V3 Coil Carbine / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

V3 Coil Carbine

first previous 5 6 7 8 next last

Move Thread

LAN_403

MrFlatox

Sun Jan 20 2013, 05:37PM

Registered Member #9349 Joined: Mon Jan 07 2013, 08:50AM
Location: France
Posts: 102

Ash Small wrote ...

Surely a capacitor accross the MOSFET will absorb the spike and protect the mosfet?

It's a fairly standard snubber device.

I have just tried this, and it makes an ugly oscilating LC circuit. I do not think that it would make things better at the end.

If my simulation is correct, then using zener diodes will never make the mosfet live because voltage spike is never under 160V, no matter the Vzener voltage I choose.

Ash Small

Sun Jan 20 2013, 06:44PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

MrFlatox wrote ...

Ash Small wrote ...

Surely a capacitor accross the MOSFET will absorb the spike and protect the mosfet?

It's a fairly standard snubber device.

Did it keep the voltage 'within limits'?

You could try putting a fast diode in series with the capacitor, to stop the oscillations. You'd could also put a high value (mega Ohms) resistor in parallel with the capacitor, to drain it after each shot. (there are other ways to drain it).

(It's just another solution that you could try simulating, and compare it with other suggestions. Did you try other capacitor values?.)

EDIT: This capacitor, diode and resistor circuit I've described is a standard snubber circuit used to protect silicon and prevent oscillations. There are other variations used, but this is one of the more common ones. It could also protect the MOSFET in the event of TVS failure.

MrFlatox

Sun Jan 20 2013, 08:42PM

Registered Member #9349 Joined: Mon Jan 07 2013, 08:50AM
Location: France
Posts: 102

It lowered the voltage spike when using high value capacitor, but the oscilations are just too big. I will try your suggestions about the diode and resistor.

As you are talking about TVS, does a TVS diode with a voltage rating of 60V would do the same thing as my diode plus zener diode in series ? It could be intresting because they can pass a lot of current, and are cheap.

Ash Small

Sun Jan 20 2013, 09:46PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

MrFlatox wrote ...

It lowered the voltage spike when using high value capacitor, but the oscilations are just too big. I will try your suggestions about the diode and resistor.

As you are talking about TVS, does a TVS diode with a voltage rating of 60V would do the same thing as my diode plus zener diode in series ? It could be intresting because they can pass a lot of current, and are cheap.

TVS's (transient voltage suppressors) and zeners do pretty much the same thing,as far as I'm aware, but I can't quote any specs or anything for either. I'm still learning this stuff too.

I've researched snubbers quite a bit though, for protecting transistors of all types from voltage spikes. Adding a fast diode kills the oscillations. A high value resistor will discharge the cap each time you fire, and not really affect the rest of the circuit,. You'd need to research'experiment/simulate the exact values/specs yourself, though. It should also protect the MOSFET in case of zener/TVS failure, etc.

Yandersen

Mon Jan 21 2013, 06:25AM

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

Guys, I'm sick hearing about your ridiculous computer simulations for circuits with 3 parts! Do any of you ever hold a soldering iron in a hand and an oscilloscope nearby? Your simulations are either wrongly done or simply away from reality. What do you think you are "learning" without practicing? For the whole month 4HV are dancing around TVS diodes antiparallel to the coils and seems like noone understands a thing.
Ah, I'm just in a bad mood. What are we talking around here?..

Ash Small

Mon Jan 21 2013, 01:15PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Yan, I don't do simulations either, I once downloaded a free version of SPICE, but couldn't be bothered to spend hours and hours learning how to use it. I just use a 'scope and a breadboard and see what happens in reality. I was working on snubbers a few months ago, but had to stop because I didn't have any fast diodes. I have some now, but my breadboard now has another project on it, so I've not got around to continuing with my snubber 'experiments'. I also use the minimum amount of maths that I can get away with. I also try to use the best quality components which are generally over rated for what I need, although sometimes I just use what I can get or already have.

I do take an interest in other peoples simulations though.

All that is required here is a capacitor with a high enough value to reduce the voltage spike to 'acceptable levels', and a diode that is fast enough to 'kill' any oscillations and handle the current, plus a resistor that only has to discharge the capacitor before the next shot.

Just using 'trial and error' will kill quite a few MOSFET's in the process, so a simulation should reduce the number of MOSFET's that have to die before the correct values are found.

All I'm doing here is suggesting an alternative to what you've suggested, and which should also protect the MOSFET in the case of TVS/diode failure. Your solution is simpler because it uses less components, but mine is arguably more reliable.

I'm simply pointing out an alternative method to achieve the same thing, ie, to reduce the voltage spike accross the MOSFET.

It could be argued that the best solution would be to implement both, if space and availability of components allow it, certainly at the breadboarding stage.

Yandersen

Tue Jan 22 2013, 03:57AM

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

Okay...

ben5017

Tue Mar 05 2013, 04:05PM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

Update:

-Added 3rd accelerator stage
-Modified coil geometry (removed 1/2 layer from each subsequent stage to compensate for decreased "on" time due to higher V initial (trying to keep amps approx same each stage to maximize power through the igbt))
-Removed diode/zener protection (replace with (3) 5kw tvs diodes in series (540V totalbreakdown voltage))
-Minimized circuit size and moved all to one board
-Remade sensors into interchangeable units
-Added Fuse box (separate fuses for each stage + injector + cap bank)

Additional Changes to be made:

-Add cap bank to increase discharge rate (from 45C (lipo) to 85C (Caps))
-Add selector switch to change between monostable and astable 556 chip to drive injector
-6s lipo instead of 5s

Test shot results.

(3 stage no injector) approx 75mm penetration into solid Styrofoam block.
(3 stage no injector) through beer can, 35mm penetration Styrofoam.

Current Bugs:

First stage works perfectly when just testing the sensor, however when I actually shoot it, it stays on, blows the fuse and the 421 inverting gate driver??? Having trouble figuring out the issue...

Also I cannot get my 556 edge triggered monostable to work correctly (I have made this circuit numerous times with no issues....)

I am starting to think my breadboard is F***** up, it has been doing some weird stuff.....

Yandersen

Tue Mar 05 2013, 09:00PM

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

I had both of those bugs. Latchup of optical gate may have many reasons. The one I have encountered was sensitivity of photodetector to the EMI from coil - when coil shuts off, sensor gives driver the white noise which ends up in continuous on-off oscillations blowing the driver and possibly switch.
The issue with 556 timer is a strange coupling between output and trigger. When I connect OUT directly to mosfet gate and mosfet is switching off, the parasitic capacitance between drain and gate forcing the gate and timer's OUT in turn. For some reason that spike acts like a triggering signal for timer. Had no such issue with NA/NE/SE/TLC-555 timers though... Add 100 Ohm resistor for the output of 556 to prevent current overloading - worked for me.

Maxwell

Wed Mar 06 2013, 12:38AM

Registered Member #8497 Joined: Tue Dec 04 2012, 06:24PM
Location:
Posts: 74

@Ben5017
Since you have optical gate triggering - why not buy a cheap arduino and count the time between gate triggers - will provide you with a pretty accurate velocity?

first previous 5 6 7 8 next last

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