concepts to product prototypes

Currently Woodsmoke is about 80% of the way to well tuned. There was SF40 to thank for the large part of that. The other important step forward though, was installing an extrudate cooler near the hot end. Everyone will have seen the improvements to single walled boxes and followed suit by now, for anyone who hasn't I heartily recommend it. [Thingiverse:6173]. My own cooler is a 12v 50mm radial flow fan, liberated from an aging ATI graphics card. it's wired to the 12v laptop supply that powers Woodsmoke's Gen6 board. The fan currently runs unfettered by resistors and fairly subdues my heated build plate to a chilly 45C. The ungoverned flow is clearly more than is needed and the amount of curling I get on tall parts is substantial. So Woodsmoke's first printed upgrade will be a governable extrudate cooler. The cooler should chill the filament enough to ensure it will not continue to sag, creating inaccuracies that add up as new layers are laid down on top. At the same time it must not cool the filament too much. As the filament cools, it contracts. If the part cools unevenly it warps and will curl away from the heated build plate. The aim of this new design is to allow me to find the Goldilocks zone for chilling newly laid down filament. 

Of course this kind of control could be achieved with a potentiometer. They didn't have any dimmers in the local hardware store and I don't imagine my parents would take too well to my  liberatiion of the living room light switch in their absence. I have also become quite attached to the idea of uploading a new part to Thingiverse called: the_Governor.stl.

The First prototype, warm off the printing plates. Who doesn't enjoy the snap and crackle of PLA  slowly separating from it's glass sheet? Already I am made aware of the need for improvements, like a square aperture for hot ends with aluminum resistor blocks and guide vanes for the Governor. Any more that you care to mention would be given good thought.

Woodsmoke

printer paranoia

First experiments with the heated chamber today. Optos still work in the semi darkness, lucky me. First off I tried the same wall component from yesterday. 50C air. 70C build plate (same low setting on the slow cooker goes higher with a higher ambient temperature. Yes, arduino and a thermistor soon enough)

Pleasingly the wall didn't curl off the build plate like it had been doing in air at room temperature, but it's no achievement given how close the plate was to PLA's glass transition temperature. Of course, the wall also started to sag heavily and the print head was pushing the semi molten plastic around. I also still have the new-perimeter adherence trouble.

Next I mean to print a series of experiments to get a better handle of how the different temperature variables affect the outcome. It is possible that for very tall, 'delicate' parts like the wall component above, sufficient heat cannot be conducted from the build plate, up through the relatively low cross-sectional-area to surface-area ratio of the part, in order to keep it from curling. This problem compounds the taller the part. There is also much less radiant heat reaching the upper layers from the build plate. This is where a heated build chamber will hopefully make a difference. In fact the build plate has a propensity for generating a distinct temperature gradient over the height of a tall part, increasing likelihood of curling. This might be a reason for printing with a heated chamber and not a heated build plate.

Woodsmoke

makeshift headed build chamber

I have the strangest hunch that a heated build chamber is a good idea. Here's the mock up before I invest in anything permanent. Using light, insulative materials for a rapid thermal response time.My second multimeter measures ambient temperature when there are no leads attached, so in to the chamber it goes. Also need to work out how my optos will respond. Do they rely on ambient light to function? or do they generate radiation of their own?

Mocked up an air vent that looks rather like a shirt collar. Up one air vent but down one cake tin and one picture frame. Perhaps in the future, restrictions on one's freedom to discard material goods and the sophistication of reprap machines will mean we frequently engage in similar on-site recycling of unwanted, out of date personal effects. Our reprap machines will extract the raw material and refashion something we do want, at that moment.

Purloined yet another electrical appliance from my long suffering parents. this time it's an air heater. At least this one won't be cannibalised. Two heat settings, 45C air on low, 60C air on high. I wonder if that will increase significantly when the heated chamber reaches a positive air pressure. Leaving outlets around the base of the chamber so that positive pressure will expel the coldest air first.

 

Woodsmoke

calibration to concept models

It's very heartening watching print errors receding in size and significance as my experience grows. But the ones which remain become increasingly difficult to rout out.

