Mark's Hadron Build - In reverse!

Post your build logs here

Mark's Hadron Build - In reverse!

Postby Preso » Mon Mar 17, 2014 12:42 am

My name is Mark Presling. I am a retired Industrial Technology teacher from the Sunshine Coast in Queensland, Australia.
This build log was only started at the point where I finally got my Ord Bot Hadron working successfully for the first time.
DSC_8433.JPG
The completed Hadron as of March 15 2014
My total build took around 8 weeks but around half of that was spent waiting for parts to arrive. So, much of this build log will be about de constructing the finished build. But first, how did I get here?
For the past 10 years I was involved in a programme called F1 in Schools. It is a technology competition in which students must design, manufacture, test and race miniature formula one cars which are powered by a CO2 soda bulb. I had moderate success with many of my student teams regularly placing in the top 3 teams in Australia. One team went on to place third outright at the 2009 World Championships which were held in London. As our experience grew it became imperative to be able to rapid prototype parts for nosecones, wings, wheels and axle parts. Initially we outsourced this to engineering companies who would use the selective laser sintering (sls) process (at significant cost to our meagre budgets). When desktop 3D printers became cheap enough we purchased two Up! Plus printers and these have proven to be invaluable for day to day printing. Having said that, both of the Up printers recently failed. Both printers had the X axis belts come adrift after the drive pulleys separated from the motor shafts. It appeared they had just been glued on with epoxy. One of the newer machines shed its heated bed after all the securing screws shook loose. I was surprised that internally, the build quality is a bit shoddy.
After I retired it became difficult to "borrow" a printer for doing private work although as I was still working as a volunteer F1 mentor, I could still fit the odd job in here and there. What I really wanted was a machine of my own. I did some research and made some tentative planning to build an Eventorbot http://www.eventorbot.com/ but I shied away from the notion of fitting all the motors, wiring and switches inside a narrow steel tube. Eventually I stumbled onto the Ordbot site and after I cursory glance I thought "This will be easy. What could possibly go wrong!" Of course things did go wrong but on reflection I was pleasantly surprised when the thing finally powered up and did what it was told.

The Spool Holder
Initially, I liked the idea of mounting the spool on the rear of the top plate. I had already fitted a Rumba main board at the rear of the plate and there was limited space to fit a spool mainly because the USB cable fits into the right hand side (looking from the rear) of the board and this is also where I wanted the spool to mount. I did print one that I found on Thingyverse but the mounting holes were difficult to fit 3mm metric screws into and I eventually abandoned the idea of fitting the spool on the rear of the plate. I had seen a few Hadrons with multiple spools fitted above the machine but I couldn't see a way to do this from printed parts without making it a bit flimsy. In my home workshop I have a Colchester Student manual lathe, a Denbigh milling machine and a Sieg MX3 CNC milling machine. My main hobby is model engineering so I am quite happy to make metal parts for my Hadron. I designed two pillars to mount on top of the vertical posts of the gantry. These were partly milled on the CNC Seig and the rest of the operations were done on the Colchester. The resulting 6 parts simply screw together and are fitted to the top of the posts with two M5 screws on each side. An axle was made up from aluminium tube with plugs in each end threaded to M5 and all the parts were powder coated to finish. Gloss black was used for the axle so that the spools will spin with very little resistance and the columns were done in translucent red sparkle. (More about the powder coating later). I can fit up to three spools on the holder and I may fit a twin extruder at a later stage so having multiple spools will be an advantage.
DSC_8430.JPG
The finished spool holder parts.

When I still messing around with tidying up the wiring, I needed to take the spool holder off fairly frequently. I found that having to find an allen key to undo the axle was a pain so I printed up two star shaped knobs for either end of the axle. These have a hex steel insert theaded M5 to fit into the internally threaded holes in the axle. The finished knobs were treated with acetone to fuse the layers together and to give them a smooth texture. I will post a picture of my method for doing this later.
Finished star knob.JPG
The finished star knob used to hold the axle.

I have found that when the machine is running the filament feeds into the top of the extruder without tangling and the spool pretty much stays where it is on the axle.

