I have the same problem, however I suspect it is due to the stir bar being too long, and the magnets not quite perfectly matching the stirbar. I wish I had a bad hard drive to gut so I could try the HDD magnet, but all of mine are currently working.
One thing that I did, is I mounted the washer on a spacer (a plastic pop bottle top and a felt "furniture silencer" pad that I had laying around) to move the magnets away from the motor. I'm not sure this is totally necessary, but considering motors utilize magnetism to function, it makes sense to keep magnets away from the motor when possible. Spacing with the cover is such that the magnets will rub if I gently press on the cover directly above the fan. This isn't an issue with a flask, as it is large enough to sit over the edges, where the cover would be reinforced. Kind of like the example in the other thread.
I also don't have the magnets glued down, they are merely stuck on the washer. They are strong enough to stay put without any adhesive, and this allows me to adjust them for the stir bar.
Here are some pics:
Cover clearance:
It isn't clear in this pic, but there is maybe 1/16th of an inch of clearance. Enough that the magnets rub the cover if I place a mason jar on top of it instead of the wide-bottom flask.
Guts:
The taped up mess in the corner is a voltage regulation circuit. This is more reliable than simply wiring in the pot, which could possibly burn out if wired directly. The pot that I used is rated for 5 watts, which means anything over a 400 mA wall wart could burn the pot out of wired direct. The fan I used is only supposed to use at most 1.2 watts, so in theory it would be "safe" to wire the pot in directly, this is just a safer circuit. The circuit consists of an LM317 voltage regulator, a 5K pot, a 0.1 uF filter capacitor, a 330 Ohm resistor, the fan, and the power supply. The wiring diagram can be found on the LM317 data sheet from TI (
http://www.ti.com/lit/ds/symlink/lm117.pdf)
With Stir Bar:
And Stir bar with magnets placed on it to locate "best" magnet placing (this is where the magnets try to end up when loose)
As you can see, the "optimal" magnet positioning is a little further apart than what I have with the washer. I could replace the washer with a larger one to improve this, but I have smaller stir bars on order. The O-Rings make the system run nearly silently, but it will run without.
BTW, the little "stick" is a spare magnet I glued on to a length of dowel I had laying around to make a "retrieval wand" This is merely a handle to hold on to the magnet on the outside, it does not go into the flask.
Running:
One thing that I would do differently if I built this again, or build another one, is I would get a small proto board for the voltage regulator circuit that I built rather than just wiring it all together as I did. It would make for a much cleaner install, instead of the taped-up mess you can see in the "Guts" picture. I may also use a power connector instead of cutting the end off an old transformer wall wart, but that isn't totally necessary IMO since this power adapter was for a broken tool.
In the background, you can see my fermentor with the second brew I have made, an Oatmeal Stout, which is happily bubbling away. I used the stir plate to make a starter from washed used harvested from my first brew, which is an Irish Stout. I cracked open a bottle yesterday, and while completely drinkable, I think it needs a little more time to bottle condition.
