Finally Finished Printer

One of the problems that I’ve found with 3D printers is that once they’re capable of producing reasonably decent objects you can very easily become side-tracked into printing all sorts of stuff whilst not actually tidying up all the loose ends and finishing the printer.

In my case this has meant several things…

Firstly for the last year the electronics has been sitting next to the printer on the table with the leads hanging off the edge of the table threatening to drag the whole lot off with one decent tug.

Secondly my fan mount for the cooling fan on the stepper drivers has been a M6 bolt screwed through a small piece of Aluminium, this has sat next to / under the Arduino MEGA, also with a tendency to slide around the table.

Finally the LCD add-on has never had its mounting case printed and so had ended up sitting under the printer where it’s pretty much pointless.

For the last  little while, whatever time I had to spend on projects has been eaten up working on the trolley but I’m pretty much at the point now where there’s not much more progress that can be made there until a new set of 12V batteries have been purchased and installed. Consequently it’s been harder and harder to ignore these final small but annoying jobs on the printer front and so this last weekend I got them sorted.

Firstly the electronics are now mounted on a printed sled / clamp deal. One of the pain in the ass parts of this was fitting the mounting screws in / around all the components on the Arduino Mega. There’s two things that are amazingly stupid design decisions on these units, the first being the header spacing (which was apparently a mistake…) the second is that it’s almost impossible to fit the heads of the screws in around the headers / SMD components on the top sides of the board. To get around this I had to stick a couple of screws in a poormans lathe (aka power drill) and turn the head down to a smaller diameter to fit. The ends result is below, The angle is a bit average but there’s a lot of cables blocking the view from above.

IMG_6201 (Resized)

IMG_6204 (Resized)

Mounting the fan was a bit easier, trawling through the “parts that didn’t print right / werethe wrong size” provided what I believe was a different version of the electronics mount that for whatever reason (maybe I forgot about it, I can’t remember) didn’t get used. Suffice to say it was perfect in that it was designed to clamp to a piece of M8 rod and had a handy hole in the end to which to attach a fan via random section of those “Helping Hands” PCB holders which then clamped onto the fan.

IMG_6202 (Resized)

Finally after a quick look through thingiverse for inspiration I made up and printed the below black clips to hold the LCD to one of the top threaded rods of the frame.

IMG_6200 (Resized)

As a upgrade I also printed several of the black belt tensioners below. Ideally the belt would be tight without the tensioner but in order to adjust this it would be necessary to remove the heated bed (which would then need to be re-leveled again afterwards). I really couldn’t be bothered doing this and the tensioners turn out to be cheap and easy to print as well as pretty effective.

IMG_6203 (Resized)

Since this printer is not marked as finished (apart from upgrades and repairs) I think I can now in good conscience thing about the next one… I3 anyone?

On the shopping trolley front the following has happened;

1) Seatbelt has been attached.

IMG_6196 (Resized)

A short section of chain extends the belt down to the body of the trolley where it is attached be a shackle to a decently thick section of the frame.

2) Cover has been printed for the throttle.

IMG_6191 (Resized)

This is a massive win for 3D printers aye, it’s amazing what they are able to accomplish in regards to bespoke shapes quickly and cheaply. The amount of filament to print this part is way cheaper than what it would have cost to have it made out of metal. Additionally once the part is designed there isn’t that much interaction on the part of the user until the printer has finished its job. The cover is made up of two sections, the first lower section bolts to two handy (yet randomly) drilled holes in the original metal of the handle bars. The top section then slots over the PCB of the controller and then screws to the bottom with a couple of M3 screws and a couple of captive nuts.

3) Driving lights have been fitted.

IMG_6198 (Resized)

Thinking about long missions in low light I realised that I needed some form of front white lighting. I really didn’t want to end up in a position where the flashing LEDs had to be used for this the whole time as it’s a bit of a waste of their potential and functionality really. Scrounging around yielded a couple of LED torches with decent output and small size so these were fitted to the front with a  couple of 35mm pipe clamps and some M10 rod. A couple of aluminium plates secured the lot to the frame.

