PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2024-11-26, 14:45:17
News: Forum TIP:
The SHOUT BOX deletes messages after 3 hours. It is NOT meant to have lengthy conversations in. Use the Chat feature instead.

Pages: 1 2 [3] 4 5 6 7 8
Author Topic: Replication of ION's bifilar  (Read 96889 times)
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Sorry to hear that ION, Thanks for the warning.

I am unable to find the requested diagram this time  :(

My pot smoked last night, too low wattage, cannot seem to get a 1 watt easily without a min £££ order, any way i will introduce your zener mod tonight, this will at least cut the high voltage down across a new pot.


Your pot should have been current limited if you used the 100K safety resistor shown in the schematic. I always add some current limit to prevent burning up parts. Hope your diac survived.

I'm going to rig up a robust variac based power supply for future testing, but will be down for a while.

I get a lot of parts from old radios, TV chassis, and monitors, this keeps the budget down and not such a loss if I smoke a part. I just ripped down an old Dumont chassis from the early 50's. What a wealth o parts, including a large 4 Henry choke, monster power transformer, numerous tuning coils,lots of pots, assorted values, tubes, and by the way, the vertical output transformer is very close to that used in the Hendershot machine.

Don't worry about the schematic, I'm close to having it running again.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Shame i am not in the TV game anymore, i wish i had kept some of the chassis i worked on during early 80's and even kept a few set's and early video going as well, we would hire a large van once in a while and take a trip to dump, looking back even some of these were real gems and no where to be found these days.

The preset pot i used was under rated even with the 100K limiting the current, I grabbed some beefy zeners and resistors this morning, luckily i ordered 5 diacs and 5 thyristors knowing my previous history  :) and they were cheap enough from ebay anyway, so will start a new build and just use my transformer.

having reviewed your early posts, i realize now my pulse was way too large something around 40uS if i remember correctly, yours is low uS so this is my first task to get this down, it maybe a 40uS pulse is quiet wide with respect to it travelling down the wire to cause localized repulsion.

Hopefully my power supply is quiet robust, i am using a 240-18V then connecting a 25-240V, so each primary is connected together, this lowers the UK rectified DC from 330 to about 250VDC and gives me isolation as well, i stick a rectifier and reservoir cap on the final winding and have primary and secondary fuses, but the variac would be best and then i could vary as well, although i may try the mag controlled dimmer to vary yet, just trying to see the effect first before i need to start adjustments, i also cannot see any reason why i could not use a cheap inductor hardened dimmer to drive my transformer primary.
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Circuit works well ION used 21K for 250VDC supply
Frequency varies between 10Hz and 1kHz
I am using 50nF for C2 to try and bring my pulse width down which it has but not as much as i had hoped @ 17-18uS under the curve.

Is it normal to get that big negative pulse at the start of the ringing.
I am getting about 5 volts neg pulse which is about 61Amps, i can see the wire wriggling slightly, maybe my amps are down because of my low C2.

that neg pulse at the start of the flyback is bigger and shorter @ 8V or 98amps.

Also when i turn the pot and start getting near the end of travel the frequency increases quiet fast up to 1kHz i can hear this audibly and the secondary amplitude starts getting bigger.

My question is how do i get my pulse sharper.

Maybe lowering my bifilar wire diameter will allow me to see more movement, maybe i need to try another transformer with more primary turns or run at a higher supply voltage.

Using my Chan2 probe not connected to anything,bearing in mind the other probe has a common ground, when i sniff the waveform by holding chan2 probe near the bifilar, i get an inversion of the waveform when near the transformer end and as i move away a few inches the waveform starts flipping over until the waveform matches the secondary waveform, very strange to see this, maybe i will remember to bring my camera home and shoot a vid of it.
   
Group: Guest
Peterae,

What you describe matches the actions found when suffering the proximity effect. This will limit the wire movements and cause higher wire resistance. When you hit the loop with a high pulse the wires may be forced together.

In that case, try hitting the wire with less energy, initially, or increasing the gap between the wires.

