Storing Cookies (See : http://ec.europa.eu/ipg/basics/legal/cookies/index_en.htm ) help us to bring you our services at overunity.com . If you use this website and our services you declare yourself okay with using cookies .More Infos here:
https://overunity.com/5553/privacy-policy/
If you do not agree with storing cookies, please LEAVE this website now. From the 25th of May 2018, every existing user has to accept the GDPR agreement at first login. If a user is unwilling to accept the GDPR, he should email us and request to erase his account. Many thanks for your understanding

User Menu

Custom Search

Author Topic: Kapanadze Cousin - DALLY FREE ENERGY  (Read 11715661 times)

itsu

  • Hero Member
  • *****
  • Posts: 1845
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #330 on: October 10, 2012, 06:34:33 PM »
to itsu

Try old 1N5408 for DSRD diode in Nano-Pulser !! Use HV capacitor for Nano-Pulser 1-2n 1.5-2kV DC, minimum.


Сергей В.

i see this is yet another variation on the Dally diagram, it now uses a PNP and a NPN transistor for driving the MOSFET :-(

I will stick to my mosfet driver (max4420) for the time being.

Thanks for the suggestion for the 1N5408 diode and HV cap.

Regards Itsu

Сергей В.

  • Jr. Member
  • **
  • Posts: 63
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #331 on: October 10, 2012, 06:57:33 PM »
to verpies

Yes, you are right !!. Sch is good for PMOS but we need NMOS. I've corrected. Thanks !!


Сергей В.

  • Jr. Member
  • **
  • Posts: 63
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #332 on: October 10, 2012, 07:31:04 PM »
to Itsu

Yes you have right about clocking NMOS. I've mentioned earlier about using high speed driver. You can conect gate directly to driver output. Then best solution is to put MOSFET as possible as you can near the output pin. When you mount you can put driver on another small board on MOSFET radiator. I think it can "fly" on MOSFET gate. On this way you will solve usual problem with parasitic oscillations on MOSFET gate.


ps. Still not news from Dally !! We all still expecting oscillograms of Dally Generator. One friend tell me he used ferrite ring from TV antenna premplifier. If you can find will be nice to try one of this rings. I haven't information about parameters of these ferrite rings!!

If you need ultra High Speed driving MOSFETs you should make something like this driver. Last picture. Use Gigahertz input trasistor for clocking push-pull MOSFETs!!

d3x0r

  • Hero Member
  • *****
  • Posts: 1433
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #333 on: October 10, 2012, 10:42:04 PM »
Re the TL949 schematic...

The only +5 V is from the clock chip to the various transistors...

So probably... which would like trim the rising edge of the wave to the first PNP (it will be open and conducting with voltage from the diode to the collector until the base is at a certain voltage, which will turn itself off... normally these aren't even conducting because the wave is low. 

I'm only getting barely over a volt on the gates of the E13009's... before the resistor I have a nice 5V rise... and I would think that during the time the pnp is not conducting, that 5V should develop on both sides of that resistor...

Or is there suppose to be a pullup on the collector of the pnp, maybe a large resistor to power?

verpies

  • Hero Member
  • *****
  • Posts: 3473
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #334 on: October 11, 2012, 12:42:28 AM »
Re the TL949 schematic...
The only +5 V is from the clock chip to the various transistors...
Where do you see +5V on this schematic?

I'm only getting barely over a volt on the gates of the E13009's...
E13009s don't have gates.
BJTs are not voltage driven like MOSFETs, they are more current driven.

Take a look at the attached schematic.
This schematic is almost the same as yours. It is laid out differently and the transistors Q1, Q4, that are inside of the TL494 Integrated Circuit, are brought out and depicted as uncircled transistor symbols.

Note that the collectors of Q1, Q4 are connected to Vcc (+12V in this case) and the main job of these transistors is to pull up points I and G to Vcc, through the diodes D1 and D2.  Their minor job is to pull up the bases of Q2 and Q5 to Vcc.

Now, consider what happens when Q1 or Q4 stops conducting: Without Q2, Q5 and R2, R5 there would be nothing that would pull down the points I and G to ground. 

That would be bad because points I and G would float and any accumulated charge appearing at those points would stay there for a long time without being discharged to ground - this would result in very slow fall times at points I and G (especially if Q3 and Q6 were transistors of the MOSFET type, that has capacitive gates).

Fortunately, when Q1 and Q4 are not conducting, Q2 and Q5 come to the rescue by actively pulling down points I and G to ground.  In other words - they complement each other (namely: Q2 complements Q1 and Q5 complements Q4).

This happens because R2 and R5 pull down the respective bases of Q2 and Q5 (points F and H) more negative than their emitters, when Q1 and Q4 are not conducting and not pulling those bases up. Note that emitters of Q2 and Q5 (points I and G) can be more positive than their respective bases, when Q1 and Q4 are not conducting, only when there is some remaining positive charge at points I and G.

