Since I’ve been doing some custom design work with embedded devices I needed the ability to make cases for them. So I did some research and taking into account the available funds I ended up purchasing the FLSUN Kossel Delta via Amazon. This package included everything you need to get started, even tools for assembly.
The box was very neatly packed and everything was labelled. Some of the bags had the writing smudged but I was still able to identify things eventually. After reading the documentation and watching the YouTube videos I managed to get it assembled over the next few days. I did end up with some “spare parts” that probably should have been used but were not in the instructions.
The version I got had all injection moulded parts and a few of the holes did not line up properly so I had to improvise a bit but not where it was important. I struggled the most with the hex keys on inside ports due to the space available. Do yourself a favour and get a set of long handled hex keys with a ball end.
The board is an “MKS Gen_L 1.0” according to the text on the board. It supports 5 stepper motors (X,Y,Z, E0, and E1) but only comes with 4 driver boards. I had to upgrade the board’s firmware. So I dug up Marlin 1.1.5 and adjusted it as per FLSUN’s documentation and my own measurements. One thing I did forget to do was calibrate the extruder so for the first six weeks I was under extruding by 40%. I eventually worked it out that I needed to set the ESTEPS to 154 to get accurate extrusion lengths. All of the defaults for STEPS were 100 but I have not touched the 3 axis settings so they could be off as well.
So I used it this way from Sept 7 until Oct 25 when I took it all apart to rebuild it with some different parts. This is due to a few minor assembly issues and the tolerances in the injection moulded parts. I had tried using a bar clamp on all the frames to get rid of the spaces but that had little effect on the calibration values.
|Item||Original||After Bar Clamp|
|Diagonal Rod Length||+220.00||same|
The first major change I have started with is swapping out the square M4 nuts for proper T-nuts. In assembling the top frame these t-nuts fit nicely but the notch on the plastic may get in the way a bit resulting in a not quite proper fit. I used the bar clamp to make sure everything was very tight and aligned. I dug up an old protractor and visually the angles are 60 degrees +-1 degree.
When I disassembled to top frame I did find a fair amount of black “dust” inside. When I removed the pulleys I noticed some oily substance as well. In measuring the dual pulley assembly it came out to almost 6mm depth. The “blank dust” may have been from the edge of the timing belt rubbing. The supplied pulleys felt kind of oily and did not have a consistent look on the ends. I intended to replace these pulleys with a notched version anyway. The supplied pulleys had a 10mm diameter and 7.95mm height. The 16 tooth replacement has a 9.6mm diameter, a 6.5mm depth, and a 8.5mm height.
When tightened the new pulley does not spin, so I ended up adding a 1.2mm spacer between the nut and the frame then tightened things so that the pulley is free moving. Since the pulley has teeth, slippage should be minimal.
In taking the bottom frames apart I came across 1 joint that I missed tightening when I first used the bar clamp a few days ago. So I have reassembled both bottom frames using t-nuts. I made sure to keep everything in the same positions so the t-nuts are the only change. The motors mount between these two frames so I am going to address the motor changes next.
The original package includes a 20 tooth gear for the motors. I once had the X gear slip on the motor’s shaft and my print literally went sideways. So I went looking and found a pulley with a D shape bore rather than the smooth bore I already had. I found this 20 tooth at Amazon but did not have any reference for the seller. I figure a D shape bore would be better as there would be no slippage. I ended up getting a smooth bore 16 tooth from Spool3D instead. I had to file the first one a bit as it would not fit the shaft. It did look a little rough around the edges so I suspect a machining glitch. Otherwise the teeth look a lot better than the original parts. Since I switched from a 20 tooth to a 16 tooth gear I will have to adjust the firmware.
I looking up the settings I came across this page which would seem to indicate the stock settings were wrong. That page has 100 for a 16 tooth gear where the stock settings had that for the 20 tooth gear. The motors that came in my package were SL42STH40-1684A which has a 1.8 degree step. So according to the calculator here my steps should be 100 with the 16 tooth gear and 80 with the 20 tooth gear. This may be partially why I was not getting the accuracy I was expecting.
According to that calculator a 20 tooth gear would get me 80 steps and results in 12.5 micron resolution. The 16 tooth gear would be 100 steps and 10 micron resolution. Just for the heck of it I checked the other smaller gear sizes:
|Step Angle||Micro Stepping||Gear||Steps||Accuracy|
|Changing board to support 1/32 micro stepping.|
|Changing motors to 0.9 degree steps|
|Changing board and motors|
That gets into some seriously small tolerances!!
It also slows down the print speed.
Reassembling the bottom frames and attaching the stepper motors took minutes when using the long handled ball end hex keys compared with the hours it took with the supplied hex keys. Definitely worth the investment in the hex keys, though I am probably going to wear out the M4 hex key.
When reconnecting the timing belt I noticed that it was rubbing on the carriage. So I am going to have to adjust the motor pulley. The instructions say 5mm from the end of the shaft, I’m going to change this to 6mm.
I also added M3 x 6mm screws to the top of the carriage. This is similar to other delta models as it makes adjusting the end stops much easier. I spent hours getting the original setup to calibrate as the square nuts just did not catch properly. Of course, the end stop nuts are really only good for course adjustment. Using the screw is substantially finer resolution. According to this page, a single twist of the screw will move 0.5mm.
In trying to use some TPU filament I found a few issues with the supplied extruder. I made an adaptor to reduce the open parts of the filament path but this only had minor improvements and the filament still bunched up after the gear. In looking at a number of pictures I have come to the conclusion that the extruder supplied is wrong and the filament path curves over the gear rather than straight across. This is not a big deal with the hard filaments. So I picked up an E3D Titan extruder with the bowden adapter. The wire that came with it is not compatible with my board. I do have a 4 pin JST connector that will fit the board if I feel the need to use both extruders. But I only have a single hot end so I do not need both extruders enabled concurrently, unless I want to get into material change for dual colour prints.
I ended up replacing 60 x M4 square nuts with the t-nuts.
45 x M4 t-nuts for frame. 5 per corner with 3 corners per level and 3 levels.
2 x M4 t-nuts for board mounting bracket.
3 x M4 t-nuts for end stops.
3 x M4 t-nuts for hot bed mounting brackets.
2 x M4 t-nuts for extruder mounting.
2 x M4 t-nuts for Titan extruder.
2 x M4 t-nuts for filament spool holder.
(total of 54 x M4 t-nuts)
3 x 16 tooth pulley with 3mm bore.
Tomorrow I’ll turn it on and run though the calibration routine.