On occasion, I have had to use an additional fan to provide cooling while printing multiple hour long prints with PLA. I’ve been doing so with this fan:
The problem with this arrangement is that the fan is 110v manually controlled by me. If it is on during the initial layer, there’s a good chance that it won’t stick to the bed and if it isn’t on for the upper layers, heat creep or the top layer creeps up until the nozzle hits. Since I’m easily distracted, I don’t always do my job right
My solution is to mount two 80x80 12v fans to the frame as pictured below (twice is not my fault). The easiest electrical solution is to wire these two fans in series and then connect them (in parallel) to the existing extruder mounted cooling fan wiring.
Is this a viable electrical solution? If not, suggestions are welcome!
Only you will be able to tell if that’s a viable solution or not. Seems like a lot of work than simply replacing the middle cooling fan with one that draws in more air. I agree the stock fan is under powered to cool PLA efficiently.
I’ve tried the desktop fan trick with no better results then it not being there. What did work for me was backing off the tension on the screw on the hobble bolt. I hit the sweet spot and am able to print PLA beyond 1.5 hours without failure. It’s still a little touch and go though, after several long prints in a row I will get a failure but still much better than what it was before with too much tension on the idler screw. I will be replacing the always on middle fan with a more powerful one that can move more than 3.x cfm of air.
Thanks for the input. I had excellent success with the external fan in the first picture except when the idiot operator didn’t turn it on or off when he was supposed to 
.
A more powerful middle cooling fan should help with heat creep but the problem I’m seeing is not heat creep, its a lack of part cooling causing the previous layer to not quite be solid when the next layer goes down. Those fans could also be replaced but I have multiple tool heads so that may not be the most cost effective solution.
This is an experiment with parts I already had on hand.
