Noteworthy things #2

It is not on Thingiverse, and it is not specifically designed to be 3D printed, but it is super cool! A detailed 3D model (actually multiple models) of the Eagle transporters from the TV show Space: 1999 (Alpha: 1999 in South Africa).

 This site has 3D models of many props from the show.  I loved the show when I was a kid (I have the entire series on DVD) and always wanted to build a model of an Eagle. I even toyed with the idea of scratchbuilding one, but never worked up the enthusiasm. 3D printing might make the difference here...

The hobbed bolt part 3

The quest for the ideal hobbed bolt just took a new turn.  On Monday afternoon I rigged up a steel "bolt hobber" (a piece of steel with a hole in it clamped perpendicular to the M8 bolt, to guide the tap).  My setup was flimsy, but showed promise, so I think I should be able to hob a bolt.

If that fails, the option of cutting grooves with the Dremel tool and sharpening the "teeth" with a needle grinding tip is always a fallback option.

However, Quentin (designer of the Morgan) just posted his solution to the problem of the Eckstruder battling with 1.75mm filament:

RepRap Morgan 1.75mm extruder on reprap.harleystudio.co.za

This latest option is a modified RepRapPro Mini Extruder, which does NOT use a hobbed bolt, but rather a hobbed drive "gear" which is tapped in the centre (see bottom photo in the article).  Quentin made his from brass on a small lathe, which I do not have access to, so at first glance, it leaves me with no option but to order one ready-made if I opt for this solution.

I might decide to stick with the hobbed bolt extruder for the moment (and accept filament jams) until I can figure out a way to manufacture a drive gear as described. 

I will be thinking about this in the coming week...

The hobbed bolt part 2

I was hoping that the blog entry "The hobbed bolt" would only have two parts.  This is not going to be the case.

Hobbing a bolt is not as easy as some youtube videos would have you believe.  I do not have a lathe, or a drill stand, and have to do with hand-held tools.  First I marked the 13 mm point from the bottom of the head, then I cut a groove in the bolt shank (using the small angle grinder and a steel cutting disc):

Then I clamped the bolt in two 608z bearings in the bench vise, and tried to hob the bolt with the tap fastened in the hand drill:

This did not result in the beautiful hobbed bolt pictured in part 1 of this post.  The tap was all over the place, scoring the bolt and generally making a mess.  The groove is roughened, and here and there clear "teeth" were cut, but in general not a successful outcome.  I'm sure the extruder will extrude, but probably not have much grip, so a new plan will have to be devised.  A picture of the result:

The ideal is a "bolthobber", Thingiverse Thing # 232511, but for that you will need a 3D printer (exactly what I'm trying to build here).  I need a way to hold the tap fixed relative to the bolt (by hand it bucks and jumps), but I'm sure I will be able to get it done.  I will report.

The hobbed bolt part 1

It would seem that the most common extruder for filament is Greg's Wade Extruder, which uses a hobbed bolt to drive the filament.  This is an M8 x 50 (or 60) bolt with a groove in the smooth part of the bolt shaft. In this groove is cut "teeth" using an M6 tap (sometimes larger, sometimes smaller).  The bolt is driven by a stepper motor via a set of gears.  The filament is pressed into the groove by the idler block via a bearing.

The exact position of the groove is not easily obtained. You can buy hobbed bolts on ebay for USD10, and I have seen grooves at distances of 21, 24, 26 and more mm from the bottom of the head of the bolt.

One improvement made by Eckertech on their Eckstruder is to reverse the hobbed bolt orientation, as the groove can run into the thread on a 50mm length bolt.  In their version the groove is much closer to the head.

Inspection of the eckstruder STL file indicates that the groove should be 13mm below the head, and this LOOKS right on ebay photos of eckstruder hobbed bolts.

Tomorrow I will attempt to make a hobbed bolt by clamping the M8 bolt in bearings in my small bench vise, and hobbing it with the tap in the electric drill.  I will first initiate a groove using the small angle grinder.  I will post photos.

