Supporting > Additive Machining

Cheap 3d printer review

(1/7) > >>

Allen Smithee:
My Better Third bought me one of these for Xmas:


A cheapo 3d printer (under £80) with an unheated bed - piece of complete junk, right? Well surprisingly I've found this machine is much better than you might expect, and with some care on the settings can actually produce good parts. There are various YouTube and website reviews of this machine and they focus on four main issues:

1. The bed is unheated
2. There is no part-cooling fan
3. It needs belt tensioners to remove play in the drives
4. The extruder is really only hot enough for PLA - you can print PETg with it but that really needs a heated bed, and the stronger plastics like ABS and Nylon need more heat.

It also has a very small print volume (100mm/4" cube) but that makes the whole machine very compact - it easily fits on the desk in my study, and can be just put on the bookshelf when not in use. But I'm getting ahead of myself.

Unlike most 3d printers this one comes almost completely assembled - no hours of trying to interpret chenglish instructions and ambiguous diagrams to  establish where screw #331 should be inserted, and none of the famous chinese cheese-head screws (made of real cheese). You just drop the vertical column into its slot (feeding the z-axis connector through a hole), connect the connector, clamp the column in place with two bolts and that's it ready to go.

I gave it the once-over and immediately noted two things - firstly the play in the print head is pretty minimal because it has cable tensioners fitted, and secondly it has a part-cooling fan and ducting incorporated into the head assembly. So it looks like the manufacturers have responded to some of the initial criticisms with upgrades (which is nice). The heated bed I'll talk about later, but the PLA limitation is just one you accept for the cost of this machine.

The bed has a flexible printing mat (think of a 4" square, 1/16" thick piece of the material they make fridge door magnets from) that sticks quite firmly to the print surface and seems like a neat idea. I have bought an aftermarket heated bed (I said more anon - patience is a virtue) but if I use that I won't be able to use this mat. I tried it without the heated bed first so I could see if was worth the effort.

The control unit has four buttons - a "home" button that moves the print nozzle to a datum height (but only height, not X/Y positions), a "+" button that fires up the extruder heater and primes it with filament, a "-" button that extracts the filament so you can switch to a different one, and a "start/stop" button which (believe it or not) starts, pauses and stops the print. Each button lights or flashes its symbol to tell you things while they are in use. On the side of the controller there's the DC socket for the power supply, a type-A USB socket which they say is for "future use" and a slot for the micro-SD card containing the file to be printed.

In the box is a USB micro-SD card reader with a card in it. This card comes supplied with some software, the manual and a couple of sample files on it. I couldn't get the supplied software to install, but that didn't really matter because there's better software freely available on the interweb. For those that don't know, 3d printers need one or both of two kinds of software. First there's some kind of 3d CAD software to create the 3d designs in (usually) a format called "Standard Tessellation Language (STL)" - if you're only going to make parts designed by others you don't need this, but I use SolidWorks. Second is what's called a "Slicer", which is a bit of software that takes the STL file and converts it into a sequence of "G-Code" instructions which tell the head where to go and how fast, while also telling the extruder when to turn on and off (and at what temperature) plus lots of other instructions for other machine parameters. Good slicer software allows you to set thickness of layers, wall thicknesses, types of infill inside the solid bits, whether you use a raft/skirt, the type of supports you use for overhangs etc and all sorts of machine parameters like using different temperature for the first layer. I like PrusaSlicer (free version) so I used this for everything.

Before use you have to level the bed by setting the head to the "home" height, pulling the power plug and then moving the bed by hand to the four corners. In each place you adjust the knurled nut to set the distance between the nozzle and the head using a sheet of printer paper as a feeler gauge (going around 2-3 times because the adjustments interact). Some people say you should do this before each print, but I haven't found the need.

The next step is to prepare the print file. I picked a "Yoda" bust from Thingiverse for my initial tests as it had lots of detail and features like overhangs to test the printer's abilities. I loaded the STL file into Prusaslicer and set it for position on the bed, 180degC nozzle temp (190 for first layer), 0.15mm layers and a 3 layer raft with 2mm skirt*. They supply a short (10m) length of PLA filament to play with, but I had read that it's pretty naff stuff so I had a 500g spool of HobbyKing matt-silver "premium PLA" to use.

Once the g-code file is prepared you just copy it onto the micro-SD card and transfer this card to the printer. The printer prints whatever is the most recent file on the card - no menu to select from or anything. So following the instructions I moved to my first print. With the power off I lifted the head around 30mm up from the surface**. Then I pushed the "+" button and introduced the fibre. The "+" light flashes rapidly while the extruder warms up to purge any old plastic from the nozzle and starts the flow or new material. After a couple of minutes the flashing switches to slow and you can push the button again to stop the purging. This was now ready, so I pressed the start button and...nothing happened.

Nothing continued to happen for about a minute, and then the head & bed moved to their start positions. Another 30 seconds and the printing started:



That first print was done with no supports, was left overnight to print and the result was much better than I expected:



One of the most impressive features of this thing is that it is extremely quiet. My "study" is a doorless 8'x4' room off our main bedroom that was originally a walk-in wardrobe when we bought the house, and I left this print running all night - The Boss didn't even notice it was running.

