I succumbed to my «need» for a larger printer, and decided to try a Delta. I found what looked to be a very good deal on AliExpress. Auto Leveling Delta 3d Printer Large Printing Size Reprap Kossel Delta 3d Printer DIY Kit with HotBed 8GB SD Card Free shipping for US$225.12 (at the time, now $258). I thought I’d share my experiences, in case anyone else is considering going down this track. Conversely, if someone else has been here before, I’d welcome your comments and suggestions.
The printer is described as 200mm diameter by 300mm tall build volume. It includes an LCD display with an SD card slot, and a filament supporter (free gift). The main structure is 2020 aluminium extrusions, with acrylic for the top and bottom platforms, and a couple of other small bits. Controller is a ZRIB V2.0 board, which essentially combines a RAMPS 1.4 with an Arduino into a single board (think PrintrBoard). There are screw terminals for the 12V supply and heaters, the rest are plugs. Hot end claims to be an E3D V6, but I’m sure it’s some kind of Chinese copy. There’s a fan and shroud to clip to the heatsink, and another fan and shroud for filament cooling. All the electronics and motors are in the top of the printer, except for the heated bed and its controller, which is completely separate. There are two 12V power supply bricks, one for the heated bed and one for the rest of the printer. Documentation is provided by downloading PDFs. these are a series images describing assembly, wiring, etc. Anyone who has assembled flat-pack furniture will be familiar with them.
The best thing I can say about this printer is that it is cheap. The quality of design and manufacture and instructions are all average at best. I think that, as supplied, this kit would result in a very poor 3D printer.
- The 3D printed parts are poor quality & weak. One of the motor mounts broke when attempting to attach the motor, another broke when I was tensioning the belt. These appear to be printed with 10-20% infill, and are about 4mm thick. I’ve epoxyed them together for now, but that part needs to be redesigned and remade. I had to scrape away overhanging lips and so on to allow the aluminium extrusions to fit against the 3D printed parts.
- Some of the screws absolutely *cannot* be done up. This is just poor design, probably combined with saving a few cents on Philips-head screws as opposed to socket-cap. The screws to attach the motors are Philips head M3x8mm. They are placed inside each top triangle piece, where there is *no* way to get a screwdriver to one them. The others are very difficult (i.e. you can’t get a screwdriver straight onto them). Obviously, no-one at Zonestar has actually assembled one of these printers. Fortunately, I had a supply of M3 socket cap screws that I could use instead.
- Bad screw selection. The M3x8mm screws supplied were all countersunk heads, whereas the holes were not countersunk. Or maybe that’s bad design of the plastic parts again.
- Some required screws are missing. I needed 6 M3x20mm screws to attach the diagonal arms to the carriages. None were supplied. I think this was porobably a mistake to due to changed design of the arms.
- The instructions don’t match the hardware. The arms are shown in the instructions as having screws attached, What I got was ball socket ends on the rods, which required screws to attach them to the effector and carriages. In some cases the instructions say to use M3 screws to attach to M5 nuts, or vice versa.
- Some instruction images are black on black. i.e. difficult (even impossible) to understand.
- The heated bed has a separate controller. This is not mentioned in the sales pitch at all, and is potentially a major nuisance, becase at the end of a print, there’s no way for the controller to turn the heated bed off. That will be a manual process, or maybe a fixed timer, or an external SSR.
- No housing/mount/buttons for heated bed controller. The heated bed is controlled by a separate controller, with its own little PCB, 7-seg LED display, and 3 buttons. This PCB is just dangling on the end of the wires from the heated bed. The buttons are little micro pushbuttons on the PCB itself. Very obviously an afterthought.
- No insulation for heated bed. Theheated bed is a 3mm aluminium plate, with a heating element on the underside. There’s absolutely no insulation on the heating element, so half the heat will be wasted. Worse yet, it will be heating the top of the table or whatever the printer is sitting on.
- Heated bed wires held in place by tape. The power wires to the heated bed are soldered on, and the thermistor appears to be in a cartridge similar to a heater cartridge. The only mechanical attachment of thermistor to bed is what appears to be ordinary electrical insulation tape.
- The LCD screen is not attached to the printer. The LCD screen is provided with a couple of acrylic legs to sit on, but not mechanically attached to the printer at all. It sits on the top of the printer. It doesn’t take a rocket scientist to foresee this falling off.
- There’s no belt tensioning mechanism. The belts go round the motor pulleys at the top, around pulleys at the bottom, and the ends are attached to the carriages. There is no mechanism for tensioning them. Given the continual motion to and fro of all 3 motors during Delta printing, slack in the belts will result in poor print quality. Again, thoughtless design. I’m thinking about a simple belt tensioner design.
- There’s nothing (apart from tension) to stop the belts from wiggling out of their slots in the carriages. I guess I’ll need to put a cover over the belt ends. Fortunately, there’s holes in the carriages that this could be screwed into.
- The «boat nuts» need to be held in place somehow. «Boat nuts» fit in a slot in the aluminium extrusion, allowing a plastic part to be tightened against the extrusion. When at right angles to the extrusion, they hold. Turning the nut through 90 degrees allows it to come out of the slot. There’s no way to stop the boat nuts from rotating, and no way to see the orientation of the boat nuts once the plastic is in place. This makes for a never-ending nightmare of placing the nuts, putting the plastic in place, and then hoping that all of the nuts stay in place while tightening screws into them. If one nut turns at any point, you have to take them all out and start again. A little thought in the design of the 3D printed parts could have prevented the rotation. In the end, I hot-glued my boat nuts in place.
- The motors are small. This is conveniently skipped over in the sales blurb. Small motors means that they have low torque, which in turn limits the acceleration and speed of printing. And fast(er) printing is one of the benefits of a Delta design. I expect I’ll want to upgrade these some time in the future.
- The fans are small (20mm) and axial. i.e. they won’t blow much air. I don’t have a lot of confidence in their likely cooling effectiveness. A too-small fan on the heatsink might lead to clogs and jams in the hot end. A too-small filament cooling fan might lead to poor print detail, and/or warping. I expect I’ll want to replace these with 40mm fans or blower (radial) fans in the future.
- The auto-levelling is done by a microswitch on the effector. I’m not sure how effective this is going to be, or how it works.
- The Bowden attachment for the extruder had been thread-stripped. The Bowden attachment screws into an acrylic piece on the extruder. it just won’t stay in. It comes pre-assembled, so whoever put that together over-tightened it, I guess.
- The documentation doesn’t mention the mechanical side of the wiring at all. for a long life, it is important that wires should be solidly anchored at the effector and control board ends, so that all flexing is done in the bundle. A handful of zipties and a long length of spiral-wrap is provided.
OTOH, having made all these complaints, I have (so far) successfully assembled the printer, and have it mostly wired up. I expect that eventually I will be able to use it effectively as a 3D printer, but also expect that there is quite a bit of work ahead of me to make it into a good printer.
I’ll post more information as I progress.
in my opinion its upside down top heavy
Check my rubbish blog for my prusa i3
I agree that putting all the motors on the top seems contrary to good sense and may make it top-heavy. Will see about that eventually.
Yesterday, after posting, I had a guilty and stupid feeling, because I hadn’t actually viewed the construction videos, because I hadn’t understood that the documentation was available on the supplied SD card. I looked at some of them last night; they consist of CAD animations, showing the parts moving into the right position. They’re a bit better than the construction PDF document, but not much. Out of interest, I looked at the video showing how to add the last corner to the bottom triangle, because that was a bit problematic when I did it. The video shows the corner piece moving *through* the aluminium extrusion. (What worked for me was to assemble two corners to two extrusions, then separately the remaining corner to the remaining extrusion, then put those 2 assemblies together).
Some other things: The controller board (ZRIB V1.0) is apparently (based on) a Melzi board. there’s a reference to that in the documents.
Plug the wires into the control board before fixing it to the top platform. You can’t plug the power connector in afterwards, and you can’t read the labels on the board. Similarly, plug the wires to the extruder motor in before mounting it.
Lower the effector all the way to make sure your wiring will have enough length. Attach the wires at the effector firmly. there’s a convenient hole in the leveling probe microswitch PCB ). Make sure that there is *no* movement in the wires between the attachment point and their various terminations. Any movement of individual wires will lead to stress points in the wires and breakages. Similarly, attach the bundle of wires firmly to the top platform, spiral-wrap all the way from one attachment to the other, and attach the spiral wrap at each end.
The connectors on the LCD are labelled EXP1 & EXP2. The matching connectors on the controller board are labelled the same. There’s no other way to know how to connect these two together (glossed over in the wiring instructions).
I’m not sure which fan is named what. the names aren’t very clear. I’ve taken a guess and plugged them in. worst case, I’ll just have to swap the connectors over.
The LCD instructions don’t match the hardware. they describe an LCD display with 5 buttons, whereas what was supplied has a single rotary plus push-pull button.
Other than this, the wiring instructions are excellent and easy to follow.
The controller board itself has all the connectors needed to control the heated bed, but according to a note in the instructions, the code to do this has been removed from the firmware. Hopefully (a) the bed heater connection is good enough to run 10 amps, and (b) the code can be added back in.
Nevertheless, I’ve powered up the controller board, and that and the LCD seem to be perfectly functional.
Next step: check the motors & heaters.
Edited 1 time(s). Last edit at 05/02/2016 06:43PM by frankvdh.
Nozzle heater & thermistor & fans are good and work fine. All XYZ top limit switches are working correctly. Heated bed heater wire separated from the bed.
The biggest problem is that I can’t get the X,Y,Z motors to move the carriages — the motors just skip. I thought perhaps the wiring was screwed up to all 3 motors, but it’s identical to the extruder motor wiring, and plugging the extruder motor into each of the X,Y,Z motor connectors in turn resulted in the extruder motor working reliably and predictably. I’ll try the X,Y,Z motors without belts to confirm. But I think now that the real problem is that there is a great deal of resistance to the carriages rolling up and down.
And I’ve broken both arm attachment lugs off one of my carriages. I’m not sure how. I wouldn’t have thought that there would be much force applied there. But, again, looks to be 10-20% infill, and maybe 1mm walls, so just not strong. Replacing these carriages would require a fair amount of disassembly, including removing the top or bottom platform. Epoxying the lugs back on.
