I have been following these leveling instructions: http://reprap.org/wiki/Leveling_the_Print_Bed and when I get the 4 corners leveled and at the right height, the center is too low. I have to adjust the z-height and it puts the corners touching the platform. There seems to be a dip in the center of my platform.
Do you have the glass on the bed? The glass is definitely flat (???). Perhaps I’m not quite understanding it. Printing this part will help with leveling:
Michael,
I have seen a similar issue with mine, this seems to pose more of an issue for me when I am attempting to do prints at higher quality settings.
Jebba,
I understand that the glass is flat but when leveling the bed it is in theory possible to flex the glass (making certain assumptions) . Do you happen to have the material properties for the specific glass that is used in the beds. I am happy to run some calculations to see if large enough loads could be applied via the clamps to cause deformation.
Michael (again),
Try leveling it when the bed is up to temp, it should not matter but Orias told me to get it a shot and it might have made my bed leveling a bit easier.
If you are ever worried about working with your bed at elevated temps I would suggest these http://www.setwear.com/HotHand.html. I got them back in college when I was doing some theatrical lighting tech and they are amazing. They allow you to wok on very hot things with great dexterity, on the flip side they also allow you to hold very cold things (i.e. dry ice). I am very happy that they have found a new life again working on my TAZ.
-Tom
The glass is Borosilicate from Allen Scientific Glass. I can get more details if you need. I’d be very surprised if it was flexing, but I work up front doing paperwork, maybe the techs have seen it happen… 
I’ve seen the exact same thing when printing (especially bed_level.gcode).
uneven heating perhaps
I’ve seen this on other small machines including 3D printers. When the X-carriage is at the center of its travel, the deflection of the X guide rods is at the maximum. Deflection in the rods results from a combination of the deflection in the rods themselves due to the weight of the extruder, and distortion of the rod mount due to the moment caused by the x-carriage.
This deflection is generally imperceptible if you were to just fiddle around with the component parts. However, with print quality being dependent on leveling to a fraction of a mm, even a miniscule deflection at mid-travel can cause problems.
Lulzbot’s over/under configuration of the rods is superior to the side-by-side arrangement that some (Makerbot) use. Unfortunately, machine rigidity has not been taken as seriously in the evolving 3DP market as it should be.
My apologies in the delay in response. I have been using the bed level gcode and printing at 110C there are issues… but I think I have found a solution.
I had been printing with the head VERY close to the platform to get adhesion. This left little room for error when parts of the glass were too high it would block the nozzle. The arms on my octopus print were very smushed because of this, but it was the only way I could get it to stick.
I have been using the ‘acetone glue’ as described in the manual for the past couple of days and that has allowed me to use a higher print height which is more forgiving to the slight planarity issue and get great adhesion.
I find that one layer of ‘acetone glue’ will last for several prints.
Ah yes, acetone glue (aka lulzjuice) makes a huge difference in getting parts to stick and not warp.
Take a look at the OP’s post, If what you are saying is the cause (deflection of the rods) then the center of the glass would be too high…but the OP is saying the center is too low, like the glass is concave…not like the rods are drooping.
‘lulzjuice’ hahaha we will start using that around the office