Add a heated bed to your 3D printer

Introduction

As you may have read I own an Arduino Materia 101 which was set-up for printing PLA on a cold bed, I got fairly good results as long as the contact area between the model and the unheated glass plate was small, furthermore it really helps to use blue painters tape that came with the Materia 101. A significant improvement is to install a silicone heater pad under the printing bed. This blog discusses how I modified the Materia 101 which I built from a kit 3 months ago. It also shows how the printed models have improved.

Modification of the Materia 101

On ebay I found that Keenovo (http://www.ebay.com/usr/keenovo) makes heater pads. I asked them to make a 4 by 4 inch silicone heated pad with a digital thermostat which will be like a WH7016D as in http://www.willhi.com/data/product_en/20140104_173339.pdf Just stick the pad under the print bed and attach the cable with a strap near the left linear bearing and warp it outside the case to the thermostat. Whatever you do, do not obstruct the z-axis, the table should still be able to move from the stop to the polycarbonate plate on the bottom. Keenovo can make the heated silicone pads for 12 volt DC or in any voltage you ask them, like 110/220 AC. The pad is a little bit overrated at 100W so that its draws 8 Amps from the power supply. My lab supply couldn’t drive it but an ATX power supply from a vandalized PC easily provides the 8 amps. The adapter of the materia is rated at 5 amp. I’m not sure what the materia needs but an ATX could drive the materia and the heated pad together if needed. On an ATX lab supply you need the black and the yellow wires, the rest can be stripped off except for the green wire (a power on signal) that turns the ATX power supply on when you connect the green wire to black. Also, all wires of the same color may be bundled together. The thermostat does not care about polarity, as long as it is DC, your Materia 101 does if you decide to replace the adapter.

One extra cable comes out of the Materia 101, you can warp it up along the upper left back, I vandalized an old ATX power supply to drive the heater and the digital thermostat that Keenovo send me
Figure 1: One extra cable comes out of the Materia 101, you can warp it up along the upper left back, I vandalized an old ATX power supply to drive the heater and the digital thermostat that Keenovo send me
figure 2: a 4 by 4 inch pad will fit under the metal plate suspended on four springs. A cable strap keeps the red cable to the support structure.
figure 2: a 4 by 4 inch pad will fit under the metal plate suspended on four springs. A cable strap keeps the red cable to the support structure.

Why does a heated printing bed help

During printing there is a temperature difference between the extruded material for which I used the recommended temperature of 220C and the print bed. As soon as hot PLA is deposited on the bed it will undergo two phase transitions which are discussed here http://en.wikipedia.org/wiki/Polylactic_acid
The temperature ranges we are interested in are the melt temperature at 173-178 degree C and the glass transition temperature at 60-65 degree C. PLA behaves like a rubber like material between both temperatures and this is where we operate heated bed. Below 60 to 65C PLA will be a glass like material, and this is what we eventually want for our models. When the PLA is deposited on the bed at room temperature the object normally tends to warp. The reason is that it the extreduded PLA goes through both phase transitions, as a result it will not stick flat surfaces like glass even when they are treated with hairspray.

So there are two solutions, one is to improve the grip on the printer bed by painters tape and this works particularly well as long as the printed object is small. Also, it helps to warp the painters tape around the glass plate because it restricts the shrinking PLA model which causes warping. But the other solution is to spray the glass plate with hairspray (please do this outside the printer). When you print with an unheated plate the shrinking of the PLA is too much for the hairspray, but, when you heat the PLA to above the glass transition temperature then it is deposited as a rubber like material that perfectly sticks to a borosilicate glass plate that is treated with hairspray. (one thin layer of hairspray on a clean glass plate is good enough).

Optimal thermostat settings

 Figure 3: On the vertical axis you find the measured temperatures on the center and near the corner of the glass plate, and on the horizontal axis is the thermostat setting. I used a non contact thermometer.

Figure 3: On the vertical axis you find the measured temperatures on the center and near the corner of the glass plate, and on the horizontal axis is the thermostat setting. I used a non contact thermometer.

There will always be a temperature difference between the center and the corners of the printer bed. We heat the plate below the center, and, ok, it is a math/physics problem but the corners will be cooler.The thermostat is right under a metal plate, and below the borosilicate glass on which we print. So you measure at the center typically a value 5 to 6 degree C below the thermostat setting. Near the corner of the printer bed lower values appear; the difference is 8 degrees when the thermostat is at 60C, and it can become as much as 50C when the thermostat goes to 130C (the glass plate survived this treatment, not sure what will happen when you would rapidly cool it down, borosilicate glass is like pyrex and it should have a low-thermal expansion coefficient, pyrex is a trade mark and borosilicate glass is a generic material). If you want the entire model to rest on a bed of 65C (which is the upper range of the glass transition temperature region of PLA) then the thermostat should be set 90C. This is printer dependent, what counts is that the Materia 101 is an open model. A second consideration is that you don’t want the PLA to be in its rubber-like state for all layers of the model which causes the model to sag. A setting of 90C is only needed for the first couple of layers, after the 2nd or 3rd layer the thermostat can be reduced to 70C.

Examples

The light blue disk that I designed for Pi day on 14-March-2015 was printed on painters tape and a cold bed. The disk has a lot of contact area and as you can see, one edge of disk deformed by as much as 3 mm which basically ruined the entire idea of the Pi disk. The print quality will deteriorate subsequent layers are printed; the extruder nozzle will hit the warped PLA and scrape off material and so that blobs will form. It results in bad prints. There are several solutions to this, one is to reduce the contact surface between the model and the glass plate; which is a significant limitation of any printer that does not come with a heated bed. But, as said before,  heat the printer bed to PLA rubber state so that the deposited material will stick the to hairsprayed borosilicate glass plate. You only need the 90C thermostat setting for the first couple of layers, because subsequent layers perfectly stick to the already deposited PLA.

Figure 4: My pi day disk was ruined by warping of PLA on a cold bed
Figure 4: My pi day disk was ruined by warping of PLA on a cold bed
Figure 5: A printed carree where the bed was continuously kept at 75C, an a bit of sagging can be seen
Figure 5: A printed carree where the bed was continuously kept at 75C, only a bit of sagging is visible
Figure 6: iPhone stand printed on a cold bed, here the warping was so bad that I decided to add three studs.
Figure 6: iPhone stand printed on a cold bed, here the warping was so bad that I decided to add three studs.
 Figure 7: iPhone stand printed on a hot bed, during the first 10 minutes the thermostat was at 85C, after that I reduced it to 65C. You don't need the studs anymore, no warping.

Figure 7: iPhone stand printed on a hot bed, during the first 10 minutes the thermostat was at 85C, next I reduced it to 65C. You don’t need the studs and there is no warping.
Figure 8: improved pi day disk , 85C initial bed temperature, reduced to 65C after 2nd layer, all warping has disappeared.

Last update: 9-may-2015.

Advertisements