Yet another coil gun prototype / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Yet another coil gun prototype

previous 1 2 3 4 next last

Move Thread

LAN_403

V2006

Tue Nov 06 2018, 10:43AM

Registered Member #61550 Joined: Thu Apr 06 2017, 03:23PM
Location:
Posts: 86

On these videos, the coil is wound with a wire of 2.6 mm (winding diameter 5.5 mm, 3 layers of 20 turns each)

Moisten at least 1.5 mm - it will be much better than 1 mm - you will immediately see.

Zuckerstange47

Thu Nov 08 2018, 01:49AM

Registered Member #61926 Joined: Wed Nov 22 2017, 04:08PM
Location:
Posts: 34

I tried to use the ESP32 as some kind of oszilloscope (not sure though, because I do not own an oszilloscope nor used one before) to measure voltages while and after activating the coil.
The top graphs show the capacitor's voltage for two activations, the first starting at 40V, the second at a lower voltage.
The bottom graphs display the voltage over the coil (the first with an initial capacitor's voltage of about 19V, the second with 30V). I don't understand why the first datapoint and the ones after deactivating the coil are 4095 (the maximum value that the ESP32 can give out) and not zero as I expected. Also I am not sure what to make with the data, but I am happy that the measurement worked.

My interpretation is, that the coil takes at least about 2000 microseconds to get full amperage after activation, but when considering the overlaying drainage of the capacitors the maximum would be farther to the right when having a constant power source. From my calculations the current should reach about 50% of the maximum in about 1500 microseconds and 90% in almost 5000 microseconds. I do not know if my interpretation and calculations are correct, but they seem to harmonize well enough.

(Unfortunately the ESP32 cannot measure voltages accurately, the output is quite noisy and imprecise, but a least it is good enough to make out a trend. I only saved a reading, when the difference to the last saved measurement exceeded a specific value, so it would be possible to get a higher resolution in regard of the time - I chose a coarse resolution because I had to read the values manually from the small blue screen).

1541641536 61926 FT183129 2018 11 08 Oszi

V2006

Thu Nov 08 2018, 12:54PM

Registered Member #61550 Joined: Thu Apr 06 2017, 03:23PM
Location:
Posts: 86

Looks impressive. But as practice shows, any calculations are significantly different from the facts.

Plasma

Tue Nov 13 2018, 06:54AM

Registered Member #61406 Joined: Thu Jan 05 2017, 11:31PM
Location:
Posts: 268

Zuckerstange47 the second and forth graph looks like you increased the resistance, ssmaller wire diameter maybe, that is forward emf after switch off, switched off current before the coil I think,
Nice work.

Zuckerstange47

Wed Nov 14 2018, 09:42PM

Registered Member #61926 Joined: Wed Nov 22 2017, 04:08PM
Location:
Posts: 34

Plasma: I do not understand why it looks like I increased the resistance, as my knowledge is very limited. To me, the first and second look almost the same. The only change between the shots was that the capacitors were not charged as much on the second and forth graph.

My next planned steps are to tidy up the construction. I designed a simple PCB to produce multiple, more beautiful light barriers easily (on the image the coil and a spring-type terminal block is overlayed).
Also I am printing an apparatus to help making coils, based on a previous design of mine (that is shown on the other images). The new one will be more stable (I hope) and the distance-adjustment should not lose it's setting by mistake, when rotating the pipe.

1542231696 61926 FT183129 Lichtschranke Test5ks

Shrad

Thu Nov 15 2018, 08:01AM

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

if you need ultra fast detection, a small laser module and a reverse biased PIN photodiode will be top gear

Zuckerstange47

Mon Nov 19 2018, 12:14AM

Registered Member #61926 Joined: Wed Nov 22 2017, 04:08PM
Location:
Posts: 34

Shrad wrote ...

if you need ultra fast detection, a small laser module and a reverse biased PIN photodiode will be top gear

I think the ones I am using are fast enough for my experiments, but thank you for the information nevertheless.

My current big problem is that I cannot build coils the way I want to. Does anyone have a tip for making a good transition between one layer and the following one? In the last image one can see that I finished the first layer, wound from right to left, and consequently want the wire of the first winding of the second layer to stay between the two leftmost windings of the first layer. Maybe the end plates (black disks) should be mounted in a more rigid way, so that the wire cannot slip into the gap...? How do professional coil-winder solve this problem?

I am afraid that my new winding-machine was printed in vain

DerAlbi

Mon Nov 19 2018, 11:51PM

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

Ger: Orthozyklische wicklung
En: orthocyclic winding

Zuckerstange47

Mon Nov 19 2018, 11:57PM

Registered Member #61926 Joined: Wed Nov 22 2017, 04:08PM
Location:
Posts: 34

Okay, I partly found a solution, but the result is still very bad. I put the coil between some nuts on a threaded rod, and adjusted it in the way that the number of windings on the first layer was right. The second layer looks okay, too, but later ones look a bit messy and probably I missed some windings. At the end I had 3 meters too much remaining on the spool (assumed the sellers specification was correct), but the outer diameter of the coil was the one my coil-calculator had generated.
The winding-procedure was not pleasant also, because rotating the printed parts seemed to result in tightening the nuts, and so the resistance got stronger after each turn, and now I got several blisters on my hand. That was one reason why I changed the first winding machine's design (the pink one) in the first place...

I do not want to unwind this thing again, but at least I learned a bit from it: Next time I will measure the wire length beforehand and mark it every meter or so. Also I might print something to have a better grip on the rod and turn the complete thing instead of clamping the pole and only turning the coil with the printed parts. I suppose the coil will work in an acceptable way even if it looks ugly, but at least I would like to know how much wire I used, and how many windings I made.

@DerAlbi: The Orthozyklische Wicklung looks very promising, but because of tight tolerances and the need for a very precise positioning of the wire it seems to be impossible to implement for me. Have you succeded in realizing this technique?

DerAlbi

Tue Nov 20 2018, 12:41AM

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

Well.. lets just say: i got so far that it seemed plausible with better equipment

I have built a 3d printed winding machine too (fully automatic) but it was way too flexible. Today i know what would need changing, and can tell you right now that your setup will never succeed. Inconsistent wire tension is the first issue - not only is the force unknown, but it will vary over time.
So what works in your case is to use a small stiff plastic piece and manually lay down every turn. But even that will result in wild winidng after a few layers. Try to make the coil former either exactly n*WireDiameter or (n+0.5)*WireDiameter, anything in between will fail. Use the highest wire tension the wire allows (can be calculated) which helps with laying down the wire consistently. Your picture shows kinks in the wire even after the wire tensioner - this may never be the case. Straighten the wire through a pulley labyrinth and pull it really hard. 1mm diameter must be pulled with around 5kg equivalent (50N). It can withstand 17kg before breaking. So dont hesitate.

However I have found that there is no substantial loss in performance of the coil, if wound wild. It is really more a cosmetic thing. Keep going. Dont invest too much time in printing. Its too far from viable.

previous 1 2 3 4 next last

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