4) Rims

IMG_6193 (Resized)

IMG_6194 (Resized)

White wheels aren’t much fun so whilst waiting for batteries I decided to bling the kart up a bit. From the parts bin came a whole lot of M10 bolts, nuts and washers, some Perspex sheeting and some bearings. Firstly the Perspex was cut into four circles, these were then drilled firstly with a 3mm pilot hole and then with a 10 mm drill bit in reverse so as to not split the Perspex. Printed hubs then had a washer and bearing fitted before being mounted to the Perspex. I then mounted the wheel horizontally to a block of wood and spun it around whist moving a grinder towards it tangentially to correct any wobble from the center holes and hubs not being dead center. The bolts were then fitted and tightened up.

IMG_6205 (Resized)

IMG_6207 (Resized)

IMG_6209 (Resized)

To mount the rims to the trolley the M6 bolt stopping the wheel from sliding off the end of the axle was removed and a length of M6 rod was put in its place. A nylock nut was then tightened down onto the rod to keep the wheel on. A second nylock nut then gets used as a spacer before the hub is slid onto the rod and held in place with a third nut.

IMG_6212 (Resized)

Result = awesome.

It should be noted that as the rear driven axle actually turns as well as the wheel (front axle is fixed and the wheel spins on bearings) it would be possible to fit blades to the rear wheels for a 21st century scythed chariot effect. Perhaps not as public friendly though

Why 3D printers are awesome

Now that I’ve managed to entice you with the title I’m going to add the caveat that they are also a right pain in the ass. Home made ones are great in that they are comparatively cheap compared to a Makerbot but at the same time because for the most part they are hacked together from components that can be brought from Mirte 10 Mega and Trademe they can also be extremely finicky. When they are in a good mood and are working though they offer the amazing opportunity to design a 3D model on a computer and then have in your hands 20 minutes (depending on the size) later a solid copy of your design at the cost of approx $1.20.

To demonstrate their usefulness and indispensability I thought I’d post some pictures of parts that I’ve printed for the Trolley project, parts that would otherwise have been impossible to produce without a number of other and on their own more expensive pieces of machining equipment.

Firstly the LED headlights on the front,

IMG_6091 (Resized)

IMG_6098 (Resized)

These are made up of 5 separate parts, a main case and lid, a backing plate sandwiched between the LED’s and the PCB that they are soldered to and 2 clamps to attach the whole bundle to the front of the trolley. The main case would have to be the part that took the most amount of time to print out of any of the parts on this project, 3 ½ hrs >_<

Next up we have the main LCD case and mount. Three parts, a front surround which the LED bolts to, a backing plate that uses the same mounting bolts and a clamp that attached to the back of all this and then bolts onto the square frame of the original handle bars.

IMG_6092 (Resized)

IMG_6093 (Resized)

Thirdly the TPTMs (trusty printed trolley mounts) a total of 14 of these have been used thus far on the Trolley and I imagine that more will be needed before the project is up. 7 minutes to make one of these.

IMG_6094 (Resized)

Forth we have the rear lighting mounts, designed to clamp onto the trolley handle bar and provide a spaced out mounting platform for the rear brake lights and indicators.

IMG_6097 (Resized)

Based upon this basic design we then have the GSM and GPS antenna mounts, These also clamp in a similar manner to the handle bar and only differ in that the mounting platform has been modified to suit whatever is being mounted (GPS mount is the square flat one).

IMG_6096 (Resized)

IMG_6095 (Resized)

We then have three more brackets that are all based upon a similar design again. Firstly the controls for the LED headlights mounted to the front. This takes two buttons and angles them down and in toward the rider.

IMG_6103 (Resized)

Based upon this we then have a slightly beefier version to which the siren controls are mounted. This version differs in that it has more bracing between the horizontal and vertical sections and also has 2 TPTMs as opposed to one on the headlight buttons.