Good luck to EM on calculating the eddy currents and wire attraction over repulsion  ;)

 
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Peterae said:

Quote
My question is how do i get my pulse sharper.

Maybe lowering my bifilar wire diameter will allow me to see more movement, maybe i need to try another transformer with more primary turns or run at a higher supply voltage.

Do not add more turns to your matching transformer primary. This is the wrong direction. Fewer turns on the primary and secondary will sharpen the pulse (again respecting turns ratio)

I suggest the next time you wind a transformer, put taps on the primary (and maybe a few on the secondary so that you can experiment with different ratio's and overall inductance.

You will not see what I have been seeing, because I think I have identified the factor that has been creating the violent excursions.

It's a bit complicated to explain, but had to do with a resonant interaction between my choke and power supply. It can be simulated with a few extra parts. I will try t0 draw something up to show what has been happening.

It is exciting because there is a real interaction between the change of inductance in the vibrating wires, some return of energy that seems to pump up a resonant circuit formed by the choke and the 1uF 400 Volt capacitor that is internal to my power supply.

This resonance would build and took out the power transistor in the series regulator circuit of my power supply.

You may be able to simulate this effect by decoupling your power supply with a diode and putting a 1uF cap to ground on the circuit side. The cap should be of the highest quality, like a Vitamin Q from Sprague, definitely a high quality film type, not electrolytic. Orange drops may work.

I'll keep you up to date and draw something up.



---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Peterae:

Here is the circuit I am now using to simulate the effect I was getting with my bench supply.

It features a decoupling diode, D5, which allows the choke and the 1 uf cap C4 to build resonance. This is the key to the bursts and violent wire oscillations.

Also features a zener pedestal that starts the CCW of the pot close to the trigger point of the Diac. You can stack some 16 V zeners or any combination to get the 32 volts for D2. D3 is not critical and determines the frequency adj range along with R3. Pot R1 can be anything that does steal too much current from D3, generally 5K or more.

R2 needs to be sized according to your desired circuit operating voltage, so that the zeners are not starved or dissipating too much power. e.g 5mA across 32 V= 160 mW of power dissipated in the zener

Also  a protective lamp on the primary, which prevents blowing a fuse should your SCR latch up. It is a handy indicator lamp that gives you a rough idea of your power draw. Of course a Kill A Watt meter on the input would also be helpful.

Feel free to ask questions. I'll have an update with more findings soon.

« Last Edit: 2011-10-16, 15:45:21 by ION »


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Quote
Do not add more turns to your matching transformer primary. This is the wrong direction. Fewer turns on the primary and secondary will sharpen the pulse (again respecting turns ratio)

I suggest the next time you wind a transformer, put taps on the primary (and maybe a few on the secondary so that you can experiment with different ratio's and overall inductance.

Thanks for this advice ION, i am in the process of winding another coil.

Quote
You will not see what I have been seeing, because I think I have identified the factor that has been creating the violent excursions.

Fabulous

   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
ION said:
Quote
You will not see what I have been seeing, because I think I have identified the factor that has been creating the violent excursions.

Bet don't be dismayed. You can simulate my bench supply by adding the decoupling diode and 1 uf 400 volt cap as shown in the schematic.

I've also found that it works better with raw pulsating DC so you can eliminate the 100 uF 450 volt shown in the schematic. There are some beat frequencies with the pulsating DC but this does not seem to hurt and makes the burst effect easily obtainable.

It also tends to prevent lockup of the SCR. A lamp on the supply side is a good indicator of lockup and gives the energy somewhere to go when the SCR does lock on, so you won't have nuisance fuse blowing. 60 to 75 Watts seems to work nicely on my setup.

The system works best when there are a burst of about ten cycles and then a relaxation time. It seems a single pulse with a relaxation time or off time (low repetition rate) does not produce the effect but is required to initiate auto trigger when the pulse amplitude is high enough.

I believe the burst occurs when the resonance in the choke capacitor combo auto triggers the SCR. Normal background pulsing at a low rate is needed to to trigger the effect.

