My school coilgun poject / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

My school coilgun poject

previous 1 2 3 4 next last

Move Thread

LAN_403

Yandersen

Tue Feb 04 2014, 08:37PM

Registered Member #6944 Joined: Fri Sept 28 2012, 04:54PM
Location: Canada
Posts: 340

Skynet, there is a banned user Dantist who asked me to send some valuable comments about your circuit.

1) The resistance of R6 (1K) is too small. This resistor discharges C1 and C3, so the voltage on high side driver will critically drop in 25ms and low voltage on the gate of Q1 may damage it. Dantist recommends to change R6 to 10K, I recommend to get rid of R6 and R7 at all.

2) The pics you posted show long wires between the IGBT's pins and a power capacitor. This adds a parasitic inductance which creates a dangerous for IGBTs and the driver voltage spikes when coil is switched off. Dantist recommends to add a small capacitor in parallel with the main cap, but connect this little cap directly to IGBTs' pins making connection wires as short as possible. By his own successful experience, Dantist recommends this type of caps:

I do not have practical experience with halfbridges, but if I would build one, I would simply mount IGBTs and diodes right on the main cap. The wires from IGBTs to coil may be long - no problem here.

P.S.: too bad that 4hv had banned the guy who built this.

Skynet

Wed Feb 05 2014, 09:04PM

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

So... I made changes added Shottkys and zeners. I did some test shot on 150V and its realy weird... look at photo of voltage on IGBT gates. And for more my TTL logic stoped working after that shot.

Yandersen

Wed Feb 05 2014, 09:55PM

Registered Member #6944 Joined: Fri Sept 28 2012, 04:54PM
Location: Canada
Posts: 340

How are you able to measure both gates at the same time?! Both drivers have different reference level ("ground")! Did you SCed COM and Vs? Obviously everything will burn then. [facepalm]

Skynet

Wed Feb 05 2014, 10:11PM

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

Ok IÂ´m an idiot :D When I want to measure it I did not realized that the ground on the oscilloscope are connected together :D IÂ´ll fix damage tomorow.

I measured voltage on resistors R6 and R7 IC, IGBTs are fine just logic have problem I made multiple shots on diferent voltages on capacitor and everything worked good.

Heres the picture of voltage on gates with 50V on cap(taken before 150V test). With 30V it worked perfectly.

Skynet

Thu Feb 06 2014, 01:42PM

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

Ok I completely remaked logic circuit (IÂ´ll post it later it still needs some fixes and I dont know how to fix them) but now it works great

At pic is gate voltage and volatge on SD pin on IC.

Video from 200V test (I had to fill a box with cardboard because projectiles were flying through. :D)

Yandersen

Thu Feb 06 2014, 09:21PM

Registered Member #6944 Joined: Fri Sept 28 2012, 04:54PM
Location: Canada
Posts: 340

Oh, very good. Congrats! :) Will you test 400V now?

Skynet

Thu Feb 06 2014, 10:08PM

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

Next week probably I need to short wires between IGBT and main cap also add small HV caps refered by Dantist. Cause I dont want do burn another IGBT its not cheap toy :D

Now I updating scheme to 1.3 revision with 555 timer "logic" its realy messy scheme but work good for 1.st stage at another stages it would be slow. So it would need big improve..hope that we dound some solution.. Issues with the old logic were caused by EMI from coil switching- I used CMOS logic =>bad idea for coilgun :D

After that I try 400V I hope for success

Yandersen

Fri Feb 07 2014, 01:32AM

Registered Member #6944 Joined: Fri Sept 28 2012, 04:54PM
Location: Canada
Posts: 340

Skynet, what about current limiting? If you equally drive both IGBT, having them ON while the optogate is dimmed, then the longer that gate is dimmed the longer the coil is ON. So first stages are overpowered and subsequent stages will have less energy stored in coil. Senseless, don't you think? Each coil must have the same energy. Think about different approach: once the gate got dimmed, both IGBTs get opened. But high side IGBT automatically closes after t microseconds and stays off no matter what. This way you get predetermined current and amount of energy stored in coil (whatever built up for t us). When the high side IGBT closes, the current loops in coil and low side IGBT. When the optogate get lighted again, the low side IGBT closes also and coil empties whatever magnetic energy left in it back into the cap. Just use a 555 timer in monovibrator mode to drive high side: trigger it with the short pulse when gate is dimmed and let it switch itself off once the specified time interval ends.

Skynet

Tue Apr 07 2015, 09:50PM

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

Soooo after realy long pause IÂ´am back!

After previous problems I decided to gain more experience for driving IGBT with best method.
I decided to construct an electrically isolated IGBT driver. Today I finaly assembled new driver and everthing seems to work fine.

Be tuned for more progress

1428443430 11734 FT158837 11147898 10200263866675361 1809043350 O

1428443430 11734 FT158837 10003074 10200242067610398 704930396 N

DerAlbi

Wed Apr 08 2015, 07:33AM

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

Great progress! I like the setup.
If you ever redesign the circuit i would suggest you include a 555-Timer as monoflop and AND-gate the input signal to the IGBT-driver. In this way you cant accidently turn the IGBT on for a longer time than safe given a certain Motor-Inductance. Ive done the same thing (..well at least similar) and its totally safe to operate- It surived every failure condition so far.
When i look into your ICs datasheet, i see you chose the configuration for parallel IGBT-Drive. (presumably for higher drive strength) I can tell from my experience that this causes more problems than it helps to avoid. A slow gate drive is not so bad as you think it is.
Youve got some strayinductance in your half-bridge circuit (specially from the Collector to the Diode above + the ESL of your capacitors and wiring.)
This strayinductance takes time to discharge and thats actually your limiting factor for your switching speed. There is no difference to slower switching and shorter active clamping and fast switching and longer active clamping. The only thing that differs is the complexity of your circuit and how much EMI problems you cause for the sourrounding circuit.

Keep it up! Its really good to see someone here whos actually producing something that works!

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