i3 bed shifting

I have a problem with my i3 variant (DIYTechshop i3 Alpha) recently. It seems the bed is fluctuating in the vertical direction quite often and I can't seem to figure out exactly why. I'm constantly levelling the bed and it keeps shifting on me and it's driving me crazy. The machine has been up for about a month now. I spent a lot of time printing calibration objects and tuning the machine and I've gotten some wonderful prints from it. I've been trying to print a new Greg's Wade extruder so I can start using my E3D hotend and put in an auto bed levelling bracket. I've printed all the extruder parts and they've all come out great, but when I went to print the extruder body on Sunday (after many successful prints on Saturday), I got a mess. It was like the nozzle was dragging on each layer too much. If you look at this photo, the ones one the left are 2 of the 3 Sunday false starts. The one on the right is a test object I printed last week and was fine with the quality. The test object is just an extruder body that I cut off in openscad above about 16 layers or so. The lines in the bad prints are not ooze strings. The filament is not sticking as it goes around curved sections and ends up just stringing the filament across the opening.

At first, I thought it was just that every time I removed the mirror plate to add a new coating of hair spray and replace it on the heated bed, the levelling would be thrown off and I would have to re-level it. But now I see it happening without doing anything with the plate at all.

I checked the bed levelling several times and each time it would be off a lot. It seems to be always the whole bed, not just one or two corners. Last night I did a test where I had the bed levelled, printed 3 small objects, then let it cool to room temp again and checked the bed. It had moved up by about .4mm all around. I'm just constantly chasing it around up and down.

When I built the machine originally, it had springs and wing nuts for connecting/supporting the heated bed and levelling the bed. Double m3 nuts under the heated bed plate hold it against the head of the m3 cap bolt. The wing nuts were not good and actually slipped sometimes, so I replaced them with regular m3 nuts and was doing my levelling that way. When I really started seeing this problem a few days ago, I went and got some much stronger springs and some m3 nylon lock nuts and put those in place. When the problem still happened, then I took out the springs entirely and now have the heated bed supported by an m3 nut on the top of the aluminium plate and the nylon lock nut on the bottom. It makes the bed levelling a real chore now, but it didn't solve the problem.

I use a .1mm feeler gauge for levelling, not a piece of paper (though I'm impressed at how accurate a piece of paper can be).

Here's what I've checked so far:

The z endstop switch appears totally solid. Checked it with a multimeter and it's either no resistance or infinite resistance with a firm click of the switch, no ambiguity at all.

The z endstop holder appears to be fine as well. I ran a test script in Repetier where I just ran the z-axis up various amounts (not moving x or y), then homed it again several times, and the distance to the bed was still correct afterwards.

The frame is tight and square, no loose nuts or anything.

The extruder and x-axis appear very solid.

The z-axis also appears very stable, no detectable slipping or skipping and it runs smoothly.

It seems to me like it's got to be something with the y carriage. The smooth rods are solid and have no play. However, I can grab the front of the build plate and moving it up and down gently there is some play there. Part of it appears to be some slight play in the LM8UU bearings and some seems to maybe be the plastic bearing holders/mounts have slits that seem to expand a bit when moving the bed up and down. Honestly, I'm not sure if that play has been there the whole time, but I do know I didn't seem to have any problem with this at first. When I built the machine I adjusted the plastic bearing holders/mounts to insure that the y carriage had smooth motion from front to back. Even with the play I see in the bed, it seems like the weight of the bed and glass should keep it from being able to move up and down under normal conditions.

Any help would be appreciated as I don't know where to go from here. I'm spending way too many hours constantly adjusting the bed and I don't think it's supposed to be that way.

Bruce

to me it looks a combination of things speed and extruding too much material, make sure your extruding the correct length of material, speed i started slow at external and internal at parameters set at 28.5 mm/s but i have wacked my feed rate up by 100% i know i can do 57 mm/s but check that your extruding the correct length and slow it down if its fast

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

Quote
chris33
to me it looks a combination of things speed and extruding too much material, make sure your extruding the correct length of material, speed i started slow at external and internal at parameters set at 28.5 mm/s but i have wacked my feed rate up by 100% i know i can do 57 mm/s but check that your extruding the correct length and slow it down if its fast

Thanks Chris,

I double checked the e-steps calibration the other night. They seem to be fine. As I said, I've been printing for about a month and none of the calibration objects showed any sign of over extrusion. From what I understand, if the bed is too close to the nozzle, it can exhibit some of the same symptoms of over extrusion.

Bruce

If you have a heated bed you need to level it when it's at temperature and check level when it's at temperature.

It's no good doing it when it's cold heat makes things expand.

Ben

Quote
plexer
If you have a heated bed you need to level it when it's at temperature and check level when it's at temperature.

It's no good doing it when it's cold heat makes things expand.

Ben

Thanks Ben,

I could do that, but every calibration page I've looked at including Triffid Hunter's page as well as the main calibration page at reprap.org say to do the calibration at room temperature. You use a piece of paper or a .1mm feeler gauge to adjust your z=0. The .1mm will be taken up by thermal expansion when the extruder heats up. Not to mention is seems calibrating hot is a recipe for getting burned a few times. And as I mentioned before, this is a recent problem.

Bruce

plexers right you need to level the bed with it heated i do it at 70 oc and with a piece of paper but i do have asbestos hands lol i done it cold but now i do it on the fly when its doing the outer brim

Edited 1 time(s). Last edit at 05/24/2014 02:44PM by chris33.

---

Check my rubbish blog for my prusa i3

up and running
[3dimetech.blogspot.co.uk]

Quote
chris33
plexers right you need to level the bed with it heated i do it at 70 oc and with a piece of paper but i do have asbestos hands lol i done it cold but now i do it on the fly when its doing the outer brim

Ok, I can try with the bed heated, but whatever I was doing cold I was doing consistently. So, if before I was getting great prints and now each time I print it's a crap shoot, it seems like something is off somewhere. For example, the night before last, I printed a calibration solid cube to do some temperature testing. That object printed just fine. I turned on the printer last night, homed all the axes and checked the clearance. The nozzle was too close to the bed and I couldn't even get the feeler gauge in there. I used my new adjustable z endstop holder to get the clearance right and all 4 corners fell into place. I then went to try and print the extruder body once again and aborted immediately when I saw the skirt was coming out completely wavy. I let it cool down, checked the clearance and the nozzle was too close again. This time I actually had to move the z endstop holder as I didn't have enough adjustment to get it back in line. When I see this happen, I grab the front of the bed, move it around up and down, side to side, but there's no obvious shift that I can make happen. I'm so frustrated because this is what I do now, just constantly chasing the bed up and down trying to get a good print.

Bruce

Check for any binds in the z-axis what kind of motor couplers do you have if you have just the tubing it's probably getting loose buy aluminum couplers

Quote
cnc dick
Check for any binds in the z-axis what kind of motor couplers do you have if you have just the tubing it's probably getting loose buy aluminum couplers

I was using the plastic tubing which came with the kit. I had purchased the aluminium couplers, but hadn't gotten around to putting them on. I put them on about 3 days ago. I put a 4.5mm BB in between the motor shaft and the threaded rod, but I don't think they are actually close enough to touch the ball. Now I need to adjust my max acceleration in the firmware I think because when slic3r adds the line G1 Zxxx F7800, the z axis whines a lot, so I'm going to try and get that to a better value. I don't really see it slipping though. I keep checking the smooth rods on the x-axis and they are still level.

I do notice that when I home the z-axis, then run scripts I setup in repetier that move to the various corners for levelling that end up moving z to 0 (not homing), sometimes the endstop clicks, sometimes it doesn't, most of the time it doesn't. The scripts exercise the bed a bit. For going to any corner, they first raise z to 10.0, move x to 100, move y to 100, then move x and y to the desired corner and move z to 0. The scripts use a speed of F150 for z moves which doesn't cause any abnormal noises from the z axis. Usually the gap on the first corner where I homed the axes still seems ok though.

Bruce

One other thing to check is make sure the z-axis is very free it has to lower itself by gravity alone if you take the threaded rods out. The reason for this is there is always play in the nuts and screws you want gravity pull it down the on the screws and nuts if there's the slightest bind holding it up would be to the top of the play in the nuts one time and then the bottom of the play in the nuts next

Quote
cnc dick
One other thing to check is make sure the z-axis is very free it has to lower itself by gravity alone if you take the threaded rods out. The reason for this is there is always play in the nuts and screws you want gravity pull it down the on the screws and nuts if there's the slightest bind holding it up would be to the top of the play in the nuts one time and then the bottom of the play in the nuts next

I will definitely check that. I think when I put the machine together, I immediately threaded the rods into the nuts and never checked the z-axis for smooth movement all the way up and down other than adjusting the stepper drivers to make sure it ran from top to bottom.

Bruce

On the couplers you have to find something that will fit in there loosely so the screw and the motor shaft touch otherwise coupler is going to act like a spring. Can be a ball bearing and something else could be two ball bearings

Edited 1 time(s). Last edit at 05/25/2014 06:46AM by cnc dick.

Quote
cnc dick
On the couplers you have to find something that will fit in there loosely so the screw and the motor shaft touch otherwise coupler is going to act like a spring. Can be a ball bearing and something else could be two ball bearings

I have two BBs in there now (.177 caliber, 4.5mm), one on each side. But I don't think the threaded rods are actually touching them because I think I have them tightened down just maybe a quarter inch into the top of the coupler. Should they be touching with the ball in between?

I disconnected the threaded rods and played with raising the x-axis up and down. At first it didn't seem so smooth. The extruder was in the home position on the left and that's the side the x motor is on also so it was quite unbalanced. I moved the extruder over to the right and things moved more smoothly. After exercising that up and down for awhile, I put the extruder back on the left and it seemed more smooth than before. I put it back together and was able to print my extruder body. It might be somewhat usable, but the base did warp (nothing to do with this problem).

So, if the movement of the z-axis up and down is not as smooth as we'd like, then what can be done? I don't think it's appropriate to put any oil or grease on the smooth rods as that would probably just gunk up the LM8UU bearings, right? The machine has only been printing a little over a month so I doubt the bearings are worn out.

Bruce

Yes motor shaft and threaded Rod should be touching inside couple which allows side to side movement but not down. If you don't have this to push on the X axis you'll see the couplers compress like a spring not a good idea

Quote
cnc dick
Yes motor shaft and threaded Rod should be touching inside couple which allows side to side movement but not down. If you don't have this to push on the X axis you'll see the couplers compress like a spring not a good idea

Ok, I fixed this. Still noticed that every day I still need to re-level the bed as it's always off. I grabbed both ends of the x axis and tried to move it up and down and I saw no play.

I think I may have a clue as to what might be happening. I think that after printing for the day, a small glob of filament is formed on the nozzle. Then when I go to check it the next day, the nozzle appears to be too close as it touches the bed. I do adjustments and then all of a sudden, it goes the other way. I believe that as I'm trying to do my adjustments, the feeler gauge may end up cleaning off the nozzle which then makes the nozzle too far from the bed.

Going on that assumption, I'm going to try making sure the nozzle is clean every day, then do a quick check to see how things sit.

Bruce