Continuous pulsing at a higher rate does not  produce the effect.

Those first few cycles in the burst produce the violent wire movement. And the relaxation time is definitely needed.



What do I think is happening? Not sure but I have a theory.

If the free electrons have time to precess and align with a external magnetic field, the effect seems to occur.

When hit with repetitive pulses, they do not have time to precess to their rest position. The violence of movement is so high, I cannot account for it occurring via normal repulsion effects especially since it does not occur with single pulses of low repetition rate..

This could be the kick SM talked about but I will have to try it with a single wire to be rule out repulsion effects.

I will also try a wider pulse that equals the width of the burst.

I know I have been changing parameters, but it has taken me some time to identify the mechanisms and effects.

Please be patient. I know this is a bit confusing. More to come.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
I thankyou ION for bothering to put the time into keeping us informed and design circuits to build, very rare anyone would bother.

You know your setup reminds me of a valve amp stage in which SM could well have found the original principle of operation, i say this because of your large series choke, your highish to low impedance output transformer or should i say driving with high voltage, i have suspected for some time that any excess energy would manifest on the primary side of the drive transformer, because of his saying the device starts to feed itself.

Be safe.

Getting a little worried the wrong people may see your posts, just in case you are on the right path.Although in my bench here, guests don't get to see it only members.

   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
I have had to take a few days away from this, but just ordered a variac which i have been meaning to acquire for many years.
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
OK picking up on this again now i have my Variac, just ordering some 32V Zeners.

For the new build.

   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
OK picking up on this again now i have my Variac, just ordering some 32V Zeners.

For the new build.



Peterae:

My most recent experiments showed that the choke can be eliminated and replaced with a power resistor. The value and power will depend on the desired repetition rate.

The triggering is now more stable since there is no choke to cause ringing.

The disadvantage is that a lot of power is dissipated in the resistor.

 For precise energy measurements  into the wire, do not include the resistor in the calculations, just the peak charge voltage  into the capacitor squared x capacitance /2


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Tinkerer
Hero Member
*****

Posts: 3948
tExB=qr

What do I think is happening? Not sure but I have a theory.

If the free electrons have time to precess and align with a external magnetic field, the effect seems to occur.

When hit with repetitive pulses, they do not have time to precess to their rest position. The violence of movement is so high, I cannot account for it occurring via normal repulsion effects especially since it does not occur with single pulses of low repetition rate..

This could be the kick SM talked about but I will have to try it with a single wire to be rule out repulsion effects.

I will also try a wider pulse that equals the width of the burst.

I know I have been changing parameters, but it has taken me some time to identify the mechanisms and effects.

Please be patient. I know this is a bit confusing. More to come.

(Disclaimer: This is a toungue-in-cheak explanation of what I have come to believe.)

During the pulse, an incredible number of particles (virtual or real) are responding to the change, each effecting the ones near to itself.  So, for a brief instant, the entire local universe responds to a small input with a large output. 
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Thanks ION no problem with the resistor, it should make life a lot easier to work out incoming and outgoing energy to the system from the reservoir cap, i cannot wait to see the + and - peaks across that series feed resistor, if i can get it working this time  O0

   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Almost built now, got the variac rigged in series with a  plug-in 40watt bulb so i can try different wattage bulbs, this then feeds my 240-24 18-240 transformers, then bridge then 2 x 470uF 450V in series smoothing caps, giving me 0-280V variable.

i have used a 1n4007 for the diode and a high quality film 1uF 1000V cap.

Zeners are 4 in series 16V 5 watt each, 21K series fed, 10K lin pot.

Just need to fit the zeners tomorrow and i should be up and running.

Not using a resistor at the moment still using my MOT.

Once i know i am up and running i will concentrate on finishing my new pulse transformer.
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Circuit is now running.

Just taken some measurements of my bifilar, as i have the setup at home now and at work i was getting noise which i believed to be interfering with the LCR meters readings.

This time i have better results
0.1710 Ohms
7.8uH @ 10kHz
5.5uH @ 10kHz separated by fingers.