Powder Coating at Home

When I was considering building my Hadron I thought I could do it on the cheap by doing all the metal parts myself apart form the aluminium extrusions. I had plenty of aluminium scrap plate, rounds and squares and I had the CNC mill so it seemed wasteful to purchase a kit. However I did like the look of the kit built machines with the anodised finish on all the plates. I have worked with anodising companies before where we were building model F1 cars. Several cars had black anodised aluminium wheels and I knew from experience how expensive it would be to outsource this part of the build. I did some research on anodising at home and by the time I would have purchased all the chemicals and a power supply it would have been equally expensive and time consuming.
While searching for an alternative I found that a few companies sold powder coating kits for home use. This looked a bit more promising. All that was required was a small oven for baking the parts, a powder coating gun and the powder. I found two suppliers of home powder coating guns and accessories, one in Australia and one in the USA. Now, living in Australia means that buying anything from a US supplier means paying unrealistic charges for shipping. I work on the principle that you can double the purchase cost to get anything sent from a US location to Australia and some US companies refuse to ship overseas at all. The kit I wanted came from Eastwood http://www.eastwood.com/powder-gun-promo-kit.html and at $149US it seemed like a bargain. I had already purchased an oven for $15 on Gumtree and I already had a compressor so I bit the bullet and purchased the kit but in an attempt to keep the shipping charges to a minimum I used a US shipping agent called Hop Shop Go. http://www.hopshopgo.com/hopshopgo/start-shopping.do?l=en&c=AU Long story short, using this service nearly doubled the time taken to get the kit and the shipping charges were still $110 US. In fairness there were winter storms everywhere at the time I ordered the kit and I did have two shipments sent at the same time. The other one was the vee wheels and idler pulleys which I ordered from the Openbuild store in the USA. I also had issues with the shipping agent holding up the delivery with the claim that the powders included in the kit were hazardous and that they would have to be removed before the rest of the kit could be sent. As it turns out the shipping agent got hold of the MSDS sheets for the powder and then changed their mind but this delayed the delivery by another three days. Would I use the service again? Probably not.
Coating the parts for the Hadron was so easy and you don't have to wait for paints and primers to dry. The resulting coating is super tough and it looks great. One of my colleagues commented that the sparkly coating looked like a pimped up hotrod but I like it and it's certainly unique.
One thing I forgot to mention; the powder coat system uses a high voltage electrostatic charge to make the powder stick and since it is a US product you will need to buy a 240V to 110V transformer to run it from "our" power supply. These are easy to buy on Ebay and the one I got does 500 watts which seems to be plenty. Any oven capable of heating to 240 degrees C will do but don't use your oven in the kitchen. Wife approval factor is usually zero for such transgressions!
Powder coat gun.JPG
Powder coating gun and power supply mounted on a board. I made the gun holder from Al sheet. Powder coated of course.
$15 oven.JPG
I used this oven. Purchased second hand on Gumtree for $15, mainly because it didn't have a shelf. I made my own.


Z screws
In the early stages of the build I purchased some M6 threaded rod at Bunnings with the intention of making my own Z nuts from brass or bronze. I went through the whole rack of threaded rod in the store and could not find one piece that was actually straight. I did buy the straightest piece I could find and took it home where I mounted it in the lathe. My idea was to straighten it by selectively reversing the kinks. However I soon abandoned the idea in favour of purchasing some 8mm trapezoidal screw shaft and brass nuts from an Ebay supplier. I had already modelled most of my Hadron using Autodesk Inventor so I knew I could make the nuts fit by machining a flat on the side of each nut and then mounting them on the gantry plates using M3 screws. Everything fitted as I had planned but I still find that leaving the M3 nut mounting screws slightly loose makes things run a bit smoother. When they were rigidly tight the Z motors would sometimes stall when the gantry got towards the lower 1/3 of its travel. Also, because I was machining my own plates I was able to custom drill the two gantry plates to take the mounting holes for the nuts.
I had to machine the ends of the z screws to create a 6mm dia spigot which would fit into the flexible couplings I had already purchased. I was able to mount each screw in a ER40 collet to accurately machine the step. The nuts were machined first with the flat and then mounted on an angle plate in my CNC mill for drilling the mounting holes and threading them to M3.
Z screw assembly.jpg
This image shows the 6mm spigot on the M8 screw.
Z nut mounting.jpg
This shows the machined flat on the nut and the method of mounting on the gantry plates


