Firming Up

In getting where I have with my printer, I have had to spend some time experimenting with and tweaking the various firmware and software settings to improve the overall quality of my prints. I have made good progress, but I am not finished yet, by any means! (For the record, I am using Sprinter firmware, Pronterface software with Skeinforge, with RAMPS 1.4 electronics.)

Here are my main firmware settings (software settings will be in a following post), which are producing reasonable prints on my machine (your mileage may vary!):

Firmware Settings – Sprinter:

The settings I have had to adjust are in the “Configuration.h” file.

Line 15: Set the Motherboard ID value to match your electronics hardware; I am running a  RAMPS 1.4 so it’s:

#define MOTHERBOARD 33

Line 25 / 26: Define your temperature sensors; I have a 100k thermistor on my Hot End, and no Heated Bed as yet:

#define THERMISTORHEATER 1

#define THERMISTORBED 1

Line 30: Calibrating the steppers; after a bit of trial and error, printing and measuring calibration blocks, etc, I am using:

float axis_steps_per_unit[] = {80.*40./51., 80.*40./51., 3200./1.25, 700.};

(This was originally {80, 80, 3200/1.25, 700} by default for a metric Prusa Mendel with a Wade’s Extruder.)

Line 45: Set the comms baud rate:

#define BAUDRATE 115200

Lines 81 – 83: Setting the build volume; as reported in an earlier post http://julianh72.blogspot.com/2011/08/assembling-y-is-there-always-something.html  , my print range is somewhat restricted, so I am using:

const int X_MAX_LENGTH = 175; // (was 200)

const int Y_MAX_LENGTH = 175; // (was 200)

const int Z_MAX_LENGTH = 110; // (was 100)

I think pretty well everything else has been left as per the defaults. Save the whole Sprinter Sketchbook project in Arduino, upload to the Arduino Mega, and we’re ready to start tweaking the software …

Preparing the Machine – Pre-Flight Checks

This post is just a quick summary of the checks that I try to remember to carry out before each group of print runs – say at the beginning of an evening of printing.

Print Bed Medium

Firstly - if you are printing with PLA onto a non-heated acrylic bed, put down a layer of “Scotch 3M Blue Tape 2090". I can’t over-emphasize what a difference it makes having a good substrate that the build will adhere to, but will separate from the printed object at the end of the job. The surface of the blue tape tends to get a bit damaged with repeated use – be prepared to replace it periodically when the marks and scores from your earlier prints become very apparent. The frustration of having a long print job de-bond and separate part-way through is MUCH greater than the cost of a  metre or so of blue tape it needs to “refresh” the print bed surface (as I found out – the hard way!) Do yourself a favour – just do it!

Level Print Bed

Check the gap between the extruder nozzle and the print bed. I am using a feeler gauge, and aiming for about 0.3 – 0.4 mm. I’m not perfect at getting an even gap over the whole print bed, but I seem to be doing OK. My print bed levelling screws are installed screw-head-down / thread-up, and have the following components (from bottom to top):

1.       Screw Head
2.       Washer
3.       Lower Bed Plate
4.       Washer
5.       Spring
6.       Washer
7.       Levelling Nut
8.       Washer
9.       Upper Bed Plate
10.   Washer
11.   Locking Nut

I am sure there are many permutations on this theme, but this works for me.

My levelling procedure is as follows:

a)      Slacken the 4 top Locking Nuts.
b)      “Home” the machine, and adjust the Levelling Nut nearest the (0,0) corner to get the gap right at the “Home” point.
c)       Move the X-axis to near its upper limit (about 170 mm in my case), and adjust the gap at that corner.
d)      Move the Y-axis to its upper limit (again, about 170 mm in my case) and adjust the third Levelling Nut.
e)      “Home” the X-Axis, and adjust the 4^th Levelling Nut.
f)       Tighten all 4 Locking Nuts
g)      Cycle through the four adjustment locations again, to make sure all is well; adjust if necessary.
h)      Move to somewhere near the middle of the print range - about (100,100) in my case - just to check the gap in the middle.

Free Movement

Check free movement of all axes – should all be quiet and free of audible “chatter”. If any of the movements is showing a bit of “stickiness” or noise, I put a couple of drops of light machine oil on the smooth rods, and move the axes through their full range a couple of times. I’m not sure how much this improves print quality – but it can’t hurt!

When the pre-flight checks are complete – you’re ready to print!

And the results:

I’m pretty pleased with what I am getting out of my machine now. Having got the machine working to my basic satisfaction, I am now concentrating on printing useful objects; that is not to say I am done with fine-tuning and calibration, but I figure I built this thing to MAKE stuff, not just for its own inherent magnificence (which is substantial!) and perfection.

In particular, I decided I need a Filament Spool so that longer prints can run continuously without me having to continuously check and untangle the filament coal lying next to my printer.

I looked at a few options on Thingiverse http://www.thingiverse.com/  before deciding on this stand http://www.thingiverse.com/thing:8290  (which clips onto the top bars of a Prusa Mendel), together with this printable spool http://www.thingiverse.com/thing:9039  , both by a contributor who calls himself 4ndy. Undertaking a substantial bunch of print jobs like this is a good test of whether my machine is basically up to the task. I'm still printing all the bits, but so far, I'm pretty happy with quality and progress:

When you can print one of these ...

... you start to believe that ANYTHING is possible!

I downloaded the STL file for the Twisted Bottle http://www.thingiverse.com/thing:12278 and pressed the print button.

Here it is underway:

And here is the finished product on the print bed:

And here is the finished article:

You've got to admit, it's getting better ...

... A little better all the time!

Well, I owe a HUGE debt of gratitude to Triffid Hunter and David Drew for their comments on my previous post http://julianh72.blogspot.com/2011/10/cheers.html . Triffid Hunter was 100% spot on in diagnosing my main problem as being backlash.

When first assembling my machine, I had no real guidance as to how tight the belts should be, and I was worried that if I over-tightened them, I could stall (or burn out ) my stepper motors, or break my machine. Triffid Hunter said "They should make a tone of 100hz or more when plucked like a guitar string. That's a low G sharp for the musically inclined." Or to put it another way, they should be "as tight as a tourniquet, dry as a funeral drum". (Google it, if you don't get the reference.)

So I set to tightening my belts. My X-axis belt was not too bad, but needed just a little bit of tightening, but my main problem was my Y-axis belt. Having had the root cause pointed out to me, sure enough, you could actually see the backlash on the Y-axis pulley every time it changed direction.

I have seen a number of clever tricks to tighten the belts on your RepRap, including this http://www.thingiverse.com/thing:10082 and this http://www.thingiverse.com/thing:4535 , but I decided to go with a much simpler solution - make a loop of cable ties around the ends of the belt, and then pull it nice and tight, and cut off the loose ends:

Simple (crude, even), but very effective, and easily adjusted without needing to tighten or loosen any clamps. If you over-tighten it, just cut it away, and have another go. So, I tightened both belts, and repeated some of my test prints - and the results are an astounding improvement (to me, anyway!):

Compare the 20 mm Calibration Cube to what I was getting earlier http://julianh72.blogspot.com/2011/10/always-listen-to-people-who-know-what.html and my second attempt at the RepRap Teardrop Shot Glass http://www.thingiverse.com/thing:11944 against my first attempt http://julianh72.blogspot.com/2011/10/cheers.html , and you would have to agree that this is a dramatic improvement.

Cheers!

Well, I think I have now got to the point where I thoroughly deserve a toast to my success and progress so far - so I downloaded the STL file for the RepRap Teardrop Shot Glass http://www.thingiverse.com/thing:11944 and printed one off - and here it is under construction:

And here's the finished article, still bonded to the Blue Tape print bed:

Separated from the print bed - note the "raft", and the internal "whiskers" where the extruder has jumped across from one side to the other when changing to the next layer:

And after some cleaning up:

No, it's not perfect - but it neither does it look like a 3D spider's web, like my first few prints. The outside layer shows a few "sagging" layers, while the inside surface actually looks pretty good, and both the top face and bottom face of the horizontal shot-glass bottom look pretty tight.

Cheers!








(Wait - what's that I hear? Someone is asking if it's actually watertight, and did I actually get a drink from it? No! It leaks like a sieve! But never mind - it's onwards and upwards from here!)

Always listen to people who know what they are talking about!

My printer doesn't yet have a heated bed, although I expect I will add one as soon as I get the machine working to my basic satisfaction. I am printing 3 mm PLA (as most resources seem to suggest that PLA is an easier material to get to grips with than ABS, and it doesn't absolutely NEED a heated bed). The RepRap Wiki and other sources strongly suggested using "blue tape" http://reprap.org/wiki/BlueTape on the print bed to get the printed objects to bond properly, with the most recommended blue tape being "Scotch 3M Blue Tape 2090". My local hardware store didn't have any Scotch brand "blue tape" when I went there a couple of weeks ago, so I bought a roll of generic paper-based painter's masking tape instead - BIG MISTAKE!!!!!

I found the PLA simply did not bond to the generic tape at all. I was able to get my early test prints to bond to paper which I bull-dog clipped to the print bed, but as I susbsequently discovered - this is (almost) OK for mucking around, but if you want to print something which is even half-decent, you absolutely MUST get a good bond between the raft / first layer and your print bed.

I went to a diferent hardware store today, and they had some genuine Scotch Blue Tape 2090, so I grabbed a roll. It was quite a bit dearer than the generic stuff I had bought earlier (although still very cheap compared to what I have spent on this project overall so far!), but what a difference it has made to the quality of my printed parts - WOW!

On the images that follow, one of my old 20 mm calibration cubes (printed onto paper) is on the left, while the same 20 mm cube printed on Blue Tape is on the right. All machine settings (firmware and software)  are identical - the only thing I changed is the contact medium on the top of the print bed.

Bottom Faces:

Top Faces:

Side View:

You can see that printing on Blue Tape gives me MUCH improved resolution overall: on the bottom face, the side faces, and the top face. I really wasn't expecting such a dramatic change to the whole print quality, just from having a better bond between the bottom printed layer and the print bed.

Do yourself a favour - if you are having any issues with the quality of your first prints (especially if you don't yet have a heated bed), I strongly suggest you use PLA filament and genuine Scotch Blue Tape 2090 - it certainly worked for me!

Getting everything squared away

Well, after my first flush of 3D printing glory, I had a “compulsory sabbatical”, due to a requirement to travel away from my machine for work purposes. (Oh, the separation anguish!) The enforced parting gave me the opportunity to stop playing and start thinking – how could I improve the design of my machine (which I only just got working, after a fashion), examining the Sprinter / Pronterface / Skeinforge settings, and so on.

In particular, I had one of those “light bulb” moments that sometimes happen when you take the time to step away from a problem, and see it from a new perspective - I worked out how to correct the non-perpendicular alignment of my X and Y axes. http://julianh72.blogspot.com/2011/08/main-frame-build-and-alignment.html

Due to the initial mis-alignment of my whole frame, for my initial prints, squares were not quite square. I had calibrated the X and Y motors so that the sides were equal in length (and correct length, as well), but the corner angles were a couple of degrees off a right angle (as I always feared). This probably doesn't matter for "sculptural" type print jobs, but for dimensionally-critical machine parts, I wanted to get everything all squared away.

Then I realised that all I need to do is move the ends of the Y-axis smooth rods across slightly at either the front or rear threaded rod until they are truly perpendicular to the X-axis smooth rods (just a couple of mm at one end was all it needed). Once I had seen the machine in action, I realised that all you need to do is get all your smooth rods for the X, Y and Z-axes mutually perpendicular to each other - it actually doesn't matter if the rest of the frame is a bit skew, as the long as the smooth rods are all square to each other.

Five minutes work with a spanner shifting the Y-axis smooth rods a bit and voila! – nice square prints!

I now realise that the design of the Mendel Prusa is actually quite clever in that each of the axis smooth rods can be independently aligned so as to be truly perpendicular to the other two axes. I am sure this is by design, not just an accident, and it is probably obvious in hindsight, but it took me this long to realise!

" 'tis a poor thing, but mine own"

Well, I sorted out all of my motor issues today, and have all three axes working smoothly, the extruder is pushing filament forwards and backwards, the end stops all work (and the machine can find Home on all axes and can "park" itself nicely), and the hot end reaches a nice stable operating temperature. I loaded a simple STL file into Pronterface, and did a "dry run" (no filament loaded) - and the machine looked like it was doing all the right things.

At this point, there was nothing for it but to push some plastic through and see what happens - and so, in the spirit of "Hello, world!" (the first program you create when learning any new programming language),  here's what happened:

(I'm just printing onto a piece of paper held onto the acrylic print bed with bulldog clips, 'coz I have no idea of whether the PLA filament would glue itself permanently to the acrylic and damage my nice shiny acrylic.)

OK, so it's not exactly "precision engineering", but I am SOOOOOO happy! It actually vaguely resembles the part I was trying to print. The most obvious defect feature is that the "voids ratio" is pretty high (it's supposed to be a solid part with a couple of screw holes, not a 3D spider's web!), suggesting I am either extruding too slowly (not pushing enough plastic through, thus creating too fine a filament for the travel speed of the print head), or I am moving the print head too fast for the extrusion rate it is capable of achieving (which is sort of saying the same thing), or a combination of both. Or maybe my Hot End temperature is too high (or too low?), or .....

Anyway, a bit of research should suggest some possible approaches to get it to print more "solid" parts. Stay tuned ...