Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Making a temperature controlled heating element

  1. May 27, 2017 #1
    I'm considering making a plastic melter / extruder for making my own 3D printer filament as opposed to buying the spools of it. The basics of the project are simple, but the hard part (for me at least) is making a heating element that has a control on it so that I can turn the heat up or down and keep it where I want to. I plan on using nichrome wire as the source of heat, and it will be wrapped around an aluminum tube (probably 1/8" thick at most). As a bonus, it would be great to implement some kind of digital display for the temperature, but if that's too complicated I can just set up a simple thermometer. I would like to be able to melt a few different kinds of plastics, so this should be able to at least get up to 250 degrees Celsius. I haven't committed to this project yet, so if it turns out to be an electronics nightmare, I'll hold off, but if it's relatively easy to understand, I'm ready to tackle the challenge. And it may also be worthy of note that my electronics skill level is basic (and I mean simple) circuit diagrams and concepts, as well as my ability to solder. Thanks in advance.
  2. jcsd
  3. May 27, 2017 #2


    User Avatar
    Science Advisor

    You can buy a digital temp controller for just a couple dollars. or for a few more. Just find one that'll met your specs.

    Many have an autotune function, which works in the same way as the autotune feature in the marlin firmware found on most 3d printers. So you don't even need to know how to tune a PID controller.

    Alternatively, there are a open source plans out there: link 1 link 2 link 3
  4. May 27, 2017 #3
    I was expecting a bit more DIY, but when options are this cheap, I'll avoid the frustration. I've looked a bit into the PID controller, and it's a bit confusing how to wire it up. I found a video of the basics, ( ) but they didn't hook up any kind of heater to the circuit. They did however have a power outlet in their circuit. Would you plug the heater into the power outlet? From what I gathered, a PID essentially regulates the temperature by using a thermistor to measure the temperature and compare it to the desired temperature, and keeps said temperature by turning the relay on and off. I think this would turn off the socket, turning off the heater, assuming it was plugged into the socket. If my assumptions are correct, please let me know, if not, please correct me.
  5. May 28, 2017 #4
    What will you be using for insulation? It will short circuit if wrapped directly around an aluminum tube.

    In an actual single screw plastics extruder, most of the heat is generated by shearing friction between the barrel and polymer. From what I've read in open source link #3, these DIY extruders externally heat feedstock to above their melting point, and use the auger (screw) to convey melt to a single hole die, and generate enough force to extrude it. This must be the case since it has a throughput of only 1 kg/hr - not much shearing there - so nearly all of the heat is generated by the heating element, and it will be the determining factor of how hot the melt gets.

    I didn't watch the entire video, but yes, the presenter was wiring the output power controller SSR to an AC receptacle.

    PID (Proportional-Integral-Derivative) is a type of feedback control loop that compares a desired setpoint value to a measured one (determining the error between them), applies a set of "PID" equations to that error, and manipulates an output to maintain as small an error as possible. It doesn't matter much to the loop, as far as PID equations are concerned, what type of feedback device is used.

    For temperature, thermocouple, RTD (Resistance Temperature Device), and thermistors are popular, but if you use a permanent magnet DC motor with an off-the-shelf DC motor controller to turn the screw, chances are it has two "PID" loops - an 'outer' speed loop (matching 0-100% speed potentiometer setting to 0 to 100% armature voltage (or, for the next step in control quality, to tach feedback) and an 'inner' current loop relating current feedback (typically, voltage drop across a small value resistor in series with the armature) to maximum bridge output current.

    I've read up some, and have learned that 100K NTC thermistors are popular for 3D printer nozzle control. Keep in mind most of the industrial controllers you'll find are designed for thermocouple and/or RTD sensing. RKC C100s are good, robust controllers (I've used them in hundreds of loops), but at those prices they've either "fallen off of a truck", are good pulls/overstock, or knock-offs. RKC is a Japanese company. If the instructions actually are in Chinese with an English translation as mentioned in the video then the chances are good it's a knock-off. Billy_Joule's 'for a few dollars more' example link to an RKC REX-C100FK02-V*AN that costs only $9.03, hmmm ... typical RKC single unit pricing is more like $200.

