Fried freewheeling diode / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Fried freewheeling diode

Move Thread

LAN_403

Benj

Sun Aug 14 2011, 08:02PM

Registered Member #4050 Joined: Sun Aug 07 2011, 04:06PM
Location: France
Posts: 17

Hi all, I've tested my brand new SCR with my capacitors charged up to 400V, and things went bad.
Actually, Three things went bad:
-The freewheeling diode put in parrallel with the coil blew up ( BY 228 ).
-The coil itself has suffered an inner spark ( its resistance fell from 1 ohm to 0.1 ohm...)
-The rectifier bridge that i used to rectify the transformer output of my mazzilli driver is dead.

the thyristor still works = ) !

I tried to figure out what happened:
-I think the current flowing through the diode when the thyristor was off ( as my circuit is ( I think ) non criticaly dampened ) was to high, thus killing the diode.
-As U=Ldi/dt, when intensity died rapidly, U reached summits...Then the insulation inside the coil has broken=> spark

About the rectifier bridge: It's a mistake coming from my previous coil gun, not to separate the charging circuit(rectified mazzili driver) and the RLC circuit. This point is easy to solve I think (I just need a switch that can withstand the charging current ).

Solution(?):
-I think i'm gonna buy an inductance meter, so I can make sure the circuit is critically dampened or at least over dampened ( according to barry's site, it would also reduce the suck-back effect). Then, no diode is needed.

Am I right?
Bonne soirÃ©e= )

Saz43

Sun Aug 14 2011, 11:57PM

Registered Member #1525 Joined: Mon Jun 09 2008, 12:16AM
Location: America
Posts: 294

You theories sound quite plausible, sir.

I would however argue with the solution by suggesting staying away from a critically damped coilgun, as these tend to be terribly inefficient due to low current as caused by the coils high resistance and a prolonged suck back due to the high inductance. Also you might save yourself some money by using Barry's coil and RLC simulators, which are accurate enough if you design with a healthy margin of error.

Turkey9

Mon Aug 15 2011, 12:12AM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

I'd go with a critically damped system. However as Saz43 suggests, if your capacitor bank is too large, the pulse will be way too long and you'll get terrible suck back. The solution is to decrease the total capacitance. If you can't (using a single large capacitor) you can increase the mass of your projectile. This will lengthen the time in the coil to match the long pulse.

Don't buy a meter. You can calculate the inductance of solenoids to pretty good accuracy. After you make the coil, you can measure it with a function generator, a resistor, and a volt meter.

Forty

Mon Aug 15 2011, 12:21AM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

how much capacitance does your bank have?

for the datasheet of your diode, i'm finding a 1500v/3A sinterglass diode. which with only a 50A current rating, it is no surprise that it broke. you need a diode with a surge rating approximately equal to your peak coil current. the diode has to be antiparallel with the coil (although i'm sure that's what you meant)

1 ohm sounds a bit high. what size wire/coil are you using? since you've got to remake your coil anyways, try to use some thicker wire.

what kind of rectifier bridge are you using? you shouldn't have to disconnect the charging circuit at all. but if you still want to, then you only need a switch that can handle the high voltage, since your charging current is actually pretty low(50-200mA probably)

with a critically damped oscillator, you'll only get 36% of the peak current that you would in an underdamped version.

Saz43

Mon Aug 15 2011, 04:11AM

Registered Member #1525 Joined: Mon Jun 09 2008, 12:16AM
Location: America
Posts: 294

Forty wrote ...

what kind of rectifier bridge are you using? you shouldn't have to disconnect the charging circuit at all. but if you still want to, then you only need a switch that can handle the high voltage, since your charging current is actually pretty low(50-200mA probably)

When I built my coilgun with a ZVS charger, I found it was necessary to disconnect the charging circuit when firing. If you don't, the charger will short through the coil. The resulting current will latch the SCR and the voltage and current spike will probably damage the charger. Also, the bridge rectifier diodes, if connected, are arranged in the same way as the anti-parallel diode, so they would have to be rated to handle the commutation current.

Benj

Mon Aug 15 2011, 10:42AM

Registered Member #4050 Joined: Sun Aug 07 2011, 04:06PM
Location: France
Posts: 17

Hi all, thanks for the answers.

@Forty:
-I'm using three 680ÂµF (theorically )capacitors in parallel, so a 2040ÂµF capacitor bank.
-Yes, I meant the diode was anti-parallel with the coil.
-actually, I'm not even sure the coil resistance was 1 Ohm..I'm using a 20-19 gauge wire ( 1mm actually ).
-I was deceived by my previous ( and first ) coilgun: I also used BY228s and they've survived so far. But the caps were only charged up to 300volts. BY399P can withstand 100A, I will make a sample request .
-i'm using this

@Saz43:
-If the current flows (partially) through the diode bridge( even if it can withstand the peak current), it's a waste of energy, isn't it? As the whole current could flow through the coil instead if I use a switch.

@Turkey:
-If only i had a fonction generator at home = ). In two weeks i'll be in engineering school, i'm sure i'll get a fonction generator there.
-I was thinking about what coilgun.ru calls "V-switch"

, so i could get rid of most of the suck-back effect... It seems a bit tricky to build one.

Forty

Mon Aug 15 2011, 01:14PM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

If you're gonna sample some diodes, then go big: fairchild's rurg5060 can take 500A
vishay's 60EPSxx can take 950A

1mm is probably 18 gauge, so that's fine.

for disconnecting your charger, you could just use a relay. have a charging button/switch that activates the relay, and use the relay contacts to connect your caps and whatnot. when you stop charging, the relay disconnects

Saz43

Mon Aug 15 2011, 08:15PM

Registered Member #1525 Joined: Mon Jun 09 2008, 12:16AM
Location: America
Posts: 294

Benj wrote ...

@Saz43:
-If the current flows (partially) through the diode bridge( even if it can withstand the peak current), it's a waste of energy, isn't it? As the whole current could flow through the coil instead if I use a switch.

As you know, if you're charging your capacitors with a Mazzili driver, you need to use a bridge rectifier on the output. This rectifier will then necessarily function as an anti-parallel diode unless you disconnect it from your capacitor bank before firing.

I was explaining this to someone earlier, and I drew this diagram which I think illustrates the point:

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