IMG_6101 (Resized)

IMG_6102 (Resized)

Mirrored off this and with some modifications we then have the mounting bracket for the LCD screen for the LED headlights. The screen mounts to the front casing which is then attached to a combined backing plate and mounting clamp. The connector for the LCD pokes out a port in the back of the backing plate, a TPTM attaches all this to the trolley frame.

IMG_6099 (Resized)

IMG_6100 (Resized)

I still have at least one more major part to re-print for this project, namely a cover for the stock throttle electronics. These are quite exposed on the handle bars so a case to protect these really is a must, fortunately should part of the electronics be snapped off the trolley should in theory stop automatically but still… Note I said re-print… the first print of this part demonstrated the finicky nature of 3D printers and how they just love to randomly stop printing 2 hrs through a 2 ½ hr job.

Reprap Update

No pictures for this one I’m afraid but don’t worry, the next update on the Mobility Scooter will have a fair few.

Last time I wrote about the Reprap I was having a weird issue with the filament drooping down the side of the prints and just generally misbehaving as well as an issue with the Z axis going out of alignment at odd times.

Ive managed to narrow the Z axis alignment issue down to thermal expansion, i.e when the extruder is cold it is slightly shorter than when it is warm. Normally this wouldn’t be an issue but the tolerances between the extruder nozzle and the print bed need to be just right, to wide a gap and the first layer doesn’t extrude properly and to narrow and the first layer squishes out under the side of the nozzle, the nozzle drags in the print and things just go generally all wrong.

To remedy this firstly I got the bed completely level in relation to the extruder. Secondly instead of trying to get the Z end stop in exactly the right place I set it so that the end of the nozzle depresses the levelling spring in the home corner of the bed. This helps prevent the filament from oozing out of the nozzle as it heats up. I realise that this will result in a small section of the glass print bed heating up a lot more than the rest and accept the fact that at some stage it may crack but it’s a trivial matter to cut and fit another. In order to get the offset right I then adjust the Z offset in the slicer settings in increments of 0.10 of a mm until it’s correct with fine adjustments (which need to happen after the 1st print of a series of back to back prints due to the thermal expansion) being taken care of by tweaking the z axis by hand until it looks right.

The second issue took a bit more time to resolve, i.e droopy prints. Firstly it looked like the temperature was too high for the PLA however lowering the temperature until the plastic was being extruded correctly left me sitting at about 150 degrees which aside from being way too low resulted in a lot of stress being put on the connection between the extruder body and the thermal barrier as the viscosity of the PLA changed. I then started looking at the filament itself. In the previous post I mentioned a colour change in the extruded filament and whilst this does happen just after you change colours I was getting it after printing meters and meters of the same colour with the black going grey for short (<10mm) sections. When I took the hot end apart I found a sticky oozy substance that had seeped up between the brass outer and the PTFE liner which was a similar colour to the odd grey sections of print. Considering how tightly these two fit together it was obvious that whatever it was had a vastly lower viscosity than normal PLA does when it is heated and extruded, almost like coloured, sticky water. I was able to clean it all off and then reassembled the extruder. I did a bit of research on the internet and discovered that PLA is actually hygroscopic (it absorbs water) now this is mentioned on the Reprap website but I hadn’t been experiencing any of the listed symptoms (puffs of steam, plastic continuing to be extruded after the extruder stops (in excess to normal ooze)).

As a trial I borrowed a dehumidifier and set it up directly under the filament reels so that it would blow the dried air directly on to them. After sealing the room and removing about 20L of water from the air over a 24hr period I had another go at printing. I knew that I would have to put a fair amount of filament through the machine to ensure that the problem was resolved but I’m please to say that after 6 months of printing under the low humidity conditions the problem has yet to reappear. Considering that it used to pop up after a couple of days of heavy printing and only go away after not printing for a couple of weeks I’m pretty happy to consider this issue addressed. Looking at the way in which the issue appeared and disappeared I have to conclude that the PLA was absorbing water which was then ending up trapped in the extruder nozzle, separated from the PLA and not seeping out gradually but rather coming out in large batches after building up for a while. Letting the printer sit for a couple of weeks then let the PLA in the nozzle reabsorb the water which was then extruded all at once (before it had time to separate) whilst printing the skirt for the first printed part when I started the printer up again.