    Pay attention to controller part numbers! For an RKC C100, 'V' means it has a DC voltage pulse (transistor) output. It could also have been 'M' (relay), 'G' (TRIAC), or a '7' or '8' (0-20 mA, or 4-20 mA analog outputs, respectively). I don't know what the 'FK02' input type is - '02' is the range, but don't know what to make of 'FK'. 'JA3' is standard thermocouple, 'DA1' is RTD, but ''FK' isn't listed in my manual (another clue it is a knock-off?).

    Fabricate a guard (1/4" Lexan should do) to enclose your extruder rig. PLA and PET are both extremely hygroscopic, and when heated can easily degrade into a brownish soup with a consistency between motor oil and gear oil. You won't want this stuff flying around - 3rd and 4th degree burns.
  6. May 28, 2017 #5
    You mention having an insulator between the heating element and the tube. I was going to ask about what to use for that. If I expect this thing to get up to 250 Celsius, what will stop the current, but act as a strong conductor of heat? Are there any materials made for stuff like this, or should I reconsider the design?
    When you were talking about motors, you mentioned having it be controlled. I wasn't really planning on doing that, it was more "wait for it to heat up, then turn the auger on." I may have misunderstood the motor lingo, as I was just expecting to use an on/off generic 12V motor.
  7. May 29, 2017 #6
    Mica and various ceramics are often used, and a good way to approach this is to learn how electrical heater manufacturers do it. I recommend Google searches on Watlow, Tempco, Ogden, Thermcraft, and other manufacturers, and reading up on the subject. Rather than building heater elements from scratch I'd investigate what was readily available, and figure barrel OD and length with this in mind. IMS started out by focusing on injection molding needs, and sells a wide range of small diameter 120V and 240V heaters.

    An over-simplification to be sure, but there are two different types of polymer flow in an extruder - drag flow (α) moving from the hopper toward the exit die, and pressure flow (ß) pushing back from the die toward the inlet hopper. Extruder output rate (Q) ≅ αN- ß(ΔP/η) where N is screw speed, ΔP is the pressure differential across the screw (hopper inlet to die outlet), and η is polymer melt viscosity. α and ß are determined by screw geometry and barrel clearance, and viscosity by polymer characteristics and melt temperature. Screw speed is the main variable you have to control output rate, and the amount of pressure at the die.

    In billy_joule's link 3 the DIY author built a functional extruder. It is a good template because he has accounted for most of the basics, including a thrust bearing (what he calls auger kick-back protection). The amount of backward thrust the screw places on the thrust bearing depends on barrel ID, and head pressure at the end of the barrel. This force is formidable in a production extruder (for a 6" machine at 1500 PSI, it is more than 21 tons), but is considerably less so with the 3/4", schedule 40 steel pipe used in his build.

    The amount of thrust generated is screw barrel outlet surface area (in square inches) times pressure (in PSI). 3/4" trade diameter steel pipe is 1.050" OD. The ID depends on which schedule the pipe is (how thick it's walls are). For schedule 40 pipe, the ID is 0.824", so it has an area of 0.533 square inches, and at 100 PSI, a force of approximately 53 pounds is placed against the thrust bearing. Mentioning it because this imposes limits to how large the barrel ID can be in a DIY project, and for the amount of head pressure it can be allowed to produce. For instance, increasing barrel diameter to 1 inch trade size schedule 40 pipe (1.049" ID) is a 27% increase in pipe diameter, but generates 86 pounds of thrust against the bearing at 100 PSI, an increase of 62%. Double head pressure, and the force will double. At some point there will be enough force to tear the extruder apart.

    You may wish to consult this University of Toledo Polymer Sci lecture note for more information on the fundamental extrusion process. http://www.eng.utoledo.edu/coe/polymer/courses/pdfs/extrusion.pdf
  8. Jun 4, 2017 #7
    So, after going over this a couple times and reading up a little bit, I've got a couple new questions. The heating elements made by companies like the ones you mentioned are almost completely isolated in the industrial field, and there aren't really many available ones for such small applications. I have, however looked at the need to be electrically insulative, as well as thermally conductive. Looking at some of the available options, Beryllium Oxide, Aluminum Nitride, Boron Nitride, Shapal Hi-M Soft, and Alumina, I think the Hi-M Soft would be the best option. I would however, have to get a part custom made, and it would would likely cost a fair bit. As for external insulation, all I really need is to make sure I don't get burned, so using some cheaper thermally insulative stuff would do the trick. With the formula you brought up, the limiting factor (I would think) would be pressure, as a DIY build most certainly won't hold up to industrial specs. Should I work on this formula using deltaP as the variable? Sorry it took so long to reply
  9. Jun 4, 2017 #8


