This study aims to evaluate and mitigate the gantry deflection seen in some vcore4 printers. The issue presents itself as a squished first layer some time after starting the print, and is a serious challenge for large prints.
The results presented here are from tests done by @tg73 and @mazas. For more detailed background, see the discussion https://discord.com/channels/582187371529764864/1293322335545196655/1293322335545196655. The pinned message in that thread is a good starting point.
In short, the bimetal structure formed by the steel rail and aluminium extrusion of the gantry cause bending as the gantry temperature changes. The absolute amount of bend at a given temperature can be minimised by relieving rail stress in a heat-soaked state. Most of the bending change can be worked around by heat soaking the printer long enough so that the changes are no longer significant.
While the bending with chamber temperature change is a known issue, we are now discussing another, lesser-known problem: bending caused by the hotend, not the chamber temperature. For this bending the printer does not need to be enclosed, nor do you need a heated bed. This hotend-caused bending cannot be removed by conventional heat soaking: when you turn on the hotend, the gantry will bow further, regardless of how long the printer has been heat soaked with chamber temperature.
To illustrate the issue, here is a measurement of gantry deflection on a stock vc4 400 as the hotend is switched on. The printer is at room temperature and front door removed. Bed is off. The values on Y are in mm, and the X axis is time in HH:MM. The distance is measured with a Dial Indicator (DI) from the top of x-rail to the bed, roughly at the midpoint of the x-gantry.
For a detailed explanation of how the measurements were done, refer to https://www.notion.so/mazas-3d/Test-methods-13bc60bcd336806e8684fb67098a85a5?pvs=25
As the diagram shows, the deflection with a stock vc4 400 is around 190um from cold to hot. Steady state is achieved in about one hour.
To verify that question is about the gantry bowing and not just a shift/offset, here are two height maps that only differ in that the latter has had the hotend turned on at 250 for 30 minutes:
First a cold heat map:
And here is after a 30minutes with hotend on. No bed or chamber heating, the scale is the same (-100 to +100um)
As can be deduced from the height maps, the gantry changed its bowing after switching the hotend on, and consequently the range decreased from 0.225mm to 0.156mm.
To verify that this happens also in a heated chamber, with a heat soaked printer, here is an example with a closed chamber, where the bed was turned to 60C for 90 minutes, and then the hotend was switched to 250, at around 18:20. The deflection increased, but not as much as without the bed heating (102 vs 190 um). The lower part of the picture shows the temperature changes, which indicates that the middle of the gantry heats up considerably more than the rest of the chamber after switching the hotend on.
The red curve in the upper diagram is the distance as measured with the beacon, the blue is measured with the DI.