But I had already spotted a snag. As supplied it has a filament spool holder which seems to be designed for something like 4" spools. All the filament I've seen comes on at least 8" spools. In the above print I dodged the problem by turning the spool-holder outwards and sitting the printer on a box so that the spool hung over the side:



This worked initially, but as soon as the filament started running down it snagged on the flanges of the spool, interrupting the print. There are lots of alternate spool-holder designs on Thingiverse, but to print one I needed a working spool feed. So I went into the workshop and grabbed a piece of 1/8" plywood, drilled four holes in it and produced this crude but workable short-term solution to bring the spool back over the chassis where it could feed more easily:



This is working well for my first "large" print - the improved spool-holder:




This print is nearly the width of the bed, and is exposing one of the known weaknesses of this machine:



The initial layer(s) of the skirt and the raft haven't stuck very well to the mat in this corner, so at the plastic cooled this corner has warped upwards. As it happens this isn't important for this print, so I've let it continue. There are four possible solutions to this problem:

1. It might just be that I need to re-level the bed. Personally I think it's more than that, but I will check it again when this print finishes.

2. I could increase the extruder temperature for the first layer to make it "stickier" (this can be dome with settings in the Slicer software)

3. I could rub Pritt glue-stick onto the bed before printing (according to some websites)

4. I could use a heated bed.

I'll try (1), but I don't think it will work. Then I'll try (2), possibly increasing the thickness and reducing the print speed of the first layer to help it "wet" to the printing matt. I really don't want to try (3) because if that's the answer it will need lots of tedious cleaning between prints to get that mess off the mat. And then there is (4)...

As I mentioned above, I bought an aftermarket heated bed when my Better Third first told me what she was getting me for Xmas. It was about a tenner, and will bolt straight on to the machine. But as supplied it has no temperature sensor and/or controller. It needs either a 12 or 24volt supply, and so I've knocked up a simple temperature controller using an LM35, and op-amp, a couple of resistors, a pot and a relay (apols to the Arduino generation - I was born in the mid Palaeozoic era so I do this sort of thing in analogue) which works well enough. On the "keep it simple" principle I will try the heated bed when I've proved that I can't achieve the same thing with just different settings. Amongst other things it needs 12v at 5A, and that will need a power supply with a noisier cooling fan that might cause the machine's indoor residence permit to be revoked.

Once this spool holder is done I will be trying the first "real" job. I want to make some structural parts for the spoiler blades mechanisms of the 1/4scale Fournier RF4 (radio controlled aeroplane) that I'm building - mainly because I can't see how to assemble the designer's drawn solution without the assistance of nanobots and telekinetic techniques.

But overall - I'm very impressed with just how good a machine this is for such a low price.

AS

*This first print got abandoned because the head drives couldn't keep up with pruslicer's default speed (60mm/sec) so I re-sliced it with a global limit of 30mm/sec which worked fine. As I get to know this machine I'll try replacing this global limit with specific feature speeds to "tune" it for optimum performance.

** I have since found out that if you don't do this the puddle from the initial purging of he extruder melts a small pit in the print mat

Allen Smithee:
So here is that spool mount completed (took about 6 hours to print):



Note the T-slot feature which slides over the flange features in the moulding on the top of the column



This is where the base warped upwards due to the first layers of fibre not sticking to the bed (as mentioned in the previous post). In this case it doesn't matter, but I'll need to find the best solution to this. I have just re=levelled the bed and it was indeed slightly low in this corner, so that may be it (but I don't believe I'm that lucky).



Here's the inside of the T-slot - the camera exaggerates what is actually just one layer of loose strands which were removed by a few scrapes with a small screwdriver. Frankly I'm stunned - this is a void about 1/2" wide which the printer has bridged without supports. I didn't add supports because I could see them being difficult to remove cleanly, and this amazing little machine has done without them!



This is how the support is used. The existing spool holder is unscrewed from the side and pushed into the slot, held in place with a short self-tapper into a boss which has been formed in the print. I used one of the screws that originally held it to the side - perfect diameter but about 3mm too long, so it broke through. I was very impressed that (a) the hole was precisely the correct diameter for the self-tapper so it screwed in easily, (b) the material was strong enough to hold the thread so I could nip the screw up to a hard stop with no feeling that it might be stripping, and (c) the boss showed no signs of wanting to split even though the "grain" (the print layers) ran along the axis of the hole. This gives me lots of confidence in the strength of PLA. As seen here I just need to unclip the spool shaft and put it on the other side.



These images show the finished item mounted on top of the column with a 0.5kg spool, showing the much simpler and more direct feed into the extruder:





The whole assembly is MUCH more rigid and generally satisfactory than the original design.

AS

Pete49:
No such thing as a bad printer just some haven't reached their potential yet.

simplyloco:
Looks neat and great value. I like my Prusa i3Mk3 but as I only play with it perhaps I should have bought what you have!
Happy New Year BTW!
John

Allen Smithee:
Update:

The next item I printed is a mount for the control unit. For this print I re-levelled the bed and then changed the print settings for the first layer (.35mm thick and 205degC extruder temp rather than the 0.2mm and 180degC I use for the rest). This completely cured the adhesion/warping problem noted in the previous posts, so I still don't have a need for that heated bed:



That made this part (the flash makes it look rougher than it is):



...which mounts the control unit on chassis rails like this, making it a much more compact and neat machine:




I'm getting the impression that a 3d printer is just like a lathe or milling machine - the trick is to learn its foibles and work to exploit them so that a skilled operator can make good parts even using cheap machine through practice and feel.

AS

Navigation

[0] Message Index

[#] Next page