3x 18mF @ 350V -> What to do? / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

3x 18mF @ 350V -> What to do?

first previous 2 3 4 5 next last

Move Thread

LAN_403

DerAlbi

Wed Jun 09 2010, 10:06AM

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

Yes, that makes things clear. Thank you for your time.. really great! I will go through your consistency-check with the 300A (or calculationg the max pulsewithd there) just to prof it

EDIT
I have done the calculations

very interesting to understand a dataseeht

300A gives a 4V drop. This is a power of 1.2kW. Defining 150K as maximum heatup this gives be a max. allowed temp. resistance of 0.125K/W. This can be read from 19.a and is a singlepulse of <12ms. Great.
This understanding will be a good help for power calculations.. but therefor i need the things below..
/EDIT

But for the next step, i need the information about what timings are realistic for the first coil for a 15gramm piece of iron with 1.0-1.5kA. The following coils should have a shorter timings so they are not critical.
Is there anything that gives me an idea? I would really like to finish my calculation for the IGBTs

So at what timevalue should i look to find a realistic temp. impedance..
Any ideas?

AND: what is the maximum temp? As you have demonstrated the Datasheet calculates the ShortCicuit-Current with a heatup of 200K! Due to the limited time.. is that an acceptable maximum? In theory I would set the maximum absolute temp to 175-200Â°C and start at 30Â°C roomtemp. that gives me 150-180.. is that too safe?

DerAlbi

Wed Aug 04 2010, 01:27AM

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

Hi again..

I have currently finished my powersupply.. my caps are all charged in parallel within 11sec to 360V. Works great. (3.3kJ in 11sec = 300W average!)
Now i have fired some test shots with the help of a random inductor without any timing-control. The fact that it issnt calculated or matched in anyway.. and given the effect-- it amazes me..

But: i am using a barrel made of brass. It is getting hot every shot due to eddycurrents - what a suprise

So the next thing is to slot the barrel.

How?
In general: is it important to slot the barrel at its whole length, or is it sufficient to slot only the areas around the inductors.. last variant is less work with higher mechanical stability. First varaint makes the barrel unstable but completly inhibits conduction.
It is also possible to make many slots that rotate around the barrel.
I would appreciate your opinions

Please dont tell me that i should search the forum.. i have read much but it is always just told that slotting is important.. not explicit HOW to slot it.

Because you helped so much: here are some photos attaced, if anyone is interested..

Barry

Thu Aug 05 2010, 01:58PM

Registered Member #90 Joined: Thu Feb 09 2006, 02:44PM
Location: Seattle, Washington
Posts: 301

You need to slot the barrel the length of the coil, plus one coil O.D. (outside diameter) at each end.
Cheers, Barry

DerAlbi

Thu Aug 05 2010, 09:55PM

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

Thanks! Is it important how many slots in parallel or is one enough..?

Currently i am designing an Overcurrentprotection and a IÂ²t-Protection. I have 50 IXGK320N60A3-IGBTs in my hands. For 6 stages this gives 8 IGBTs in parallel. Each is capable of 700A for 1ms.
Can someone please give me an idea of how long the IGBTs have to be turned on?

If the pulses are too long for my IGBTs i could use SCRs and turn them off with my IGBTs in parallel. Wouls this be the better choise?

Barry

Fri Aug 06 2010, 06:24AM

Registered Member #90 Joined: Thu Feb 09 2006, 02:44PM
Location: Seattle, Washington
Posts: 301

Electrically, one slot is enough. All we need to do is break the skin somewhere in the circumference. Mechanically, there may be some advantage in other patterns to help it hold its shape. I think one person used overlapping slots, probably so that each slot wouldn't be too long as that might allow the sides to cave in a little.

For timing, you'll probably find the first stage needs a few msec 'on', perhaps in the range of 1 to 4 msec. And then subsequent stages will have progressively shorter duration.

If you can arrange a sensor to turn off the IGBT when the projectile reaches a certain position, that would work better than a pre-programmed 'on' time. My experience has been that closed-loop control always works better than open-loop.

Cheers, Barry

DerAlbi

Fri Aug 06 2010, 04:49PM

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

I will slot the barrel twice in the region you toldme. this gives also space for optical triggers.

Due to my large caps i will handle high currents. The biggest problem is to protexct my IGBTs..

So i am thinking about to use SCRs (found some Hockey Pucks on Ebay 240A rated). With 8 of my IGBTs in parallel i am theoretically capable of switching 5.6kA for 1ms.
Due to your given timings this seems not enougth. so i am thinking about using these SCRs and use my IGBTs for turing them off.

I want to use small Caps(200uF?) @ 350V to produce a current pulse that should stop current through the SCR. So effectively it should shut down..
Did anyone tried this for timing control? Is this possible to simulate? (Need LTSpice Libs for SCR and IGBT)

DerAlbi

Wed Aug 11 2010, 01:51PM

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

Beeing more specific for now:
I got some big SCRs on Ebay with 120kAÂ²s. Turnoff time is 60Âµs worstcase. So my IGBTs have to be turned on for around 120Âµs to be sure its off.

Problem: IGBTs Vsat and Voltagedrop about the SCR is nearly equal.. so current over the SCR might not go to 0.
I know that some people in this forum already have solved this problem....
Some Ideas..?

klugesmith

Thu Aug 12 2010, 09:34PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Hint: use small capacitors and/or inductors.
For example (I haven't actually done this):
SCR fires to energize Main Load and charge the small value capacitor C.
Then IGBT or another SCR fires, attempting to pull the main SCR anode negative, and stealing ALL the main load current momentarily. The length and energy of the commutating pulse are limited by the small value of C.
Small C needs to withstand peak current. Spikes of very high voltage are unavoidable if you need to quickly stop current flowing in an inductive load.

V+      same_or_lower_positive_supply
|         |
MainLoad  R
|     C   |
------||---
|         |
SCR    IGBT
|         |
Gnd--------

DerAlbi

Thu Aug 12 2010, 11:45PM

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

Sorry.. i do not see how your circuit works.
1) If both SCR and IGBT are not on: the left side of the C is connected to 350V through the coil. Left side is also connected to 350V through R.
2) If the SCR fires: the left side is driven to 0V. C is loading, Current is limited due to R.
3) IGBTs is tunred on:
It only discarges the Cap thorugh the IGBT. then the IGBT will discarge the 350V source with a current limited by R.

So this circuit only produces currentspikes and heat

Did I miss something?

DerAlbi

Fri Aug 13 2010, 12:08AM

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

Sorry for doublepost. Editting a post to attach a file after the post was posted without a file seems not possible.

My idea was this:
1.turn the igbt on: the cap is carged to 350 though the coil.
2. turn igbt off.
3) Fire SCR. the left side of the cap is pulled to 0V.
4) then trigger the IGBT. The left side is then again at 350V and should supply the current.

Problem here is:
On Step 3, when the left side f the cap is at 0V the right side is at -350V. This is is not soooo good for the collector of the igbt. Connecting the emitter to the Cp and the Collector to GND, would solve this problem. But Gatetriggering would be highly comlicated due to 350V potential difference. :-/

first previous 2 3 4 5 next last

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