I'm currently attempting to construct a simple 1:200 scale concept model of my architectural thesis. To overcome FFF limitations, my proposal will be printed in component parts for further assembly. Walls, floor plates, roof trusses etc. And to best employ FFF signature printing artifacts, such as the horizontal strata of successive layers, I am printing walls upright on the print bed, so the final part will suggest the successive layering of masonry construction. This is proving pretty ambitious.

One wall component. Given these are supposed to be concept models, I turned off fill in Skeinforge to save time and plastic. Unfortunately this aggravates an issue Woodsmoke has while beginning new layers. A layer is finished and the extruder retracts the filament to prevent oozing. The print head rises 0.4mm and the extruder returns to its previous position. A tiny delay occurs before the print head sets off around the new perimeter. The result is a splodge of plastic that interferes with subsequent laying down of new filament. Having moved beyond the splodge, new filament soon re-adheres to the previous layer. But without infill to assist with this re-adhesion, small errors will rapidly add up to a ruined print.

Would Oozebane solve this issue? Certainly the default settings produce no noticeable difference. Another artifact was recently under discussion on [reprap.org] forum. Small changes in the tension of printer belts can cause a noticable waivering in straight perimeter edges. Woodsmoke was exhibiting these symptoms on the Xaxis. I added bearing lubricant to the belt and upgraded the pulley from a reprapped part to a profi-printed part and the problem went away. At least on the Xaxis. This wall component was printed diagonally on the print bed. The waiver is back. Further investigation will establish whether this is X, Y or both axes at fault.

Woodsmoke

toaster < bread maker < slow-cooker.

That's my highly scientific equation for Macgyvers out there considering home-brewing their build plates. I'll cover the serial assimilation of those most unfortunate electrical appliances, not to mention just exactly why in a post to follow. Meanwhile:

Slow cookers house a heating element band, wrapped around a steel skirt in which the ceramic pot sits. Keeping the appliance's electronics intact, I flattened this band and kapton-taped it and it's thermistor to to the underside of the build plate. The most basic slow cookers have two settings, low and high. On low the build plate heated up to 60C before thermal losses equalled the thermal gains and the temperature levelled out. (85C @ high). The glass mirror I placed on top of the build plate reached 55C. Great, unfortunately this was without the cooling fan in operation. Turning the fan on brought the temperature plateau down to 41C. While switching to high didn't raise this reading noticably. I inked black the interface between glass and aluminium and added insulation beneath the build plate. This brought the temperature plateau up to 53-8C, with the fan enabled. So finally I can leave Woodsmoke printing unattended! And I can do it with a heated build plate that costs the same to run as a single light bulb! Or at least that's what slow cookers were reputedly run for, back when they were released. Now it's more like an urban legend. Meanwhile the 4 2.2KOhm resistors sit in their bag.

Woodsmoke

a design blog of my own.

Greetings!

I am a final year student in Glasgow, just about to complete a Diploma in Architecture. Besides this time spent constructing architectural scale models, I am a self-styled inventor with pretensions of engineering and product design. The Reprap project seemed ideally suited to my ongoing design explorations. Happily I have begun documenting this technological odyssey.

I acquired classic Mendel printed-parts on Ebay in October 2010. Then bought the balance from [mendel-parts.com] during the same month. What followed was a lot of waiting! Eventually I began construction around Christmas last year. Been 'calibrating' on the odd day in between architectural deadlines ever since. Finally found several consecutive for surmounting that starting-from-cold syndrome that had held me back this spring break. Now at last 'Woodsmoke' my printer is printing a Driven_gear in preparation for it's predecessor's possible failure. Pretty pleased with how far things have come.

Woodsmoke's current setup

• GEN6 Electronics
• Sells Mendel parts in white ABS
• Profi-printed pulleys
• Adrian's geared extruder 59:11
• V4 (mendel-parts.com) hot-end
• Salvaged CPU cooling fan
• Mirrored glass sheet upon a heated aluminium build plate - backsides inked black.
• Salvaged slow-cooker heating element - kapton-taped onto the build plate.

Woodsmoke