The purpose of the diodes D2 and D1 is to prevent points I and G from pulling up points H and F, respectively.
Due to D1 and D2, only Q1 and Q4 can pull up the bases of Q2 and Q5 (points F and H).

In summary, the transistors Q2 and Q5 act as active pull-downs, because the internal TL494 transistors Q1 and Q4 are only capable of pulling up the points I and G to Vcc (e.g.: +12V).

The active pull-downs allow the bases of Q3 and Q6 to be quickly jerked up and down (Vcc and Gnd), however because of the way BJTs work, the voltage between the base and emitter of Q3 and Q6 rarely rises above 0.7V.  Read this for an explanation - why.

Does that help at all?

P.S.
It is much more interesting to measure the voltage between the emitters and collectors of the Q3 and Q6 transistors, while these transistors are conducting.  If this voltage does not fall to zero then it means that Q3 and Q6 are underdriven and are not conducting fully.
« Last Edit: October 11, 2012, 12:11:19 PM by verpies »

mihai.isteniuc

  • Newbie
  • *
  • Posts: 40
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #335 on: October 11, 2012, 05:55:46 AM »
@d3x0r
 
See the schematics in the attachment. It's an AT computer power supply that uses TL494. you can see how this configuration works with other types of transistors. btw TL494 it's almost a must in the AT-ATX power supply units.
 
Or google it with AT power supply schematics -> images.
 
The first result: http://www.pavouk.org/hw/en_atxps.html
 
Use a different setup for this part. There is nothing exotic in this part of Dally's schematics. Go with the solutions in the AT-ATX to achieve your goal. The result should be the same. If you have a broken AT-ATX extract the transistors and the transformer from there cause they are a perfect fit and use them. They have been design to work together.
 
Mihai
 
PS I have modified the post because the google link, was breaking down the design of the page. Nothing was lost because of this.
 
Mihai
« Last Edit: October 11, 2012, 10:43:13 AM by mihai.isteniuc »

verpies

  • Hero Member
  • *****
  • Posts: 3473
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #336 on: October 11, 2012, 09:57:22 AM »
Or google it with AT power supply schematics -> images.
The schematics of the ATX Power Supplies are unnecessarily complicated for this purpose. 
Often they use advanced methods of driving the primary winding, e.g. a half-bridge or even full-bridge topologies.

mihai.isteniuc

  • Newbie
  • *
  • Posts: 40
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #337 on: October 11, 2012, 10:41:05 AM »
The schematics of the ATX Power Supplies are unnecessarily complicated for this purpose. 
Often they use advanced methods of driving the primary winding, e.g. a half-bridge or even full-bridge topologies.

I agree with you. The schematics should be consulted not copied to see how they are design and maybe to match some transistors (pnp or npn, and the rest of the resistors diodes and caps) with the transformer already present in the source. Also don't take into consideration some advanced designs, more like the "chinese simplified" ones. :)
 
At some point maybe a reaction circuit dependent with the load ... will be necesary ... if the device will ever work ... also a syncronization one if I'm right. Will see ...

d3x0r

  • Hero Member
  • *****
  • Posts: 1433
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #338 on: October 11, 2012, 12:59:59 PM »

I agree with you. The schematics should be consulted not copied to see how they are design and maybe to match some transistors (pnp or npn, and the rest of the resistors diodes and caps) with the transformer already present in the source. Also don't take into consideration some advanced designs, more like the "chinese simplified" ones. :)
 
At some point maybe a reaction circuit dependent with the load ... will be necesary ... if the device will ever work ... also a syncronization one if I'm right. Will see ...


But, the output of the TL949 goes to power, not to any sort of drivers...


I reduced the cap size I had across the neons and am able to get higher voltage faster... but not enough to keep the neons lit and run the pulser...


Hmmm My part seems to be reverse polarity of what I needed...
http://www.semicon.panasonic.co.jp/ds4/2SB1414_BED_discon.pdf   vs [size=78%]http://www.onsemi.com/pub_link/Collateral/2N5401-D.PDF[/size] ...


But anyway...


I changed my load from a incandescent bulb to a string of LEDs end-to-end... I'm up to about 38V output... that's with everything running... there are very narrow bands where side A and side B are tuned to be a good frequency, and can move both together.... Although I can also not run the nanopulser and get as much power out from just the L1 oscillation... but always when I turn on my nano pulser there is not enough power to light the neons

mihai.isteniuc

  • Newbie
  • *
  • Posts: 40
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #339 on: October 11, 2012, 02:22:43 PM »
I was making a wrong interpretation ... never mind

verpies

  • Hero Member
  • *****
  • Posts: 3473
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #340 on: October 11, 2012, 02:41:50 PM »
... but always when I turn on my nano pulser there is not enough power to light the neons
That's not surprising.
We still have not seen a nice sawtooth waveform measured between points A and B.

