Thanks for posting in the string you typed. I tried doing a binary compare of your string against the one provided by duolingo, and I noticed this difference in how they represented the "ทำ" in unicode (utf8):
- Duolingo: ท (0xE0B897) ํ (0xE0B98D) า (0xE0B8B2)
- Yours: ท (0xE0B897) ำ (0xE0B8B3)
I.e., for Duolingo's string, the 2nd letter is called Nikhahit ( ํ), which is followed by Sara aa (า):
This is in contrast to your string which simply (and more correctly) uses sara am (ำ):
So the original contributor of the q/a used two separate characters to form the equivalent of sara am. Not sure why. I would have thought the correct/better way would be to have just used sara am, as you have done.
Oh well, glad to finally figure out what went wrong on this occasion. Hopefully the course organisers will take note of such findings and fix them in future releases.