Replication of ION's bifilar

[Peterae]

To keep your thread clear of my attempt, i am starting here.

I see the problem you mention about me driving this with my fet stage, i have very little current drive ability, and as you say i would need to construct a transformer and drive with a high voltage.

I am now thinking that maybe i can use 2 IXTK80N25 in parallel and drive from a fet driver, these fets are rated at 250V 80A,
Connect the source pin straight to one side of the bifilar, connect a low esr across the drain(GND) and the power supply and other end of the bifilar.


Can i ask, can you see any problems with me doing it this way.

EDIT Is there anything i need to do, to stop the smoke, i will heat sink both together


For a heat test, it would be possible to place the bifilar in a hose and blow air down one end and measure the temperature of the air at the other, we could then compare this to another tube same length with a resistor and then pump power into the resistor until we get the same temperature out the other end.

[ion]

The part you mention is good for 250 amps in pulse mode, limited by the junction temperature T jm. I would start with a single device, 1 uS pulse width  and keep the pulse repetition rate low as you monitor the junction  temp. !50 C is the maximum.

Paralleling the FETS may cause current hogging in one unless equalizing resistors are used.....a waste of pulse energy.

I would play it safe and use a matching transformer. My CMC with 10 turns added equaled 13.3 to 1 current transformation. This is far less stressful to the driving device

You have some good ideas for the heat test.  Good luck with your build. If you could whip up a schematic, I can comment further.

Off to do more research.

[Peterae]

OK lashed out on some 20 SWG enamelled wire this is a bit thicker than 20 AWG equivalent to 19 AWG

SWG  CSA                Resistance   Diameter
20    0.6567mm²    0.026 3    0.914mm

I have now measured the resistance of my bifilar at 0.0817 Ohms

I will use a ferrite ring to wind my transformer, i will try to match my secondary turns resistance to  the bifilar resistance, once i know how many turns that takes i can work out my primary turns ratio.

Still looking for a thyristor, not had much time yet.

My core is OD 30mm ID 20mm height 22mm

OK i have 8 Turns for the secondary which measures 0.0803 Ohms
106 turns secondary
Having measured this i seem to have 1.865 Ohms which is a ratio of 23.225 so not quiet sure whats going on here as i have rechecked the primary resistance and it's correct. How can i be so far out(maybe there was 7 turns on the primary but even this would be wrong, i suspect theres something going on with this meter).I used the same spool of wire.

So using your 13.3 ratio i would need 106 turns for my primary

I will probably need to wind my own supply choke so i get 200 Ohms, probably use another core the same for this.

It looks like my weapon of choice is the BT151-500 Thyristor

Diacs and thyristors are now ordered.

[Peterae]

This LCR meter of mine uses 1 of 4 frequencies to measure at 100hZ, 120Hz,1kHz and 10kHz, it would appear i measured the resistance of the wire when not on the ferrite core and cut this to match the resistance of the bifilar, now i have wound this on the core it's reading 0.148Ohms

What is going on with this meter, on 10kHz setting i am getting 312 Ohms Grr bearing in mind i don't have it set for impedance but resistance.

I'm confused, it would appear i have a meter that cannot read the correct resistance of a wire.

[ion]

This LCR meter of mine uses 1 of 4 frequencies to measure at 100hZ, 120Hz,1kHz and 10kHz, it would appear i measured the resistance of the wire when not on the ferrite core and cut this to match the resistance of the bifilar, now i have wound this on the core it's reading 0.148Ohms

What is going on with this meter, on 10kHz setting i am getting 312 Ohms Grr bearing in mind i don't have it set for impedance but resistance.

I'm confused, it would appear i have a meter that cannot read the correct resistance of a wire.

Probably be best to just calculate it from the length, as it is not that critical. You don't need to perfectly match the resistance of the secondary to the resistance of the bifilar. It will cause the ohmic losses to be matched meaning half the power will be spent in the ohmic resistance of the bifilar, the other half in the ohmic resistance of the secondary.

For our test purposes, we just want a reasonable current step up into the bifilar. A lower resistance heavy guage secondary will put more of the power into the bifilar, so I would use the heaviest guage you have lying around for the secondary and leave enough room for the primary.

To keep things reasonable, the primary can have an ampacity 10 to 20 times less than the secondary.

I'm going to rewind my secondary with heavier guage, still minding the turns ratio (approx). This will put more of the power into the bifilar as less will be soaked up with secondary resistance and wind up as heat.