Also, we still don't know the current-draw of the nano pulser.

verpies

  • Hero Member
  • *****
  • Posts: 3473
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #341 on: October 11, 2012, 03:07:28 PM »
I will stick to my MOSFET driver (MAX4420) for the time being.
ACID TEST: if you can obtain 100ns wide pulses with 25ns rise/fall times on the drains of your N-Ch MOSFETs driven by the MAX4420, then you should be fine. 

A 5MHz square wave would have those 100ns pulse widths, so this is not cosmic technology ;)

itsu

  • Hero Member
  • *****
  • Posts: 1845
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #342 on: October 12, 2012, 10:58:18 AM »
ACID TEST: if you can obtain 100ns wide pulses with 25ns rise/fall times on the drains of your N-Ch MOSFETs driven by the MAX4420, then you should be fine.

I tested my nano-pulser using a MOSFET driver.
I have now a (on 12V) 100ns pulse using a 220pF cap. on the nano-pulser 74HCT00.
Video here: http://www.youtube.com/watch?v=olsr1m_DHKE&feature=youtu.be

Next step i did was hooking up a 2e PS to the drain of the MOSFET to see how it behaves on 200V.
However, hooking up a 2e PS to the drain (return lines of the both PS's connected) completely distorts the pulse coming from the MOSFET (lots of ringing / nasty pulses).
Even if i substitute the 12V from the drain by only 12V from yet another PS (my dual PS) its distorted.

I blew up both the MOSFET and the driver in the process.
Not sure why the pulse gets distorted, so need to find out.

Regards Itsu

verpies

  • Hero Member
  • *****
  • Posts: 3473
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #343 on: October 12, 2012, 02:24:36 PM »
Next step i did was hooking up a 2e PS to the drain of the MOSFET to see how it behaves on 200V.
However, hooking up a 2e PS to the drain (return lines of the both PS's connected) completely distorts the pulse coming from the MOSFET (lots of ringing / nasty pulses).
Even if i substitute the 12V from the drain by only 12V from yet another PS (my dual PS) its distorted.
This most likely happens because the wires from the power supply are too long and/or your 100Ω drain load resistor is inductive.
Long wires behave like coils.
When the MOSFET starts conducting, the Source to Drain current ramps up through that 100Ω load resistor you have in series. (the current ramps up linearly at first,  and later - asymptotically up to the V/R limit).  See here.
If your MOSFET can withstand the 200V/100Ω = 2A current limit and the Source-Drain path can withstand the voltage of 200V, then nothing bad can happen at this stage.

However, when the MOSFET stops conducting, the parasitic inductances of your resistors/wires will try to maintain that 2A current flowing in the same direction at all cost (even if it means increasing the voltage in this circuit to 10kV). 
Obviously your MOSFTET will not be able to handle that 10kV flyback pulse and will break down (likely taking its driver with itself).

The solution is to suppress the amplitude of this high voltage flyback spike by:
1) Minimizing the wire lengths between Ground, Source, Drain, the Load Resistor, and Vcc (think millimeters!)
2) Making sure the Drain Load Resistor is not inductive.
3) Putting good bypass capacitors across the Source and the Vcc, at the point where the Load Resistor is attached.
4) Providing a low impedance path for the flyback pulse:
    a) 1kΩ resistor in series with a HV 1nF capacitor, across the Source and the Drain (or across the Drain and Vcc)
    b) A high voltage Transil across the Source and Drain
    c) Recover the energy in that flyback pulse, just like we did with T2 a while ago
    d) Use the energy in that flyback pulse to energize the remainder of Dally's circuit (the primary of T1 and the diode in parallel with it)

P.S.
Please try to put a 10kΩ trimpot in series with a 47Ω resistor (R12) located next to C17 (the swappable 220pF capacitor of the U3 monostable multivibrator, that determines the 100ns pulse width now) and see how much you can adjust the pulse width using this trimpot.

Also, your 100ns pulse rise-time (actually off-time) is suspiciously long.
This has nothing to do with the operation of the U3 monostable, but most likely it is caused by the HV flyback pulse returning to the Gate of the MOSFET through the Drain-Gate capacitance (see: Miller Effect) and not allowing the MOSFET to turn-off quickly.  There are some better MOSFET drivers (such as the fast UCC27511) which have a higher sinking (8A) current than the sourcing current, to combat exactly this bad effect. However this problem should disappear once you get rid of the HV flyback pulse on the Drain, by one of the methods described in pt.4 above.
« Last Edit: October 12, 2012, 03:49:18 PM by verpies »

Black_Bird

  • Jr. Member
  • **
  • Posts: 81
Re: Kapanadze Cousin - DALLY FREE ENERGY
« Reply #344 on: October 12, 2012, 02:28:58 PM »
@itsu

Probably the toroidal pulse transformer has a very low inductance and current is growing beyond expected in the drain. I had to increase the number of turns to avoid that. Had exactly this problem before.