Re: Glass Bed Thickness
Perfect upgrade. Only problem is the position of the clips that hold the glass to the bed. The purple glue stick has never failed me one bit either.
Solidoodle 2 Pro - 3mm Glass Bed
27 2014-07-20 05:46:36 (edited by wardjr 2014-07-20 05:49:02)
- wardjr
- Moderator
- Offline
Re: Glass Bed Thickness
I've said it before and I'll say it again. The clips aren't needed, in fact they can defeat the very purpose you've switched to glass in the first place. If your bed is warped the glass helps with adhesion by providing a flat surface. Glass can bend and the clips will do exactly that. Spray the underside of the glass with some hairspray and run it through 3 heat cool cycles and it won't move until you want it to. No clips to hit your nozzle 
Printit Industries Model 8.10 fully enclosed CoreXY, Chamber heat
3-SD3's & a Workbench all fully enclosed, RH-Slic3r Win7pro, E3D V6, Volcano & Cyclops Hot End
SSR/500W AC Heated Glass Bed, Linear bearings on SS rods. Direct Drive Y-axis, BulldogXL
Thanks to all for your contributions
3-SD3's & a Workbench all fully enclosed, RH-Slic3r Win7pro, E3D V6, Volcano & Cyclops Hot End
SSR/500W AC Heated Glass Bed, Linear bearings on SS rods. Direct Drive Y-axis, BulldogXL
Thanks to all for your contributions
28 2014-07-21 13:11:42 (edited by Tomek 2014-07-21 19:56:54)
Re: Glass Bed Thickness
I use 2mm glass and I'm happy with it. It was cheap. It does not appear to meaningfully bend, and has enough contact with the slightly bendy aluminum bed that it gets good heat transfer.
I do not think people who say thicker is better "because it holds more heat" are really correct. The rate of heat loss from glass to air will be the same irrespective of glass thickness, it's a material & condition problem. The temperature of the glass at the surface of the transition can be dictated as proportional to the thickness of the glass and the temperature of the aluminum plate, and how fast heat is lost from the glass surface. I'm assuming the aluminum plate is a uniform temperature so this is some sort of 1-D problem.
There is some "heat capacity" to heavier materials and more thickness, but it is not relevant to our application because we're holding the aluminum plate at a fixed steady-state temperature. There are no moments in which heat capacity would be helpful as a buffer because there are no fluctuations in heat loss from the build plate. The exception to this is the first layer, when my printer head is closer to the surface and the cooling fan might be cooling it quickly. In that case I use a higher initial layer temp to compensate.
Sorry wanted to chime in without having read everything here, but there are what I believe to be some incorrect parts.
29 2014-07-22 01:18:16 (edited by grob 2014-07-22 01:27:26)
- grob
- Member
- Offline
Re: Glass Bed Thickness
Tomek wrote:
I do not think people who say thicker is better "because it holds more heat" are really correct. The rate of heat loss from glass to air will be the same irrespective of glass thickness, it's a material & condition problem. The temperature of the glass at the surface of the transition can be dictated as proportional to the thickness of the glass and the temperature of the aluminum plate, and how fast heat is lost from the glass surface. I'm assuming the aluminum plate is a uniform temperature so this is some sort of 1-D problem.
Agreed. I did a quick spreadsheet to work out what the difference in surface temp would be for a 200mm square glass plate, with a bed temp set to 110°C, with glass thicknesses of 2, 3 and 4mm, assuming the chamber is sitting at about 40°C.
The theory is that there is heat loss from the top surface by convection and radiation (for which I found this handy calculator). Glass has relatively poor thermal bulk thermal conductivity of 0.93 W/mK (see wikipedia). The heat being lost from the top surface has to be conducted through the thickness of the glass, which necessarily requires a temperature difference between the bottom and the top. With this, we can find a surface temperature at which the conduction through the glass balances the loss from the surface, which I think would be a pretty good estimate. This assumes there is no re-radiation, the bed surface is clear and there is perfect contact between the glass and aluminium bed. See the below/attached graph for results, the working spreadsheet is also attached.
In short, if you're using 4mm glass (quite common), then when you set your bed temp to 110°C you should expect the actual temp of your print surface to be closer to 105°C. While not a large difference, it is appreciable, and using thinner glass can save up to 2°C. Certainly specifically choosing a thick insulator as suggested previously in this thread will have a significant impact on the surface temp, which you should be aware of.
Also interesting is the c. 40W of power lost by convection and radiation, just to keep the temp steady - this is part of what heats up your enclosure.
This should also highlight why good & even conductivity from the aluminium plate to the glass is important (i.e. air bubbles are a big issue). Which for a warped bed like mine, remains a significant problem.
SD3. Mk2b + glass, heated enclosure, GT2 belts, direct drive y shaft, linear bearings, bowden-feed E3D v5 w/ 0.9° stepper
Smoothieboard via Octoprint on RPi
Smoothieboard via Octoprint on RPi