I’m still having the occasional issue with the extruder body stretching where it connects to the thermal barrier but this only happens every 4-5 months and just means that I need to keep a couple of printed bodies as spares. When they do need replacing it only takes about 30 minutes to swap everything onto a new body and provides a good opportunity to readjust everything and clean it all out.

 

More extruder randomness

Recently I’ve been having a great deal of success with the printer, aside from the occasional Z calibration it seems to have been working well.

Over the weekend I was able to print the remainder of the parts for the lower section of a MendelMax with little drama, there was consistency in the print quality and I was able to bang one out after another stopping only to remove the completed parts from the bed.

Yesterday I started on the top MendelMax frame sections, expecting to continue my run of complication free printing. I fear that this (if it already hasn’t) may soon come to an end. As I had taken the printer out for the weekend I knew that the Z axis would need to be tweaked to get the first layer to adhere properly, this however took a lot longer than normal as the resulting distance between the end of the extruder and the bed varied in ways which would not be expected based upon the adjustments made. Furthermore after printing for about 1/2 an hour the offset between the two needed to be increased by .15 of a mm in slic3r in order to obtain the initial, correct gap. I would expect some expansion of the hot end as it heats but I would have thought that this wouldn’t have resulted in such a delayed increase. (note this also happened again when I fired the printer up tonight).

Once I finally got the first layer printing properly I started having a problem that I never had before, the outer perimeter wouldn’t sit properly on the previously printed layer but would droop and hang down the side of the printed part (as on the black printed part below). I stuffed around with the settings in slic3r which dealt with the speed, retraction, z-lift and extrusion multiplier all to no avail.

I then noticed tonight that occasionally the color of the extruded material would change from a light blue to a darker blue as would the consistency, from what I have always experienced, to something what was more akin to 50% PLA mixed with 50% water (like the stuff you get if waterbased paint in a tube separates or is not mixed properly). You can see the areas where this has happened on the print below to the left of the right had bolt hole, below the left had bolt hole and on the left center side of the lower face.

image

Initially I thought that I was something in this section of the PLA, some contaminant or change in plastic brew so I swapped out the blue for some purple and tried to print a model of a tornado.

image

Greatly improved but the outer face still had some imperfections in it that I hadn’t got on my earlier prints. I also had another test piece behave slightly strangely with the purpler and not adhere to the bed on the outside edge of the extruded path when in-filling the first layer. It almost looked like the printer wasn’t extruding enough plastic for the distance it was covering.

image

I then swapped out the purple for some black where once again I had the droopy issue. The final 10 mm of the model (at the bottom in this pic) I dropped the extruder temperature 5 and then 10 degrees to 175 which weirdly seemed to stop the problem.

If it is a case of the temperature being wrong I’m not sure why it would happen all of a sudden and especially why it would be out by 10 degrees. I understand that different colors will extrude at slightly different temperatures but if the poor quality with the blue filament is solved in the same way as with the black (i.e a temperature drop) then what has gone wrong to require this. Nowhere have I heard of the temperature having to be adjusted by such a large factor across approx 50m of filament.

I noticed that the base of the extruder body seems to be deforming slightly in the same way as the first one. There also seems to be some burnt material seeping from between the joint of the PEEK and the brass sections of the hot end. If the base is deforming slightly as it heats up then this would explain the expansion issue that only presents after a period of printing but at the moment I’ve got no idea what the burnt stuff could be.

I guess tomorrow I’ll have to rip apart the extruder and hot end and have a look inside to see how things are, perhaps there are answers in there as to why all this strangness is happening.

3D printers are cool, really cool, but man they can be a pain in the ass. Badly behaved, they don’t tell you how they are feeling and then just when you think everything is fine they go and crap themselves. Bugger it all

Extruderplosion

Took the extruder off the carriage tonight to have a look at how bad the damage was. To be honest I was hoping that it’d be a simple matter of screwing the brass back into the PTFE but no such luck.

