Right Z axis lead screw slipping

[EDIT: Ignore this message - have really tightened coupling hard, seems to be OK for now]

My M90 is operational. (Everything worked first time!)

I'm trying to level the bed along the X axis.

Levelling the carriage by twizzling the right Z coupling results in the coupling slipping round the lead screw.
(Twizzle,twizzle,twizzle, CLONK!)

I degreased the lead screw thoroughly before assembly. The coupling is tight (no gap between the shells) on the lead screw, ~2mm open on the motor tubing.

I want to put some heatshrink on the lead screw to improve the grip. Is this a good idea?

Edited 1 time(s). Last edit at 06/17/2013 08:52AM by threedyprinter.

I don't understand. When you turn the coupling the torque is between the motor shaft and the coupling. The screw would be no more likely to slip than when the motor is turning it. Are you sure it is not the motor shaft slipping? In which case tighten the bottom screws until the gap is about 1mm. Make sure the flat on the motor shaft is facing towards the gap.

I don't think adding heatshrink to the lead screw will cause a problem, I think somebody else mentioned doing that but I don't see why it would be needed.

---

[www.hydraraptor.blogspot.com]

You're correct, it doesn't matter whether the coupler is turned manually or by the motor.

In fact, I have just printed the android test piece and, as it was finishing and winding the carriage up to the top, both left and right Z couplings started clonking. Each clonk was accompanied by the carriage dropping by a mm or so on the side that clonked. When it reached the top, the carriage was visibly lower on the right.

The flats on the motor shaft are indeed facing the gap; the motor side of the coupling doesn't appear to have slipped at all.

I suspect you must have the coupler upside down if it doesn't grip the threaded rod. The smaller end is for the rod.

---

[www.hydraraptor.blogspot.com]

The couplers are the right way up. BUT: some oil had seeped down (I oiled the lead screws prior to testing as per the manual) and must have been a bit over-zealous. Anyway, I degreased, fitted heatshrink and tried again. Perfection. Not a hint of slipping.

I'm not quite out of the woods yet, the first few layers get mushed a bit (partly because the bed is bowed in the middle by .1mm) and I haven't got the Z Home point entirely sorted. (I've just been shimming the Z axis end-stop rather than faffing with Z_HOME_POS). That apart, the second test android has come out really well.

Ha, I would regard shimming the endstop as faffing and changing one line of code as not.

Which way is the bed bowed and how have you measured it? There are several ways it can appear to be bowed when actually it is flat.

---

[www.hydraraptor.blogspot.com]

> Ha, I would regard shimming the endstop as faffing and changing one line of code as not.

Harrumph, I would say cutting up a bit of plastic as opposed to fixing up a development environment when your dinner is on the table is pragmatic!

>>Which way is the bed bowed and how have you measured it?
I used a dial gauge; removed extruder assy. and tie wrapped it in place. It sits snugly in the square extruder hole. (This is a faff! It's only got a short probe so has to sit underneath the carriage and you have to bend double to see the wretched needle.)

The bed appears concave in the middle; not hugely - about .03 mm max. I was surprised that, even after fixing the Z lead screws, the initial outline on the test print was good on the left but smeared on the right. Obviously some more levelling is required.

I'm going to fix the software first, then level again more carefully.

Other things I noticed:
1. Every so often the extruder seems to extrude a little too much plastic (using the Faberdashery filament from the kit). My android man gets a few curls in his hair.
2. In the manual, the bed clips are illustrated at the sides of the glass. My clips foul the gantry if I place them there so I moved them to front and back. Now the fan casing fouls them so I've removed it temporarily.
3. Printing from Pronterface works perfectly with baud rate @ 250000 with Mint 14. Problems only arise if the /dev/ttyUSB0 connection goes down (I switched on a "noisy" desk lamp and crashed the Melzi). udev then brings the connection up again as, say, /dev/ttyUSB1. I suspect the FTDI has not been programmed to supply unique "Melzi" vendor / id to use in a udev rule.

threedyprinter Wrote:
-------------------------------------------------------
> > Ha, I would regard shimming the endstop as
> faffing and changing one line of code as not.
>
> Harrumph, I would say cutting up a bit of plastic
> as opposed to fixing up a development environment
> when your dinner is on the table is pragmatic!

