Weird Height Variations!

Discussion in 'Calibration, Help, and Troubleshooting' started by Jasons_BigBox, Aug 21, 2016.

  1. Jasons_BigBox

    Jasons_BigBox Well-Known Member

    Joined:
    Mar 25, 2016
    Messages:
    248
    Likes Received:
    71
    Afternoon all,

    I was involved in Jet's post "Inconsistent Printed Height Variations" as I thought I was having similar issues but I'm beginning to believe that's not the case. I printed a model tank earlier in the week (1:100 scale Sherman Firefly for those who like them) but the supports that were needed to build the tracks/wheels were too difficult to remove without damaging the print. I edited the model to allow me to print the tracks separate to the body and set the print running. The result was very poor so I re-levelled the bed and made sure the extruder was correctly calibrated (it was).

    I began the print but cancelled it midway when I found this:

    DSCF1143.JPG

    The top of the print was being squashed by the nozzle and you could hear it as the nozzle rode over the previous layer, a rumbling noise (the pic was after I'd removed it from the bed, it was printed parallel to the x-axis). Looking at the LCD, the z height was higher than what I later measured the model as. I went back to my previous prints and found this:

    DSCF1141.JPG

    Now, both of these were printed at 0.1mm layer height and I've jacked the silver one up so that the underside of the chassis' are level. Notice the difference in quality but more importantly, the difference in height! The silver one is about 80% in z but the same in x and y.

    I went back to first principals and printed a 20mm cube using the same S3D settings as in the silver tank above. The results:

    DSCF1145.JPG

    The silver cube is 16.4mm tall yet when the print was completed, the display said 19.99mm. It was printed at 0.1mm layer height compared to 0.25 for the yellow one which was printed a few weeks ago. Now, I have just changed over to Alex's firmware with mesh bed levelling and I was beginning to wonder if something had gone wrong and the levelling was interfering with the z axis somehow.....

    Today I decided to print another cube but with the same gcode I used to print the yellow one.
    DSCF1157.JPG

    Now this is where it gets interesting! The silver cube is 18.8mm high, the yellow one 20.3mm. same gcode but different filament. Now, see those layers that are sitting proud of the surface? I think there are six of them (one near the bottom is a bit fuzzy). If the z-axis didn't move on those layers then 18.8+6x0.25 = 20.3mm! Hmmm, interesting.

    OK, one last print. I re-printed the same gcode as the first silver cube (16.4mm) but the print failed immediately as the nozzle was ploughing the glued surface and no PLA could escape. I re-ran an autohome and restarted the print. This time I monitored the print closely and photographed the LCD and the print every few minutes to see if anything odd happened. Well something odd did happen, I got what is arguably the best print I've made so far. It's almost flawless, see below:
    DSCF1179.JPG

    So, the two on the left came from the same gcode file albeit a month apart and using different filament.
    The two on the right were printed a day apart using the same gcode file.

    The 3rd one from the left is the new 0.1mm layer print (2 perimeter shells, 5 bottom, 7 top) and it is pretty much 20.0x20.0x20.0. The LCD said the last layer was at 19.99mm.

    So, given that after the no.2 cube (in the pic above) I had the nozzle running into the bed, can I safely assume that the issue here is that my z-axis is somehow not moving on random layers even though the gcode requests it (I've checked the file and it does for every layer)?

    Does a sticking z-axis also explain the issues I've had with unreliable z homing too?

    I hope I've solved the cause of these issues, now for the solution? I re-greased the screws and the z-guide rods yesterday so maybe it's all starting to sort itself out. We'll see tomorrow when I run a few more prints through.

    Cheers all,


    Jason
     
    Michael Frampton likes this.
  2. moshen

    moshen Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    101
    Likes Received:
    13
    You might have skipping Z motors. Try moving the bed up and down and see if the motors run smoothly, also add more voltage to the z stepper driver. Say 0.1v more and see if it fixes your issue.
     
