Ok...looks like the rods are glued in place.  Scrap that idea.  Try this:

Hey, IronMan! Thanks for taking your time to help me out! Will try it out over the weekend.

Drop the Z-axis enough to put 3 small machine screw nuts on the bed on a wide triangle to create a 3-point support.  Don't heat the bed.

Put a pc of 2mm glass on the nuts and make sure there is no rocking.  Now bring the indicator (attached to the carriage) to touch the glass and make sure all 4 corners are adjusted to the same reading on the dial.

Now, when you run the X-axis from side to side, that indicator should be almost dead-on.  If not, then it is probably your guide rod.

I'm not at my 'Doodle and I don't recall how the extruder guide rods are held in place on the side carriers, but see if you can spin the extruder main guide rod 180 degrees.  Then if the situation is reversed, there is your answer.

I'm suggesting that your glass may be warped as well; while unlikely due to the manufacturing process of glass, but it could happen...I have seen warped glass...

Ah! Sorry for misunderstanding you.

In this case, I really don't think the glass is warped. I tried it turned 90 degrees, too. Same result.

IronMan wrote:

0.21mm will probably be OK for .3mm printing.  For .1mm you might want to consider checking the flatness of your glass, and getting some 0.55mm (or slightly thicker) metal shims to install under the edges of the thermo-pad to help equalize the bed supporting the glass all the way across.  Then clip the glass on the upper left corner and frontmost (toward you) left and right corners.

Your reply hints that the 4 mm glass is bending by force of gravity alone over a 10 cm length. I need some solid math to believe that to be true.

IronMan wrote:

BTW, how is the thermal transfer on the 4mm glass?...that seems pretty thick.

Not that good. Normal heat up time is ~13 min. It takes ~24 min for the thermostat to indicate 95 C in the bed, but at that time the glass top side is just ~87 C. Ten minutes later, it's still below 90 C. At that time, I did a test run. It failed: the adhesion was way too low. I guess I'll have to scrap that idea. Oh, well...

I'm suggesting that your glass may be warped as well; while unlikely due to the manufacturing process of glass, but it could happen...I have seen warped glass...

I have an SD3 and the bed is warped .015" center to edge so when I tried to use Glass the  their was too much air gap between it and the bed. I thought about using thermal paste to fill the gap and permanently mounting the glass.

You may want to look at thermal conductivity pads instead of thermal paste. Easier to handle and less messy.

0.21mm will probably be OK for .3mm printing.  For .1mm you might want to consider checking the flatness of your glass, and getting some 0.55mm (or slightly thicker) metal shims to install under the edges of the thermo-pad to help equalize the bed supporting the glass all the way across.  Then clip the glass on the upper left corner and frontmost (toward you) left and right corners.

Your reply hints that the 4 mm glass is bending by force of gravity alone over a 10 cm length. I need some solid math to believe that to be true.

Not that good. Normal heat up time is ~13 min. It takes ~24 min for the thermostat to indicate 95 C in the bed, but at that time the glass top side is just ~87 C. Ten minutes later, it's still below 90 C. At that time, I did a test run. It failed: the adhesion was way too low. I guess I'll have to scrap that idea. Oh, well...

BTW, how is the thermal transfer on the 4mm glass?...that seems pretty thick.

So I got me some glass plates, one of which is a frosted glass mouse pad almost 4 mm thick, the others 2 mm picture frame glass. (Going for frosted glass seems like a smart idea, should give me lots of adhesion?) To get an optimal heat distribution, and to make thermo-contact over the entire heat bed, I placed a thermal conductive pad between the aluminum bed and the glass plate.

After re-measuring with my indicator dial, I find that it is still not perfectly level. It is still a ridge, but with less difference between the highest and the lowest points: 0.21 mm.

(25,200)     (105,200)    (185,200)
   -0.03           +0.12           +0.04

(25,100)    (105,100)     (185,100)
  +0.01          +0.18           +0.03

(25,50)      (105,50)       (185.50)
  -0.06          +0.12             -0.03

My main questions are: What are the required tolerances for the heat bed? Is 0.21 mm difference too much or will I be fine with this?

Another question is: I'm still getting a ridge, with clips holding the glass in the center only. I that indicative of bent rods? Or is there some other explanation?