My school coilgun poject / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

My school coilgun poject

first previous 3 4 5 6

Move Thread

LAN_403

DerAlbi

Wed Apr 25 2018, 06:27PM

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

Mighty.
Still dont like the kick-back though...

Whats in the box? This is currently one of my bigger pains.. stopping the projectile.. I use some modeling clay mixed with Kevlar fibers on a hard foam board to stop it.. but that leaves the projectile dirty and penetration is more than one projectile length.. so recovery is frustrating.

Unfortunately i cant read your diagram titles.. is the 3rd picture the individual coil contribution to kinetic energy or is it the overall result?
Also your efficiency worries me.. are you just adjusting voltage or are you also optimizing the switching times for every voltage level [according to the projectile speed]?

Skynet

Wed Apr 25 2018, 09:22PM

Registered Member #11734 Joined: Thu Mar 21 2013, 08:44PM
Location: Brno Czech republic
Posts: 35

Kickback is quite problem now I agree but hopefully some glue shoud fix that damn coils to table :D
In the box are multiple layers of thin cardboard and my scarf to dampen it just a little :D

And in 3rd picture is overall result so if you want just coil 2 you need to deduct energy of coil 1.

About voltage I just giving more charge to capacitors. My switching times are strictly based on projetile position to coil. Dropping effiency at more power is quite weird. It could be relied to poor magnetic coupling or something like that but now I dont have much time to make new coils :D

DerAlbi

Thu Apr 26 2018, 04:58AM

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

Well efficiency in coilguns is a bitch to master :-/

There are a lot of reasons why your efficiency can be messed up.
1) Positional switching is not good enough. It should be more sophisticated. Higher voltage leads to faster current rise times which energized the coil before it can do its most energy transfer.
2) Saturation of course.. based on your geometry and my gut feeling i would expect a coil of yours be efficient up to maybe 5J to 7J energy transfer (however and this is what you are currently doing but already experiencing heavy degradation)
3) Kickback. Moving/Accelerating your heavy coils (4 of them) backwards also consumes energy that is mechanically dumped into your build... that consumes energy that is not in the projectile.. Although the coil movement seems springy the recovery (forward motion) occurs only after the projectile exited, so the energy is truly lost.. i experience the same in my coilgun.. when i shoot from my hand the gun performs measurably worse due to my mushy human body compared to my stand its usually shot from.

Do you have your coil data available? (Turns, Inductance, Resistance?, Switching times of your IGBTs)
Have you made a Spice-Simulation what currents you expect? Just a ballpark figure is enough to help reasoning...
Alternatively you can just watch your cap voltage on an digital oscilloscopic and calculate the current that way...
Or use a shunt.. but that is hard to implement (due to your build)

Skynet

Sun May 20 2018, 05:52PM

Registered Member #11734 Joined: Thu Mar 21 2013, 08:44PM
Location: Brno Czech republic
Posts: 35

Hi,
another update there ;) Ive just managed to fire up all 4 stages :p Still sorting some problems with current sensing to be able use full power. So far best result 41,25 m/s with 22,1 J of energy with coils eff. C1: 6,9% C2: 10,9% C3: 19,4% C4: 45,3%.

Still a lot of work to do but slowly getting to the point that I want hopefully 200m/s would be achievable speed :D

For Albi:
My coils have these parametres:
Coil1: L=663uH, R=240mohm, N= unknown
Coil2: L=385uH, R=170mohm, N= unknown
Coil3: L=267uH, R=129mohm, N= unknown
Coil4: L=242uH, R=110mohm, N= unknown
Switching times are strictly dependend on singnal from front optogate so: dimmed=on free=off. Coil is switched on 2,5mm before projectile reaches coil and 2,5mm before projectile get to center of coil.

I have my simulation and current get to that high value that I need current regulation (easily over 400A) Ive included result ;)

1526838768 11734 FT158837 Wp 20180520 00 06 31 Pro

1526838768 11734 FT158837 Wp 20180520 06 42 38 Pro

1526838768 11734 FT158837 Wp 20180520 14 51 36 Pro

]test.pdf[/file]

DerAlbi

Mon May 21 2018, 05:05AM

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

Thanks for sharing your results. It is getting very interesting; I dont see the 200m/s as realistic but anyway you reached the same level of performance as my gun does. (strictly energy wise). Your coils are noticeably better than mine - your L/R time constant is nearly twice as long as my coils which should give you an edge.
Unfortunately i cant really interpret the first photo because i dont know if thats before or after a shot and if its after i dont know the start voltage nor the capacitance. It would also be interesting to have an overall efficiency (Ekin/Ecap_all) for better comparison - if your first coil has only 7%eff but pushes only 10% of the energy into the projectile the 7% dont mean as much if 90% Ekin is pushed by a 40% eff coil, i hope you understand.