Using my 10K pot set to min frequency i was getting bifmin1 & bifmin2
using my 10K pot set to max i get a dim lighting of the 40watt bulb in series with the variac and get bifmax1 & bifmax2
I was still using my MOT for the inductance and not yet using a resistor, i have a 1uF film 1000V cap after the diode

Still not yet seeing any violent movement of the bifilar wire.

I am still yet to finish my new pulse transformer.

So bifmin is 4.2V so I=4.2/0.1710 = 24.56 Amps  :-[ @ 307Hz
& bifmax is 5.6V so I=5.6/0.1710 = 32.75 Amps  @ 1.75 kHz

I now think that something was scewing my LCR measurements at work, may have been the 20 or so high wattage CFL's i run in the gallery.

So it would appear it is now important for me to work out why my current is low, just checked and it looks like i have c3 @ 33nF so this maybe a bit low, i must try and find a few more higher voltage caps to drop into here, this cap must control the energy that's dumped into the primary of the pulse transformer.

   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Peterae:

The violent movement of the wires in my setup was caused by my SCR, an 800 Volt device being triggered into avalanche mode at around 1000 volts.

Where did this voltage come from since my power supply was set to a much lower voltage? It was identified as the peak voltage created by a normal trigger pulse shocking the inductor and 1.0 uf capacitor into a ringing condition.

The very first peak of the ringing cycle was over 1000 volts and  avalanched the SCR with a burst of pulses until the energy stored in the LC tank was reduced below avalanche threshold..

The signature of the pulses was the dead giveaway as they were neatly centered around the ringing peak and each pulse of the burst showed a mark to space ratio modulated by the rounded first peak of the ringing waveform.

I have since destroyed a few SCR's trying to get a continuous avalanche condition using just a higher voltage current limited power supply. Obviously the SCR's  I am using cannot withstand the severe avalanche duty. They were 400 and 800 volt devices. Fortunately I have a quantity of these on hand so could afford to waste a few.

During avalanche, current was 3X normal for the pulses into the wire as could be expected.

I have drawn this out in a timing diagram posted below. Waveforms are approximate, just trying to get the idea across.

V thr1 and V thr 2 represent two different values of power supply setting, V thr 2 taking the system closer to avalanche.

Hope this helps to understand the mechanism causing the violent excursions. Oddly I did not see this effect with continuous pulsing at high current avalanche mode. Nor did I see it at lower current levels sweeping through the same range of frequencies.

  I am inclined to believe a relaxation time is necessary in the wire, hence this is an interesting accidental discovery, and supports the notion "when current first flows in a wire, there is a kick". More research is certainly needed to prove or disprove this effect.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
So if my understanding is correct, you are firing your Thyristor this is causing ringing in the choke, the first ring is of such a high voltage that it is causing a re-trigger of the thyristor, because the voltage is much higher the second re-trigger is 3 x the original current.

So is my first ring negative and opposite to your first ring being positive?

I can just feel the urge to have 2 thyristors firing with a variable delay between them, or a delay in one leg of the bifilar  :)

edit I have just located a 100nF HV cap so will try that when i get home for C3

Quote
I am inclined to believe a relaxation time is necessary in the wire, hence this is an interesting accidental discovery, and supports the notion "when current first flows in a wire, there is a kick". More research is certainly needed to prove or disprove this effect.

A delayed second pulse  :)
« Last Edit: 2011-11-01, 15:49:16 by Peterae »
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down

From Peterae:

Quote
So if my understanding is correct, you are firing your Thyristor this is causing ringing in the choke, the first ring is of such a high voltage that it is causing a re-trigger of the thyristor, because the voltage is much higher the second re-trigger is 3 x the original current.

So is my first ring negative and opposite to your first ring being positive?

At the first pulse, it is the inability of the inductor to keep supplying current that causes a dip in the power supplied to the SCR. This is the perturbation that causes the ringing. The start of the ringing is from the negative as seen in your scope shot and my drawing. It must go positive and exceed the avalanche point of the SCR in order to get successive retriggers. Naturally this must occur on the positive ringing first peak.

