Coaxial LC design for pulse compression using a moving short / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

Coaxial LC design for pulse compression using a moving short

1 2 3 4 next last

Move Thread

LAN_403

Shrad

Wed Mar 04 2015, 07:57AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

The idea is to discharge a capacitor in an inductor using a ferromagnetic moving part to short the circuit and discharge the energy stored in the capacitor

The principle would be that it we can manage to make the discharge duration equal to the time the moving part will take to move from one end to the other of L, and L would generate a field which would suck the moving part along L, the moving part would accelerate during the discharge time due to L value reducing and field augmenting

That is valid if the circuit used to make the discharge time long enough for the projectile to move to the end of L keeps a somewhat constant energy flowing during the discharge

Would it be practical to design such a circuit? my opinion is yes, with L and C in a PFN design
Would it be able to produce a sufficient current during all the discharge? my opinion is no, but a sufficient current might be achieved with high C and clever design (thus coaxial)
Would it be applicable to a coilgun design? in my opinion, a coaxial L and C with tubular design is close enough of a barrel that it would be realistic

If my idea is correct, the pulse compression would accelerate the projectile to the end with the L center always leading the moving part thus no suckback, and the final EMP would be of inverse direction to a moving part then placed after the field, so being further repelled from the barrel end

Anyone wants to discuss this constructively?

Added bonus if you can form the pulse to propagate in the L C coaxial line at the projectile speed ;)

DerAlbi

Wed Mar 04 2015, 09:01AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Ok, lets be constructive. (...and state the obvious again..)

The principle would be that it we can manage to make the discharge duration equal to the time the moving part will take to move from one end to the other of L

Ok. Basic multistage design.

Lets think about the mechanics:

using a ferromagnetic moving part to short the circuit

Thats describing a mechanical switch with the Projectuile as the closing mechanism.
The plasma and arcing of such a rubbing switch leads to rapid degeneration of the device equal to the Railgun-Problem. It also causes some friction. The flexibility of the whipers make the projectile additional sussceptable to being asymetric inside the coil wich causes additional friction (in massive proportions) and may even make the switch unrelieable. (When the projectile is forced to one side of the barrel, you need to make sure the other side still connects). To circumvent this, you could make the contact wipers tighter, thus increasing friction and wear&tear again.

Would it be practical to design such a circuit? my opinion is yes, with L and C in a PFN design

See above. The Switch is allready a big ESR. Its therefore not only a transmission line, but a lossy transmission line. (comparable to an oscilloscope probe cable). Math gets more complicated. So lets avoid it ^_^ Also you state yourself that you give bonus to everyone who could slow down the Pulse enough. And still you say "YES, its practical", with the knowledge that this problem is extremely hard to solve. This allone speaks against "practical".

Would it be able to produce a sufficient current during all the discharge? my opinion is no, but a sufficient current might be achieved with high C and clever design (thus coaxial)

Coaxial is not the best formfactor for high energy density (that would be Multilayer). You want high C and tend to something that inherently gives a quite low C per volume and weight? You wanted to be constructive! Please explain.

Would it be applicable to a coilgun design? in my opinion, a coaxial L and C with tubular design is close enough of a barrel that it would be realistic

As i understand you would wrap the coax around the barrel as we do it with copper to form the coil? If you speak of PFN, you would know that you might be confronted with contradicting goals here... the geometry of the Coax forms the pulse, so you cant be allways be effectively close to the barrel, since you need to vary in thickness while the optimal thicknes for couppling would be 0.
Next thing is that Coax guides a EM-Wave INSIDE the coax. there is no external magnetic nor electric field to attract the projectile. The external magnetic field would depend on the ESR, which you need to make big to have an outside effect. Yet you actually want 0 ESR. Hmmh.

Anyone wants to discuss this constructively?

Sry dude, constructively is different from "say what i want to hear". It IS productive to smash the idea.

Added bonus if you can form the pulse to propagate in the L C coaxial line at the projectile speed ;)

Formulas for this are on wikipedia. Present your ballpark calculations and lets discuss them! You said allready such a circuit would be practiacal. Sure this opinion is based on something solid?

See i was nice. You please be nice too and solve the problems above with reasonable answers WITHOUT guessing. Maybe if i got the Coax-Coil deisgn wrong, explain it further along the solutions for the mechanical switch problems and capacitor design reasons.