Maybe if you are in Linux. On Windows you can just click on the arduino exe on the SD card and it just works.

>
> The bed appears concave in the middle; not hugely
> - about .03 mm max. I was surprised that, even
> after fixing the Z lead screws, the initial
> outline on the test print was good on the left but
> smeared on the right. Obviously some more
> levelling is required.

The spring on the dial gauge bends the glass slightly in the middle making it appear dish shaped when it isn't, see [forums.reprap.org]

> 1. Every so often the extruder seems to extrude a
> little too much plastic (using the Faberdashery
> filament from the kit). My android man gets a few
> curls in his hair.

Is this printing from SD as that is what happens if the PC inserts a delay before the retract?

> 2. In the manual, the bed clips are illustrated at
> the sides of the glass. My clips foul the gantry

Yes sorry about that. We got some clips that have handles 2mm longer than we had previously. I assumed they would all be the same for a 19mm clip.

> if I place them there so I moved them to front and
> back. Now the fan casing fouls them so I've
> removed it temporarily.

You may have to remove the top handles when using the fan duct. Some sit flatter than others.

> 3. Printing from Pronterface works perfectly with
> baud rate @ 250000 with Mint 14. Problems only
> arise if the /dev/ttyUSB0 connection goes down (I
> switched on a "noisy" desk lamp and crashed the
> Melzi). udev then brings the connection up again
> as, say, /dev/ttyUSB1. I suspect the FTDI has not
> been programmed to supply unique "Melzi" vendor /
> id to use in a udev rule.

The vendor ID and product ID will be unique to FTDI but should be the same across all instances of the same chip. They have a unique serial number so they get assigned a unique fixed COM port number under Windows. If they move around on Linux it must be a driver bug.

Edited 1 time(s). Last edit at 06/18/2013 09:13AM by nophead.

---

[www.hydraraptor.blogspot.com]

>Maybe if you are in Linux. On Windows you can just click on the arduino exe on the SD card and it just works.
In the end that's pretty much the case for Linux.

# cd path/to/Mendel90/Windows/arduino-1.0.1
# arduino (or path/to/wherever/arduino/is/installed/arduino)
open path/to/Mendel90/Marlin/Marlin/Marlin.ino

Must confess I was a bit amazed when the whole thing worked (including uploading) first time.

>The spring on the dial gauge bends the glass slightly in the middle...
Good point, I did set the gauge height to use just the very end of the probe travel (min. force) though.

>curls - Is this printing from SD as that is what happens if the PC inserts a delay before the retract?
It was printing from the PC, so I tried printing from SD. The result was worse, but see * below.

>FTDI ... vendor id ... If they move around on Linux it must be a driver bug.
No, kernel USB drivers are good now, it's just the way Linux is intended to work. That unique number (A601U2Z2 on my FTDI) has to be written into a rule for udev to say "create (e.g.) /dev/ttyMendel90 when the kernel detects a FTDI chip with that serial". Any road up, it's not important.

* I didn't think the PC or electronics were a problem so: I resliced the android using slic3r (which I think you don't like, but can't remember why). Changed nozzle temp to 210, altered filament width to 2.88mm, extrusion multiplier to 0.9. Result: not a curl in sight, really getting impressive now. Only snag is that android man has solid antennae (without the groove in them) but that seems is what slic3r intended. It builds the outline nicely then deliberately draws a line on top of the groove. Under investigation.

En passant: My resliced android has the default honeycomb fill which drove the M90 into a frenzy (but it coped well). The X and Y belts were bouncing around much more than usual so I think a bit of tightening is required. It occurred to me that there should be some sort of 3D equivalent of the old BBC test card to help get things tuned up. Maybe there is?

Thanks for your help Chris. I'm nearly home.