I wouldn't think this was possible, but 2 dead NDSL's is pretty convincing.
Has anyone had this happen to them?
I've never had it happen, but I'd be asking for them to send the two R4's they haven't used back for inspection - with the "killer" one clearly marked. I doubt there'd be an issue with the third cart, but it's better to play it safe.
On a DS cartridge, fingers 1, 7 and 17 are GND. 8 is VCC. 1, 8 and 17 extend further towards the edge than the others, or at least they do on original carts and every flashcart I've seen (so far - R4, DSTT, M3Real, EZ-V) - this is so the ground and supply connections are the made before the signal ones when the cartridge is inserted, and is a feature that any flashcart designer would probably be wise to copy.
FWIW, I suspect that there might be a short between VCC (pin 8) and GND on the board (e.g. a solder bridge across a surface mount cap) or a short inside one of the components. Alternatively, and especially if it's a new R4 without the guides/dividers, there may be a manufacturing issue that's resulted in the board being mis-aligned - and the contact for either 7 or 8 might be shorting the fingers for 7 (GND) and 8 (VCC) by contacting both. The rest of the lines are signal lines of various sorts - I doubt that a short between any of them and either VCC or GND would result in enough current to fry a fuse, but it might be possible.
In the interests of customer relations, you might end up having to get them to ship you the dead DS's for inspection; hopefully you've got some contacts who could replace the fuses (note: REPLACE, rather than just bridge - RS Components ought to be able to help with surface mount fuses, but they might not be cheap and there may be a minimum order quantity) and maybe try a known-good cart and the killer one on one of them before it's closed back up - that way, you might be able to identify for sure whether the flashcart killed them or whether the customer is trying one on. Plus, you might actually be legally liable for any consequential damages caused by a product you've sold - and even if you're not, postage both ways and some repair costs are almost certainly going to be cheaper than any loss-of-goodwill associated with hanging the customer out to dry if your product really was defective.
You might also be able to test the killer cartridge with a fused 3.3V supply - anything that would blow the internal fuses in a DS should almost certainly be able to blow a 125mA fuse, whereas I wouldn't expect a good cartridge to draw very much at all. That might eliminate the need to try the suspected killer in a DS at all - if it makes the magic smoke leak out of fuses, and good R4s don't, then you've found the cause. Mis-alignment won't be picked up by this, but a close visual inspection of the killer and a known-good cart ought to indicate if this is a possible cause.
That said, the customer could be trying one on. A friend-of-a-friend knows a lady who bragged about buying a new pink DS Lite just before Christmas, putting a fifteen-month-old pink DS Lite back in the box, taking that back to a major chain store (with the receipt) and saying "I got a used one in the box" during the post-christmas refund/exchange confusion - the customer service desk staff didn't bat an eyelid, knowing that some of the other shopmonkeys don't check returned goods properly before putting them back into stock, and called the appropriate area to bring a new one up the front for an exchange. I used to work in retail myself, so I know full well of the kind of things that some customers will try and the kind of lazy aresholes who sometimes end up working behind the counter - and how damn hard it can be to disprove a well-thought-out and plausible scam like that.
Good luck, and may you get to the bottom of the matter - whatever it is.
(and as an aside, have you got any good prices on GC modchips, and replacement top housings for full-sized discs?
)
(edit: changed from 1,7,17 as long fingers to 1,8,17 - GND,VCC,GND make contact before the signal lines)
