PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2024-03-29, 05:16:23
News: Registration with the OUR forum is by admin approval.

Pages: 1 ... 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 40 41 42 ... 48
Author Topic: Akula0083 30 watt self running generator.  (Read 927500 times)

Group: Administrator
Hero Member
*****

Posts: 3866


Buy me some coffee
Fantastic thanks for letting me know they have arrived O0
   

Group: Moderator
Hero Member
*****

Posts: 2711


Buy me a cigar
Dear Verpies.

I am using inverting ( single transistor ) Mosfet driver.

BTW. why? WTF !!  ???

Cheers Grum.


---------------------------
Nanny state ? Left at the gate !! :)
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4056
Both of these waveforms appear to have sufficiently steep falling edges (>18V/ms) to trigger the comparator in TL494 at pins 1&2.
The 2nd waveform (with 'STOP' switch closed) also has -500mV negative excursions, that would trigger the second comparator in TL494 at pins 15&16.
Do these comparators take notice of these pulses?  Can you see the evidence of this on pin 3 ?

I wish I could see scopeshots like this from Itsu and Grumage.

P.S.
Are you using an inverting MOSFET driving circuit topology like this or non-inverting like this?


I can try that tonight.

By the way, i am using an inverting MOSFET driving circuit topology, but looking at the screenshot i published last sunday,
the signal from pin 11 (TL494 put) see the purple trace, has the same polarity/shape as the gate signal, see the yellow trace.

So to me that means it is NOT inverted; how could that be?

Also notice the signal shape is identical,  including the Miller plateau signature, how could that be on pin 11?

Another question i have is about the (unnamed) diode right after the start switch (D1 in GL diagram).
In my case it is an 1n4007 (1A), but after all my tests showing 10A going into the circuit, it still measures OK.
Guess its a matter of time this diode will pop,  right?  Need to use a more sturdy one here?

Regards Itsu
   
Group: Guest

P.S.
Are you using an inverting MOSFET driving circuit topology like this or non-inverting like this?

Verpies,

Like the first schematic (Akula's original schematic) with the single transistor.

Waveform 14 is scoped at pin 3 to ground with the 'STOP' switch open and LED fully lit. Waveform 15 is scoped at pin 3 to ground with 'STOP' switch closed.

Hoppy
   

Group: Professor
Hero Member
*****

Posts: 3354
the signal from pin 11 (TL494 put) see the purple trace, has the same polarity/shape as the gate signal, see the yellow trace.
So to me that means it is NOT inverted; how could that be?
Because the signal is not inverted between pin 11 and MOSFET's gate.
The signal is inverted between the NOR gate and the MOSFET's gate.

Also notice the signal shape is identical,  including the Miller plateau signature, how could that be on pin 11?
Because only a 7Ω resistor separates pin 11 and the gate of the MOSFET.
« Last Edit: 2014-04-17, 01:01:33 by verpies »
   

Group: Professor
Hero Member
*****

Posts: 3354
Another question i have is about the (unnamed) diode right after the start switch (D1 in GL diagram).
In my case it is an 1n4007 (1A), but after all my tests showing 10A going into the circuit, it still measures OK.
Guess its a matter of time this diode will pop,  right?  Need to use a more sturdy one here?
Yup
   

Group: Professor
Hero Member
*****

Posts: 3354
I am using inverting ( single transistor ) Mosfet driver.
I know.
The allegedly successful replication in this video uses the non-inverting TL494 -->MOSFET driving circuit topology that matches this schematic.
...and this schematic is equivalent to it signal-wise.

BTW. why? WTF !!  ???
Because, gates of N-ch enhancement mode MOSFETs need to be positive in reference to the source terminal in order to turn on.
Also, where from are you getting the negative voltage in reference to ground ?
« Last Edit: 2014-04-17, 01:00:53 by verpies »
   
Group: Guest
I found a bit more time for playing today and I now have 3 x 9W LED lamps in parallel as a load and have found that the duty pot can now be adjusted to bring the LED light level down to given level or completely off. The current draw is still high at a supply voltage not much above 9.5V before my bench PSU clamps and with a 12V battery, the draw is over 10 Amps.

Hoppy
   

Group: Professor
Hero Member
*****

Posts: 3354
Waveform 14 is scoped at pin 3 to ground with the 'STOP' switch open and LED fully lit. Waveform 15 is scoped at pin 3 to ground with 'STOP' switch closed.
Waveform 14 shows no significant activity.  The deviation from a flat line seems to be just HF interference.

Waveform 15 shows significant activity since 3.2V on pin 3 is sufficient to decrease the output duty cycle of the TL494 to 3% at the NOR gate.
DTY = 3% at the NOR gate after inversion yields DTY = 97% at the MOSFET, so it is not surprising, that the supply current consumption skyrockets.

In other words, increased voltage at pin.3 causes increased duty cycle at the MOSFET due to the polarity inversion after the NOR gate.
The feedback loop of TL494 appears to be positive.  Positive feedback loops are unusual in PWM power circuits because they create run-away duty cycles.
   