    User Avatar
    Science Advisor

    All of those are hard (non-deformable) and, with the possible exception of Shapal Hi-M, brittle. That's why Mica was sugggested. Another possibility is Kapton film. It's more pliable than Mica and more available; even available as a tape or sheets. It is used as a solder mask when wave soldering. You probably ought to check the maximum operating temperature before commiting to it.
  10. Jun 4, 2017 #9
    I was under the impression that kapton tape was thermally insulative, but I suppose it isn't. Seeing as how I could simply wrap that around a tube, as opposed to custom ordering a part, I think that's the better option. I was looking into that formula for extrusion speed, and looked up some other ones for clarification. The one I found for drag flow was QD = (π/2)WHDNcos θ ≡ αN where Barrel Diameter D = 2R, Channel Depth H = R − Ri, Screw Helix Angle θ, Screw Clearance h = R − Ro, Screw Pitch B + b ,Channel Width W ,Screw Rotation Speed N (RPM), and Flight Width w. For some of these values I need to work with a specific auger, and I've yet to find one that would work for any reasonable amount of money. If anyone knows where to buy single augers for a reasonable price, please let me know. In the mean time, I'll be looking for augers.
    Edit: So, I'm wondering if a have to use an multi-hundred dollar industry feed screw, or if it would be okay to use a $20 auger bit for a drill. For my purposes, I'm thinking that the drill bit will work fine. If you know that it will make any major differences, please let me know.
  11. Jun 4, 2017 #10
    I did some digging and found the formula you brought up. They say that β is a proportionality constant. What parts of the screw geometry would that be a part of? If I knew this, I could begin playing around with the formula to find the pressure.
  12. Jun 5, 2017 #11
    Powered alumina is often used as electrical insulation within rod and cartridge heaters, but it isn't very suitable for a band heater. Beryllium oxide has good thermal conduction for something that is electrically non-conductive, but is expensive, and requires special handing (once cast it's safe, but dust from filing or machining operations cause berylliosis). I like Tom's suggestion of Kapton film, although I've never seen it used in a conventional heater band.

    I'm wondering, why not simply buy an off-the-shelf heater? Ian McMill's DIY project used a 200W, 120V, 25mm dia. x 30mm long mica band heater he picked up on eBay. Searching there kicked up several suitable Chinese band heaters at less than $10 a pop, and they are also sold through Amazon. IMS part# 134606 is a 1" diameter x 2" long 120V/200W mica band heater for $9.70. After all, your primary interest is in producing polymer filament suitable for 3D printing (which, in my opinion, is enough of a project in itself). Buying a band heater represents a lot less time, effort, and cost than building one from scratch, and leaves you more time to invest in your ultimate goal.
  13. Jun 5, 2017 #12
    Q ≅ αN- ß(ΔP/η), and QD = (π/2)WHDNcos θ ≡ αN are useful at providing insight how extrusion process variables relate to one another, and I mentioned the former to show why screw speed control was necessary.
    Channel depth H, channel width W, and flighted screw length L.
    Qp (pressure flow) ≅ (W*H3/12η)*(ΔP/L)≅ß(ΔP/η)

