Mar 242014

I have two 3D printers, both are the FDM types that squirt out plastic filament and build up an object layer by layer. My first printer is MendelMax 2.0 cartesian printer I built from a kit a couple years ago and the second is a RostockMax delta style printer I’m in the process of building.

The MendelMax has been a good printer but I’ve always had mixed results. I get good prints one time and horrible prints the next. I built the printer on a table in the family room, so I could work on the printer but still be around the family, and did all of my test and tuning prints sitting on the same table. Results were consistently inconsistent. Eventually I got frustrated, and tired of having the printer sitting in the family room so I moved it downstairs to my workroom. My workroom is a bunker style room in the basement with no windows, no vents, and only one door. I immediately noticed a huge improvement in print consistency. Prints might have errors but the errors were repeatable and reproducible. The number of errors were significantly smaller when printing in the basement than in the family room and the errors were consistent. It finally occurred to me that being in the family room the print nozzle and print surface was subject to the whims of breezes and temperature changes that come with people entering and exiting the room, windows and doors being opened and closed, uncontrolled variations that changed the print quality every time the environment changed. Being in the “bunker” in the basement the environment was mostly stable and undisturbed. With no windows or vents and only one door the basement room is a pretty consistent temperature and there are very few random breezes. A more stable environment led to more stable printing.

After moving the printer to the basement I could finally get consistent prints, not error free prints, but repeatable prints. When I did have a recipe for an error free print I could print the same object again and again and finally expect the same results. Yahoo! Life is good! But, you know the old saying, if you give a mouse a cookie…. Anyhow, now that the major obstacle was overcome, just being able to get a repeatable print, it was time to start looking at secondary failures.

The next biggest class of failures I experience are the saggys. Saggy prints, bridging that droops, overhangs that fall and loop, glooping and bumps, all failures indicative of printing too fast, that is, not letting the material cool before trying to lay down another layer.

Here are some images of the Pink Panther Woman printed with one or two perimeters, in ABS with 0 to 40% infill. Mostly with one perimeter and 0% infill.


You can see multiple places in my prints where the filament sags and droops when trying to print an overhang.






I found that if I print really slow, like really really slow, I could get a great print. The nozzle takes so long to traverse the print the previous layer had time to cool before being visited by the printhead again. This worked but the print time was increased by a factor of 3 to 10. That was clearly unacceptable.

The next solution was to add a print cooling fan. The kit I bought was the MendelMax from I have been super happy with the kit, it is well made, well supported, and super complete. All of the parts needed to assemble the kit, plus extras was included. A nice organizing case was provided for the screws and small parts. Cables came pre-terminated. I had to do a little soldering but for the most part the kit was plug and play. A very clean build and with what seem to be quality parts. I’ve not had a bit of trouble with the printer to date.

One thing the kit did not ship with was a fan duct for filament cooling at the printhead. It did ship with the fans, RRD Fan Extender, and cables to add up to two filament cooling fans, but no fan duct.

The first step in figuring out how to add a fan was deciding if my software and electronics could support an extruder fan. I’m running Marlin as the printer firmware so a quick google search confirmed Marlin’s support for an extruder fan. Now how about the electronics. The printer is driven by a RAMPS 1.4 on top of an Arduino Mega clone with Pololu style stepper drivers. All in all a good electronics platform but it turns out the RAMPS 1.4 out of the box does not support an extruder fan as a standard option. Some might take issue with this last statement given the unused D9 port, but D9 was intended for a second extruder, not an extruder fan.

In order to drive a fan a high power output is required, higher than can be driven by the bare io pins from the Arduino. The RAMPS board provides three high power outputs, D8, D9, and D10.


D10 is used to power the heater on extruder 1, D8 is used for the hot bed, and D9 is intended for extruder 2. On a single extruder printer D9 is unused, and therefore could be used to power and control an extruder fan, and indeed there are multiple internet responses supporting doing just that, wiring a fan into D9. So indeed it is reasonable to say that the RAMPS board supports an extruder fan out of the box…. assuming you never want to add a second extruder. For those who think they might want to have a second extruder D9 is off limits and another solution is required.

Obviously I’m not the first guy to want dual extruders AND extruder cooling fans on the RAMPS electronics. To satisfy the limitation in the RAMPS board of no high power outputs for extruder fans some enterprising souls created the RRD Fan Extender board. The RRD board mounts directly onto the RAMPS board and adds two channels of mosfet controlled power output. These two channels are not nearly the current capacity of D8, D9, and D10, but are more than enough for running a fan. The RRD board is able to take input power from 5V or 12V so it can support 5 or 12 volt fans. The RRD board has two channels of output supporting two extruder fans for dual extruder setups.

Something to keep in mind when buying a RRD Fan Extender. The servo port brings out pins D11, D6, D5, and D4. Some RRD boards bring D11 and D6 out to drive the two fan mosfets and some RRD boards bring D4 and D5 out to the mosfets. You need to check which RRD board you have because the printer firmware, Marlin in my case, needs to be configured properly to tell it which pins the fans are on.

If your RRD board uses pins D4 and D5 no software changes are required because the stock Marlin configuration for a RAMPS 1.4 (motherboard 34) is already configured to use D4 as the extruder fan 1 control. If your RRD board uses D11 and D6 you will need to edit pins.h and change your FAN_PIN assignment.

Like I mentioned above, the 3DPrinterTek kit shipped with a RRD, multiple fans, and all of the cabling required, but no fan duct. Not really a big deal, there are plenty of fan duct models out there, just download one, print it, and wa-la, instant extruder cooling.

Step one was find a fan duct compatible with my printer. A quick search of Thingiverse turned up this duct for a MendelMax.


I printed the duct, mounted a fan to the duct with 3mm screws and nuts, zip tied the duct to the print carriage, and wired it up. The RRD mounts onto the servos port of the RAMPS and power is jumpered from either the 5V or 12V aux power. My fans are 12 volt so I jumpered to the 12V aux power pins.


The RRD supports up to two fans but I’m only using one so I only have one of the outputs cabled.

Here is the printed fan duct zipped tied to the printer carriage.





Notice the fan only points at the back of the print so I expect I may still have problems with the front side of the print.

With the fan duct installed, the RRD on the RAMPS servo port, and Marlin configured to use pin 4 as the FAN_PIN I ran a test with the same model.



Notice the sagging on the back of the butt is nearly gone and the sagging on the upper back is completely gone. The quality of the prints, at least from the back, has jumped tremendously. The next step is to either get a fan pointed at the front of the print or print a wrap around cooling duct.

Maybe like this wrap around fan duct.