Z-Height issues

[brucehvn]

I built an i3 variant a little over a year ago and have been using it ever since. I've noticed that my prints are now not coming out at the proper Z height. It was losing about .25mm per 10mm of height meaning I print something 10mm high and it comes out about 9.75mm. I was printing at .25mm layer height and switched to the more exact (per the calculator) .2480, but that didn't help.

The printer uses the standard M5 threaded rod for the z-axis. The motors are coupled using aluminium couplers.

Yesterday, I ran the z-axis up and down and cleaned the threaded rods, then re-oiled them and the problem actually got worse. Now I appear to be losing about .5mm per 10mm of height.

As near as I can tell, moving the carriage manually using Repetier moves the proper amount. It's kind of hard to find a consistent place where I can put my calipers and try and get measurements, but moving 10-20mm seems to be giving me accurate movement +/- about .03mm.

The real weird thing is that I still have calibration objects I printed months ago and every one I measured averages the exact height they are supposed to be or very close to it.

The prints I'm getting now are very good. I really have no complaints about the quality, it's just the height is off. I'm building a new printer and I need to print parts for it that are pretty precise but this is a problem now.

My relevant firmware settings are below. I've upgraded to the latest Marlin, but that didn't appear to help anything.

The things I've tried so far:

• Oiled the Z threaded rods
• Since I use ABL, checked to make sure the Z probe offset hadn't drifted. Seems dead on.
• Changed the Z feed rate from 2 to 1 and also tried 3, but the Z motors made a lot of noise at that setting.
• Increased the stepper voltage from .41 to .55v
• Double checked the e-steps and they seem almost perfect.
• Upgraded the Marlin version to the latest.

#define DEFAULT_AXIS_STEPS_PER_UNIT   {80,80,4000,724.394}
#define DEFAULT_MAX_FEEDRATE          {500, 500, 2, 25}    // (mm/sec)
#define DEFAULT_MAX_ACCELERATION      {9000,9000,100,10000}    // X, Y, Z, E maximum start speed for accelerated moves. E default values are good for Skeinforge 40+, for older versions raise them a lot.

#define DEFAULT_ACCELERATION          3000    // X, Y, Z and E max acceleration in mm/s^2 for printing moves
#define DEFAULT_RETRACT_ACCELERATION  3000   // X, Y, Z and E max acceleration in mm/s^2 for retracts

// The speed change that does not require acceleration (i.e. the software might assume it can be done instantaneously)
#define DEFAULT_XYJERK                15.0    // (mm/sec)
#define DEFAULT_ZJERK                 0.4     // (mm/sec)
#define DEFAULT_EJERK                 5.0    // (mm/sec)

I'm printing ABS with an E3Dv5 hot end @240C. I've tried slicing with both Slic3r 1.2.9a and Cura 15.04 with the same issue. Slic3r indicates 40 layers which seems to add up right. 40 * .2480 = 9.92mm + (1st layer is .32, diff .072) = 9.992mm.

I'm pretty much out of ideas at this point. I hope someone can offer up some suggestions on what I can try.

I've done many temperature tests in the past before settling on 240, but if the temperature were too high, could it cause the layers to squish down more, possibly leading to this?

Bruce

[printbus]

The fact that the Repetier moves appear to be correct suggest there's a problem accelerating/decelerating the Z axis. The Z numbers in your parameters seem conservative to reasonable for an i3 with M5 z-rods. Do you have EEPROM enabled? If so, maybe the EEPROM settings are more aggressive than what is in your configuration.h file.

[brucehvn]

The fact that the Repetier moves appear to be correct suggest there's a problem accelerating/decelerating the Z axis. The Z numbers in your parameters seem conservative to reasonable for an i3 with M5 z-rods. Do you have EEPROM enabled? If so, maybe the EEPROM settings are more aggressive than what is in your configuration.h file.

Hi printbus,

I am using the EEPROM, but the settings there are the same. I refreshed them after upgrading Marlin and verified in Repetier that they were correct. I use the EEPROM settings for temporary testing of stuff like e-steps, but once I like a setting, it moves to the Configuration.h as well.

Bruce

[printbus]

OK, good to know you understand how EEPROM fits into things.

Trying to reduce the value of 100 for Z in DEFAULT_MAX_ACCELERATION is the only idea I have. This is again based on your observation that a single manual move of 10 or 20 mm measures out right. A manual move like that would only involve one acceleration and deceleration on Z, while a print would require acceleration/deceleration on every layer height adjustment. You might not notice a few missing steps on a single 10 or 20mm move, but the error would add up if that same number of missing steps occurs every layer shift.

I think there have been a few reports of the M5 threaded rods and nuts on MakerFarm i3v printers wearing out after a lot of use, especially on printers with automatic bed compensation. I don't recall whether the symptoms involved print height issues, but it's something to consider.

If you haven't already, I'd pull the threaded rods out or otherwise unconnect them from the x-carriage and make sure the x-carriage slides up and down smoothly and consistently.

Good luck.

[brucehvn]

The problem turned out to be my z offset for the auto bed leveling. I had checked it and thought it was correct, but when I went back to level the bed manually, I found the offset was off by about .4 mm. Once I corrected that, then I now get the proper height on my objects. I thought it was crappy filament the way it seemed like the first layer would be squashed so flat. The skirt used to look like a flat ribbon. Now it's more like I remember it used to be.

But at least I didn't have any problems with warping or objects not sticking to the bed