Hope this helps

[Peterae]

Thanks ION i definitely have 13.33 turns ratio from primary to secondary, but have used thinner wire on my secondary, so i will switch it over to 20 SWG in that case.

[Peterae]

Here's the coil.

18Swg Primary 106 Turns
20 Swg Secondary Now 9 Turns was 8 of 18Swg

[Peterae]

ION your input supply choke must be huge something like 5H at a guess, these are immensely expensive, can you see any reason why i could not use the primary of a mains transformer for this, only an idea, i will do some measurements tomorrow when i get to work and see what resistance and inductance they are.

This bifilar reminds me of something i once had seen on a work bench, it was audio related and if i remember had long springs down it's length, some a bit older than me and into audio will know what it was, maybe echo,delay, reverb unit or something, i seem to remember it being really long maybe 3 feet, mind you this is years ago.

[ion]

I don't have a way to measure it accurately right now. I suppose I could resonate it with a known value cap and look at the peak resonant rise vs. freq, then work backwards from there.

Edit: It looks like 3.2 H, 60 Hz resonant with a 2.2uF cap

Yes you can use the primary of a small mains transformer, that would work fine.

My choke is very small, like the kind of audio transformer you find mounted on a small PM speaker, 2 x 1.5 x 1.5 inches approx.

By the way the primary of a small audio transformer like this would work just fine.

A larger choke with less resistive loss would be more efficient, but I'm not worrying about circuit power dissipation just yet as I can optimize that later. The important thing is the energy delivered to the load, and this is easily calculated from the capacitor charge, cap value, and pulse repetition rate. Just scoping the input to the bifilar can tell you a lot about the energy transfer.

Here's a pic of the rusty devil

[Peterae]

That's great news ION, I will do some measurements when i get to work.

Thanks for going to the trouble of measuring your transformer.

OK i have found out what the spring acoustic device was i had seen, i don't think it's relevant but was just interested as i knew it used acoustic vibration, turns out it was a reverb unit.

http://www.amplifiedparts.com/tech_corner/spring_reverb_tanks_explained_and_compared

[Peterae]

I measured the inductance of my bifilar

with the meter set to 1kHz setting
L=210uH and Q=0.062 then separated i get L=84.2uH and Q=0.1602

Then i switched to 10kHz and got these results
L= 3.048uH Q= 0.565 then Separated i get 2.94uH Q=1.609

I found it impossible to get a stable reading when the meter was set to 10Hz maybe because of the mains in the room at 50Hz

So although the impedance goes down when separated the Q goes up

[ion]

I measured the inductance of my bifilar

with the meter set to 1kHz setting
L=210uH and Q=0.062 then separated i get L=84.2uH and Q=0.1602

Then i switched to 10kHz and got these results
L= 3.048uH Q= 0.565 then Separated i get 2.94uH Q=1.609

I found it impossible to get a stable reading when the meter was set to 10Hz maybe because of the mains in the room at 50Hz

So although the impedance goes down when separated the Q goes up

Those readings are counter-intuitive and defy theory for sure, e.g. there should be minimum inductance when the wires are parallel and touching, maximum inductance when they are separated to form a loop or rectangle.

What gives?

[Peterae]

Sorry ION

I will recheck tomorrow, maybe i should not have used my fingers to separate the wire, either that or it's the meter, i will try measuring a known inductor just to see whats going on with this dam meter.

EDIT how are you holding the end tight, i have a loop and then come back and twist around the wire, i then thread through this loop and pull, is that acceptable to do it this way

[Peterae]

Not really sure whats going on then, i have rechecked and v=made a video of how the inductance meter reads during separation of the bifilar.

Are my end OK to be terminated in the way i have done, i have used copper rings screwed into a bench, i attach 2 pictures, 1 of each end, i use a second ring at the far end with string to tension the whole length of the bifilar.Maybe it's the rings or my twisted wire that's causing the problem.

PS i have not really tensioned as much as i need to, the wire is really thick and takes quiet a bit to tension it totally, i will tension them really hard when i run experiments.

[BEP]

May I suggest both of you compare the test signals from your inductance meters using the same type and value inductor.

According to the rules these outputs should not make a difference. They do.

Output wave shape and duration differences can show a swap in increase/decrease but only if the relative wire length is a multiple of a 1/4 wavelength. I can't imagine why the two results are opposite.