As you can see the actual body of the extruder has deformed and cracked. I wouldn’t have thought that the PLA would buckle this easily but I guess that the heat from the hot end will have risen and softened the plastic a bit. Anyway looks like a new body will have to be sorted before repairs can proceed. To prevent this happening in the future the only thing I can think of would be to decrease the limit on the extrusion speed to reduce the pressure in the hot end.

Average

I actually thought I was getting somewhere with this…

Managed to get one good print of something other than a box the other night, namely the knob for the display panel that will eventually be mounted on the front. Since then however its been a complete mess. Firstly there’s been no end of trouble getting the bed level, mostly this has to do with its design and the decision to have 4 mounting holes (one in each corner). Whilst this may be great for support its a right pain in the ass when it comes to adjusting them all. Secondly I wonder if the perspex that the heated bed is mounted on is too thin and flexes too much. This could explain why after I’ve spent 20 minutes leveling the whole thing when I go to print its sits too close to the extruder and appears to not be level again.

Tonight I also noticed an issue with the actual extruder. During another attempt at printing the knob I noticed that the large gear on the extruder seemed to be stalling at times that made no sense, i.e in the middle of a long pass (not near the end, when moving or any other times when you would expect the machine to not want to extrude material). Initially I thought that either the current limiting on the motor controller was set too low and the motor was stalling or that the limit was set too high and the controller was shutting down from overheating. Neither of these turned out to be the case, different current settings didn’t help and the heat-sink with the added fan was defiantly below the + 150 degree limit.

It turned out that the screw attaching the smaller of the gears to the motor shaft was slipping which left the motor still turning during these times but the gears and extruder stationary.

This was duly tightened.

I then gave the extruder a test, extrude 100mm of filament so I could check that it wasn’t slipping anymore and so I could reset the current limiting and this was when it all went wrong.  Whilst fiddling with the motor controller I heard several clicking sounds coming from the extruder. I then noticed the brass lower section descending from the PTFE rod that attached the hot end from the extruder body. By the time I was able to hit the emergency stop button the brass section was most of the way out of the rod and only held on by the leads coming off the thermistor.

Needless to say this is a real bugger of a thing to happen, looks like tomorrow night is going to be spent ripping apart the x-carriage and extruder to see how bad it all is, hopefully the threads in the PTFE rod will be fine and the brass can just be screwed back in.  At the moment the brass section is completely out of the ORFE rod and hanging from its power cables above the heated bed.

I’m not sure how to prevent this happening in the future however, I’ve had a look at the graph and the temperature never dropped below 183 so I know that it wasn’t an issue with the filament clogging up from being too cold. It looks like it just came down to too much filament being forced in the hot end and it not all being able to exit the nozzle quickly enough.

Proper printing

The couplers between the Z-axis motors and the threaded rod that acts as a screw arrived on Tuesday night. On the couplers that I ordered, the holes in either end had a diameter of 6mm so in order to connect the rod to the couplers and the motors some machining was required. Peter came to the rescue once again by turning down the ends of the threaded rod to 6mm and turning a spacer to fit on to the 4mm shaft of the motor to enlarge the OD of it to 6mm.

I only needed two of them but at $1.50 each ordered 2 extra, Probably a good thing as the machining on them is less than accurate. You can see on the top of the one on the left the threading for the grub screw has broken through the top face.
After a few firmware tweaks to get the Z-axis moving at the right speed with the right steps per unit this was produced

One very rough 20 x 20 x 10mm cube.

Both the belts on the X and Y axis need to be tightened, the bed needs to be leveled and the frame squared up still but as a first print with all this alignment / calibration stuff to do I’m pretty stoked. Looking at the top of the cube it looks like the extruder may have stalled a few times, I suspect this was because the motor controller was overheating and shutting down, Since mounting a fan above the RAMPS board this hasn’t seemed to have happened again.