Under Bed Insulation -> Faster bed heating and better temperature control

I finally got around to adding under bed insulation to my Robo. I’ve also added a purpose built Solid State Relay to avoid over heating of the RAMPS polyfuse and the mosfet. In my case the bed is the kickstarter 3mm thick borosilicate.

Before the under bed insulation and SSR were added the heated bed was taking more than 10 minutes to get to 115C and struggling to hold temperature. To verify this @Galaxius has done a test on a unmodded Robo which shows it took 14 minutes to reach 115C and needed full power to hold temperature.



The next screenshot shows that after the mod the bed temp profile from ambient (23c) to 115c is 6 minutes. Power to hold the bed at 115c is reduced to approximately 60 - 65%. TBH I was surprised by the dramatic improvement.



The insulation material I used is a type of reflective foil used widely in the building industry (in Australia anyway). Brand names in Australia are Sisalation, HardieWrap and WallWrap.



The foil is thin (less than 1 mm) but very tough and is attached underneath the bed foil side up using kapton tape. The reflective surface improves thermal performance without the need for any kind of bulk insulation (eg such as cork) which would interfere with the Y belt. The foil also has a high fire rating (ie low flamability)

To get this better performance I also had to recalibrate the Marlin heated bed PID values in configuration.h. With the old PID values for the bed, the firmware paused the faster temperature rise way too early.

To update the bed PID, make sure #define PIDTEMPBED is uncommented in configuration.h and use gcode M303 to run bed PID tuning and display the new heated bed PID values Kp, Ki and Kd. Update these values

#define DEFAULT_bedKp ?
#define DEFAULT_bedKi ?
#define DEFAULT_bedKd ?

in configuration.h, recompile and reflash the firmware.

I’m very interested to hear any feedback from anyone using similar foil insulation.

 

SSR is my own design. It is opto isolated, has very low on resistance and just gets warm under the full load of the heated bed.

An SSR is not absolutely necessary, but it means the polyfuse on the RAMPS is no longer a problem with over heating and the RAMPS mosfet does not get hot. Overall, having the SSR saves a few watts (which are better used heating the bed) and avoids any bed cut outs due to overheating of the polyfuse.



I am getting some professional PCBs made for the SSR design as they are useful in many projects.

Re the insulation material:

First requirement is the insulation needs to be thin (ideally < 1 mm). Reason is there is very little clearance under the bed and above the Y axis pulley, idlers and belt. The sisalation (brand name) I am using is very thin 0.11 mm (and consequently very light). I think similar material in the US is called ActicFoil.

Second I am convinced reflective insulation (radiant barrier) gives a better result than thermal resistance, conductive type insulation (eg fibreglass, cork, exhaust wrap etc) in this particular under bed application. This demo

​

shows the incident radiant heat is being reflected back by the radiant barrier - which is exactly what we want under the heated bed.

 

I think the problem with the roll of foil (radiant barrier) tape is the tape would stick to the bed foil side out. You actually need to have the reflective surface facing the under side of the bed.

One possible solution is shiny aluminum foil lightly glued onto thick paper (like some drawing paper or cardboard). Place this foil side facing the underside of the bed and use kapton tape around the edges to secure it to the bed. You actually want some air gap between the bed and the reflective foil surface so the radiant barrier works properly. I haven't tested this, but the principle is exactly the same as the "sisalation" I used or ActicFoil.

 

Not exactly the same but this product would work well.

But rather than buy a roll, I suggest you try the aluminum foil (cooking foil) stuck on thick paper first. This should be just as effective.

 

@Steven Stowell

SSR circuit is as in the eagle schematic attached.

Mosfet is type IRL1404 http://www.irf.com/product-info/datasheets/data/irl1404.pdf
Similar type is ok but make sure the on resistance is at least as low as the IRL1404

Opto Coupler is 4N25 http://www.vishay.com/docs/83725/4n25.pdf but any similar type eg TIL111 would be ok.



I had some SSR printed circuit boards made by OSH PARK (What a fantastic service that guy offers!!) - I am testing these to make sure they can handle the bed current.