I have a serious hobby coilgun project powered by a 6S (25V) LiPo Battery. Currently i have the LiPo, capable of at least 600A burst current connected to a ~75mOhm circuit for some 300+A of peak current. I have built said system with a Mosfet switch (AUIRFB8409 ~1.4$ each from AliExpress). These have worked, but if the current stays on even a little longer (or maybe vene from just wear) they break. The first switch is exposed to some ~4-8 ms of current, which according to my calculations would only rise to the 300A at the 4ms mark or so (pure speculation).

My question: What are my best alternatives to switch the current?

As i see it:

  • Parallel 2-4 of those mosfets with strong thermal coupling. Cheapish, would probably work?

  • IGBT? The high voltage rating is wasted and would not work properly with my 25V? and they seem to either have same current handling as Mosfets or be super expensive and huge

  • Something else entirely?

Thank you for your time

  • \$\begingroup\$ 8 ms is a lot of time and a D2PAK is not capable to handle it. The 195A of package limit is a IR marketing invention. The device blows up at that current in realistic condition. irf.com/technical-info/appnotes/an-1140.pdf \$\endgroup\$ – matzeri Aug 10 '16 at 19:59

You can use something like IXTN600N04T2. That is N-MOSFET built for 600A continuous drain.


according to datasheet, there is 200A external lead current limit. So it seems this MOSFET is actually not capable of continuous 600A drain.

  • \$\begingroup\$ aliexpress.com/item/… like that? Is this for real? how is it so cheap and good? wow thanks! \$\endgroup\$ – maxx Aug 10 '16 at 19:26
  • 1
    \$\begingroup\$ I don't know anything about that chinese seller and its quality. It's 15 USD for 20 pieces. Genuine item costs like 20 USD a piece. And do not forget to put big heatsink to that. Just calculate heat dissipation. \$\endgroup\$ – Chupacabras Aug 10 '16 at 19:37
  • \$\begingroup\$ I will give it a try as my experience with the cheap chinese has been very positive so far. I would upvote but my rep is too low. \$\endgroup\$ – maxx Aug 10 '16 at 19:41
  • \$\begingroup\$ Yeah, I would try that chinese one too ;) That price is worth experimenting ;) \$\endgroup\$ – Chupacabras Aug 10 '16 at 19:44
  • \$\begingroup\$ The datasheet says 600A is the silicon limit. The package is barely 200A (may be). \$\endgroup\$ – matzeri Aug 10 '16 at 19:55

The best improvement is to add a series inductor, that will block the rise of curent trough the battery bank at discharge. The capacitor will supply the required energy pulse.


simulate this circuit – Schematic created using CircuitLab

  • \$\begingroup\$ I use a LiPo because my coilgun is full-auto so there is no time as far i understand to charge a capacitor between shots. I should have mentioned that. Or did i fail to understand your explanation? \$\endgroup\$ – maxx Aug 10 '16 at 19:28
  • \$\begingroup\$ @maxx But you have also forgot to add some schematics, indeed the coilgun has a coli thus inductivity, the resitance of the coil (75mohm) has less importance. \$\endgroup\$ – Marko Buršič Aug 10 '16 at 19:34
  • \$\begingroup\$ Yes, i was hoping for an trivial solution like the one Chupacabras provided, so i didn't think of posting a schematic. My schematic is essentially yours without the capacitor, with the coil as the inductor. \$\endgroup\$ – maxx Aug 10 '16 at 19:44
  • \$\begingroup\$ @Chupacabras According to you, all the arc welder inverters, and other inverters are wrongly made. \$\endgroup\$ – Marko Buršič Aug 10 '16 at 19:57

8 ms is almost equivalent to DC case. The weak points are the internal Aluminium source wires.
Consider 120A as current capability of a good D2PAK 3 pin, as the one you are using, and 180A for a good D2PAK 7 pin.

More than that there is only the TO-Leadless as SMD device, like https://www.fairchildsemi.com/datasheets/FD/FDBL9401_F085.pdf
that is rated 300A and to blow up need ~ 500A.

To reach a 600A capability you need to put several devices in parallel with a layout very symmetrical to guarantee a good current sharing between the several mosfets specially during the turn-on that is the critical moment.

The package limit is due to the number, thickness and length of the Al wires plus the size of the source contact. The number of source wires depends on the die source area and the area of the leadframe available for the bonding.

A D2PAK has up to 3x500 um Al wires with a length around 6 mm.
A D2PAK 7 pins has up to 4x500 um Al wires with a lenght around 6 mm and 4 source pins.
A TO-Leadless has up to 5x500 um Al wires with ~ 4 mm length and a large source contact.

The SOT227 is designed more for high voltage than high current http://www.j-rep.com/download/ixys-appli-book/S23_27.pdf

I doubt has more that 4 wires and may be not of 500 um but just 350-380 um.

  • \$\begingroup\$ I see. How about SOT-227? they are much larger so they should be much better? \$\endgroup\$ – maxx Aug 12 '16 at 9:17
  • \$\begingroup\$ I see. Where can i find the wire details for a specific model? Also, can i improve the package limited current with a heatsink? Thank you. \$\endgroup\$ – maxx Aug 15 '16 at 16:57
  • \$\begingroup\$ No supplier put the bonding info anywhere. It is only available to big customers if they ask details. The heatsink improves the power handling but has minimal effect on max current capability. \$\endgroup\$ – matzeri Aug 18 '16 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.