Your simulation seems to be lacking detail (circuitry) - it plainly shows that your capacitor voltage goes negative during your on-cycle which is not possible in a half bridge. Anyway, i get the point.
Edit: stupid me. Its the coil voltage i was complaining about, thats ok! I was confused for a second and judged; i apologize - the cap voltage is fine

I doubt your switching is optimal yet. Its a nice start, but it may be beneficial to experiment more - i would suggest to use the timings the current shots have.. just log them through the optogates, then make it a fixed sequence and then vary parameters. I figure there is a lot of headroom for improvements - which would be very interesting because if you reach the efficiency of my design with much smaller technical investment that would mean a lot (mainly that i am an over-engineering, time wasting idiot, lost in his topic of interest, working under false assumptions, suffering wildly the dunning-kruger effect)

I am not sure if you really should go for current sensing for current regulation because every shot is kind of the same as any other shot but not really [due to initial projectile position variation]. A regulation scheme would only introduce more unknowns and would add exit velocity inconsistencies. I would again suggest you use a fixed (maybe still triggered by the optogates?) switching sequences to simply avoid excessive currents by shorter but repeated on times [PWM]. Current sensing is mainly interesting to compare simulation to reality and to study the impact of the projectile to the real current waveform.

Again, those a very interesting results. Specially the 4th coil amazes me. From what you say i wouldnt have thought it possible to have >40% coil-efficiency with the setup and switching scheme you propose. It kind of humbles the achievements in my gun. Stop working on it! I pay you off; oh wait, i spent all my money on a more sophisticated version of the same thing

Skynet

Tue May 22 2018, 08:08PM

Registered Member #11734 Joined: Thu Mar 21 2013, 08:44PM
Location: Brno Czech republic
Posts: 35

Hi,
dont border about first photo its just showing function for cap carger monitor ;) Yup I understand but for now is quite nice progress for me. From measurements that Ive done it seems that effiency is exponencialy depended on entry speed. So for next think to do is changing parameters for 4th coil and try to determine what to do with other coils to get same results ;) Ive included exel file with all data so you can look for anything you want.

About optimizing switching I kinda want to do it eventualy but for now I need to get some faster MCU because Arduino is on its limits for now :DDD

And current regulation shoud be there only just as failsafe for IGBTs in normal operation I want just on-off state.

Wish you luck with your gun man ;)

Edit: Shiet this forum doesnt like exel files... :D So here you go: https://mega.nz/#!l51T0KiZ!4B03ez6BowlADIy2mCXu26DvUViGinGNsWexc-HxTf0

DerAlbi

Wed May 23 2018, 04:54AM

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

And current regulation shoud be there only just as failsafe for IGBTs in normal operation I want just on-off state.

I solved this security issue by using 2 microcontrollers. One for the shot (timer-interrupt driven state machine) and one supervisor-controller that actually reads the pins of the shot-control and performs continuous matching to preprogrammed (by communication between the cpus, not hard coded!) sequences. If a fault is detected the IGBTs are shut down by the supervisor. I drive my optocouplers towards the IGBTs with a line driver (because STM32 can only supply 2mA). This line-driver has a convenient output-enable pin, so this it is easy so shutdown everything - and even to avoid accidental IGBT turn on for whatever reason as long as the supervisor is not configured.
I do actually calculate the current and capacitor voltage waveforms, heatloss of the coil and IGBTs in real time by calculating the differential equations every 10Âµs. I configure the algorithm by providing the coil properties and security boundaries such as "max 20J heat loss per coil, max current is xxxA".
What i want to say is: you dont need to measure the current to know how high it is in a simple setup such as an Capacitor + Inductor with ESR setup. Remember: any additional hardware is an additional point of failure.

Because Arduino is on its limits for now

For the shot control Arduino might not be the best thing, but the underlying microcontroller should be good enough. Just use a real programming language like C++ and everything is fine. It will consume time.. but leaving the arduino platform is a necessary step sooner or later as things become more complex as you develop your skills.

Regarding your Excel sheet, i see now why your 4th coil is so efficient now

. You push only 3.5J with it - thats very much on the low side given your geometry. That eases my mind in comparison to my gun *puuh*
On further inspection your coils pushe 7J, then 7.6,J then 4J then 3.5J, so you get diminishing returns the more stages you have. (further making 200m/s impossible)
Your guess that individual coil efficiency depends on entry velocity might not be fully correct then (it is partly)

there are a lot more correlations in your excel sheet than this statement. => The less energy you push, the more efficient it will be, thats well known. However for this to be true, the switching time must be optimal of course.

Wish you luck with your gun man ;)

Yes the excel sheet restored my overly inflated self-perception. ^_^
BTW, check your ICQ

first previous 3 4 5 6

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