  3. Jasons_BigBox

    Jasons_BigBox Well-Known Member

    Joined:
    Mar 25, 2016
    Messages:
    248
    Likes Received:
    71
    Thanks, I'm beginning to think the same. I've checked the travel a few times and both motors seem to be moving smoothly. I've measured the movement manually and they seem fine. I'll try tweaking the voltages up a little. They were at the minimum when I set them up.

    We'll see what tomorrow brings, cheers

    Jason
     
  4. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    529
    Likes Received:
    169
    Yup almost certainly something like skipping steps, probably the threshold changes a bit with ambient temp vs driver temp vs stepper temp. So that might explain why you get different results on different days. Z needs to be a little higher than the others in my experience due to it driving two motors.

    Mine is set to 0.65
     
    mike01hu likes this.
  5. Tom De Bie

    Tom De Bie Well-Known Member

    Joined:
    May 7, 2016
    Messages:
    191
    Likes Received:
    35
    Something you can also try is the following...

    Step 1: Check if the PSU has a potentiometer next to the connectors. Mine did... If not your out of luck ;)
    Step 2: Heat up your bed to the maximum temp (ie 110°C)
    Step 3: Heat up all of your hotends to 240°C (they should be able to handle that temp.
    Step 4: While both are heating up check the voltage output on the 24V line of the PSU with a digital multimeter. If the voltage is below 24V turn the potmeter until the voltage is at 24V.

    This assures that when under load the PSU gives the exact amount of power... Now if you turn off the headbed and hotends you will notice that the PSU will deliver 24.1 - 24.2 Volt.
     
  6. Jasons_BigBox

    Jasons_BigBox Well-Known Member

    Joined:
    Mar 25, 2016
    Messages:
    248
    Likes Received:
    71
    Thanks for the suggestions guys, I just came back to say that I think I've solved the problem but by going a little unconventional!!

    I wanted to know for sure so I shot a video of the stepper couplings with pointers attached to them to see what was happening. Sure enough, both steppers were stuttering at the same time which kind of ruled out some kind of mechanical issue.

    I began moving the z-axis in 0.1mm steps and every now and again they stuttered. 1mm, 10mm or full stroke didn't, I assume because of no need to overcome static friction.

    I dialled the voltage up to 0.65 but it didn't help so I turned it up a but further to 0.75 and everything got WORSE. Stuttering all over the place. Soooooo, I turned the voltages DOWN to 0.45 and this happened:

    DSCF1187.JPG
    DSCF1189.JPG

    It's pretty much perfect, as are the other prints I've run since.

    I cane to a similar conclusion that you did Tom, that the steppers were demanding more current than available so wouldn't move. I thought it may have been a limit on the drivers but a PSU limitation would also make sense. I think greasing the rods and screws a lot has reduced the resistance too, that will help.

    I'm not an electrical engineer but turning them down if there's a current limitation I think would help, I=V/R ??

    Well, I'll keep an eye out for missing steps in future. Looking back, A few of my better prints have the odd line around them so it may have been getting worse for a while.


    Thanks for the ideas everyone, now on to the ABS that I received in the post yesterday!!


    Jason
     
    #6 Jasons_BigBox, Aug 24, 2016
    Last edited: Aug 24, 2016
    Tom De Bie likes this.
  7. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    589
    Likes Received:
    100
    My initial thoughts are the stepper drivers are overheating. Did you touch them or even better measure the temp?

    I have different stepper drivers but heat sinks are never wrong: https://goo.gl/photos/DTb8KecaN1NsH7Nx7

    The problem with many drivers is that the chips are made to dissipate heat from the bottom (the PCB) and need a thick and large plane to get rid of heat. The next best thing is to add a heat sink attached to the PCB like in my pic. The third best thing would be to add heatsinks to the chips themselves.
     
    cez likes this.
  8. Tom De Bie

    Tom De Bie Well-Known Member

    Joined:
    May 7, 2016
    Messages:
    191
    Likes Received:
    35
    Another advantage of tuning the PSU is that your heatbed will reach it's temperature faster. I print a lot of ABS and set my heatbed to 110°C. With the tuning of the PSU I get there in about 8 minutes. ;)
     
  9. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    529
    Likes Received:
    169
    What do you mean by tuning the PSU Tom. You mean setting the output voltage ?
     
