First build teething problems: Z-axis out of sync

Hi there folks

First post and my first printer I have built. After a load of looking about I went for the Prusa i3. I have built it from scratch sourcing all the parts from across the planet so its not a kit. I have manged to get it all together and Im happy with the build rigidity and quality of the materials. Its a single Alu plate frame from a 6 mm sheet og aluminium. Im using an arduino mega paired with a RAMPS 1.4 board and stepper drivers. I have run all sorts of tests on the electronics before the NEMA 17 motors were placed into the printer.

I'm using Marlin as my firmware, with pronterface and Sclic3r x64-1-0-0-RC1 (when I used x64-0-9-10b I got all sorts of problems as was observed in another post, i changed Sclic3r version as was recommended in the post and it solved those issues). .

In Pronterface and can make the axes go in any direction and do what I want and all seems fine. No skipped steps or the like and the z axis is perfectly in sync.

When I undertake a run in Pronterface just as it starts the z axis motos go mad for a fractrion of a second and buz, they then stop and printer does some warm ups, then they buz again for a fraction of a second. I noted that the G-code Sclic3r was placing in had a stupid high rate for the Z axis. Upgrading to the latest Sclic3r and re-entering the settings for the printer seemed to sort that out.

Current Sclic3r settings
Perimeters 30mm/s
Small perimiters 30mm/s
External perimiters 70%
Infill 60mm/s
Solid infill 60 mm/s
Top solid infill 50 mm/s
Support material 60mm/s
Bridges 60mm/s
Gap fil 20mm/s
Travel for non-print moves 80mm/s
(note Im still doing calibration so stuck with these till I can sort out the step skipping issue)

My problem is this though, as the print runs the two Z axis motors are running out of sync now and then. Not all the time, just now and then. At first I thought one of the stepper drivers was overheating but all is fine when I touch it .. check with an IR temp gun and its not above 30. So I thought that my Ramps board may have a dry joint so I built a small board and placed both motors on the one Ramps board output. Thing is that they still go out of sync. It only seems to happen when its lifting to the next layer during a print, so was wondering,
- do you think that its an acceleration issue ?
- do you think that one of my motors are duff ? but it only happens when printing, when not and just moving about in pronterface they are all angels
- do you think that something is overheating ?

I did notce that my Y axis has now started to skip a step every few layers. Im getting a small fan delivered to actively cool the board. However until it arrives I thought I would ask about the strange Z azis issue I have there.

Any thoughts would be more than welcome.

-Me

Be sure that your max feed rate settings, max acceleration settings, and max jerk settings in the firmware all make sense. That's where you want to put your limits. The slicing software and the loading software (Slic3r and Pronterface) have no idea at all about acceleration and jerk. There are calculators around the web that will give you a pretty good idea of what is and isn't reasonable on acceleration.Max Z feeds around 2-4 and max accelerations in the 5-10 range make sense.

Hi there

Thank you for the pointer. I originally had my settings at

#define DEFAULT_MAX_FEEDRATE          {500, 500, 2, 45}    
#define DEFAULT_MAX_ACCELERATION      {9000,9000,100,10000}    
#define DEFAULT_ACCELERATION          3000   
#define DEFAULT_RETRACT_ACCELERATION  3000

which I canged to

#define DEFAULT_MAX_FEEDRATE          {500, 500, 2, 45}   
#define DEFAULT_MAX_ACCELERATION      {450,450,10,10000}   
#define DEFAULT_ACCELERATION          450    
#define DEFAULT_RETRACT_ACCELERATION  450

In addition I included a fan blowing on the drivers/RAMPS/Arduino board. The buzzing definitely has stopped and I can actually print I managed to print my first few things.Not pretty as there is some blobbing but its now going through to completion. I printed a second hollow calibration cube, bridging not as nice as the one I did with the first settings but I'm sure it's just all down to tweaking and calibration which I will now go into with the printer.

Im seem to be getting a "clunk" sound from one of the steppers. Either the X or the Y. Doesnt seem to be having an effect to the print but will play about with the numbers now that I have a setting that works. My drivers dont have any heatsinks so may invest in a set of memory heatsinks form just in case.

Thanks for the heads up. I had turned the Z speeds back but definitely not back far enough from what settings you proposed.

Lesson learnt just need to sus out this random clunk.

Cheers for the help, much appreciated.

-C

There's also a setting called "jerk" that's just below the feed rates and accelerations in the file. It's more or less "how fast can you turn a corner". Fiddling them might be worth doing. I have not messed with the settings that came with the kit.

I'd bet the clunk noise is something (like a set screw on a gear) loose. It could be a loose linear bearing. It might be a loose belt (do they go "twang" when you pluck them?).

Thanks for getting back to me. Sorry about the lag. some events of late came as a bolt from the blue and all took a back seat to that.

Time on my hands again and looking into the clunk problem. I have found that I get a shift on the X and the Y axis still. Its not often ... about once on each axis a print. Just enough to make me want to eBay the printer and seek an alternate hobby. Put a load of effort into the problem and still no resolution.