Group: Guest
Waveform 14 shows no significant activity.  The deviation from a flat line seems to be just HF interference.

Waveform 15 shows significant activity since 3.2V on pin 3 is sufficient to decrease the output duty cycle of the TL494 to 3% at the NOR gate.
DTY = 3% after inversion yields DTY = 97% at the MOSFET, so it is not surprising, that the supply current consumption skyrockets.

In other words, increased voltage at pin.3 causes increased duty cycle at the MOSFET due to the polarity inversion.
The feedback loop of TL494 appears to be positive.  Positive feedback loops are unusual in SMPS.


So this would make some sense of the NC 'STOP' switch if we had no polarity inversion and explain why Akula revised his schematic to show the push-pull driver.
   
Sr. Member
****

Posts: 336
Waveform 14 shows no significant activity.  The deviation from a flat line seems to be just HF interference.

Waveform 15 shows significant activity since 3.2V on pin 3 is sufficient to decrease the output duty cycle of the TL494 to 3% at the NOR gate.
DTY = 3% at the NOR gate after inversion yields DTY = 97% at the MOSFET, so it is not surprising, that the supply current consumption skyrockets.

In other words, increased voltage at pin.3 causes increased duty cycle at the MOSFET due to the polarity inversion after the NOR gate.
The feedback loop of TL494 appears to be positive.  Positive feedback loops are unusual in PWM power circuits because they create run-away duty cycles.


Verpies,

If we connect the gate resistor (6,8R) to pin 8 on TL494 instead, will the feedback be negative then?

GL.
   

Group: Professor
Hero Member
*****

Posts: 3354
So this would make some sense of the NC 'STOP' switch if we had no polarity inversion and explain why Akula revised his schematic to show the push-pull driver.
...and later the schematic with the integrated non-inverting driver IXDD609.
   

Group: Professor
Hero Member
*****

Posts: 3354
If we connect the gate resistor (6,8R) to pin 8 on TL494 instead, will the feedback be negative then?
No :(
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4056
Here the screenshot of the signal on the Junction R5/R7 (purple trace) together with the gate signal (yellow) and pin 3 (Blue):

Stop switch normally closed

Regards Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 4056

Here the same with the stop switch opened:


Regards Itsu
   
Group: Guest
For those that have not yet populated their PCB's and wish to go straight for the push-pull version 2, I have attempted to list out the stages of modification: -

1)  Cut pin 9 from IC or holder if used.

2)  Bend out pin 10 and wire to R1 pin on VT2 side. Do not fit a resistor in R1 position. Link IC pins 11 & 12 on underside of PCB.

3) Build push-pull circuit (New PNP transistor Q3, 100R, 2 x 1K, 15V mosfet gate zener and 1N4148 diode - across 10R) on a small piece of strip-board and connect into circuit with input to R1 pin on Q1 side (to connect with new wire link).
    Connect new PNP transistor (BC556) emitter to Q1 emitter and its collector to a conveniently located ground connection. Connect output to the mosfet gate (Q2).

4) Do not fit C12, R3 & R19. Replace Q1 with a BC546 transistor.

5) Cut PCB track between junction of R1 / Q1 base and junction of D2 cathode.


I have not done this myself yet, so it will need checking for accuracy.

Hoppy
   

Group: Professor
Hero Member
*****

Posts: 3354
Here the screenshot of the signal on the Junction R5/R7 (purple trace) together with the gate signal (yellow) and pin 3 (Blue):
Yup, the pulses are so short that the comparators (pins 2 & 15) are not even noticing it.
Most likely your R5 and C5 are too good of a quality ;) , ...or your MOSFET switching is weak, inverted or both.
   

Group: Professor
Hero Member
*****

Posts: 3354
For those that have not yet populated their PCB's and wish to go straight for the push-pull version 2, I have attempted to list out the stages of modification: -
And for those already having a non-inverting integrated MOSFET driver such as the IXDD609 or UCC27511 or TC4428 and wanting to go to ersion 3, the list of modifications is below:

IN REFERENCE TO Akula's DIAGRAM:
1) Remove R16, R17, R18, C12, VT2, VD1.
2) Connect pin 11 of the TL494 to pin 12.
3) Bend out pin 10 out of the socket and connected it to where the junction between R19 and C12 used to be.

   NOTE: at this point you can connect the junction of R19&C12 to the gate of the MOSFET even without the integrated driver.  The downside of this is that the MOSET will not switch as fast as with the integrated driver.

4) Bypass the power supply pins of the integrated MOSFET driver chip with a 470nF (or greater) ceramic capacitor (directly soldered across the supply pins at the underside of the driver chip)
5) Superglue the integrated MOSFET driver chip on its back to the power MOSFET.
6) Connect the positive power supply pin of the integrated MOSFET driver chip to the output pin of the three-terminal 12V linear voltage regulator (or to the positive terminal of C6) with at least 0.5mm dia. wire (short & stranded  wire is better)
7) Connect the negative power supply pin (a.k.a. GND) of the integrated MOSFET driver to the ground pin of the three-terminal 12V linear voltage regulator (or to the negative terminal of C6) with at least 0.5mm dia. wire (short & stranded wire is better).
8) Connect the non-inverting output of the integrated MOSFET driver chip to the gate of the MOSFET through a 4.7Ω carbon resistor.
9) Connect the non-inverting input of the integrated MOSFET driver chip to where the junction between R19 and C12 used to be
« Last Edit: 2014-04-16, 07:26:58 by verpies »
   
Sr. Member
****

Posts: 281
...and later the schematic with the integrated non-inverting driver IXDD609.

Hello All

Interesting note regarding the IXYS series of MOFET driver ICs:

Some of the IXYS series MOSFET drivers  utilize integrated Schmidt triggers for wave forming logic.  I remember on a couple of projects we used a few of them with excellent results...... gives 10s of nano-second leading edge rise times.  Miller plateau all but eliminated.

I have not had time to check the spec. sheet for the IXDD609 series, (as posted in schematic by Verpies) to see it they have the integrated Schmidtt trigger logic and/or Active Miller clamping.

When my board arrives I will try the Schmidt trigger mod.



take care, peace
lost_bro

EDIT Yes, it does have the integrated Schmidt trigger for waveform shaping..... see attachments
« Last Edit: 2014-04-15, 19:51:46 by lost_bro »
   

Group: Professor
Hero Member
*****

Posts: 3354
For those that have not yet populated their PCB's and wish to go straight for the push-pull version 2, I have attempted to list out the stages of modification: -
And for those already having a non-inverting integrated MOSFET driver such as the IXDD609 or UCC27511 or TC4428 and wanting to go to version 3, the list of modifications is below:

IN REFERENCE TO Groundloop's PCB:
1) Remove R1, R2, R3, C1, Q1, D2.
2) Connect pin 11 of the TL494 to pin 12.
3) Bend out pin 10 out of the socket and connected it to where the junction between R4 and C1 used to be.

   NOTE: at this point you can connect the junction of R4/C1 to the gate of the MOSFET even without the integrated driver.  The downside of this is that the MOSET will not switch as fast as with the integrated driver.

4) Bypass the power supply pins of the integrated MOSFET driver chip with a 470nF (or greater) ceramic capacitor (directly soldered across the supply pins at the underside of the driver chip)
5) Superglue the integrated MOSFET driver chip on its back to the power MOSFET.
6) Connect the positive power supply pin of the integrated MOSFET driver chip to the output pin of the three-terminal 12V linear voltage regulator (or to the positive terminal of C7) with at least 0.5mm dia. wire (short & stranded wire is better)
7) Connect the negative power supply pin (a.k.a. GND) of the integrated MOSFET driver to the ground pin of the three-terminal 12V linear voltage regulator (or to the negative terminal of C7) with at least 0.5mm dia. wire (short & stranded wire is better).
8) Connect the non-inverting output of the integrated MOSFET driver chip to the gate of the MOSFET through a 4.7Ω carbon resistor.
9) Connect the non-inverting input of the integrated MOSFET driver chip to where the junction between R4 and C1 used to be
« Last Edit: 2014-04-16, 07:28:53 by verpies »
   

Group: Professor
Hero Member
*****

Posts: 3354
If somebody wants to use the TC4428 integrated MOSFET driver, below are its connections:

The bypass capacitor C1 should be soldered similarly to the photo below, but on the underside of the chip and across the power supply pins 3 & 6.
« Last Edit: 2014-04-15, 20:32:59 by verpies »
   

Group: Professor
Hero Member
*****

Posts: 3354
...and the same for the UCC27511 integrated MOSFET driver chip:
   

Group: Professor
Hero Member
*****

Posts: 3354
In this video the author of a similar and allegedly self-running LED flashlight circuit, shows the internal construction of its transformer.
Such videos are uncommon.  I've never seen a transformer gutted like this by its constructor.

A schematic diagram that was reverse-engineered by magpwr directly from this video, is attached below:

« Last Edit: 2014-04-15, 22:20:31 by verpies »
   
Group: Guest
In this video the author of a similar and allegedly self-running LED flashlight circuit, shows the internal construction of its transformer.
Such videos are uncommon.  I've never seen a transformer gutted like this by its constructor.

Quite revealing verpies:

Two copper shield layers noted.
Outer winding appeared darker in color.  Would that be from heat or is that wire a coated black annealed iron wire?
Cardboard spacer between center of pot core.
Teflon tape between layers.

Anything else noticed and relevant?
   
Group: Guest
For the hell of it, I'll initially try the ICL7660 inverting driver as a comparison to the two versions of the bipolar transistor drivers. I'm currently right out of non-inverting drivers.

Hoppy
   
Pages: 1 ... 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 40 41 42 ... 48
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-03-29, 05:16:23