Tuesday, Sep 4 2018
I’ve done several things to my CR-10S in an attempt to get better quality prints. The issue, I believe, is inconsistent extrusion. The biggest improvement came when I changed from the stock filament spool holder to this one but I’ve also changed extruder springs, software, blower fang, slicer software, printer firmware, etc.
A few days ago I updated the firmware to stock Marlin 1.1.9 using the supplied CR-10S configuration plus enabling LIN_ADVANCE. I calculated the K-factor value using generated G-code. I had a hard time printing the test due to under extrusion. I cleaned the nozzle and hot end and re-calibrated the extruder. The calibration test said I needed to increase the steps/mm by about 11%. The next test said I was over extruding by about 5%. Every test was somewhat different. I narrowed in on a usable value and got OK results. Mostly.
I even got varying results before feeding filament through the hot end. Time for a bigger change. I ordered the Bondtech dual extruder for the CR-10S.
Look what arrived today
No instructions
Extruder, CR-10S mount, and hardware
That is what arrived today. There were no instructions in the package. I found a PDF of the install instructions on the Bondtech web site. Seems simple enough. Let’s get started.
Heating nozzle to remove filament
Filament removed
First things first: heat up the hot end and remove the filament. With that done I turned off the control box and got to work.
Disconnect filament sensor and bowden tube
Disconnect stepper
Remove wiring to the filament sensor and the stepper motor.
Remove filament sensor bracket and extruder tension arm
Remove extruder and stepper
I removed the filament sensor bracket out of the way. Next was removing the extruder tension arm with spring so I could get to the screw that was under the arm. Holding the stepper so it wouldn’t fall I unscrewed the stepper motor adn extruder housing from the CR-10S bracket.
Remove filament sensor from stock bracket
Filament sensor in Bondtech bracket
At my desk I removed the filament sensor from the stock bracket and installed it into the Bondtech bracket. Two of the original screws are used to hold it it place. It looked like the sensor switch lever was too high to be triggered by filament. Nope. Once the top was screwed in place the sensor sits low enough to do its job. I verified that with an ohmeter and a short piece of filament.
remove stock drive gear
Remove burrs from shaft
Install Bondtech drive gear
I felt a burr on the stepper shaft where the second set screw held the stock drive gear. I gently filed the burr down until the Bondtech drive gear would easily slide down the shaft. It is held by a single set screw on the flat of the shaft.
Mount extruder and stepper to bracket
CR-10S mounting screws
The stepper motor goes on the bottom of the bracket, held in place by screws going through the extruder on top of the bracket. The stepper drive gear meshes with the white extruder gear. Once everything is in place I tightned the extruder mounting screws and installed the drive tension screw and spring assembly.
I went looking for the screws that attach the stepper/extruder assembly to the CR-10S. I knew I’d seen them someplace. It took me at least 5 minutes of searching before I thought to turn the assembly bracket around to find the screws threaded into the bracket. I removed the screws and took the assembly over to the CR-10S.
Mount unit to CR-10S
Other side
The stepper/extruder assembly was attached to the CR-10S. I like that the filament path is no where close to the Z-axis lead screw.
Install bowden tube
Connect filament sensor
With the extruder in place I installed the bowden tube and the filament sensor wiring. There is a clip on the bowden tube connector that must be removed to install the bowden tube. It helps keep the tube in place once installed.
Stepper wiring change
Ready to connect stepper
This is the hardest part of the job – and it is not very hard. Because the extruder is driven by a gear the direction of rotation has to change. This is done by swapping two of the wires in the connector going to the extruder stepper motor.
I used a very small jewlers screwdriver to remove the two pins from the connectory. The pins are pushed back into the connectory in swapped locations.
Connect stepper
Feed filament
With the wiring modified the stepper can now be plugged in. That’s it. I turned on the control box and started heating the hot end before loading filament.
While the hot end was heating up I made the needed firmware configuration change. The geared drive reduces the amount of filament moved per step. The recommended setting is 415 steps/mm vs the stock 95/mm. I used Octoprint to enter m92 e415 to set the number of steps and then m500 to save the change. If using the stock CR-10S firmware you’d have to change your slicer to send the m92 command at the start of a print.
I loaded some filament until it was in the bowden tube then marked off a 120mm section and told the machine to feed 100 mm. When done I measured 20 mm to my mark. Perfect. I repeated the test. Perfect again. This is already better that what I was seeing with the stock hardware.
I printed a Benchy as a test. It came out OK, with a hint of stringing. The stringing is somewhat new, something that started when I switched to Marlin 1.1.9. Slicer changes to my retraction settings may be needed. The bottom of the Benchy was not perfect. I need to check my bed height and first layer settings.
Benchy
Finished Benchy 1
Finished Benchy 2
Finished Benchy 3
Finished Benchy 4
Finished Benchy 5
Finished Benchy 6
Finished Benchy 7
This Benchy printed in 1:32 including the time it took to bring the bed to 60 °C. It was sliced using Slic3r Prusa Edition using a 0.2 mm layer height profile.
Thursday, Sep 6 2018
Test cube
This is the bottom of a test cube printed with one bottom layer and one shell after I adjusted the bed height (it was off by a bit) and tweaked the slicer settings, particularly the extrusion multiplier. I set the multiplier to 1 now that I seem to be getting consistent extrusion. Cube X/Y dimensions are good (within 0.02 mm). The Z dimension was a short by 0.1 mm. No zits or blobs in walls or base. I’m happy with the way the print turned out.