I have not definitively found  out if a series of sharp kicks followed by a relaxation time is the only way to get the violent excursions. A precision burst pulse generator would seem to be in order. We can gate a transistor fed from an adjustable diac pulse generator. The gating could come from an external pulse generator.

Factors to investigate:

1) Required relaxation time (off time of the gating pulse)

2) Number of pulses required during the gating period.

3) Amplitude and duration of the pulses in the gating period.

4) On time of the gating period.

5) A precision way to measure the amplitude of wire vibration.

A lot on that plate, but as I say, worthy of research.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
A lot of work there ION

I have now fitted the 100nF in place for C3.

2 waveforms as follows bifmin3 which is the pot set to minimun, and bifup1 which is when i turn the pot up slightly, the squareness of the waveform i guess is saturation of my core.

What i am seeing now is a linear separation of the bifilar wires as i turn the pot up and the frequency increases, the bifilar starts getting hot.

So bigmin3 4.8V I=4.8V/0.1710 = 28.07A
bifup1 14.8V I=86A

The first ring of the bad waveform is 17.6V = 103A

Wow that's quiet an increase in current when i turn the pot up and get the bad looking waveform.
« Last Edit: 2011-11-01, 22:21:32 by Peterae »
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
Quote
At the first pulse, it is the inability of the inductor to keep supplying current that causes a dip in the power supplied to the SCR. This is the perturbation that causes the ringing. The start of the ringing is from the negative as seen in your scope shot and my drawing. It must go positive and exceed the avalanche point of the SCR in order to get successive re-triggers. Naturally this must occur on the positive ringing first peak.

Thankyou for explaining this to me.

Interestingly the Avalanche voltage can be controlled by biasing the gate with a dc voltage
« Last Edit: 2011-11-01, 22:01:00 by Peterae »
   

Group: Tinkerer
Hero Member
*****

Posts: 3948
tExB=qr
Do SCR's have a secondary avalanche mode that increases current like transistors do?  (they are destroyed if the current is too high)
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Do SCR's have a secondary avalanche mode that increases current like transistors do?  (they are destroyed if the current is too high)

I have not studied this in any depth so cannot answer your question.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
Group: Elite
Hero Member
******

Posts: 3537
It's turtles all the way down
Thankyou for explaining this to me.

Interestingly the Avalanche voltage can be controlled by biasing the gate with a dc voltage

I have not tried this so cannot comment. Usually the SCR will triggrer on when the gate exceeds the cathode by  around 0.5 to 0.6 volts. I have not applied a bias to see if the avalanche can be controlled.

The specs for most SCR's specify a gate pulse risetime and gate hold current to sustain turn-on.

Also there is the minimum holding current required to latch the SCR "on". All of these parameters vary depending on the size and current carrying capacity of the SCR.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   

Group: Administrator
Hero Member
*****

Posts: 3960


Buy me some coffee
OK got the new coil wound, 210 turns primary, 10 turns secondary.
you can see the core i used here click specification to see it's dimensions, it's part code is HEM3016 the bottom one big ring
http://www.maplin.co.uk/ferrite-rings-and-beads-32792

I cannot explain the waveform, mostly due to core saturation i guess, anyway i certainly have secondary current now  C.C

weird how the ringing is at the start of conduction  C.C

anyway i now have 38V on the secondary and i make that I=38/0.171 = 222Amps  :)

I have had latch up once as well, the bulb works well thanks ION that's one Thyristor saved.

I am still not getting movement of the bifilar though, i have seen some strange oscillation burst a couple of times but have not worked out how to get them to appear, i was playing with core saturation with neo's on the toroidal core at the time, i was getting port corruption as well on my laptop and DSO at the time so it may have just been some bogus data.

I must bring my HV probes home and see whats happening on the primary side, don't want to risk a x10 probe on there as i don't know what to expect right now and could blow a Chan if it's high.
   
Pages: 1 2 [3] 4 5 6 7 8
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2024-11-26, 14:45:17