    A re-purposed 3/4" auger bit will work fine for a DIY project of this limited scope, in fact, a "real" pressure generating extruder screw requires a suitable barrel (a hunk of 1" pipe will no longer do), and a lot more in the way of support structure (a couple of wooden blocks held together with steel angles won't cut it, either).
  14. Jun 5, 2017 #13
    When it comes to DIY, I have a bad habit of over DIYing, so thanks for keeping me in reality land. I didn't think those heating bands came in small enough sizes, but the 1incher will likely do. This paired with a 1 inch auger drill bit, an aluminum pipe, and a PID temperature controller will likely do. My main issue now is knowing what kind of motor to use torque and rpm wise. I'll play with some calculations to figure that out.
  15. Jun 5, 2017 #14
    1" heater ID means a pipe (extruder barrel) OD of 1". You'll be needing a 3/4" auger.
  16. Jun 5, 2017 #15
    Looking at the heaters, they seemed slightly adjustable, but if not, then I suppose I will need to make the auger smaller. How much clearance should I have between the auger and the ID of the tube? Should it nearly miss it by 1mm, or should there be some generous space?
  17. Jun 5, 2017 #16
    Edit: I was thinks no about what motor to use, but I just came up with an idea. If I simply wire the motor to a potentiometer I can just adjust the rpms until it runs smoothly. Does this sound okay, or do you think I'll end up damaging something?
  18. Jun 5, 2017 #17
    Rule of thumb minimum clearance for a new extruder screw is 0.001" (1 mil) per side per inch of diameter. Maximum new clearance is 33% greater than that. For instance, on a 6" extruder the barrel bore diameter is ideally 6.000", and screw is 12 mil (6" dia x 1 mil x 2) to 16 mil smaller in diameter, or from 5.988" to 5.984". As clearance increases, so too does the amount of polymer pushed back over the flights due to pressure flow. Keeping that general rule in mind, a build using a drilling auger and a length of water pipe are unlikely to provide these tolerances.

    A complicating factor here is if you use an aluminum tube rather than a steel pipe for the barrel. Although other alloys are used, particularly if the polymers involved are corrosive, abrasive, or both, extruder barrels and screws are usually made of 4140HT alloy steel, and both have the same CTE (coefficient of thermal expansion). The CTE for steel is 6.78 micro-inches/inch/°F, and the CTE for aluminum is about twice as much (13 micro-inches/inch/°F). You'll want to read up on CTE, run the math, and determine how it will affect your project.
    You'll have to sketch this out and show me what you mean. If it's what I think you mean the answer is no ... it won't work, and something will be damaged.
  19. Jun 6, 2017 #18
    You mentioned using steel instead of aluminum. My dad works as a fabricator so most of the scrap I'm free to use is aluminum, that's why it's my go to. I think they may also have steel on hand (although I'm not sure what type). Also, if the auger needs to be that close to the ID of the pipe, I may have to re-evaluate materials. Or, do you think if I used a 3/4" auger in a 3/4" tube, the manufacturing error would give me that hair of clearance I need? Without knowing the specific geometry of the auger, how would I determine how fast/strong the motor has to be?
  20. Jun 6, 2017 #19
    Tomorrow I will get a legible drawing together showing what I'm imagining in my head so you guys can visualize what I'm trying to get at.
  21. Jun 6, 2017 #20
    From post #6,
    Engineering Toolbox pipe dimension table

    For schedule 40, 0.824"-0.75" = 0.074" =0.037" clearance. Schedule 80 steel pipe ID is 0.74" (3/4" auger won't fit), but chucking it up in a lathe and running a 49/64" (0.7656") twist drill through it provides approximately 8 mil clearance. Near as I can make out, Ian McMill used schedule 40 pipe in his proven build - a lot of slop between barrel and screw means less efficient pumping with more backflow over the screw flights, and may be one of the reasons it worked without blowing up. Another thing - the more efficient the screw (the more mass is moved per screw revolution) the more power it takes to run.

    Buy a 3/4" auger, sketch it, take dimensional measurements, annotate the sketch, then plug these values into the throughput calculation. However, don't expect the answers to bear very much relation to throughput in actual practice ... there's a lot of difficult-to-quantify stuff going on in an extrusion process.

    This article in PT (Plastics Technology) magazine supposes we're talking about real world production extruders, and not a miniature model, but keeping that caveat in mind it provides a calculation to estimate screw power requirements. Using the specific heat for HDPE, and the same melt temperature as used in the article, an extrusion rate of 1 kg/hr (2.2 lb/hr) works out to about 1/4 HP. You'll want to look up specific heats for the polymers you intend to use, and run the calculations.
  22. Aug 24, 2017 #21
    I haven't had a chance to read every post in this thread, but did you figure out the heating/controls? I actually make heating elements (and sensors, and controllers) for a living, so I could give more guidance if you let me know where you're currently stuck.

    We wrap nichrome wire around a ceramic -- usually magnesium oxide (MgO). To reach 250C, you'll likely have to wind more wire than you'll enjoy doing by hand. Most 3D printers use a cartridge heater, btw. You can probably find them on ebay for pretty cheap.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted