Tuesday, May 15, 2012

Simple control of heated bed...


Printing without a heated bed has it's limits. Larger objects begin to curl at the corners and lift off the bed. The plastic just doesn't sit well. I had reached the limits of what I could print without a heated bed so decided to build one. There are a few designs out there, just search under "heated bed" on the RepRap wiki. I chose an aluminium plate with heat resistors as a starting point. I had some scrap aluminium to hand.

The Gen6 electronics has no heated bed control, so I was faced with a challenge, like many other Gen6 owners. (Gen6 Delux board has been introduced in recent months which does have heated bed controls.)

I started with this popular hobby thermostat kit.(also visible here.) In it's standard operation it has a temperature range of 5DegC to 30DegC, and uses a 10k NCT thermistor, and you can manually vary the cut-in / cut-out temperature setting by adjusting the trim-pot. You can change the behaviour of this little thermostat by changing resistor values. They have printed tips directly on the board, e.g. Decrease R5 to increase Max temperature.


With limited electronics knowledge and help from my friend Google, I figured out that changing R5 from 120K to 33K would raise TMax well above the range I'd need for PLA (somewhere between 40 and 60 Deg C), and swapping the 10K R4 resistor for a 100K resistor would allow me to use a 100K Thermistor instead of the supplied 10K one (the 100K resistor (R4) is soldered to the back of the board so looks missing in the photo - don't ask! ) The hysteresis can be changed by changing R7.

From what I had read the little realy on the board wouldn't be up to switching the heated bed, so I salvaged a MOSFET from a scrapped UPS to act as the switch. I had read in Adrian's Prusa Notes how a MOSFET could be used to switch large DC loads that would be drawn by a heated bed. I also watched this mosfet tutorial on how they work and how to wire them up.

The mosfet has 3 pins, Gate, Drain and Source, seen above from left to right. The white wire supplies a voltage, managed by the thermostat, to the Gate pin of the mosfet. The mosfet then allows a more powerful current to flow through from the Drain pin to the Source pin switching on the heated bed. The mosfet can take a much greater load than relay contacts, which would burn out from the switching of a large current. Note: The mosfet will stay 'on' even without a voltage applied to the Gate pin unless there is a 'drain' resistor linking the Gain and the Source pins. This is typically a very high value resistor. My resistor is not visible in the photo but is soldered on the underside of the board.

The next thing I had to do was beef-up the power supply situation. I kept tripping out my salvaged power supplies, so I bought one of these 12v 20A supplies. It's the business! (pict below)
I bolted four heater resistors (2.2ohms) to some 3mm aluminium sheet, two pairs in parallel, then in series. That came to an over all resistance of 2.2ohms and with 12v gives 65Watts roughly. I figured that would do me as a starting point for use with PLA at least. Resistor layout shown below.
I used M3 countersink bolts, tapped the resistor holes and secured them down with some heat conductive paste ensuring good heat transfer to the aluminium sheet.

I thought I might be able to get things to heat faster if I put a higher voltage through the resistors... Above is what happened when I reached19v! There was a popping sound and a puff of smoke! Now I know what's inside these little resistors!

As part of the new heated bed I printed a thumb wheel to make the bed easy to level. I found that springs from some clothes pegs worked nicely on each levelling leg. The little thumb wheel hangs nicely over the edge and makes levelling very easy.

For my first run of the heated bed I placed a piece of glass from an old halogen-lamp, on the bed, covering it with kapton tape. It worked well as a test. It heated nicely. I measured the heat at various points around the top surface using a laser thermometer. I ran some test prints and observed how the PLA was sitting.

The tempered glass was undersized and quite heavy, so when I read how much success RichRap was having with cheap mirrored glass I got a �2 mirror, cut it to size, and clipped it on to the aluminium heated bed. It worked a treat! I'm only heating it to about 60DegC and so far it's surviving nicely. I'm only clipping it on in two places so it's not constrained or over stressed. Hasn't budged yet. I wanted something flat to test print on the mirror so I printed this box. It came out very well. Seen in pict above.

Here's the under side of a z-motor bracket I printed on the mirror heated bed. It's smooth, uniform and perfectly level. It stayed firmly attached throughout the print, and just clicked off once the bed cooled down.


Here's a view of the top-side of the same part.I'm very pleased with the finish. Tech note: 1.75mm PLA, .35mm nozzle, .25mm layer height, 20% fill, 40mm/sec perimeter, 60mm/sec in-fil, 80mm/sec travel.

Through trial and error I've established a position on the thermostat trim-pot that gives a good print bed temperature for printing PLA. I've market that position so I can easily return to it. The thermostat has a convenient red LED that clicks on/off as the bed temperature is reached, drops slightly, then is raised to it's target temperature again. I've found the current hysteresis pattern of the thermostat to be satisfactory but if I wanted it to respond more quickly to falling temperature I could simply swap out resistor R7 for a different value. 

Conclusions:
Low number of resistors gives uneven heat distribution and possible hot-spotting. This is only been observed by taking temperature reading from various points around the board after the heater has been on for a while. We'll see how things go as I try print larger objects or multiple objects.

The little thermostat controller board is a convenient semi-manual way to manage a heated print bed. A mosfet is a key component for switching larger DC loads.

The cheap mirror makes a good print surface for PLA. I might get a second one so I can remove one job and start a second.

Thanks for viewing. Comments and questions welcome. I'm also contactable on the RepRap forum under the username "NumberSix".