  10. Tom De Bie

    Tom De Bie Well-Known Member

    Joined:
    May 7, 2016
    Messages:
    191
    Likes Received:
    35
    Like I said before. Set the heatbed to 110°C and the hotend to 240°C and while heating check the voltage on the PSU. If it's lower then 24V. Then adjust...It's that simple ;)
     
  11. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,561
    Likes Received:
    483
    Is there a dial on the PSU then for that?
     
  12. Oliver

    Oliver Active Member

    Joined:
    Aug 15, 2016
    Messages:
    26
    Likes Received:
    0
    There usually is a poti near the connector side of the PSU.

    I have the same issues but they mostly only showed them self in the MBL FW, maybe the Z Axis has more work with that FW.
     
  13. cez

    cez Well-Known Member

    Joined:
    Apr 4, 2016
    Messages:
    133
    Likes Received:
    23
    Do you have a link for those heat sinks?
     
  14. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    589
    Likes Received:
    100
    Spoon Unit likes this.
  15. EpicFail

    EpicFail Active Member

    Joined:
    Apr 22, 2016
    Messages:
    42
    Likes Received:
    4

    Where do you touch what with the multimeter to test your voltage on the PSU?


    Images please?

    Also, I assume I cant check any V+ V- That already has something connected to it? ... for sparky reasons.
     
  16. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    589
    Likes Received:
    100
    You can check wherever you want as long as you do not short circuit anything. I would start at the main wires on the RUMBA from PSU I think you are talking about :)

    Just be sure you are using a Volt-meter and not an Amp-meter :D
     
  17. EpicFail

    EpicFail Active Member

    Joined:
    Apr 22, 2016
    Messages:
    42
    Likes Received:
    4
    Well why i ask is , I followed along with some visuals, and ended up with sparks going across the contact point.

    So, as not to say, kill myself, or fry some electronics. I am curious HOW you would say, check the voltage on the PSU, and weither or not you can do that with everything plugged in. WHen I attempted to do as I saw, the difference I was aware of was I already had lines attached to the + and - , So if there are connections being made, how do you test?
     
  18. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    589
    Likes Received:
    100
    If you see sparks when metering voltage stop immediately. You are doing something wrong yes. Difficult to say what but I thing we need to start with the basics. How to use a multimeter measuring voltage for a starts. I am at work now so I can not make take any pics or anything.

    So can you please take a pic of you multimeter with it configured as it was when you saw sparks (including the leads)? With a functioning volt-meter you should never see sparks except if you accidentally touch more than one pin and short them. And if you are unlucky doing that anything can blow.

    You are not in any way in danger of getting electrocuted as long as you are metering on the RUMBA. The 120V/230V connections at the back of the printer going to the PSU is another story.

    https://learn.sparkfun.com/tutorials/how-to-use-a-multimeter/measuring-voltage
     
  19. Stian Indal Haugseth

    Stian Indal Haugseth Well-Known Member

    Joined:
    Sep 11, 2015
    Messages:
    589
    Likes Received:
    100
    Here is the mains from the PSU and where you want to meter, the yellow circle. This should show about 24V and is not dangerous for you in any way.
    overview_rumba_metering.PNG

    You should measure both the mains input and the heat bed input.
    detail_rumba_metering.PNG
     
  20. Miasmictruth

    Miasmictruth Well-Known Member

    Joined:
    Sep 4, 2015
    Messages:
    804
    Likes Received:
    118
    Like Stian said, you need to make sure your probes on on the voltage side, here is a picture off Amazon, your probes need to be in this configuration. If you have the probes in a different holes you will be measuring Amps which would be like creating a short circuit with the meter. Mostly likely if you have done that then the only real damage will be the fuse in the meter.

    You need
    Black Probe: COM - common
    Red Probe: V - volts
    Multimeter.jpg
     
    Stian Indal Haugseth likes this.

Share This Page