A dodgy thermistor caused my head to go up in smoke, got an E3D v5 print head. Solved almost all my oozing issues. Different league of head compared to my J-Head. I will see if I can get some pictures of my past failed prints and post them on here.

I have no heatsinkson the drivers but they are cold to the touch, Have a high CFM turbo fan on the drivers none the less. I have tried upping the current to the motors, they arenow about neutral to the touch rather than cool to the touch as they were before. The skips come at random. Its not after a heavy infill section or the like. Im going to go back to utter basics as I have not been able to print anything but calibration cubes from the thing.

Belts are all OK, nice and tight but not too tight. I could possibly do with pilling one more tooth of a tightening to my Y axis but its going so slow now (Marlin setting) that Id be surprised if it was that. Still at this point Im willing to give it a shot.

Unless you have very unusual motors, they should be hot (40C) to the touch after printing. I suspect you have your current set way to low. Dig out your motor spec sheets and see what they are rated for. Set the Vref on the stepper drivers to 1/2 of the max rated current on the motors. I suspect your skipping issues will go away.

Edited 1 time(s). Last edit at 12/28/2013 03:11PM by uncle_bob.

Will have a look into upping the pot as soon as I am able to locate a datasheet. I have sourced all the items myself (mostly from the orient via a certain auction site) so will contact the supplier for the sheet.

I followed one of the calibration guides that stated to connect the motors all up and to move the pot till the motors turn happily then add one quarter turn.

Will definitely have a look and report back my findings.

Thanks for the help

Ummm ..... errrrr... that's not a real good way to set the pots. They are not very accurate little beasts. You very much need to do it with a DVM. I've been tweaking this and that for > 50 years and I can't set them by hand.

Fully agree, though thats what the guide said. Finally got disheartened and decided to see what I could push through the steppers. Got the datasheet for the motors out and start checking on what I should be looking for using this guide.

The motors that I have are as follows :-

NEMA 17 Stepper Motor Model 17HS19-1684S - Specifications

    Frame size 42 x 42mm - NEMA 17
    1.8 degree step angle
    Length 47mm
    Holding torque 44 N.cm (4.4kg.cm, 62oz.in)
    Voltage 2.8V
    Current/Phase 1.68A
    Resistance/Phase 1.65ohm
    Inductance/Phase 2.8mH
    Inertia 68g.cm
    Bipolar, 4-wire
    Weight 0.35kg
    Step Angle Accuracy ±5%(full step,no load)
    Resistance Accuracy ±10%
    Inductance Accuracy ±20%
    Temperature Rise 80?Max.(rated current,2 phase on)
    Ambient Temperature -10?-+50?
    Insulation Resistance 100M?Min. 500VDC
    Dielectric Strength 500VAC for one minute
    Shaft Radial Play 0.02Max.(450 g-load) 
    Shaft Axial Play 0.08Max.(450 g-load)

So, in using the quick calculation to calculate the VREF I should take the Amp rating of the stepper (1.68A) and multiply it by 0.4. This would give me a voltage of 0.672V. Is my logic sound in this regard? Will admit that I placed all up to about 0.6V and the motors were at about body temperature. The drivers were still cool to the touch (small turbo style centrifugal fan is shifting a good bit of air across and under the drivers) though I still await delivery of the little heat sinks.

I have attached a picture of two of the portal companion cubes that I am using to print (small object that is easy to churn out and quick to show skipping steps). I have tried using the hollow calibration cubes but they usualy print without issue. One on the right hand side is at approximately 0.6V one on the left hand side was at 0.47V though please ignore the large shift at the top as that was me doing something daft. Thought that I had nailed it with the 0.47V cube on the left after the first skip. The it started skipping at the top.

I have had not too bad results with a 0.47V setting as was recommended by another that had the same motors as I do. One thing I have noted though is that (and I still have to verify this as I only noticed it at the end of the last print). Is that the skips would apear to coincide with when the heat bed reaches its temperature and the power is cut. I'm using a 203W modified Xbox power supply, though I have several DL360 Prolient servers that have recently been retired. I removed one of the redundant PSUs and it has a 37A 12V rail. I can get it to power up but as its not got much in the way of cooling, I will have to jerry rig something up for it. I note that the Xbox 203W PSU only provides about 16.5A on its 12V.

Was wondering if there was a power spike as the current cuts from the heat bed resulting in a fraction of a second overload on the rest of the circuitry.

The adventure continues . . .

If you have a heated bed on the printer, that supply isn't big enough.....

The voltage for a 1.68A motor would be 0.672V. If you wanted to drive the Z axis (2 motors) at full current, you would set it to ~1.3V.

Any setting over about 0.45V is likely to cause the stepper drivers to cut out from over temperature. Exactly what the trip point is depends on your pcboard design, the heatsink on the chip, the ambient temperature, and the amount of forced air.

A setting for 1/2 current should be about .34V. The motors should be hot to the touch after an hour of printing at that setting. If they aren't see if the windings are 1.65 ohms. You may not have the motors you think you do.