The Extruder
I was not keen to be building my own extruder. I have had a look at the various "flavours" of extruder out there and I did print the mechanical parts of the EZStruder however it was the hot end that I was not a big fan of. It seems that most solutions end up looking a bit tatty with kapton tape being used to insulate the heater and the thermistor. I did read a post by gyrogearloose who was in favour of a Chinese made extruder which is for sale on AliExpress http://www.aliexpress.com/item/Hot-Sale-100-brand-New-0-4mm-Nozzle-Extruder-Print-Head-for-3D-Printer-Free-shipping/1011128915.html Considering the price and the fact that is a complete unit off the shelf, I decided to get on on the basis that if it was rubbish I could at least modify it. The vendor will fit it with a 100k thermistor instead of a thermocouple (the thermistor was more compatible with my Rumba board) and the hot end is insulated with a silicone shroud. Mine came with a 0.4mm nozzle and I must say the overall build quality is very good. It mounts to the X axis plate with a single M6 screw and with just a moderate tightening torque it seems to stay put. I had already machined my extruder plate without the horizontal bend as per the original Hadron just because it was easier. When I hooked up all the wires and ran it from Repetier Host it did run, albeit backwards, way too fast and although it was heating up I could not get it to extrude anything. After lots of tweaking, I reversed the feed direction and managed to get the steps per mm dialled in fairly accurately. The heating issue was partly related to selecting the correct thermistor in Marlin configuration H. I had no idea what type of 100k thermistor I had and the vendor did send a specification sheet but it was no use to me as it was mainly a log of temperatures vs resistances.
In the end I went with this one:
// 1 is 100k thermistor - best choice for EPCOS 100k (4.7k pullup)
It seems to work OK. I set my extrusion temperature to 235 degrees in Slic3r and my extruder to 88 steps per MM (in the eeprom settings). It now works reliably and I have never had an issue with jamming filament. The print quality is satisfactory however the first layer seems to have quite large gaps between the extruded lines but that could be a Slic3r setting that needs looking at.
All things considered, for $69.50 US and free shipping I think it is a good solution. There is a twin version of this same extruder but if I need to upgrade later I will just buy a second one and mount them with the fan facing the extruder plate.
Chinese extruder.JPG
My handy dandy off the shelf Chinese Extruder.


The Build Plate
My build plate was machined from a 3mm thick slab of aluminium sheet, custom drilled to suit the wheels I was using. It was powder coated like the rest of the machine plates. I chose a Mk2A heater and a piece of borosilicate glass for the build surface. During the first couple of test prints, I noticed that the heater/glass assembly was wobbling around on the springs during high speed X and Y movements. I dismantled the heater and found that the three mounting holes were 1/8" and the screws I had used to mount the heater to the build table were 3mm so there was around 0.2mm clearance around each of the three screws. I machined up some new screws with a shoulder at the top of 3.2mm and the thread at 3mm metric. I had to ream out the holes in the heater with a small needle file until the shoulder of the screw just fitted through and then reset all the springs and nuts and levelled the table again. Now the glass surface is rock solid but can still move vertically if necessary.
build table screws.JPG
Old vs new mounting screws for the heater.

During the first two print runs I found that the printed part separated from the glass by the time it got to the second layer. I was cleaning the glass with acetone as described in various posts but logic told me that it wasn't going to work. After lots of reading about ABS "juice" and various types of tape (blue, kapton, silicone etc.) I got lazy and just smeared the glass with glue stick. It works! I rubbed on one application and to date I have done around 6 prints on the same surface and it still looks OK for more prints. The parts come off easily afterwards. I usually don't even unclip the glass to get the parts off.

