temperature control system for 3D printer enclosure

 The Problem:

 My 3D printer is in a basement room which, in the winter months, stays well below reasonable printing temperatures. The house is literally 115 years old, so fixing the room temp isn't going to happen.

 The First Half of the Solution:

 Fortunately, a very good friend gave me a heat-resistant 3DP enclosure.  The printer alone dissipates over two hundred watts between the bed heat, hotend, and motor drivers.  So I measured temps after the printer had been in "preheat" mode for 30min, and it plateaued just below where I needed to be, while the room temp was on the "high" side of its low range.  Well poop. 

 My tent needed a heater, but obviously if I just stuck a small heater in there, there would be nothing to stop the temperature from climbing so high it cooked the printer's electronics.  It was time to consider a proper temperature control.

 The first bit of good news is that I've done this sort of thing before so I already knew how easy it really is.  It's just a matter of buying a few Commercial Off The Shelf parts and connecting them together with a few simple screw terminal connections.  Usually, you don't even need to solder.

Complete kits containing the PID controller, a "solid state relay", and a heat sink can be found online for under $30, so I did that because I was broke and cost had to trump quality this time.

  (for the reason 'SSR' is in quotes, please see this Big Clive video, in which he tears down an identical unit)

 So it's a fake SSR in the strictest sense, a sense we don't care too much about, but the screw holding the thermopad / TO220 style device inside to the back plate was literally sticking out of the heat sink surface! I had to file and sand all of that flat to get a good mate to the all-too-thin Chinesium heat sink.

 Fortunately, the thing has astonishingly low forward drop compared to the devices I am used to; it didn't get warm at all under a 20A load, and the device is nominally rated to 40A.  I flatly did not believe that number before I tested it... but it might be possible!  In any case, the load for this application is only 2.5 amps.

 The next item I needed was an enclosure.  Because I needed to preserve as much desktop real estate as I could, I decided I wanted a deep enclosure with a small front panel.  After poking around the usual places, I was astonished to find that I couldn't find one to buy at any price, so I decided to make one.

 A quick review of available materials and a bit of brainstorming assisted by a bowl of the finest Longbottom Weed, I realized a scrap of extrusion which has been knocking around my shop for literally decades had finally found its final fate. All I had to do was cut it in half, patch up a bunch of ugly holes, mill a bunch of not-ugly ventilation slots, oh and figure out how to attach the two cut halves to each other in a permanent and durable way, without being able to weld aluminum.  Visions of Rosy The Riveter, hashtag 'we can do this'.

 This we then do.

 

  I drilled matching holes in the thicker bits where the two halves would meet, epoxied little aluminum pins into the holes, applied epoxy to the roughened edges, and "glued" the two halves together. 

 When done, I lacked confidence in those pins and their small surface areas, so I added three machined 'clips' and epoxied those to the inside, around the thick edges.  NOW I am semi-confident that the two pieces will remain one piece... provided nobody drops it.  I wish I had thought to take a picture of them before I painted everything black but I didn't so there it is.

 Next I needed a front panel for the controller and a power switch, and a back panel for power entry module, fan and heater output receptacles, and thermocouple input terminals.  That effort looked like this:

Blogger won't let me put multiple photographs on the same line, isn't that amazing?

 The back panel was a bigger PITA than the front, because it had more fiddly holes in it, but nobody sees it.  

 The front panel needs to look good, but it's only got two big square holes and the four small panel mounting holes.

 I have to assume this extrusion was intended for making things like this.

I also have to assume there is clever hardware for joining the halves which I do not have on hand.

 I decided to make the front panel unnecessarily thick, so I could unnecessarily counterbore the panel screws and use socket-head cap screws for added sexiness.

 This is why some of my simple projects take a long time, but I was motivated to do a better job than I usually would (for myself) because I planned to blog about it here, and I didn't want to embarrass myself.

Fiddly, fiddly. Measure thrice, cut once.

 At least with the front and back panels, if I scrapped one, I could replace it.  That wasn't true for the enclosure halves, and I was tense about that a few times when I was cutting on them.

 

Milling thin materials - or rather, fixturing thin materials so you aren't also milling your machine's table... can be interesting.

 I saw no reason to put much effort into the finish of the back panel, so I just ran it over a convolute wheel on the buffer a few times.

 Then came paint.  On a lark, I decided to coat the enclosure with Rustoleum™ Texture paint (rattle-can).  I bought it with my employee discount at the store where I work.  This won't be an endorsement, nor a condemnation, because while I like the finish, Rustoleum seems to have a problem with their cans; since they now spray in any orientation, it's not possible to clear the nozzle & valve per the instructions still printed on the can.  

 As a consequence, you can get two, maybe three spraying sessions out of a can before the valve clogs.

"image enlarged to show texture"
( for scale, that's a countersunk hole for a #10 screw )

 Seriously, this stuff has a lovely appearance and feel, at least for my aesthetics.  The feel / texture is not "rough" in that it does not drag against the skin the way, say, sandpaper might, as the peaks are not sharp, but it is noticeable.

 The character of the texture changes somewhat with additional coats.  I did a few tests on other bits of aluminum, and it adheres under abuse, and once baked on with heat (always do this if you can) it becomes quite scratch and abuse-resistance, as much as any paint I suppose, but I think the texture helps to hide marks.

 I could have hidden the seam lines in the enclosure using epoxy or bondo and a bunch of sanding and/or milling... when I got to  that decision point in the project, it seemed like a lot of additional work, and I decided to try to reign in my OCD at that point to save myself much frustration, and remind myself of the wisdom of Voltaire: "Perfect is the enemy of Good".

Ze Back Panel.

 The controller can use either T/C or RTD input, and I have brought both out to terminals on the back panel.

 Ground wire is connected to the shield of the T/C cable.

 Yes, there is a missing screw. We don't talk about the missing screw. That was a Very Bad Day.

Done.

 The controller is the auto-tuning PID sort, very easy to use if you are patient and follow the instructions.

 It's not as easy to use as an Omega, but an Omega is $200 whereas this thing cost me $20 including a SSR and heat sink.

 Sometimes Chinesium wins. :(

Which reminds me, I don't know where the instructions are any more. >_<

 note, later:

Now I just need to make something like it all over again for our espresso machine, which ships from the factory with small, click-disk type thermal switches. The inside of the unit is a dog's breakfast of wiring and water tubing.  Just replacing the switches ought to be loads of fun. •cough•wheeze•

 It would literally be easier to disconnect the thermal switches and leave them in place, epoxy a thermocouple to the boiler, connect the boiler leads to the temp controller and cover it back up.

 I'd better start thinking about another enclosure...