Shrad

Wed Mar 04 2015, 11:44AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

I was not mentioning multistage design

I was also refering to the inductor being coaxial (in mechanical terms) to the capacitor, so that the magnetic field would be generated by the inductor

for the capacitor lets imagine two cylindrical electrodes of 5 and 3 centimeters in diameter and 20 centimeters in length which are coaxial mechanically speaking... that makes 1.7nF according to http://www.ajdesigner.com/phpcapacitor/cylindrical_capacitor_equation.php with a permitivity of 710 according to http://www.physics.usyd.edu.au/teach_res/db/d0006c.htm

that is, for me, a practical high voltage capacitor with high C

for the inductor I took a diameter of 2.5cm and a length of 20cm with 50 turns of 3mm diameter copper wire which according to http://hamwaves.com/antennas/inductance.html gives an inductance value of 6.8ÂµH approximately (and 8.9*10^-5 ohms R)

that makes for a decent inductor which can be placed inside the inner capacitor electrode in a coaxial position

now that capacitor, at 50kV, will store approximately 2 joules in an ideal world, so lets say 1.6 joules for the 80% storage which is more realistic

discharge that in a 6.8ÂµH inductor which I simulated here :

internal resistance of the capacitor is derived from a flat bar of width equal to half the cylinder cross sectional perimeter (as if it was flattened) to ease calculation, and I will take 1mm width for the material

outer = 15cm perimeter so 7.5cm wide and 2mm width which gives 0.0000222 Ohms
inner = 8.8cm perimeter so 4.4cm wide and 2mm width which gives 0.0000379 Ohms
total is 0.0000601 Ohms (source for bar resistance is http://www.eeweb.com/toolbox/trace-resistance which applies nicely)

so far I'm trying to simulate the capacitor discharge in ltspice but I have strange behavior if I set .IC V(capacitor node) 50k as it gives me a damped oscillation at the resonant frequency of gigavolts, so it will have to wait for further reasoning ;)

DerAlbi

Wed Mar 04 2015, 12:03PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

I compare it to a multistage design, because its also the conept of a moving magnetic field.

Back to Topic: first of all, your C=1.7nF and L = 6.8uH gives you a resonant frequency of about 1.5MHz.
Please use

and put in 1500 kHz and press calculate. Then redesign your material and reevaluate your resistances.

For LTSpuice: make sure you use "uic" in the Run command. you need no voltage sources.
your cap capacity entry should be "1.7n IC=50k".
Please also notice that LTSpice assumes a default ESR in Caps and Inductors of 1mOhm if you leave this field empty. Replace this with the values you recalculated using the equivalent skin depth.

Also notice that your capacitor wont work as expected. since the coil is at the potential of at least one plates/zylinders the shpae you use to calculate is wrong. The other plate is effectively only as far away as the coils diameter is away from the plate.

Shrad

Wed Mar 04 2015, 01:05PM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

DerAlbi wrote ...

Back to Topic: first of all, your C=1.7nF and L = 6.8uH gives you a resonant frequency of about 1.5MHz.
Please use

and put in 1500 kHz and press calculate. Then redesign your material and reevaluate your resistances.

how would you reevaluate that resistance based on which redesign criterias? do you mean a 0.1mm wire would be the one that matches the R I have in mind?

If I recalculate the inductor for 1.5Mhz wihch I missed first I have a reactance of approx. 60 ohms and an equivalent resistance of approx. 0.3 ohm

Can you please explain the effect it has on the discharge?

DerAlbi wrote ...

For LTSpuice: make sure you use "uic" in the Run command. you need no voltage sources.
your cap capacity entry should be "1.7n IC=50k".
Please also notice that LTSpice assumes a default ESR in Caps and Inductors of 1mOhm if you leave this field empty. Replace this with the values you recalculated using the equivalent skin depth.

I'm OK with the default values ltspice will give and will change that, thanks

with those parameters I have a 800A current through L, but I still have ringing which, I think, would not occur if the mobile short was progressing inside the inductor

Do you have any idea of the way I could decrease L in time according to the moving short in a practical way?

DerAlbi wrote ...

Also notice that your capacitor wont work as expected. since the coil is at the potential of at least one plates/zylinders the shpae you use to calculate is wrong. The other plate is effectively only as far away as the coils diameter is away from the plate.

I don't get what you mean there

I'm totally open to have my idea crushed if it's in a constructive way, as you can see, and will use this to progress ;)

DÃ©dÃ©!

Wed Mar 04 2015, 02:28PM

Registered Member #4932 Joined: Thu May 17 2012, 01:42PM
Location:
Posts: 59

Shrad, why would you use such a large distance between the inner layer and the outer layer of the water capacitor? distilled water has a dielectric strength of 65-70kv/mm so a few millimeters in difference should do it. If you also decrease the size of the inner tube and use multiple layers for the capacitor, you would get a way higher capacitance.

Shrad

Wed Mar 04 2015, 03:10PM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

that would also do the trick, I must admit I chose the values totally arbitrarily... and also I think my mind did the shortcut of a single layer capacitor due to size and weight constraints of a barrel, but a couple layers wouldn't hurt...

what is not practical though is that your endcaps have to be of the same dielectric strength (maybe PTFE insulator material would suit this use, but not standard machine grade)

DerAlbi

Wed Mar 04 2015, 04:01PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Skineffect or better the skindepth gives you an equivalent thicknes of the conducting layer. No matter how thick your copper is, if you have 1.5MHz the resistance of a round wire is then the same as a holow cylinder with a wall thickness of the skindepth. (for a round conductor of course.
On a PCB-ground plane for example if the frequency is high enough there is no point in calculating trance resistance based on the full 35Âµm copper.. its the skindepth that counts. Same with your 3mm wire. Its more like a 3mm cylinder with a quite thin wall.
For the wire in the coil you could counteract that by making many parallel (isloated) wires. But you dont get the copper fill that way so your resistance is at least 130% of that what you expect.- practically its maybe 200%.
But it has huge impact on your capacitor ESR. There you cant have many parallel plates in your coaxial design which makes it way more resistive.
in other word:
"outer = 15cm perimeter so 7.5cm wide and 2mm width which gives 0.0000222 Ohms"
that 2mm assumption is nothing more than a wish! use the skin depth.

Your capacitor is calculated wrong, because the coil in beween the plates will have a potential too! The effective capacitance is then not caluclated by the distance of the 2 plates, its caluclated by the distance of one plate to the nearest inductor outline.
If you still dont understand: think about it this way: you have a conducting stuff inside the dielectric. That must do something weird to your capacitor and cant be ignored therefore.

Now lets think about the concept a little bit more: youve got 2 Joules stored and your half wafe is 0.3us. for a 20cm length coil. That makes 10cm acceleration distance (neglet the moving short, thats a bonus). That proposes a constant acceleration (if the world would be ideal

) of 1.05e6 m/sÂ² and an end velocity of 0.315m/s. LOL

if you have a 100% conversion efficiency you would launch a projectile that is as heavy as 40kg.

Of course thats only with the most perfect ridicules assumtions which are far off the practical world. But as you can see: the ballpark figues are completely ridicules.

The problem is here, tha you start completely wrong.
You should define a Projectile and an energy specification first. Then work out which coil shape would be the best for the projectile and the required energy transder and THEN think about how to apply your concept. However i think the last step wont be so successfull. The stuff you can realize in your setup is by it own far away from that whats known to be efficient.

Andy

Thu Mar 05 2015, 01:22AM

Registered Member #4266 Joined: Fri Dec 16 2011, 03:15AM
Location:
Posts: 874

Shrad
This might work for the timing, the outer coils are open circuit when the inner coil fires they build up voltage triggering the spark cap, feeding power back into the discharge.

1425518525 4266 FT169480 Img 20150305 142244

DerAlbi

Thu Mar 05 2015, 02:21AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

A spark gap is the same as a SCR. thats now the common multistage design. The only difference is that the sparkgap lossier, louder and degenerating rapidly.
Thats a completely unecessary step as long as you havent done the ballpark calculations stating that you cant achieve your goals with 6.5kV SCRs and REALLY need more than this.

Sry, to throw in practical considerations again.. starting to be the ass for the second time -.- . Proofe me wrong before complaining please

1 2 3 4 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob