Tiger's Blocking Generator replication

Author Topic: Tiger's Blocking Generator replication (Read 36127 times)

Verpies	Re: Tiger's Blocking Generator replication « Reply #50 on: 2021-09-27, 09:15:27 »
Group: Administrator Hero Member Posts: 4527	Quote from: Vasik041 on 2021-09-27, 05:58:01 It's 6v but capacitors charged to 12v. Overvolting electrolytic caps outgasses their electrolyte and damages them permanently even if their dielectric is not punctured.

Itsu

Re: Tiger's Blocking Generator replication « Reply #51 on: 2021-09-27, 09:40:09 »

Group: Moderator
Hero Member

Posts: 4704

Quote

To prove/disprove the Dielectric Soak/Absorbtion hypothesis, force the supply voltage to slowly increase over time and note whether the shape of the "PRF vs. Voltage" graph is the same as when this voltage is decreasing.

I am using a 12V battery as input now, but will see what a variable PS does as i manually increase the voltage.

Concerning the 6V cap as C3, here the 12V from the battery / PS does not "reach" it, doing a quick test i see it reaches 3.7V or so (depending how long i hold the 12V on the input) then decreasing.
Could be the L3 / L4 coils are not according to the specs (not bifilar), so will see if that increases after changing them.

I do have another 9 of those 10000uF / 6.3V caps in series (3) and parallel (3), so will have 10000uF @ 18.9V.

Itsu

Itsu

Re: Tiger's Blocking Generator replication « Reply #52 on: 2021-09-27, 13:59:41 »

Group: Moderator
Hero Member

Posts: 4704

Quote from: Itsu on 2021-09-26, 21:24:13

OK, i will rewind L3 and L4 that way tomorrow.

This evening, with the above modifications, it ran for 1 Hour 19 Minutes, starting at 12V, ending with 0.62V and with a PRF around 1.82Hz for ¾ of that time.

Voltage steadily decreasing over time, so no increase noticed.

Tomorrow i will show the Excel graphs.

Itsu

Excel data / graph from this test run yesterday:

Itsu

Tiger's Blocking Generator replication

Modification test run.png (78.74 kB, 1598x943 - viewed 795 times.)

Itsu	Re: Tiger's Blocking Generator replication « Reply #53 on: 2021-09-27, 15:06:52 »
Group: Moderator Hero Member Posts: 4704	I increased the supply voltage from 2V to 4V in 0.1V steps, but the PRF stayed the same value as noted down yesterday (xls graph's above). Itsu

Verpies

Re: Tiger's Blocking Generator replication « Reply #54 on: 2021-09-27, 15:45:34 »

Group: Administrator
Hero Member

Posts: 4527

Quote from: Itsu on 2021-09-27, 15:06:52

I increased the supply voltage from 2V to 4V in 0.1V steps, but the PRF stayed the same value as noted down yesterday (xls graph's above).

I was looking for a difference between two graphs like in the attachment below (i.e. hysteresis), that would have to be present for the amplitude to bounce up like in the Tiger's video. The dielectric soak cannot be responsible for the amplitude bouncing up like that when such hysteresis is not present. Without the hysteresis it can be only responsible for stabilization without bouncing up.

This reasoning hinges on the assumption the the peak amplitude of the ringdown is proportional to PRF.
To be certain, you'd have to directly measure the peak amplitude of the ringdown vs. supply voltage (in both directions).

The red graph is for the PRF when the voltage is decreasing (retyped from your latest spreadsheet, arrows hand drawn)
The blue graph is for the PRF when the voltage is increasing (all hand drawn - fabricated).

Tiger's Blocking Generator replication

PRF Hysteresis.png (20.8 kB, 702x641 - viewed 704 times.)

Itsu

Re: Tiger's Blocking Generator replication « Reply #55 on: 2021-09-27, 16:26:03 »

Group: Moderator
Hero Member

Posts: 4704

Ok, i think i understand, my below graph shows the 2V to 4V part of the data from the excel sheet from yesterday which is with a decreasing supply voltage.
This morning i used my variable PS to measure the PRF with an increasing supply voltage from 2V to 4V in 0.1V steps and the result was that the PRF data was the same as from yesterday.

