »
Group: Guest
Quote from: Groundloop on 2014-04-12, 22:42:20
Hoppy,

If you cut the track then you need to add a 10K resistor from gate on MOSFET to ground.
This because with the track cut, there is no current path for the gate to discharge to ground.

Also, do you have a o-scope image (DC) of the gate voltage to the MOSFET?

GL.

GL & all,

I've now re-instated the track and as requested attach some scope shots taken from mosfet gate to ground (DC coupled, x10). Shot 06 shows the waveform well before it breaks into audible ferrite core oscillation by adjustment of the 'Duty' pot. Shot 07 shows the waveform just before it breaks into audible oscillation. Likewise, shot 08 shows the waveform as scoped at cathode of D6 well before it breaks into audible oscillation. Shot 09 shows the waveform just before it breaks into audible oscillation. I have not found it possible to get a stable shot of the waveform whilst in oscillation (tried all trigger modes).

At all points of pot adjustments, the LED cluster shows no sign of illumination. However, during periods of ferrite core oscillation (sounds loud, like a squealing pig!) cap C10 gets hot. Originally, I had 2 x 1000uF, 35V in parallel and both of these started to blow and leak. I've now replaced these with a single 3,000uF, 63V cap.

Hoppy
« Last Edit: 2014-04-13, 10:31:39 by Hoppy »
   
img  Groundloop
Re: Akula0083 30 watt self running generator. « Reply #598 on: 2014-04-13, 10:38:45 »
Sr. Member
****

Posts: 336
Quote from: Hoppy on 2014-04-13, 10:04:55
GL & all,

I've now re-instated the track and as requested attach some scope shots taken from mosfet gate to ground (DC coupled, x10). Shot 06 shows the waveform well before it breaks into audible ferrite core oscillation by adjustment of the 'Duty' pot. Shot 07 shows the waveform just before it breaks into audible oscillation. Likewise, shot 08 shows the waveform as scoped at cathode of D6 well before it breaks into audible oscillation. Shot 09 shows the waveform just before it breaks into audible oscillation. I have not found it possible to get a stable shot of the waveform whilst in oscillation (tried all trigger modes).

At all points of pot adjustments, the LED cluster shows no sign of illumination. However, during periods of ferrite core oscillation (sounds loud, like a squealing pig!) cap C10 gets hot. Originally, I had 2 x 1000uF, 35V in parallel and both of these started to blow and leak. I've now replaced these with a single 3,000uF, 63V cap.

Hoppy

Hoppy,

Thank you for providing the o-scope shots.

>>the LED cluster shows no sign of illumination

This indicates a problem in you circuit. Also the capacitor going warm indicate a high current draw to ground somewhere.
Have you used a voltmeter and checked what voltage you get in C10? Is the LEDs connected the right way? Do the LEDs
have a 1 Ohm resistor current path to ground? Any shorts in your soldering? Etc.

GL.
   
img  verpies
Re: Akula0083 30 watt self running generator. « Reply #599 on: 2014-04-13, 11:27:48 »

Group: Professor
Hero Member
*****

Posts: 3499
Quote from: lost_bro on 2014-04-13, 02:02:01
As pointed out earlier, because the configuration is common emitter, sinking current ( with pull-up resistor), when the pulse is NARROWER @ pin 11; then the inverse of the signal would correspond to a LONGER pulse at the MOSFET gate,
Correct and because of this inversion, the functions of error amplifiers (pins 2 &15) are also inverted and their activation (pin 3 high) causes the lengthening of MOSFET's ON-state.  
Longer MOSFET conduction time translates to more average current in the transformer/inductor.

Good catch :)
   
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 ... 48
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-26, 22:30:05