Odd layer shift problem in my "micro" delta

Interesting problem with my custom scratch-built makerbeam XL micro delta printer.

I have a custom "micro" delta printer that I designed as a portable table-top-toy-printer. 70mm diameter and 90mm height (yeah, LITTLE). I can print anything that I want as long as my print area is under 35mm on the bed. Above that and I get layer shifts at about 10mm up from the bed.

My print speeds are between 10 and 40 mm/s at various types of features (walls, infill, tops/bottoms, support, etc.) and movement at 120 mm/s (non-printing).
I have checked belts for gunk, pulley teeth for gunk, free motion (very little friction in the action)... The mechanical things.

I discovered that the problem greatly reduced, perhaps mostly eliminated if by reducing my non-printing moves to 40mm/s. Fiddling with other printing speeds yielded no great benefits. This makes me think that I am moving too fast with the non-printing move if there is a large movement to be made across the piece. As you can see, there is very little room under the bed of this printer, which only houses the NEMA14 steppers. I also can't get my fingers in there to see if the steppers are getting hot. I can make things up to 80 mm tall, if they are narrow, and I can make tall items in various places around the bed near the extreme print edges. I just can't print anything wide.

I am next going to put a fan under the printer blowing up through the mount plate to get to the space housing the steppers. If this works, I'll report it - if not, well, I'll keep experimenting.
[edit 1]
Well, the fan didn't work, still get the layer shift.
Next idea... Up the stepper current a little bit

I thought that I'd report on my progress in the spirit of "paying it forward" for anyone else that might be seeing odd stuff like this!

DLC

Edited 1 time(s). Last edit at 09/10/2019 12:57PM by dlc60.

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

It looks like increasing the current was the trick. This is no great surprise. What WAS a surprise was to see that I had the bias voltage on my 9855 stepsticks at .3V, which translates to about 550ma. I increased that to .39V, about 710ma. So far so good. I am glad that I put the fan blowing into the stepper cavity, the steppers might get a bit warmer...

So here is some anecdotal evidence that a weird mid-print layer shift can be caused by too low stepper current. Which shows up when you push the steppers too fast. In my case, when doing non-printing moves.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

No, it isn't the trick. It just moved the needle a bit.

Here are a couple of pictures of what happens. I do not understand what is going on with this and am wondering if there is something wrong with the design.



It looks like one layer does not get printed, and the print offsets a little bit and then continues on. This happens whenever a print is wider than some, as yet unknown, limit. If I print a tall narrower model, it does not happen. If I print a wide but shorter than about 10mm model, it doesn't happen.
I have raised my stepper current about 50%, it seemed to extend the diameter of the print before the problem happens. Also, slowing down the print tends to move the error up a little bit in the Z axis.

As I mentioned, this is a custom design and I am wondering if I did something incorrectly. The build volume is about 60mm x 60mm by 90mm high. My rods are about 105mm long.

I am baffled, so any thoughts that anyone has, I am open to suggestions.

Thanks,
DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

I found the problem.
Any guesses?
The heater block was loose.

When cold, the block did not move, when hot, you don't reach in there to move it... But after one frustrating failure, while the hot end was still hot I used my brass brush to clean it off and the heater moved.
Duh.
What caused this was was interesting. In general, I secure wires tightly to keep them from moving, especially on a delta where that end effector moves a LOT and often very quickly. I had these wires nicely laid in the gap between two arms, out of the way of any moving part. The "bad" part of this was that there was a torque applied by the wires against the current position of the heater block such that over time it gradually forced the block to unscrew.

The fix is the rotate the hot end in the end effector so that the wires no longer applied pressure in the "unscrew" direction.

Sheesh.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.