[Peterae]

Hi WW
I am not sure what process this meter uses to calculate Inductance, in the manual it shows a formulae that it uses, i have scoped across the Test leads while it's measuring and cannot see that much, there is a pulse but very low voltage, my scope is set to 50mV/Div and 5uS/Div timebase, but this is background noise because it's still there when the Inductance meter is disconnected although a little less amplitude, looks to be about 6.6kHz so must be a switched mode psu in action somewhere or something similar, maybe my PC in another room or maybe even the scope?

So basically i cannot see any sort of driven wave from the Inductance meter

EDIT ah this is interesting when i dosconnect the inductance meter and just scope the bifilar, i still get the noise and if i separate the bifilar wires the signal get a bit stronger, so maybe this noise is confusing the meter reading.

[ion]

Peterae

I borrowed that exact meter to run the inductance test on the 58" parallel wires. I don't remember the frequency setting so I will have to borrow it again and re-run the test  at different settings.

Glad to see you are using the Kelvin clips, very important at low values of LC or R.

My wire are looped at the far end around a tiny piece of wood like a toothpick.

I cannot at this time explain the differences, but not to worry, you will have some fun when your SCR's arrive.

We'll have to backtrack over this soon.

[Peterae]

My thyristors have arrived now and the diacs, just need to pick up a pot in the morning and knock it up on a bit of vero, i will try to join as close as possible to the thyristor legs, my biggest problem is time in the next couple of days, flat out at work for the last 2 weeks now 

Not sure if i will be ready until the weekend

My transformer primary measured quiet low under half a Henry, so still not sure what i am going to do here, going to have a rummage at home now to see what other transformers i have.

That meter has 2 versions, it's only the A that does 10kHz, not sure if the other one does 100Hz,120Hz and 1kHz i seem to remember it does.

I was also trying to track that 6kHz signal, i switched of everything i could think of, PC, CCTV supply, credit card machine etc and it was still there, but on the way home i realized i have about 20 high power cfl spot lights on all the time, so will check these in the morning.

[EMdevices]

I'm glad you jumped on this experiment as well Peterae,   may I suggest something for both you and Ion?   

Keep the two wires that repell each other spaced apart by a short distance so they don't hit each other, maybe 1 cm.   The magnetic field generated under those amps will be strong enough to repell them, that should not be the worry.   The worry should be not to have them banging into each other when they recoil.    We want to have a type of stringed instrument where the strings can vibrate at their natural frequency based on the stiffness and mass of the wire.

EM   

[ion]

My thyristors have arrived now and the diacs, just need to pick up a pot in the morning and knock it up on a bit of vero, i will try to join as close as possible to the thyristor legs, my biggest problem is time in the next couple of days, flat out at work for the last 2 weeks now 

Not sure if i will be ready until the weekend

My transformer primary measured quiet low under half a Henry, so still not sure what i am going to do here, going to have a rummage at home now to see what other transformers i have.

That meter has 2 versions, it's only the A that does 10kHz, not sure if the other one does 100Hz,120Hz and 1kHz i seem to remember it does.

I was also trying to track that 6kHz signal, i switched of everything i could think of, PC, CCTV supply, credit card machine etc and it was still there, but on the way home i realized i have about 20 high power cfl spot lights on all the time, so will check these in the morning.

Peterae

Don't worry about the choke so much, nearly any reasonable choke in range of 1-3 Henries will work. I have not tried others, but I'm sure they will work also.

The main function of the choke is to allow the SCR to be reset by the back EMF or ringing in the output inductor. If there is not enough inductance in the choke, the SCR will latch on and you will have to manually interrupt power to get reset. This is one of the drawbacks of the circuit.

I suggest you start at a low voltage value from your power supply say 50 Volts. Also would be valuable to limit the current to around 100 mA. This will protect the SCR for initial tests. You can run the voltage up when you are sure all is working well.

EM

Thanks for the suggestion about the wires spacing.

[Peterae]

I suggest you start at a low voltage value from your power supply say 50 Volts. Also would be valuable to limit the current to around 100 mA. This will protect the SCR for initial tests. You can run the voltage up when you are sure all is working well.

That's good advice thanks ION, i can only go to 125V on my supply anyway so may end up putting more turns on the primary yet.

I will do some initial tests and see where they point me.

