Thursday, 25 March 2010

Heated bed MK3

My first heated bed worked OK but it was slow to warm up and hard to remove objects.

The second one was only ever intended for experimenting with vacuums and magnets but I ended up printing most of my Mendel on it. It worked well but limits the build area.

I have now replaced it with a full size version, using the lessons learned from the first two.

The first bed was the same size as HydraRaptor's table (200 × 200mm) but the build area is only about 150 × 150mm. The warm up time and power are both proportional to area, so I made this one just big enough, i.e. 150×150mm. Removing the 25mm border nearly halves the area!

The other innovations were to make it easier to build. I replaced nine AL clad resistors with four TO220 resistors. These are rated at 50W and 155°C, so are actually totally within spec when I run the table from 48V giving 188W. Instead of having to tap two M2 holes for each resistor these only need a single M3 hole. That is much easier to tap as the tap is a lot sturdier.

They are also lower profile of course. I just noticed that 47Ω ones are less than half the price, so I should have used four of those in parallel instead.

The thermocouple is mounted with a clamp made from PTFE.

Since this bed has a steel plate on top none of the holes need to be blind. That makes drilling and tapping easier as well.

On my previous magnetic bed I placed the magnets in blind holes that were almost all the way through. That required a milling machine to get flat bottomed holes. On this version I just drilled almost all the way though, leaving a lip to retain the magnet.

This is the top side. The magnet in the middle was done with an alternative technique. I drilled a through hole and then jammed the magnet in with a few strands of copper wire. That gets it flush with the surface, giving maximum magnetic force, but it pulled through on first use. I will have to glue it with high temperature epoxy I think.

After a suggestion by Enrique that wool was a good high temperature insulator my friend Steve gave me some carpet underlay made from wool. I used it to insulate the underside, thanks guys.

For the steel plate on the top I used the cover of an old CD ROM. It is only 145mm wide unfortunately. I think it is mild steel with nickel plating. Not as good as the stainless steel springy piece I got from inside a toaster.

So here is the finished article with the biggest bit of Mendel built on it. It was quite hard to remove. I had to remove the steel plate and bend it a little as intended. I had found that things I built recently on the small table could be removed without lifting the plate. I think the Kapton gets less grip as it ages. I tried cleaning it with alcohol and sanding it with very fine emery paper, but that seemed to make it worse if anything. It seems that shiny Kapton gives more grip than matt.


  1. Captivating read! Now the question is, how do you fit that on the Mendel?

    (BTW: this is the blog post pace that we want! Thanx!)


  2. Well the first bed is the right size for Mendel. A bit heavy though.

  3. The resistors seems terribly expensive. Most TO220 N-FETs ($1) should handle required temperature (175C juction, 0.5K/W junction-case -> 150C@50W) and power (300W).
    Gate driving circuit would be neccesary, but simple .7ohm shunt, npn transistor and zener diode(to protect gate) should be enough to build constant-current sink. (somethink like

  4. "It was quite hard to remove."

    NopHead, did you try to let it cool down to 40ºC before removing the piece? -- I can remove very easily when I let it cool to 60ºC or less... (I am using Kapton tape from DealExtreme).

  5. ledvinap,
    Yes FETs are a lot cheaper but I am using AC. For a DC version fed from a PC PSU FETs would be a good solution but its hard to get enough power for a quick warm-up on a Mendel sized table.

    For my Mendel I plan to use 10 50W 47R resistors in two chains of five across the mains giving 500W for about £25. The advantage of AC is it is nearly 100% efficient, just a couple of volts lost by the SSR.

    Yes I always cool to 40C with a FAN at the end of the build. With new shiny Kapton ABS sticks very well but as you pointed out it gets less grip after it has been used a lot.

  6. Your have 220 vac hooked direct to the resistors?
    500W! Sounds fair, the rest of Mendel should use only around 60W.

  7. Yes that's the plan, I ordered the resistors today. Actually its is 240 vac here.

    It will only use 500W when warming up. Once it reaches temp it is only about 40W average to keep it there, so heating up quicker actually saves energy. I would like to use 4mm AL but I haven't found a source of the right alloy to ensure it is flat.

    I will add a thermal cutout and a couple of earth straps to make it safe.

  8. So I ASS-U-ME that this 150 mm x 150 mm table is currently 6 mm thick like your first version?

    I have a piece that is 185 mm x 244 mm and 10 mm thick! I want to see if I can get a company that does water jet cutting to make it somewhat smaller and then create 2 - ~ 5mm thick pieces out of it.

    The big question is can I afford to pay for it...

  9. Yes all three hot beds have been 6mm.

  10. Something like may work on AC, but it is ugly and quite unsafe with live 230. FETs will have to be isolated from plate. Probably not worth the effort ...
    One small advantage over resistor is NPN's negative base threshold voltage coefficient - colder areas get more power.

  11. Looks great.
    But I am still confused.
    Which is better: more grip or less grip?
    Which is better:
    Kapton that is fresh, new, and shiny, or
    Kapton roughed up with sandpaper until it is matt?

  12. More grip is better, otherwise I get the odd corner lifting during the build.

    Fresh shiny Kapton has the most grip and it gets less with use and becomes more matt in appearance where plastic has been stuck to it and peeled.

  13. Wonder if it is surface damage, or a film of degraded/oxidized abs? Have you tried cleaning it up with solvent?

  14. I tried isopropyl alcohol in case it had got greasy but I suppose I should try acetone.

  15. Hello again nophead,

    I was thinking of making a heated bed, I have a couple of questions. I don't want a quick release plate, I want to just print directly onto the bed with just some PET tape on top so:

    1) Does the bed need to be milled flat tooling plate, or will 6mm thick Aluminium Rolled plate be okay? e.g Will this do:

    2) I would like to use 12v instead of 48v, So I need resistors with less resistance I was thinking that these would be okay:

    Running them at 12v would push 43W through them which is double thier rating of 20W (which seems to be a rule of thumb that you used for other heaters)

  16. I think it is pot luck with rolled aluminium. Sometimes it is flat, sometimes not. I recently bought 6mm C250 tooling plate which has specified flatness.

    I used some cheaper 50W TO220 resistors on my Mendel and they short circuited to the tab when I ran them at 50W, 140C. If you look at the datasheet they de-rate to 0 at 155C. It seems you can't get away with abusing these resistors, so I am going to go back to AL clad.

  17. Have you tried press fitting your magnets? I've had occasion to do this at work a couple times recently. For 1/4" rare earth magnets I made the hole around .008" under for composites and more like .002" under for aluminum. A small adjustable hand reamer makes this an easy thing to do, then I used an arbor press to push them in, perfectly flush and very firmly in there. You could probably get away with using a vise to press fit them. I've had trouble in the past getting any kind of glue to stick to their plating with consistency.

  18. No I haven't tried that as I don't have one of these reamers. They might get loose as the aluminium expands when it heats up.

  19. that carpet underlay looks pretty nasty lol