Printing resumes

OK, now the Prusa is working once again (with Yellow ABS this time) I immediately began printing the next set of blades for the VAWT project I am working on, again encountering the malformed leading edge issue I had spotted previously.

As it turns out this does not appear to be anything to do with my retraction settings but more to do with trying to print overhangs so close to the hotbed in ABS, the overhang stays soft and is mechanically deformed by the next layer printed resulting in curling on the leading (or trailing) edge, I have tried several different techniques to resolve this issue none successful as yet and I am communicating with Quentin (the VAWT creator) to see if we can resolve this issue, it is very difficult to tell if this is due to the VAWT model needing tweaking so that it is printable in ABS on a hotbed or if it is my particular printer setup that is causing the issue , I will have to investigate overhangs & warping/curling thereof in the wider reprap community to see if anyone has any other suggestions of how I may solve this.

I have already tried

1)     Inverting the blades to see if the problem was only present on the leading edge (it wasn't) but it also showed up on the trailing edge.

2)     Printing several (3) blades next to each other to see if additional layer cool time would help (a very slight improvement but not significantly better)

3)     Lowering the hot end temps and hotbed temps as low as I can get them without the ABS becoming  unstuck, still no significant improvement.

I may have to resort to getting Quentin to add a support structure to the first 1-2cm of the leading/trailing edge (leading or trailing depending on which is overhanging, which in turn depends on which end you are printing the VAWT from)

Other alternatives I will try are to experiment with active cooling, by either changing the hot bed temps (although my previous tests suggest this won’t work) or by force cooling with a fan, (this would most likely work but would also likely cause the part to warp and become detached from the hot bed so not a good solution either)

Alternatively I could switch to PLA which needs a lot less heat (I will try this anyway just to see)

Anyway regardless of the poor print quality I forged ahead with printing the remaining parts of the VAWT after all I can always repair the malformed leading edge with bits of leading edge from other failed prints by hand with a hacksaw & glue!

More ABS please

Purchased another reel of ABS (yellow this time) as I am getting very close to the end of my 210meter reel of black ABS, this was also so that I could get some replacement hot end resistors on order.

ABS arrived with hot end resistors, and by strange coincidence so did the IMU for the balancing bot (I had completely forgotten about this part, and is the last part needed to start the balance bot project yey)

Set about replacing the hot end resistor and immediately spotted problems, the resistors I had bought were slightly bulged in the center causing me to have to drill out the heater block so that they would fit, after much fiddling I did get the Parcan hot end back on the Prusa and working again.

This issue has highlighted the need for me to get building my own hot ends & heater blocks, this is hopefully coming soon as my local EV dealer has a great workshop with drill press & lathe which I will be visiting  soon (he also wants a 3D printer for his EV projects and has offered the use of his workshop, cool)

Darwin Upgrades

As mentioned in the last post, I have been making some mechanical changes to the Darwin in the hope of making it into a useful printer once again (I had previously butchered it to get the first child working as the extruder stepper on the Darwin had been previously borrowed from the Prusa parts for the Darwin so that I could print the plastic parts for the Prusa (1^st Child)).

1)     Removed X carriage completely drilled out  X ends (motor & idler) to 10mm from 8mm so that I could fit the induction hardened rails I had bought in error for the Prusa to the Darwin, I really need to print better X ends later as they were designed for 8mm rods so this is a fairly major bodge.

2)     Fitted 10mm linier bearings and new custom quick-fit X carriage to X rails (this is when I realized that the belt clamp positions were also different on the Darwin so I drilled 4 extra holes in the carriage to bodge another temporary solution.)

3)     Fitted 608 bearing to idler to replace the “creaking” (yes it was audibly creaking in use) resin cast idler wheel, this required a selection of different sized washers to effect a belt guide 2 of which I had to cut to fit the idler housing, hacksaw job.

4)     Fitted 8mm linier bearings to Y carriage rails and modeled out the holes in both the idler and motor carriage ends to fit using a sharp knife, I will need to come up with some way of securing the bearings to the carriage ends until I can print replacements.

5)     Fitted 608 bearing holder to the Y axis motor end rod (far end from the motor) this replaces a cast resin bush hopefully further reducing friction in the system, I could also print 2 more of these for the idler end of the Y carriage (will do these later)

6)     Reassembled the whole shebang and ran up to Darwin Pronterface setup I had last used several months ago to see if it would move, and yey it lives.

Now I just need to purchase another stepper motor and make a better hexnut nozzle and th darwin should be capable of printing once again.

I still have a whole host of mechanical issues with the Darwin (mentioned in previous posts) and I am planning to upgrade the electronics using  the Gen 3 plus solution (don't want to dump the Gen3 electronics just yet)

I should be able to rewire the extruder stepper so that it is directly connected to the gen 3 mobo, remove the RS-485 extruder controller completely (may try and repurpose this as a filabot extruder controller) and connect the hot end heater and temp probe directly to the gen3 mobo, this should allow me to load the sprinter/marlin (sailfish?) firmware and that will be a major upgrade to the Darwin’s capabilities.

I am also planning to remove two of  the Z axis corners and run the print bed similar to the Prusa (just with 1 motor and a timing belt, rather than 2 motors) this should significantly decrease the friction in the Z axis and I am hoping improving reliability.

I may also add 608 bearings to these two remaining Z corners to improve the mechanical stability and accuracy of the Z axis.

Bowden Blowout!

Had my first Bowden tube failure the other day the tube became detached at the extruder motor end due to internal pressure I guess, and in my frenzy to get it fixed I broke my hot end’s heat resistor, unfortunately I had no stock of these resistors but I did have the old Darwin hot end that I had retro fitted with a new heater block purchased on eBay (it used to heat using a coil of nichrome wire) I had already fitted my only spare heat resistor to this hot end so I thought whilst I was waiting for the replacement resistors to arrive in the post I would fit the old Darwin hot end to the Prusa and see what it could do.

After much fiddling around I managed to get this “working” the quality was poor compared to the Parcan hot end as the nozzle diameter is approx. 0.9mm on this old hot end, also this hot end has a brass hexnut style nozzle so it would be fairly easy to replace the hexnut with one with a smaller nozzle hole, this was all done so that I could test the possibility of reusing this modified Darwin style hot end on the Darwin once I get it back up and running.

In fact I was actually in the middle of printing a new X carriage for the Darwin when it broke, the new X carriage is based on the 10mm lm10uu linier baring derivative of RichRaps great Quick-change X carriage (ref).

I had already printed one of these 10mm Quick-change X carriages but it turned out that the Darwin has 42mm X rod spacing unlike the Prusa which has 50mm X carriage rod spacing, so I loaded Sketchup and modified the X carriage (virtually) chopping 8mm out of the center and bringing the 2 halves back together and exporting as STL.

Printed this modified Quick-change X carriage using the old Darwin hot end (with retrofitted heater block) this worked remarkably well (well at least better than I had expected) and after a few hours of fiddling the Darwin is moving again.

Purchasing for more projects

Bought the parts for a balancing bot (inverted pendulum) ref
got the motors / motor driver board just waiting for the IMU to arrive and then I can get started on the balancing bot project hopefully using the reprap as the system glue to give the project that professional look and give me some practice with openscad.

I have also decided to take the plunge on building myself an EV (Electric Vehicle) found a very local (purely by chance) EV builder and researcher who is keen to get me up and running so I have bought myself a donor vehicle (a Honda Beat) as this project is going to be complex and large I will start another blog dedicated to this.

http://electrobeatev.blogspot.co.uk/

Don't forget your software upgrade’s

Whilst trying to print the planets  for the tiny planetary gear set I found that they would not fill correctly, after reading on this people were suggesting that the extra shells settings were the cause after much fiddling to no avail I finally got the impression (again from more reading) that maybe there was a fault with the software so I bit the bullet and did a complete update of both Pronterface and SFACT to the latest versions and voila the problem vanished.

Print frenzy round two

After my safety improvements (mosfet move & PSU connector upgrade on the sanguinololu) the overheating issues have completely vanished, along with the newly insulated hot bed & moving the entire printer into a heated bedroom rather than an unheated workshop (need to build a chamber for the workshop) the printer is behaving very well build quality and speed are all better and I have taken the plunge to print a complete set of new parts for the 1^st grandchild.

6 x - Vertex – ref
8 x - Barclamps - ref
2 x - Rodclamps - ref
2 x - Beltclamps  - ref
1 x - Y carriage lm8uu bed mount - ref
2 x - Z motor mounts  - ref
2 x - X carriage’s - ref (rich raps universal quick change for lm8uu & lm10uu derivatives)
1 x – Gregs X carriage (a spare for the one I am currently using)
1 x - X motor end
1 x - X idler end

most of these parts worked perfectly and other than a small problem with very thin (and hollow!) walls on the X ends I am very happy with the results, they are leagues ahead of what the Darwin could produce, and all in ABS!

I have also spent some time printing demo pieces to show to colleges such as the tiny planetary gear set (ref), spirograph set for the kids, several holly wands again for the kids.

Bowden issues

The addition of the Bowden tube has introduced some fairly major hysteresis into the system, this required an increase in the retraction distance to overcome the stringing and blobbing that has appeared since the Bowden got put in place.

Carried on with the Atoms of the VAWT (3 required for a complete VAWT) unfortunately my overzealous retraction settings had an undesired effect on the leading edge of the VAWT wing section at the bottom where the print head moves from the wingtip to the centre

Printed all 3 Atoms of the VAWT regardless of the quality issue purely as a build volume test and so that I could tweak the retraction settings to try and eliminate the malformed leading edge issue (and although I saw a slight improvement I think I may need to shorten the Bowden tube as much as possible to avoid or at least improve upon the hysteresis issue)

Chopped out approx. 70mm of Bowden tube and tested, slight improvement in the hysteresis suggesting I am correct in the assumption that the Bowden tube is causing the loss of quality.

As I am still getting issues with bed leveling/flatness I decided to upgrade to a glass bed on top of the heated bed PCB, this cured the leveling and flatness issue completely but introduced a large delay in the heat up time of the hotbed, from approx. <5 minutes to around >10 minutes, this is annoying as it considerably slows the whole process down, I must find some thinner glass, (maybe from an old picture frame)

After much testing of many prints, I remembered the other reason for upgrading to a Bowden tube, speed so I began experimenting with pushing up the speed from 20mm/second to 30 then 60 then 100!

Yes 100mm/second and everything seems to be working flawlessly, wicked!

In fact at these speeds I start to notice mechanical wobble appearing in my y axis print bed, need to shorten the screws mounting the hot bed above the wooden build bed.

I also noticed that I am getting warping again possibly due to the fanning effect of the rapidly moving Y axis.

I am currently finishing the hotbed upgrade to shorten the screws and properly insulate the hotbed underneath so that the hotbed holds its heat better.

Almost finished this last night 21/09/2012 other than the need to re-level the bed properly and fix it in position properly.

If this does not solve the warp issue I will probably need to move the whole printer into a heated room as I am currently doing all my printer stuff in an unheated room and now that summer is ending and winter is on its way it is starting to get cold again.

last year this had major warp implications with the plastic I was using at the time, which I had assumed was PLA but may actually have been CAPA (I am still not exactly sure what plastic this is as It was gifted to me with the original Darwin machine)

It took me a long while to realize that this was due to the local temperature conditions in the room as when summer arrived the problem promptly vanished proving that local temperature conditions in the room need to be warm not cold.

Need to investigate a heated build chamber or at least temperature monitored.

The hot bed should be good enough to heat the chamber assuming it is fairly small and well insulated but I would need to figure out where to connect the chamber thermistor to the sanguinololu electronics set, more web research required here.

I also fell fowl of trying to implement the pronterface control of a fan on small layers recently, as when I looked into this I discovered information (ref) suggesting  that the hotbeds Mosfet on the Sanguinololu board was not rated for use on a PCB of this fine a track width.

Following the advice, I removed the Mosfet from the board completely and rewired it as suggested, basically connecting the hotbed’s Mosfet directly to the ATX power supply and just wiring the gain to the PCB, this means that the Mosfet draws most of its current from the PSU rather than routing it through the Sanguinololu’s PCB tracks.

It was suggested that this is a possible safety issue with the Sanguinololu’s design, At the same time I also replaced the screw terminal power connector with a proper 4 pin ATX 12V PSU socket doubling the copper connecting the PCB to the PSU.
I had noticed also before these upgrades that the power regulator (on the other end of the PCB from the PSU connection) was getting extremely hot and was in danger of meltdown (in fact it was manifesting as the sanginololu resetting itself during prints, which is how I first spotted the problem.

Bowden Extruders are go for launch!

23/08/2012

 

Last night I spent quite some time implementing a Bowden extruder mechanism using parts (ref extruder motor end) (ref nozzle end) I did have to mod the extruder end somewhat as the part was designed for a 16 mm hot end and my current (bought) extruder has a 13mm hot end fitment, so seeing as I am not yet versed in the ways of OpenSCAD I used a stanley knife to whittle down the 15mm hot end plug to a 13mm one and unfortunately due to slicing issues the nozzle end part did not slice correctly (warnings in Pronterface of this problem but I went ahead and built it anyway), as it turned out the Bowden nut hole was missing a corner of the hex nut so again I simply hacksawed one corner off of a nut so that it would fit, both these mods although fiddly (and technically unnecessary as I should have been able to tweak the parts in a CAD package or design my own from scratch) were relatively simple and meant that I could use what would have otherwise been wasted parts.

 

I must start getting my head around OpenSCAD.

Tested the Bowden setup once last night with a very small part (ref belt clamp) and again this morning with another attempt at the first layer of the VAWT bottom part I am still trying to get built.

I discovered during this test that I will need to upgrade my extruder idler to a “Guidler” variety to stop the filament slipping out of the hobbed bolt.

I already have one of these printed in ABS (early ABS test of the new Premium Black ABS I am now using) and again although the quality/ surface finish is not great it should be perfectly functional.

Plan for tonight is to implement the Guidler and attempt a print of the bottom part of the VAWT, this should provide an excellent test of my build volume improvements.

I also took a look at the newly arrived SDSL (SD ram card adaptor for the Sanuinololu set) discovered it uses a “micro” sd card I will source one of these today so I can do some SD print tests at the weekend.

SD Ram card adaptor for Sanguinololu arrives in post (Whoopee!)

22/08/2012

This will give me the advantage of being able to load the stl to be printed into ram that is directly accessible by the machine and not via the USB connection to the PC, I have read that this greatly improves quality as however smoothly you think your printer is printing trust me it is not as smooth as it could be if you were printing directly from ram, I am of the understanding that this can completely remove the surface digital artifacts that appear on some larger prints as these are caused by delays in the PC communicating with the printer via USB, I proved this quite effectively by just loading several other applications on my PC during a print and more or less immediately the printer starts stuttering whilst It waits for the next packet to arrive over the USB connection.

Can’t wait to get this working.

At the same time I will be implementing the Bowden extruder mods to give me that extra build height I so desperately need to complete the Savonius & Gorlov VAWT.

I have also been formulating a plan for a major mod to the prusa that will incorporate a ABP (Automated Build Platform)

I am thinking of a design that incorporates the Y axis as an ABP, I am hoping that the addition of a couple of aluminium tubes with printed skate bearing holders at each end (4 bearing’s in total) will allow me to use the Y axis motor as the ABP motor allowing the ejection of finished  parts using a simple Gcode sent as the last line of Gcode in any particular part printed.

That Gcode will simply rotate the whole build surface by one complete revolution, there would be a requirement for no max end stop to be present on the Y axis and it would also remove the need for a Y axis drive belt (as the build belt would become the drive belt.

The printed tube ends would be studded with teeth much like an old 8mm film projector and the build belt would be peppered with a row of holes that would mesh with these teeth.

There are obviously a lot of technical details to iron out but I think this would be a significant advancement over the existing ABS’s that seem to bolt onto the build plate and require an additional motor/electronics for ejection.

Build volume improvements

20/08/2012

1)     Stripped out the hot bed (again!) in order to reposition the Y bed slider bars below the front and rear bars that tie the triangular ends together (this is to give approximately 20mm extra build height) this also required that I pack the build plate mounts so that the build plate just clears the front and rear bars that tie the bottom of the triangular ends together, this basically just puts the bed as low as it can go without fouling on anything. Also whilst I was rebuilding the build plate I repositioned the mounting bushes (lmu88 bearings in my case) more central as I had rushed this part initially not realizing the advantage of keeping everything central to the build plate, this means that I am now able to reposition my Y end stop to utilize the whole hotbed in the Y direction. (every little helps when it comes to increasing build volume a few mm can make all the difference to a part failing due to it being just slightly too large to fit in your printers build volume).

2)     Began printing parts for the Bowden extruder mods (Bowden extruders are just where you separate the hot end from the geared extruder & motor via a long PTFE tube that connects the two) Bowden extruder mods are being touted as a way of reducing mass of the X carriage and therefore increasing speed and also giving the possibility of mounting multiple hot ends, but they also have the obvious advantage of giving considerably more build height on a RepRap although this does not seem to get a mention in terms of Bowden mod advantages! Strange as I would consider it a great advantage.

The print frenzy begins

19/08/2012

Printed universal size spool end caps (ref Part)

Printed spool holder top mounts with integral handles (ref Part)

Fitted spool mount to top of printer for continuous printing without the need to monitor it closely for filament tangles.

Printed Sanguinololu fan adaptor part  (ref Part)

Fitted Sanguinololu Fan adaptor (stripped fan from Darwin as it was the correct size for this part.

Attempted another go at Savonius & Gorlov VAWT http://www.thingiverse.com/thing:16504

Three tries later (2 fails due to overheating of the x & y axis motors, it just happened to be one of the hottest days this year and this shows the need for heat sinks on my motors.)

Bootstrapped a heat sink to the x axis motor, setup another fan to cool the y axis (this proved to be a mistake as the wash from this fan caused the large VAWT part to start warping again, adjustment of this fans direction cured this temporary bodge fix warp issue until I can find and fit some good heat sinks possibly fan cooled)

Third attempt at Savonius & Gorlov VAWT http://www.thingiverse.com/thing:16504

Proved very successful (more luck than judgment as will be explained later) (Photo) (video on youtube)

Dove straight in to print the other parts to this cool wind turbine blade and discovered that my build volume was leaving a lot to be desired.

Upgrades are the future!

16/08/2012

Last night I completed a strip-down & rebuild of my 1^st Child Prusa printer, and to be quite honest I was stunned by the resultant improvement in print quality to the point where I am not quite sure exactly which of the improvements made the difference but the difference is enormous.

Here is a list of the improvements I made over the course of a couple of days (evening’s).

1)     Stripped out partially melted PLA X carriage and replaced it with my new (not great print quality) ABS X carriage, (which I broke slightly whilst fitting it due to delamination issues mentioned above, but this should not affect its performance)

2)     Stripped off the hot bed and remounted with a center fixing bolt to prevent the hot bed from warping down in the center, this was causing issues with getting the first layer right without constant tweaking of the Z end stop position.

3)     During the X carriage replacement I broke the thermistor (again) and decided to switch to a much simpler thermistor sourced from Maplin’s UK (ref part number) it is much larger than the micro sized glass bead thermistors that are recommended for Reprap’s, this required me to drill out the brass heater blocks thermistor location point hole to 3.5 mm, The Maplin thermistors are quite large in comparison to the glass bead type but they are a third of the price (£0.99 each) and a hell of a lot easier to deal with as they have decent thick leads rather than hair like leads for connection and are therefore much less prone to breaking.

4)     Whilst I had the whole thing apart I thought I would add a couple of electrical connectors one for the thermistor leads and one for the heater leads, this way when I need to strip down the hot end for any reason I can simply unplug the leads, snip the filament, unscrew the 2 mounting screws for the hot end and run the extruder motor (with the heater off of course)  until the hot end drops out.

5)     Also whilst I was rebuilding the heater block with the new thermistor I wrapped the whole block in Kapston tape, I have seen other people do this before on various RepRap blogs and presumed it was just to stop the melted plastic from sticking to the heater block, but I now remember back to the days of the Darwin how insulating the hot end was important in maintaining a constant temperature without putting undue strain on the heater electronics, this may be part of the reason for the sudden quality improvement.

6)     I also reduced the layer height to 3.4/6 mm and increased the temperature to 250 Degrees C (these were going to be my next software tweaks anyway before I decided it was time for a complete strip down.

7)     Adjusted Y end stop switch position to give me full access to the whole print bed size, needed to do this as the last thing I was trying to print before I decide on the strip down was the Savonius & Gorlov VAWT http://www.thingiverse.com/thing:16504 this also demonstrated the need to fix the warping hot bed issue.

Considering the massive improvement in print quality I am not going to tweak anything more until I have taken measurements of anything that may affect the print quality.

 

E.g.

 

Measure the exact temperature of the hot end and hot bed as I am fairly sure they don't exactly match the Pronterface settings and this should give me an indication of exactly what this Premium Black ABS likes to be printed at.

 

Measure the exact length of the idler screws, this should give me an indication of how tight I need to clamp the filament so that it does not slip or strip.

 

Save the default configuration (yes I have just been tweaking the default config rather than creating new ones, I have already done this last thing after the extremely successful print, I immediately saved the config as a new separate one and named it sensibly (Premium ABS Black) this way I can begin to build up a set of config’s for each type of filament I have at my disposal.

Premium “Definitely” ABS Tests

I received my new roll of Premium Black ABS the other day an got straight to it testing with a small belt clamp print test to see how it printed, The quality of this plastic is totally different to the reel I had that I thought was ABS (it is probably HDPE) and has its own set of peculiarity’s.

After printing several test pieces I decided to go for a very large print in order to test possible warping issues, after that test failed for different reasons this is what I learnt.

Warping is non-existent it appears that running this premium Black ABS at 220 degrees C and the hotbed at 110 degrees C eliminates warping completely, as I was attempting to print the Darius self-starting wind turbine which fills the build platform almost completely (can’t build larger than that without a custom build printer with an enlarged print bed).

The reason it failed was due to delamination of the ABS layers from one another, on further reading this may be due to my hot end not being hot enough and or possibly my layer height being too large.

Another issue is that as the hotbed on my Prusa heats up is warps the entire bed downwards, further reading on this issue turned up a simple fix of using the center hole as another fixing point and using a countersunk machine screw to fix the center of the hot bed in place this supposedly resolves the warping hot bed.

One of my ABS test prints was a new X carriage and although  the quality is not great (slightly warped on one corner, as I was running the hot bed at only 90 degrees C at the time, since turned up to 110)  it looks like it would be useable.

I need to strip it down and fit the hotbed fix and ABS X carriage, I have also noticed some considerable flex in the X idler end and I am not sure if this is causing some backlash issues or not but I will probably print a better X idler end soon as well.