Tiger's Blocking Generator replication

[Itsu]

Here i will attempt a replication of Tiger's Blocking Generator.

Itsu

[Itsu]

i quickly put something together (yoke i had, MBR4045 dual diode, MJE13009, 600nF base cap, 2nF L2 cap, 1M base resistor, schottky diode and the coils like mentioned etc.) and it seem to be working right away when powered by 6 to 12V.

I do see that after removing power, it keeps on running for half a minute or so with a similar signal as in the video, then drops voltage to 0.5V and
then keeps on oscillating with some pulses for some additional 5 to 10 minutes (i have a 1000uF 50V electrolytic cap at the input).

I did reverse L1 leads as when in the circuit diagram.

But it does work as a blocking generator (joule thief) as expected, so some tuning and different components could further improve on it.

Itsu

[Itsu]

Hello Itsu,

I looked up some Data:

For KT805 there are A, B and C-Types so there might be a difference

https://alltransistors.com/search.php?search=KT805


The Diode you choosed is a general rectifier, they are slow in the reverse recovery-time, but above all they have a
high instantaneous reverse current of 1 mA which gives proof of beeing slow,so you loose energy, should be in the range of 50 to 100 myko-amps.

I have found that it depends on the application- circuit which choice you make : either general recovery time ( myko-seconds) and low Forward-Voltage-drop or higher Forward-Voltage-drop >1 V and fast recovery time in the range of 20 to 50 nsec.

I had a flyback-recovery-circuit for a transformer-primary ( pulsed and the switched off)  and tested fast Diodes ( 50 ns , Vf 1,1 V) versus an AA143 Germanium Diode and the AA143 outperformed the differend fast diodes I had available.

Concerning the inductance:
pay attention to the reversal of magnetism losses but I do not have much experience in selecting the right core. His monitor-chocke-core might be low on these losses ( depends on the airgap and material)

Its a good start though, thank you Vasik 

Also look up the end paragraphs of my post #35 above about capture of flyback-pulses which might be of importance.


Mike

Hi Mike,  thanks for the reply.

I put something together without taking notice to much of what i got, so any NPN transistor.
I have some KT805AM on order and will start using them to get as close to the circuit as possible.

I have some Germanium Diodes which i could try instead of the Schottky's.

I have my doubts about the way the coils suppose to be wound, i think i have the L3/L4 wrong (like a push pull now), will try different methods.

Presently i use a TV yoke i had with a 0.1mm Mylar insulation inbetween the gaps.

Looking into your post #35 later today, and i also will make some screenshots of the signals (talking Collector signal yesterday) then.


Itsu

[Vasik041]

Here another document from Tiger, I think it is relevant to the topic

[Itsu]

Thanks Vasik, that document could come in handy.


Mike, good information on paralleling a MLMC (MultiLayer Ceramic Capacitor) to the electrolytic ones to overcome the inertness to fast pulses.



Here a video on how my present setup functions.

I measure the collector of the transistor and the time it takes to show these signals after removing the 12V battery from the circuit.

The fist 20 seconds show a similar signal i think as shown by Tiger in his video.

Then the voltage suddenly drops from 12V to 2v or so and the pulse train becomes differently for some 6 minutes.

Video here:  https://youtu.be/cFdOazjelQs

Regards Itsu 

[spark2]

Hi Itsu
your video is in private navigation
right click on utube link and open
thank you also to Vasik for his many pdf, very interesting
well done Itsu thank you
Spark2

[Vasik041]

Thanks Vasik, that document could come in handy.
Mike, good information on paralleling a MLMC (MultiLayer Ceramic Capacitor) to the electrolytic ones to overcome the inertness to fast pulses.
Here a video on how my present setup functions.
I measure the collector of the transistor and the time it takes to show these signals after removing the 12V battery from the circuit.
The fist 20 seconds show a similar signal i think as shown by Tiger in his video.
Then the voltage suddenly drops from 12V to 2v or so and the pulse train becomes differently for some 6 minutes.
Video here:  https://youtu.be/cFdOazjelQs

Regards Itsu

Nice work Itsu 
You can try add more capacitors in parallel (with smaller capacitance) after rectifier like we see in the video, this will reduce ESR and may improve circuit performance.

Regards,
Vasik

[Vasik041]

Just a historical note: this setup was quite popular before "Tesla coil rush" started
Many people experimented with it and had interesting results.
More similar schematics in the attachment.

-Vasik

[Itsu]

Hi Itsu
your video is in private navigation
right click on utube link and open
thank you also to Vasik for his many pdf, very interesting
well done Itsu thank you
Spark2

Hi Spark2,

my video's suppose to be "hidden", so everyone which has the link should be able to open/see it.
Checking on this latest one, its also "hidden" and works for me when clicked on.

Are you sure its says "private" to you?

Nice work Itsu 
You can try add more capacitors in parallel (with smaller capacitance) after rectifier like we see in the video, this will reduce ESR and may improve circuit performance.

Regards,
Vasik

Vasik,


thanks for the info, i will add some MLCC's parallel to the rectifier cap and experiment with different transistors.


Those new diagrams remind me of the "joule thief" days where there were many threads all with different joule thief (blocking oscillator) circuits all able to light up leds (some hundreds) and with the intention to let is selfrun.

Itsu

[AlienGrey]

Itsu and others can some one try  a torroide with a air gap in the loop, called the magnetic diode!
it's supposed to cancel #~~~Lenz law.

Why don't I try it ? I dont have a grinder  and I have no gash cores to experiment with and i dont play with let is, 
I mean is it worth a go or not ?

Regards Sil

[picowatt]

Thanks Vasik, that document could come in handy.


Mike, good information on paralleling a MLMC (MultiLayer Ceramic Capacitor) to the electrolytic ones to overcome the inertness to fast pulses.



Here a video on how my present setup functions.

I measure the collector of the transistor and the time it takes to show these signals after removing the 12V battery from the circuit.

The fist 20 seconds show a similar signal i think as shown by Tiger in his video.

Then the voltage suddenly drops from 12V to 2v or so and the pulse train becomes differently for some 6 minutes.

Video here:  https://youtu.be/cFdOazjelQs

Regards Itsu

Itsu,

How does the circuit behave with the single diode omitted from the circuit (clip lead across it, etc)?

PW

[NickZ]

Hi Spark2,

my video's suppose to be "hidden", so everyone which has the link should be able to open/see it.
Checking on this latest one, its also "hidden" and works for me when clicked on.

Are you sure its says "private" to you?


   

















Vasik,


thanks for the info, i will add some MLCC's parallel to the rectifier cap and experiment with different transistors.


Those new diagrams remind me of the "joule thief" days where there were many threads all with different joule thief (blocking oscillator) circuits all able to light up leds (some hundreds) and with the intention to let is selfrun.

Itsu

   Guys:  As far as I know, Tiger never had any OU devices, nor one that would self run. Even after all these years.
   Yes, like in the joule thief days. But, Joule Thiefs never could self run.  Wishful thinking at best...

   This current circuit is just running off of the input cap, (rectifier cap).  Take out the cap, or just replace it with a lower value, and you'll see that it's only there to fool you, without having any load on. What's the point?

   I once asked Nelson Rocha, if he actually had what appears to be a self runner.  He said no, it's actually just recycling the stored energy in the caps and such, for a while. But, it ran for much longer than one would expect, which, made me think...

   Just trying to let you know what I think, if it matters to anyone.  As I did spend some time working on this type of thing, but, to no avail.
  You can't get something from nothing. A joule thief type of circuit is not an energy conversion device, nor an open circuit, it may be more efficient, but that's it. Perhaps that's why we have wasted so much time, barking up the wrong tree. Just saying...

    NickZ

   P.S.  Itsu your last link to the video is opening up just fine for me on YT.
   

[Itsu]

Itsu,

How does the circuit behave with the single diode omitted from the circuit (clip lead across it, etc)?

PW

PW

I added a 0.1uF MLCC across the rectifier cap (1000uF / 50V).

Still using the MJE13009 transistor and signals as shown yesterday seems the same.

With the single diode omitted from the circuit (clip lead across it) the first stage (20s at 11V with single pulse plus ringing) is omitted as it directly seems to go to this tapering of pulse train signal at 2V decreasing to zero.

But there does seem to be a difference.


Itsu

[Itsu]

Changing the transistor has a big impact on the signals and time the oscillations keep on going as expected.

A 2SC4632 was not changing a lot, a MJE3055 increased the time of oscillations well over 8 minutes and a present BF871 is still running after 13 minutes (stopped after 14 minutes) producing only the signal shown in the 1st 20s of the MJE13009 earlier (peak plus ringing).


But as Nick mentioned, they all slowly stop oscillating, but its fun to try to extend this dying out as long as possible.

Itsu

[Itsu]

With this BF871 running at start, there is about 1V on the rectified side (anode) of the single schottky diode and a (decreasing) 11V on the plus (kathode) side.

This of course needs to be reversed for any selfrunning to occur.

Shorting out this diode only increases the loading from the input and shortening the oscillation times me thinks.

Itsu

[Itsu]

Looking at the below diagram, i see the ringing frequency should be around 30 to 40KHz.

L2 together with C1 is the frequency determinating LC.

With C1 = 1.5nF, and a ringing frequency of 35KHz, then the L2 inductance (40 turns) should be 13.7mH.

My L2 (and L1) are 40 turns and measure 450uH, so my resonance frequency with C1 = 1.5nF is way higher 193KHz.

To get to 35KHz, i have to increase C1 to be 45nF

Where does this difference come from, the ferrite core used?

Itsu

[Kator01]

Itsu,

do you have Germanium Diodes like AA143 available ?

Then replace all diodes with this type. AA143 has a peak reverse Voltage of 20 V so this will work.

I did some tests with energy-recovery from collapse-spike lately and these germanium-type outperformed all.
Much of the energy dissipates because of the forward-voltage-drop. Even power-schottky diodes never go below 0.7 Volt,
Some of the AA143 even go below 0.3 V

I also think that it would be worth while to use a germanium-Type transitor if you have saved some from the transistor-radios of the eighties.

Now here you can see one of the main influences concerning the ringing-frequency. He leaves the base open so you can see it clearly. The beat-frequency is just another story

https://www.youtube.com/watch?v=Mz4VYcuXL5E

I documented this in 2011 and we are lucky it's still there.

There are more influencing factors which I forgot ( have to revisit my memory-palace ), but base-collector-capacitance strongly determines the ringing and its obvious why. Base-collector capacitance together with the collector coil is a LC-series
configuration. Also battery-capacitance must play a role in this setup.

Mike

[Vasik041]

Looking at the below diagram, i see the ringing frequency should be around 30 to 40KHz.

L2 together with C1 is the frequency determinating LC.

With C1 = 1.5nF, and a ringing frequency of 35KHz, then the L2 inductance (40 turns) should be 13.7mH.

My L2 (and L1) are 40 turns and measure 450uH, so my resonance frequency with C1 = 1.5nF is way higher 193KHz.

To get to 35KHz, i have to increase C1 to be 45nF

Where does this difference come from, the ferrite core used?

Itsu

Itsu,

Do you have a gap in the core ?
If not, then it is probably different ferrite.

-Vasik

[Itsu]

Kator,

i used some 1N34 Germanium diodes for both the single schottky diode as for the dual schottky's rectifier, but with the BF871 transistor which worked the best up till now (16 minutes) i have comparable running time, so no improvement using the Germanium diodes.

I have some AD161 transistors which i tried, but with the germanium diodes in the circuit, it did work, but only for 1.5 minute.

So although the various components do make some difference (mostly the transistor), i think the main activator is something else, probably the used ferrite in combination with a certain frequency.


Vasik,

i have 0.1mm mylar in between the both halves of the yoke.



Itsu   

[Vasik041]

Vasik,

i have 0.1mm mylar in between the both halves of the yoke.

Itsu,

you will get significantly higher inductance if you remove/reduce gap between halves of the core.

-Vasik

[AlienGrey]

Itsu it wants to be like a 'C' not 2 half sections

Sil

[Itsu]

I have removed the (mylar) gaps and rewound the L1 and L2 coils to be 40 turns each around one whole circumference of the yoke measuring both 562uH @ 10KHz.

The L3 and L4 coils are 13 turns each and both measuring 66uH @ 10Khz.

Still with the 3 germanium diodes, the BF871 and C1 being 2nF it still runs for 16 minutes.

Below screenshot shows the following signals:

Yellow:          collector voltage (AC coupled!).
purple / blue:  Anodes L3 and L4 diodes.
green:           L2 (collector) current.



The delta 145.8KHz ringing frequency in the upper right box of the screenshot very closely match the calculated (http://www.1728.org/resfreq.htm) parallel resonance frequency (150KHz) of the used 562uH @ 2nF.



So, gap or no gap, L3/L4 more or less turns (10 v 13), germanium or schottky diodes, the thing oscillates for a similar time, so i conclude that those components are hardly influencing its behavior.

Itsu

[Itsu]

I received my KT805AM transistors and used it in the above mentioned circuit.

The circuit was running for 14:30 minutes, so better then most transistors used, but not better as the BF871.


I will try to sweep the L2 coil to see if there is a sweet spot the ferrite responds to and use that frequency to get L2 and C1 in resonance.

Itsu

[Itsu]

A 10s sweep from 1KHz to 300KHz using a single turn around the yoke does not show any specific peaks in this range when scoping across C1 / L2.

So finding any sweet-spot will be on trail and error base i think.

Itsu

[Verpies]

For KT805 there are A, B and C-Types so there might be a difference
...
Mike

That is a good point even if the difference is only the package and not the silicon structure inside, the parasitics can be different between the transistor versions.