oneTesla

I tested my completed oneTesla for the first time yesterday, and like some others I blew the IGBTs and two of the MUR460 diodes. While fiddling with testing and desoldering these components today, it got me thinking:

Why not put the components that are likely to break but difficult to replace (mainly the IGBTs) in some sockets or screw terminals. That way they can be swapped super easily.

I've seen on this forum that a lot of people break their IGBTs during early testing, and the wording in the manual makes it seem like they can easily overheat and break even during normal use. If they're so fragile (or poorly protected?) shouldn't they be easier to replace?

Besides, desoldering and resoldering several times is likely to mess up the solder pads so that the whole circuit board must be replaced. It would be great to reduce this risk too.

Is there any technical reason why this is a bad idea? Please let me know your thoughts!

I used to use screw terminals for all of the IGBT's on my projects, but they're quite a hassle - I've had countless failures due to poorly secured screws, sketchy terminals, etc. During initial testing, if there spark is not present, noisy, or doesn't grow as you turn up the power, you should stop and check your soldering.
Once the coil gets working, its fairly hard to break during operation, as long as you stay away from songs over a couple minutes long.

Thanks for your reply.

If that's the case, I think I'll try my luck with some screw terminals and see how they work out. Removing the IGBTs while trying not to break anything else on the board was quite an ordeal; don't want to go through that again if I don't have to.

I hindsight, I think I may have had a "noisy" spark, (if that means what I think it means) but not having any basis for comparison I didn't realize it was a problem before I tripped the fuse (tripped the 16 A circuit breaker, the 10 A fuse on the board never blew --weird)

I get your point about being careful; and that's what I did wrong in this case. I don't agree that only two minutes of safe run-time counts as "hard to break" though A device which operates for only two minutes at a time would count as extremely fragile in my book.

That begs the question; is there any way to improve runtime and reliability?
- Beefed up cooling?
- More powerful IGBTs?
- Protection circuitry to prevent failures, or at least prevent cascading failures once one component blows?
- Temperature monitoring?

Would be pretty cool to have indefinite run-time Even 10-20 minutes would be a great improvement. Dunno how realistic it is though.

Well so the "couple minutes" runtime is (1) a recommended limit; in actuality we've run it for far longer without breaking anything and (2) purely a thermal limitation; as long as the heatsink doesn't get too warm you're good to go.
You're free to run it forever and ever if you can keep the heatsink cold enough; anything less than 60C should be good.
The idea behind the recommendation is that it is very conservative, a couple minutes is a fairly long song, and it keeps down ozone levels and doesn't deafen your audience too badly. We strongly advise against leaving the coil on for tens of minutes for safety reasons.

Thank you for clarifying. I think I'll look into adding a temperature indicator to make things easier. I found some stickers with thermochromic indicators that I might try.

Not a socket but it makes igbt repair easier.

download/file.php?id=10

Interesting. How does it cope with the weight of the heat sink?

Slartibartfast wrote:How does it cope with the weight of the heat sink?

Its hot glued to the top of the bus caps. Its just a scrap piece of thick pcb with six strips etched.
Very easy to change igbts but I dont think it will fit in their case.