Zonestar support has responded promptly to my requests and provided the source code for the firmware (Repetier Firmware) and the STL files for the 3D printed parts.
To do: Look more closely at why the carriages are hard to move. Try X,Y,Z motors without belts. Fix heated bed and test. Try extruding some filament. Look at how the Z-min endstop is used for auto-leveling, and check that it’s working.
More progress. extruded my first filament, no real problem. It was a bit difficult to feed the filament through the Bowden connector at the extruder end though. I had to disassemble things to feed it through one piece at a time.
All the motors work without belts on. It would have been a good idea to check all this out before putting the belts on. But I’d done the mechanical assembly, followed by doing the wiring.
Turns out part of the problem with the carriages is that the carriages and/or wheel spacers aren’t (sufficiently) dimensionally accurate. When the screws attaching the wheel to the carriage are cranked up tight, it presses the wheels laterally against the extrusion, increasing the friction. After I backed off the screws by a turn or so, they become much easier to move. Another part of the problem is that the wheels aren’t round. They have little ridges on the surface the runs against the extrusion, which causes the carriage to move in clunky steps. I scraped away those ridges, and then vigorously ran the carriage up and down the extrusion by hand.
So I have one carriage now that moves under program control But it moves in the opposite direction to what I expect. A Home command moves it down, *away* from the end-stop microswitch. Requiring a quick pull of the power before it crashed into the bed. Coincidently/fortunately/suspiciously, the wiring instructions include a section on how to pull the pins out of the connector to reverse their order and consequently reverse the direction of movement. And, from experience, I don’t like the idea of pulling those pins. it will be difficult to do that without damaging the connectors. That’s *really* irritating, because a *tiny* piece of work by the developer would have saved every purchaser from having to pull all those pins to reverse the direction of the motors. (a) They could surely have provided cables that were wired up correctly, or (b) a quick glance at configuration.h in the source code shows that there are settings for motor direction and homing direction. I’ll be looking at recompiling the software myself, with the right settings. a bit of trial and error to figure out the right values for those two things.
Hi thanks for this post, could you share these parts ? they hasn’t sent them to me
ho and I’ve got a D810 too. (I’m french so my english could sucks a bit)
I didn’t have any problem with the wiring, everything works at first try.
1 — Parts : definitely need to reprint carriage parts, with a bigger infill. Also I’m gonna modify the carriage in order to be able to tense the belt.
2 — Power supply : it’s sucks, the 5A adapter doesn’t even deliver nearly 5A, so the board doesn’t works. And there’s also another adapter for the hotbed . I’ll get rid of these and use an ATX power supply, put a main switch and wire all of that to the power connector on the ZRIB.
3 — Case : the case for the LCD suck need to reprint one fixed to the printer. Also, no case for the hotbed board :/ need to print that too.
4 — MCU : the ZRIB 2.0 is a good board, but the documentation given with the printer is for the old generation (1.1), and there’s a big gap, the old one is based on sanguino, the new one on ATMEGA2560. Its flashed with repetier firmware. Compare to the original repetier firmware, only few constant declaration on pin.h are made due to the exotic origin of the board (not officially referenced by repetier). There’s something that tingling me, it’s seems that pullup resistor is directly sodered to each endstop connector, which means (almost sure about it) that pullup parameters in firmware has no effect on the board. need to confirm that point.
5 — Extruder : it’s a really good part, and the filament support is well design, a very good point.
6 — Effector : Let’s be quick, the Z probe mechanism sucks, it smashed the hotend to the bed in order to trigger a mechanical MakerBot 1.2 sensor on the effector. It’s the first thing that I’m going to replace, I’ll go for inductive sensor (an NPN one). Also, results from probing are not consistent with this mechanism due to the mobile part of the hotend support. Definitely need to re-print a better support for the hotend where it can’t slightly move.
7- Contact support : the guy behind the support is definitely a real hardware engineer, he always answer my requests and questions, even if they are really specifics. Of course he’s chinese, so sometimes it’s difficult to understand perfectly it’s english, on the other hand it’s the same from me to him (french)
8 — Firmware settings : Default settings are good, good print without any heavy tweak to the settings, just the usual stuff. But there’s a potential critical issue with parameters of the P1-P2-P3 coordinates. These coordinates are use for auto leveling procedures. P1-P2-P3 coordonates in the firmware settings draw an isosceles triangle, which is good for a G29 command but very bad for a G32. G32 build a grid from P1-P2-P3 coordinates, which end to a parallelogram with these settings instead of a square. Because P1-P2-P3 coordinates are quite close to the border of the bed, a part of this parallelogram is out of the bed. I almost crash my hotend yesterday, hopefully I was watching and saw that the pattern was somehow strange. Personnaly I use the following coordinates (X/Y) : P1 -60/-60 ; P2 60/-60 ; P3 -60/60 which draw a right triangle, not as good as an isoceles triangle for G29 command, but definitly safe coordinates for G32 commands. Because my english could be bad I’ve made a drawing of what I mean :