[Peterae]

OK Well i built the board and we are up and running, but a couple of things yet, i only had a 330n Cap to start with, and the transformer choke is under 1H and reads 25 Ohms so must find the time to get a better one.

The frequency can be quiet high when i adjust the pot, i am not getting a big kick on the wires but can see them jiggling slightly, so must fix the above problems yet, but at least i am running.

A couple of times at power up i have seen the wire jolt quiet hard.

Run out of time for now

From the scope shots at the lowest frequency i am running at just under 50Hz
Red trace is across the bifilar, the yellow trace is across the primary of my coil.
Red Trace is 1V/DIV & 5mS, Yellow Trace is 50V/DIV on BIF1

On Bif2 the timesbase is at 25uS same voltage settings.

Thyristor is running cool at my max voltage of 125V

EDIT my bifilar is 0.0817 Ohms, looks like i am dumping about 2.5 volts on the scope about 30 Amps  

You know this has really convinced me i was on the right track with the fast delayed pulse stuff i was doing, if i hold a small magnet near the wire with the 30Amps i get a very little tug, well in my pulsers i get a massive tug and don't necessarily have a tight wound coil to feel this, what does that say about the power in my kicks, it will be interesting to compare 100amps when i get there

[Peterae]

I have now changed C2 for 660nF, i only just noticed ION switched to a lower value than the circuit diagram 500nF so it looks like i may have gone the wrong way.


Any way here's 2 scope shots of the waveforms Bif4 & Bif5, Both traces are 50V/Div Bif4 is zoomed in on the timebase.
Red trace is at the choke/capacitor/anode junction with respect to ground
Yellow trace is C2/Transformer primary junction with respect to ground(across primary)

Bif6 is across the secondary 1V/Div

I am still running at a minimun setting which gives me just under 50Hz.

From what i can see the Choke and Capacitor are charging and discharging as expected Bif6 shows i have a 4V pulse which means i am now up to 49Amps and still only seeing a little bit of movement.

What do i need to do now to get my current up, is my choke too small, am i running at too higher frequency, do i need thicker wire on my primary and more turns?

I am running at just below 50Hz min adjusted, What frequency does yours run at, i had the impression you were in the low Hz

Sorry for the questions but the circuit seems to work well, maybe my problem is my supply only goes to 125V

[Peterae]

Bif7 is across the Choke 50V/Div
Bif8 is across the primary 50V/Div

[ion]

Hi Peterae

you asked:

What do i need to do now to get my current up, is my choke too small, am i running at too higher frequency, do i need thicker wire on my primary and more turns?

I am running at just below 50Hz min adjusted, What frequency does yours run at, i had the impression you were in the low Hz

Sorry for the questions but the circuit seems to work well, maybe my problem is my supply only goes to 125V

I will try to answer some of these questions in more detail as time permits, as I have been away from the lab and the forum for a while.

When I get back to it, I will try to make some more exacting measurements of PRR and applied voltage.

Recently I changed the discharge cap to 0.18 uF (measured) and have been able to operate between 300-400 volts and able to tune in to the strange burst mode.

Earlier I was operating at around 162 volts with a 0.47uF cap when I was getting the burst mode, so it appears a certain energy is required on my setup to initiate it.

Note that this will be different for each setup because of wire tension etc.

Also it worked (burst mode)  with my 3 H 200 Ohm choke but not with a lower resistance choke.

I am operating at a higher voltage, but perhaps you can use a larger discharge cap to get more energy, just be careful not to exceed the peak current rating of your SCR.

The burst mode will occur even without the wire attached, but produces extremely powerful vibrations when the wire is attached.

I've concluded the burst mode must be a function of the BEMF recharging the cap or a saturation phenomena of the choke.

In either case, I'm sure there are many ways to pulse the wires that do not involve such a temperamental oscillator system, and perhaps there are better ways of doing this such as your original noise testing setup.

As I recall somewhere between 2-4 msec PRR would trigger the burst mode.

My entire reason for doing these tests was to produce a vibrating wire setup that could later be formed into a circle so that I could try to reproduce the washboard and gyro effects with perhaps multiple PRR. I believe those floppy interior wires held in a foam like soft structure in the SM15 that was cut apart were the vibrating wires that are the source of the physical effects.

In effect I want to work backwards from the real physical effects, not to try to get OU out of the chute so to speak.

Be patient, I will try get more info as my schedule permits. Meanwhile try some higher resistance chokes if you have them handy.