Rumba Board and Wiring.
General view.JPG
General view of the Rumba and wiring
I selected a Rumba board, actually it came as a kit http://www.robotronics.com.au/web/products/rumba-3d-printer-controller-drv8825 consisting of fully integrated Arduino board, complete with six Cool Drive stepper motor drivers, six heat sinks, a professional ceramic screw driver, an SD card adapter and a 40mm fan plus all the wiring for steppers, thermistors, end stops etc. It was the single most expensive part of the build but it was available in Australia and apart from the actual stepper motors and the extruder it formed the backbone of all the electronics. I printed a pair of standoffs in ABS to mount the board on the rear of the Ordbot handle plate. I have put these on Thingiverse http://www.thingiverse.com/thing:250324. The Rumba also came with a SD card adapter however the only way I could get it to work was to mount it on the header pins upside down. Some posters have stated that the adapter has the silk screen pin names put on incorrectly although it seems strange that the board is inverted outside the footpirnt of the Rumba board. A bonus though is that the SD card can be inserted from the front of the machine. I mounted the board and roughly wired up all the motors, endstops, heaters etc. and then set off on my knowledge quest to understand how Marlin firmware works. I can tell you that the learning curve here was near vertical but after lots of reading and tinkering I got it all to work. As an aside I had absolutely no joy with the Repetier firmware. I could get it to upload to the board but Repetier Host would have nothing to do with it.
After getting everything to work, I had to tidy up the wiring which, frankly, was a rat's nest. I really didn't want to disassemble the complete machine to drill holes in the V Slot so that the wiring could be run up the void in the centre of the rail. Instead I used 10mm split corrugated conduit (available at Supa Cheap Auto). The advantage with this stuff is that it can be assembled over the completed wiring even if you have already terminated the ends. I also put heat shrink tube over the ends to stop the splits in the conduit from opening. Only do this after you are completely satisfied that everything works as you would like it to. I printed some clips that fit into the grooves in the V slot. These are open so that the conduit can simply be squeezed into them or removed if necessary. I found some clips on Thingiverse but I ended up modelling my own.

Using V Slot
When I was planning this build I was keen to purchase supplies from Australian sources - you know; "Buy Australian" and all that. The first problem was that I could not find a supplier of the Makerslide that was featured in the original Hadron design. I could get V Slot which to me looked like a simple alternative http://openlab.com.au/product-category/openbuilds/vslot/ and I could get compatible vee wheels to suit the V Slot. Too easy!
After assembling the rails and wheels it soon became clear that the spacers supplied with the vee wheels were too short to clear the homing switches, the toothed belts and their clamps. I did take the precaution of modelling most of the build using my 3D CAD programme so I could calculate how long the new spacers would need to be. These were duly machined up from aluminium rod and fitted along with longer M5 bolts. Now everything seemed to work OK. Later in the build when I actually started to fit the belts and home switches it became clear that the heads of the M5 bolts were clashing. Not by much but enough to make the whole assembly fail. I resolved this by inserting 3mm aluminium spacers between the Y axis plates and the vertical rails and between the handle plate and the back of the vertical rails. See the linked image. The green bits are the added spacers. A test run showed that I could get the full range of movement from the complete assembly but it is frustrating to do multiple assemblies and teardowns to get something right that I should have spotted in the 3D model. My CAD programme does have interference detection tools but I did not use it at the time.
https://dl.dropboxusercontent.com/u/676 ... pacers.jpg
Last edited by Preso on Fri Mar 21, 2014 4:12 am, edited 1 time in total.
Preso
 
Posts: 7
Joined: Thu Jan 30, 2014 6:58 am

Re: Mark's Hadron Build - In reverse!

Postby flickerfly » Wed Mar 19, 2014 12:37 am

Hi Mark, it sure is looking pretty. That red powder coat came across beatiful and the attention to detail made me think, "Oh, of course he has winning F1 cars!" :-) Thanks for sharing! Hope it serves you better and longer than your last couple bots.
flickerfly
 
Posts: 247
Joined: Sat Apr 27, 2013 10:04 pm
Location: Allentown, PA


Return to Build Logs

Who is online

Users browsing this forum: No registered users and 31 guests

cron