Optimising maths for speed

Depending on the settings of your compiler, some maths might already be optimised.  However, I believe most people are using the Arduino development suite as it is installed (I don't even know whether you CAN optimise the Arduino compiler).

You can save time (increase calculation speed) by not calculating "constant" equations every time (an optimised compiler will do this for you).  For example, in the Morgan code, the following line

SCARA_C2 = (pow(SCARA_pos[X_AXIS],2) + pow(SCARA_pos[Y_AXIS],2) - 2*pow(Linkage_1/1000,2)) / (2 * pow(Linkage_1/1000,2));

contains two instances of a "constant" that can be pre-calculated:

2*pow(Linkage_1/1000,2)

and inserted in this line.

The second optimisation I tried was to replace the "pow(x,y)" math function (which raises x to any power y) with the "sq(x)", or square function.  This is significantly faster than the generic first function.  The code becomes:

constcalc = 2 * sq(Linkage_1/1000);       // just after where Linkage_1 is defined

SCARA_C2 = (sq(SCARA_pos[X_AXIS]) + sq(SCARA_pos[Y_AXIS]) - constcalc) / constcalc);

These small changes make about a 30% difference in execution speed.  These are only part of the total calculation for Morgan, however.

Another optimisation (which I have not investigated) is to replace the "atan2" function, which calculates the arctangent from two parameters, with a pre-calculated lookup table.  There is more than enough memory available to have a comprehensive (read accurate) table, but again, the complexity added might not be worth the increase in calculation speed.

 

Withdrawal symptoms

I am starting to suffer from "withdrawal symptoms", the Morgan parts are sitting there, looking at me, and I can't progress...  At least the Kapton tape and thermistors have shipped, but will not be here any time soon. 

Quentin should be sorting out his backlog of Morgan printed parts orders this week so I expect the printed parts towards the end of the month, at least then I'll be able to progress.  In the mean time the stepper driver boards should be here soon, and I have started looking at the Morgan software.

Quentin reported that the Arduino runs out of steam to when he tried to up the number of delta calculations per second.  Currently the software is configured for 200 calcs per second on my hardware.  I have committed to look at improving that, but ideally I need moving steppers to test (the problem manifests as loss of quality and steppers actually slowing down when it runs into trouble).

Being new to RepRap but an old hand at embedded software (I used to write embedded software for a living) I feel this is one area where I can add value.

The Morgan software builds on top of the Cartesian calculations for the other RepRap machines - the x and y movements are translated into theta and psi angles by inserting an extra calculation cycle.  The maths is fairly standard, there is a lot of squaring, some square roots, and a few arctan functions being performed every time.  Everything is done in floating point maths (which is slower than integer maths in a processor).

Staying within the current scheme (it is possible to completely overhaul the maths and not do the Cartesian to angle conversion), the first step would be to optimise the equations for speed.  If more speed is needed, convert the whole lot to integer maths, but the gains vs. effort might not be worth it.

The next post will be slightly technical on the first things I'm trying.

Last few vitamins

Yesterday, on my regular office Monday in Johannesburg, I ALMOST succumbed to the temptation of picking up the last vitamins for my Morgan build at RS Components.  I need Kapton tape and a bed thermistor.  Prices at RS:  Kapton tape (25mm wide) ZAR470, EPCOS thermistor ZAR50.

Luckily my Monday was exceptionally busy and I only got out of the office late.  Today I ordered the items from ebay (shipping included - I learnt my lesson on the stepper motors).  Kapton tape (50mm wide) ZAR70, 10x thermistors ZAR100.  The thermistors are not EPCOS as far as I can see, but I compared the data sheets and the error should be no more than about 5 degrees on 110 degrees as is.

I intend covering the heated build plate with Kapton tape.  As far as I can see that is the easiest way to ensure adhesion of both PLA and ABS - apparently a glass bed does not do well on both plastics.  50mm wide tape will make lining easier.  There was a roll 200mm wide for ZAR230, but I decided against that.  It WOULD make for a nice uniform platform surface though...