There seems to be a basic misunderstanding with some members as to how the tramming tool works and how to properly calibrate it.
Part of this stems from SPI and Starrett making the instrument convenient by providing the base of the square as reference surface. That surface is "pre-calibrated" to be perpendicular to the shank of the square and the indicators are "calibrated: when they read zero when the tips are in the same plane as the base. Well and good..... as long as you don't disturb the alignment by bumping or dropping and as long as there is no angular runout in in the spindle mounting scheme. If either or both of these conditions exist, the calibration is useless.
Enter the Edge Technology spindle square and others like it. It is actually calibrated in the machine prior to use. The square is calibrated by zeroing the indicator on a fixed vertical reference surface.
In the following illustration, the square is mounted in the spindle and the first indicator is zeroed when its tip contacts the reference point "B" on the table. the square can be rotated and the circle the zeroed indicator tip describes is concentric with the true spindle axis and is contained in a plane which perpendicular to the spindle axis. Now rotate the second indicator to reference point "B" and zero it. If the square is rotated, it also sweeps a circle which is concentric with the true spindle axis and in a plane which is perpendicular to the spindle axis.
Since both planes are perpendicular to the spindle axis and both planes pass through the calibration reference point "B". The planes have to be coincident. To put it another way, if point "A" is on the true spindle axis, distances AC and AD have to be equal and the only way that can occur is if the line CD is perpendicular to the spindle axis.
Note that there are no assumptions made about the indicator mounting bar or the shank of the square being straight or there being zero angular runout. The bar and shank of the square simply are a means to rigidly position the indicators.
But what if there is runout due to a bent shank or angular runout in the mounting system? All the happens is the swept circles of the two indicators have slightly different diameters. As long as their contact point with the reference "B" is at the same z height relative to the spindle, the calibration is still good. If reference "B" is actually a small cylinder, the calibration will be good. Say that the spindle is out of square by .010" in 5" or .002"/". If the reference surface "B" is .25" in diameter, the error will be at most .0005". With any reasonable care in manufacturing the square, this error should be much less.
As to the issues with possible errors arising with using a cheap indicator, as long as the indicator will repeatedly return to a reading, there should be no problem. Whether a division is reading .001 or .0015 makes no difference. You really don't even need a scale on the indicator to use it. You can just adjust until both indicators read the same as your calibration point.