So the graph lines (decrease and increase) would be the same also, so no hysteresis, see the graph below.

I can do that again (decrease and later-on a manual increase) for the whole 12V range, but i think again the PRF data would be the same.

In the mean time i have removed my L3 and L4 series coils and replaced them by L3 and L4 bifilar coils (18 turns each).
A run time test is running now (34 minutes and going).

What i did notice just now is that the supply voltage (12V momentarily from a battery) drops very quickly (seconds) to 5 / 6V then slower to end at 0.6V or so, but the ringing amplitude increases the first seconds.

I will try to capture that on video later today.

Itsu

Tiger's Blocking Generator replication

Hysteresis.png (12.15 kB, 623x516 - viewed 695 times.)

Verpies	Re: Tiger's Blocking Generator replication « Reply #56 on: 2021-09-27, 17:42:22 »
Group: Administrator Hero Member Posts: 4527	Quote from: Itsu on 2021-09-27, 16:26:03 I can do that again (decrease and later-on a manual increase) for the whole 12V range, but i think again the PRF data would be the same. I think we already have this data but you could do the Ringdown ampliude vs. Supply voltage measurement (in both directions) which constitutes a more direct proof of this hysteresis.

Itsu

Re: Tiger's Blocking Generator replication « Reply #57 on: 2021-09-27, 20:24:22 »

Group: Moderator
Hero Member

Posts: 4704

The below video shows the start of a cycle by touching the input cap with a 12V battery.

It shows the input cap voltage in blue and the collector signal (ringing) in yellow.

The blue input cap voltage can be seen dropping from 12V to a few volts in seconds, but it also shows the ringing amplitude signal increasing in those first few seconds, then stabilizing.

As you also can see in the below video is that capturing the ringing amplitude in data is not so easy as the PRF of 2Hz or so probably is to slow,
so getting the "Ringdown amplitude vs. Supply voltage measurement (in both directions)" will be hard.
Putting the scope in "normal mode" helps, so i will try to capture that data.

Video here: https://youtu.be/N41Z_BdcDq4

Itsu

Verpies	Re: Tiger's Blocking Generator replication « Reply #58 on: 2021-09-27, 20:40:49 »
Group: Administrator Hero Member Posts: 4527	Quote from: Itsu on 2021-09-27, 20:24:22 so getting the "Ringdown amplitude vs. Supply voltage measurement (in both directions)" will be hard. I agree. You'd have to force the supply voltage with a variable PS in both directions and keep it there for the amplitude measurement.

muDped

Re: Tiger's Blocking Generator replication « Reply #59 on: 2021-09-27, 23:47:01 »

Group: Tinkerer
Hero Member

Posts: 3324

When it comes to the "switching" action of the Transistor in this particular circuit, which is preferred: Soft or Hard Switching?

Should the efficiency of the switching device be as high as possible or does it matter?

Have any switching characteristics been identified which may enhance circuit performance?

Is the switching Pulse Width an important consideration?

---------------------------

For there is nothing hidden that will not be disclosed, and nothing concealed that will not be known or brought out into the open.

Verpies

Re: Tiger's Blocking Generator replication « Reply #60 on: 2021-09-28, 00:31:40 »

Group: Administrator
Hero Member

Posts: 4527

Quote from: muDped on 2021-09-27, 23:47:01

When it comes to the "switching" action of the Transistor in this particular circuit, which is preferred: Soft or Hard Switching?

I think Hard Switching because a half-open transistor with current flowing through it dissipates a lot of energy as heat. Fully closed or fully open transistors dissipate the least of energy. Thus half-open durations should be minimized.

Also, since the transistor feeds an inductance and current in an inductor cannot change quickly by definition, then pulses with short risetimes are smoothed out by the inductor and don't cause losses. Short falltimes will generate high reverse voltages when no path for the current in the inductor is provided when the transistor turns off quickly and that can lead to the loss of energy which was stored in the inductor. A capacitor connected in parallel with the inductor can recycle this energy.

Itsu	Re: Tiger's Blocking Generator replication « Reply #61 on: 2021-09-28, 16:22:22 »
Group: Moderator Hero Member Posts: 4704	The change yesterday from L3 and L4 in series (13 turns each) to L3 and L4 bifilar (18 turns each) has resulted in a decrease in run time from 1 Hour 19 minutes to 1 Hour 6 minutes. Itsu

Verpies	Re: Tiger's Blocking Generator replication « Reply #62 on: 2021-09-28, 23:57:08 »
Group: Administrator Hero Member Posts: 4527	With different number of turns it is not an apples-to-apples comparison.

Itsu

Re: Tiger's Blocking Generator replication « Reply #63 on: 2021-09-29, 13:47:50 »

Group: Moderator
Hero Member

Posts: 4704

Below graph shows the supply voltage versus ring-down voltage for both decrease (blue) as increase (orange) of the supply voltage.

They are very closely matched.

The supply voltage of 12V will cause the circuit to start oscillating, so the 15.1Vpp ring down voltage mentioned could be off.

Itsu

Tiger's Blocking Generator replication

V-supply v V-ringing.png (59.59 kB, 1376x943 - viewed 581 times.)

Verpies	Re: Tiger's Blocking Generator replication « Reply #64 on: 2021-09-29, 16:54:26 »
Group: Administrator Hero Member Posts: 4527	Looks monotonic. I expected some extrema based on your video. Hysteresis looks minimal. I don't think this data can explain the bounce back shown at 3:10. Is Smudge here?

muDped

Re: Tiger's Blocking Generator replication « Reply #65 on: 2021-09-30, 03:41:44 »

Group: Tinkerer
Hero Member

Posts: 3324

You've tried a lot of variations with some very interesting results.

I wonder, for the DC Feedback portion of the circuit; have you considered configuring it as a Voltage Doubler to more efficiently extract energy from the oscillations?

such as this configuration:

Would it make any difference at all?

---------------------------

For there is nothing hidden that will not be disclosed, and nothing concealed that will not be known or brought out into the open.

Itsu

Re: Tiger's Blocking Generator replication « Reply #66 on: 2021-09-30, 09:45:11 »

Group: Moderator
Hero Member

Posts: 4704

muDped,

i was wondering about "the DC Feedback portion of the circuit" myself too.

It seems to me that the relation L2 to L3/L4 should be equal at least (now 40t v 10t) to have enough energy to keep on going.
I already increased L3/L4 to 18 turns, but that seemed to have not enough influence.

Your Voltage Doubler could be an option, but at first i would like to stay as close to the original circuit as possible.

Thnaks, itsu

Verpies

Re: Tiger's Blocking Generator replication « Reply #67 on: 2021-09-30, 12:00:33 »

Group: Administrator
Hero Member

Posts: 4527

Quote from: muDped on 2021-09-30, 03:41:44

I wonder, for the DC Feedback portion of the circuit; have you considered configuring it as a Voltage Doubler to more efficiently extract energy from the oscillations?

I don't think so because that would entail capacitor-to-capacitor energy transfers which are inherently inefficient.

Capacitor-to-inductor (and vice-versa) energy transfers are very efficient and IMO that is the principle of the improvement suggested here.

Also, that circuit probably has currents with two frequency components circulating through it (just like that other device).
I'd venture that the energy contained in only one of these components is the result of the gain mechanism and should be looped back to the input of the device. This would mean that a frequency-selective feedback circuit could be of benefit, too (as described in the 3^rd paragraph here and here).

Verpies	Re: Tiger's Blocking Generator replication « Reply #68 on: 2021-09-30, 13:16:26 »
Group: Administrator Hero Member Posts: 4527	Quote from: Itsu on 2021-09-30, 09:45:11 ...but at first i would like to stay as close to the original circuit as possible. So you should not omit the large area aluminum heatsinks, which can act as HF capacitor plates even if they are thermally superfluous.

Itsu	Re: Tiger's Blocking Generator replication « Reply #69 on: 2021-09-30, 15:42:15 »
Group: Moderator Hero Member Posts: 4704	Quote from: verpies on 2021-09-30, 13:16:26 So you should not omit the large area aluminum heatsinks, which can act as HF capacitor plates even if they are thermally superfluous. Right, was thinking about putting back the schottky diodes and attach heatsinks to them and the transistor. Itsu

muDped	Re: Tiger's Blocking Generator replication « Reply #70 on: 2021-10-01, 02:50:42 »
Group: Tinkerer Hero Member Posts: 3324	Verpies, I agree with your prognostication that it may not work as hoped. I was principally wondering if anyone had given the concept any thought. Often we are surprised by how things actually perform when put to the test. --------------------------- For there is nothing hidden that will not be disclosed, and nothing concealed that will not be known or brought out into the open.

Itsu

Re: Tiger's Blocking Generator replication « Reply #71 on: 2021-10-01, 09:37:21 »

Group: Moderator
Hero Member

Posts: 4704

I made some changes these last days like:

increase the choke from 1.5mH to 9mH
increase C3 from 10000uF to 15000uF to 20000uF.
decrease Cap in series with L1 from 600nF to 450nF
increase ringing frequency from 29KHz to 35KHz by decreasing C1
changed the rectifying germanium diodes to a dual schottky diode and put it on a heatsink.
installed a heatsink on the KT805 transistor.

These changes where done one at a time after which the circuit was monitored for a while.

But is amazing to see how little effect those changes have on the basic operation of this circuit.
Only the run time is influenced by the value of C3, and changing the ringing frequency makes little difference while the PRF stays around 2Hz.

Itsu

Verpies	Re: Tiger's Blocking Generator replication « Reply #72 on: 2021-10-01, 11:43:41 »
Group: Administrator Hero Member Posts: 4527	Looks like you have precisely replicated Tiger's circuit (including the winding layout). At this point, the only parameter that I can suggest further experimentation with is the ringing frequency by varying the C1 and/or the air gap between ferrite's halves.

Itsu

Re: Tiger's Blocking Generator replication « Reply #73 on: 2021-10-01, 15:36:32 »

Group: Moderator
Hero Member

Posts: 4704

Yes, it seems C1 (ringing frequency) is a parameter which can be tweaked to sync f.i. with the ferrite resonance frequency

Concerning the below diagram, when putting 12V on the input cap C2 (right), it takes some time for this 12V to "reach" output cap C3.
When tapping only, it will not load C3 more then a few volts.
It will take some 10s or so for C3 to reach 6V.

This loading of C3 is done via inductive coupling from L2 to L3/L4 as the blocking diode in the plus lead prevents direct loading of C3.

So to see in Tigers video his voltage rise / stabilizing to 6/7V with only a short connection of the battery on C2 (is there a C2 there?) raises some eyebrows here.

A quick way to load C3 to above 6V is to connect the battery directly across C3 and let it run from there.

Itsu

Tiger's Blocking Generator replication

Tiger diagram mod.png (886.54 kB, 1064x611 - viewed 506 times.)

Centraflow

Re: Tiger's Blocking Generator replication « Reply #74 on: 2021-10-01, 18:33:59 »

Group: Experimentalist
Hero Member

Posts: 3914

I know you want to keep to the Tiger circuit, but why has he put the choke where it is.

Please see the attached

Regards

Mike

Tiger's Blocking Generator replication

IMG_20211001_192910.jpg (931.01 kB, 1840x3264 - viewed 469 times.)

---------------------------

"All truth passes through three stages. First, it is ridiculed, second it is violently opposed, and third, it is accepted as self-evident."
Arthur Schopenhauer, Philosopher, 1788-1860

As a general rule, the most successful person in life is the person that has the best information.