Hello, reprappers,

I built a custom H-Bot, solid steel frame, 300mm x3 build volume, heated bed, bowden extruder, print cooling fan, run by Marlin-1.1.0-RC6 on RAMPS 1.4.
I've been having some really strange issues with it, when I print a large volume on the XY axis, like a 40mm cube, it gets a layer misalignment problem. This doesn't happen with smaller objects, or even less of the same object. I've attached an image to illustrate the problem.

I'm printing some clamps with PLA, and if I print 2 at once, no layer misalignment of any sort. However, if I print 3, then it fails as early as layer 2-6 in the print. I haven't changed any settings between the prints, and I've tried this many times. I didn't have issues with large objects earlier, this just cropped up one day, and hasn't left since. I've also tried the following, among other attempts to fix.

• I've swapped the stepper drivers
• Tuned Vref for max torque
• Tried running it slower
• tightened and checked all screws and pulleys
• all electronics are fan cooled, stepper motors are below body temp
• tightened and adjusted belt
• no friction on any axis

There's no sign of any nozzle collision or belt slipping, it just starts suddenly printing it with an offset, and it's the same direction every time. This issue's has been bothering me for over 2 weeks now, so any advice or suggestion is appreciated.

Thanks for your time,

when this happens to me it is always a nozzle hitting on the print makes it lose steps...! do you watch the print from the beggining to the end ?

---

Delta Printer
Duet 0.8.5 firmware 1.19

The most significant difference between two or three clamps is the longer travel distance with three clamps.
You could try to reduce travel speed and accelerations in general.

Quote
Gaou
when this happens to me it is always a nozzle hitting on the print makes it lose steps...! do you watch the print from the beggining to the end ?

I've carefully watched my printer, and it doesn't seem to hit anything or make any sounds to indicate jamming, impact or contact with obstruction. It just starts suddenly printing it offset from the previous layer.

Quote
o_lampe
The most significant difference between two or three clamps is the longer travel distance with three clamps.
You could try to reduce travel speed and accelerations in general.

I've already tried reducing the speed and acceleration. Speed is currently ~30mm/s for printing, and 60mm/s for traveling. The acceleration is 20mm/s2. I've printed a much larger and longer object before, around 16cm x 5cm x 20 cm tall, and I didn't get this issue. Now, even if I print a 40mm calibration cube it gets this misalignment between layers.

The misalignment seems to be in perfect 45° direction?!
If a H-Bot has same kinematics as a CoreXY, this indicates one of the steppers competely fails. I'd check the hardware and wiring for cold solder joints or broken wires.

Edited 1 time(s). Last edit at 03/26/2017 03:11AM by o_lampe.

Quote
o_lampe
The misalignment seems to be in perfect 45° direction?!
If a H-Bot has same kinematics as a CoreXY, this indicates one of the steppers competely fails. I'd check the hardware and wiring for cold solder joints or broken wires.

I'm not sure if that's the case, but I don't think I can rule it out entirely either. I've had prints fail early as the second layer, to as late as 20 min in during a 40mm hollow cube. Printing two clamps is always successful, and takes 40 min, so I'm not sure it's related to time or just chance. I'll double check the solder joints regardless though, just to be safe.

Quote
o_lampe
The misalignment seems to be in perfect 45° direction?!
If a H-Bot has same kinematics as a CoreXY, this indicates one of the steppers competely fails. I'd check the hardware and wiring for cold solder joints or broken wires.

Quote
A_Designs
I'm not sure if that's the case, but I don't think I can rule it out entirely either. I've had prints fail early as the second layer, to as late as 20 min in during a 40mm hollow cube. Printing two clamps is always successful, and takes 40 min, so I'm not sure it's related to time or just chance. I'll double check the solder joints regardless though, just to be safe.

Ok, I opened up the XY steppers and there doesn't seem to be anything visually wrong in there. I also realigned the entire Z axis (I had some new mounts and proper leadscrews to install). I have a little issue with juddering on the Z axis, but nothing concerning.

The XY axis still has the same error as before. I printed another 40mm calibration cube, deviation is in the same direction again. It seems to deviate a lot initially, then reduce and stabilize somewhat. That's the confusing part, if it were a loose stepper, then it would be random throughout. If it were poor hardware, it would be failing at specific areas. I'm rather stumped at this point.

Anything else I can try or any other ideas?

Have you tried a print with 'avoid crossing perimeters' set or maybe with 'Z hop' set. This would remove the possibility of the nozzle snagging on the centre piece.

Sorry if this is a stupid idea

Quote
obewan
Have you tried a print with 'avoid crossing perimeters' set or maybe with 'Z hop' set. This would remove the possibility of the nozzle snagging on the centre piece.

Sorry if this is a stupid idea

At this point, I'm desperate enough to try anything, this issue has been on my head for a month now. Earlier, it didn't make any sounds or movements to indicate nozzle catches, so I dismissed it fairly early on. Just now, I enabled Z Hop in cura, and it hops perfectly with my new leadscrews. On another note, I'm using a SD card loaded with .gcode, and a discount LCD controller.

The misalignment problem is still there though. Here are some photos to illustrate the issue better.

Today, I printed two 40mm filleted hollow calibration cubes, and the result was this-

It looks ok for the first few layers, then it misalignes, and gets misaligned some more, then stabilizes and continues building while still maligning every now and then.
I enabled Z hop, and then this happened-

Someone exorcise my printer please.
The XY axes' motion is slicker than an oiled baby down a waterslide, there is low friction, no jitters or vibrations, no slipped belt, no skipped steps, nothing to indicate anything out of the ordinary, and the print head moves back and forth like it's business as usual. However, the lines end up weirdly like that.
Without powering down my printer, I just rotated the glass and selected the same object to print again, and this happened.

It got worse...
It started getting misaligned from the second perimeter wall line itself, you can see how it started moving away. Then each consecutive line started moving further and further away downwards in the image, and then it went up and started printing somewhat weirdly in the other half of the square, like nothing even happened earlier...

Now what...
I opened up my electronics box, glared at every solder joint, nothing seems particularly out of place or burnt, so I'm still stumped. Anything else to check, short of replacing the entire electronics package?

Edited 1 time(s). Last edit at 03/30/2017 09:26AM by A_Designs.

... still looks like loosing steps - have you tried to rise the motor current?

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Quote
VDX
... still looks like loosing steps - have you tried to rise the motor current?

The current's at its max, and I've used new stepper motor drivers just in case, but still the same issue. I'm microstepping at 1/16 with A4988 drivers. It also doesn't make any sounds like missed steps or impacts when the deviation shows up.

... there were similar effects with slipping gears, not enough fixed on the motor shafts ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Quote
VDX
... there were similar effects with slipping gears, not enough fixed on the motor shafts ...

My stepper's have standard D shafts, and the pulleys are locked on tight. They're definitely not the problem, and I've put marks on the pulley to check if there was any slipping, and it's rock solid. I thought it might be the endstops falsely triggering, but I get no messages about that, and I tried triggering the endstops with my fingers during a print, and my printer kept going as if it didn't care, so it's not that either.

Im still going with current needs to be higher. I have to have mine REAL high and the motors burning hot otherwise ill miss steps.

Quote
Floyd
Im still going with current needs to be higher. I have to have mine REAL high and the motors burning hot otherwise ill miss steps.

Missing steps creates a audible sound, so I would have noticed it. I tried tuning my drivers with vref, but that turned out to be way too low, so I tuned it until I didn't get any more torque when increasing, and only more noise.
Besides, it was working earlier with the same settings, so I don't see why it would suddenly kaput itself.

Alright, I removed the belt and checked the axes again, friction is still negligible. I checked the [reprap.org] page, and crossed every item off the list. The only remaining possibility is that my controller's gotten somehow messed up, either the arduino or the ramps board. I conducted a visual inspection of both boards, and I didn't see anything burnt, poorly soldered, or otherwise damaged.
Is there a chance that somehow something got damaged in either of these boards? and would there be a way to check for that rather than buying a new pair?

Just in case, I also updated my marlin firmware from RC6 to RC8. A little head scratching and bug stomping later, I got my printer working, but again, the issue persisted. I'll try to swap out either the arduino or ramps boards with another and see if that fixes it.

It looks like both X and Y axes are affected and at the same time--as if the Arduino was overloaded, the RAMPS board couldn't produce enough power, or the SD communication was losing packets.

Why the system would add (or subtract) to both X and Y though, seems odd.

Maybe try to reduce the processing load on the Arduino? Check the power that's available throughout the print? And check the buffer sizes?

Quote
kentmcn
It looks like both X and Y axes are affected and at the same time--as if the Arduino was overloaded, the RAMPS board couldn't produce enough power, or the SD communication was losing packets.

Why the system would add (or subtract) to both X and Y though, seems odd.

Maybe try to reduce the processing load on the Arduino? Check the power that's available throughout the print? And check the buffer sizes?

Ok, how should I check if that's the case? I've tried printing at really slow speeds, and I've set my SD card speed to halfrate just in case.

How do I check the processing load and buffer on the mega? If you're referring to power as in power supply, then I've got a 12v 50a PSU hooked up to a computer UPS, which is hooked up to my building's UPS, so it should be rock solid.

Quote
kentmcn

Why the system would add (or subtract) to both X and Y though, seems odd.

Because it is a H-bot and one motor moves at 45/135 degrees?

You said the motor is getting hot. That means the current is high, which means the driver module is probably overheating and shutting itself down to protect itself. After a brief interval it cools enough to turn on again and then promptly overheats again, missing steps along the way. That would explain why it's OK for a little while and then starts failing. Those little driver modules are just a crappy design. They do not have sufficient surface area to cool the chips, and taping a tiny little heatsink onto the chip doesn't work. Turn down the current.

I am running a pretty massive coreXY mechanism with two little 64 oz-in motors at 1000mm/s^2 acceleration and up to 150 mm/sec print speed and they literally don't get more than 5C above ambient temperature. If you need a lot of current to move the mechanism the mechanical alignment of the rails and bearings is probably off. It may have been fine before you put the belts on but the belt tension puts a lot of force on the mechanical stuff and may have pushed it out of alignment enough to cause binding or at least a lot of resistance to motion.

[vimeo.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
sigxcpu

Quote
kentmcn

Why the system would add (or subtract) to both X and Y though, seems odd.

Because it is a H-bot and one motor moves at 45/135 degrees?

HBots and coreXY's are a pain to troubleshoot, because if only one motor is turned, then the head moves diagonally, and if the other is turned, then it goes diagonally perpendicular to the first one. If one is turning, and the other skips, jumps, or stutters, then it'll cause the head to get misaligned, and depending on which direction the head was Supposed to be moving in, it gets misaligned one way or the other. I tried recording it, but it's such a small deviation, and it happens so quickly, that you can't really notice it's in the wrong spot until it's printing in the air.

Quote
the_digital_dentist
You said the motor is getting hot. That means the current is high, which means the driver module is probably overheating and shutting itself down to protect itself. After a brief interval it cools enough to turn on again and then promptly overheats again, missing steps along the way. That would explain why it's OK for a little while and then starts failing. Those little driver modules are just a crappy design. They do not have sufficient surface area to cool the chips, and taping a tiny little heatsink onto the chip doesn't work. Turn down the current.

I am running a pretty massive coreXY mechanism with two little 64 oz-in motors at 1000mm/s^2 acceleration and up to 150 mm/sec print speed and they literally don't get more than 5C above ambient temperature. If you need a lot of current to move the mechanism the mechanical alignment of the rails and bearings is probably off. It may have been fine before you put the belts on but the belt tension puts a lot of force on the mechanical stuff and may have pushed it out of alignment enough to cause binding or at least a lot of resistance to motion.

[vimeo.com]

I checked the thread again, and I don't think I've mentioned my steppers getting warm. They're pretty cool, just a bit above ambient temperature, and below body temperature. I'm pretty sure that the current setting isn't the fault here, considering that I've set it correctly and tuned it, and my setup was working earlier, until one fine day when it decided to start printing with a misalignment. My drivers are the same temp too, and they're actively cooled by an 80mm case fan pointed diagonally at them to blow air under the ramps board and over it.
I don't even remember if I had made any changes to my setup around the time it started misaligning layers. I've also mentioned that the mechanism is smooth with the belt removed, and even with the belt on, I only need to apply a little more force on the stepper's pulleys than you'd exert when cleaning your ear with a Q-Tip. (I possibly need a better metric than that, but that's as universal as I can think of).

I got a spare RAMPs board from a friend, so I'll try that out and see if that alleviates the issue.

Oh, it was someone else's comment about motors running hot...

Check the motor cables from the controller all the way to the motors. Pay attention to the connectors at both ends- those things are also fragile and easily damaged. An intermittent wire/connector will damage the motor driver just as pulling the motor cable from the controller board will when the motors are energized.

If the cables look OK, try swapping in some fresh motor driver modules. Those things are easily damaged, and they don't always fail completely.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Day 40, hope is starting to dwindle, frustration building. I borrowed a RAMPS 1.4 and mega from a friend, and installed marlin RC8 onto it, and hooked it up to my printer. It started up, began printing, and got misaligned as usual. Same error again, even with another board. Now I'm really getting stumped on what's the issue. I think I'll try swapping two stepper motors and seeing if that somehow fixes the issue.

We got the same shiftinglayer issue on our six mendel90:s at my work. They all shift. Very unpredictable.
After fiddling with steppers, (voltage and cooling) , belt tension and so on i found that the follwing seem to give the most promising results:

In Cura i activate enable accelleration Control. I got the two settings lowered to about 900 and 500. No shifting on the last eight prints.
I will report here again after more testing prints.

Right, I realized the Y axis rods weren't perfectly aligned, so I went and got some supports lasercut out of MDF. I realigned the frame, the Y rods, and the stepper motor & idler mounts, and finally, after 2 months or trying to fix my printer by procedurally checking every possible thing short of a complete stripdown,

The misalignment is now at regular steps. Pretty much perfectly, there's the exact same deviation with each new layer, it almost looks intentional.
Side view=


Is there a "misalign layer by x and y" feature in cura or marlin? The initial ten or so layers are fine, then it starts deviating it a bit each time...

What's left for me to try out to fix my printer?
Maybe I could consider adding a rotary encoder to give me positional feedback on the idler pulley, and therefore make this a closed loop control system?

Alright, I just did two test prints, 3 cubes in a line, one print oriented to the X axis, and the other oriented towards the Y. Both prints got misaligned, and although the X drifted earlier in the print, I haven't repeated this test multiple times, so the results seem inconclusive. However, I noticed that sometimes, the print would get misaligned during the skirt itself. The initial loop would come out fine, but the second loop would drift a few mm, so it might not be a drift every layer, but possibly more frequent in nature.

Since the layers print OK, what happens if you reduce the move speed to the same as the print speed? That should rule out speed related issues.
I don't know much about H-Bots but I assume motor synchronisation is critical.
Assuming the firmware is configured correctly then differences in the drivers may have an effect. Just speculating but wire size and length differences?

In a corexy mechanism diagonal lines are produced when only one motor is turning. I don't know if h-bot works the same way. In a corexy machine that torque-slips on the 45/135 degree lines, I'd turn down the acceleration until it stops slipping.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
MCcarman
Since the layers print OK, what happens if you reduce the move speed to the same as the print speed? That should rule out speed related issues.
I don't know much about H-Bots but I assume motor synchronisation is critical.
Assuming the firmware is configured correctly then differences in the drivers may have an effect. Just speculating but wire size and length differences?

I'm currently going at 20-30mm/s, and all print and travel moves on the first layer is at 20mm/s, but sometimes even the two loops of the skirt gets misaligned, with the second loop exhibiting the same drift as seen above in this thread. The motors being in sync is indeed critical, which is why I bought all 5 motors used in this build from the same vendor at the same time, the XY cables also have the same length and same drivers, set to the same settings, so short of electromagnetic induction, I don't think that's the case.
I was wondering if the fact that the X and Y carriages have a different number of bearings would affect the build, but I find that unlikely, given how freely it moves either way, and how my printer was working ok up to 8 weeks ago.