Huxley Duo and Firmware 1.04 [axis homing, bed leveling, orthogonal axis calibration, heating] -- my experiences and questions.

Hi,

I am struggling with the commissioning and calibration of my Huxley for some time now and had several issues. Here I share with you what I came across so others with the same problems know that they are not alone and maybe one of the pros in this forum has an idea on how to solve them.

Z-Axis Homing

1. IR-Reflectors: Z-probe calibration is not documented properly (for the Huxley). It took me a while to figure out that you should put pieces of white paper under the probe or between kapton tape and probe (to improve reflectivity for IR light). Later I removed these reflectors again for reasons described in the bed leveling section.
2. Position: The position for the Z-axis homing is obviously just copied from the Ormored. I changed line 4 in homez.g from 'G1 X60 Y0' to 'G1 X40 Y7 F2000'.
3. Speed: Traveling to the position where the z-homing is done was enormously slow after I upgraded to the new 1.04 firmware. I had to specify the feedrate in line 4 (F2000 see above) to make it faster. This was especially a problem if I started z-homing from the other corner of the bed. It would take 5 minutes or so just to reach the probing position!

Z-Probe calibration instructions for the Huxley
Z-Probe calibration instructions for the Ormored
[edit] All these issues solved by the dc42 fork.

---

Bed Leveling

1. Positions: Again the positions where the bed is probed are just copied from the Ormored. Therefore the carriage just crashes in the sides with the default configuration! Since I extended my body image to the Huxley this would physically hurt me!!! I changed the lines 3, 6, 9 and 12 in bed.g to hold the positions (40, 7) (40, 130) (130, 130) and (130, 7) instead of (60, 0) (60, 185OUCH!) (230,185 OUCH!OUCH!) and (230 OUCH!, 0). In addition to that I set the four probe points using M557 accordingly. [edit] Solved this by using the dc42 fork. It has proper default set and it will even try to prevent some crashes.
2. IR-Reflectors: After learning about the white paper snippets for z-axis homing I put 4 of them in the corners used for bed leveling. Maybe the were too small. At least I did not get consistent readings from the z-probe in all four corners which made the whole bed leveling process completely useless. Putting them below the kapton tape would not help either. So I just removed them again and recalibrated the z-probe to work with the plain heatbed. The results are better now but I still get up to 0.2 mm inaccuracies between the four corners.

Bed leveling instructions for the Huxley

---

Orthogonal Axis Calibration

1. Printing: When printing the calibration pieces I found out that the Huxley would somehow not wait until everything (especially the bed) was heated. So I had to set the bed temperature manually in advance (to 65° C) and wait until it reached that before I started the print. [edit] Solved this by using the dc42 fork. Together with the new web interface of zombiepantslol it is awesome btw!
2. Unclear Documentation: My limited understanding of the English language did not allow me to understand how to measure the S-value to be supplied to the M556 calibration command. The 78mm were obviously far too large for the part I printed (the longest side was shorter than that). Again this value was obviously copied from the Ormored documentation. For me I put in 55mm because that made sense. It is the distance between the outer corner of the calibration piece and the center of the screw.
3. Z-Axis goes crazy: After I put the values I measured into my config.g using M556, the z-axis would go crazy after finishing the bed leveling procedure. It would suddenly think it is -400mm in the ground and would would drive upwards constantly (I hit reset before it would crash in the top). Others have reported the same problem with the Ormored but apparently there is no solution. I found out that if I leave Y and Z at 0 in the M556 command it would not show that behavior. But then two of the planes (XZ and YZ) are not calibrated. [edit] dc42 also fixed this.

Orthogonal axis compensation instructions for the Huxley.
Other forum post describing a crazy z-axis

---

Heating

• After figuring all this out I tried to print the axis calibration piece again (knowing that just the XY plane would be corrected) only to find out that apparently the new firmware reads too high temperature values (around 50°C at room temperature) for both the bed and the extruder. Therefore they won't heat up sufficiently and no plastic leaves the nozzle that then would not stick to the bed because it is not properly heated. I payed no attention to this before because I read, that the temperature probes are inaccurate at lower temperatures.

[edit]OK I figured out that the newest config.g had the lines for the 4k7 series resistor commented out. This was the reason for the bad temperature readings.

Other notes

• Yes, you have to calibrate the feed rate. The default value is bad and it DOES matter.
• Clean the bed surface before the first print. This is a fairly easy source of error to remove. There will be plenty of others to keep you busy.

dc42 Firmware Fork
I did read good things about the dc42 Fork of the firmware (but mostly be the user dc42 himself) and I hope it would fix the crazy z-axis problem. I would like to try out this firmware but it does not ship with a sys folder for the Huxley and I could not find too much about people using this fork with the Huxley. I probably could look at the diff between the Ormored and the Huxley sys folders of the original firmware and try introduce the same changes to the Ormored sys folder that ships with dc42 but given the *huge* amount of success I had with this printer so far I do not trust myself to not screw this up.
[edit] OK since yesterday afternoon there ARE sys files for the huxley. This post made them happen.

OK thanks for reading this long post. I hope I could help others with my description and I would be super happy about any advice on any of the topics.

Have a nice day!

Teleman

Edited 3 time(s). Last edit at 03/28/2015 11:36AM by teleman.

Teleman,

In reguard to the DC42 fork, as of last night it now includes a sd image for the Huxley Duo, I know cos I wrote it for my Huxley and passed it to DC.

The 'Going crazy' after doing the M556 command on the Z axis - I had this also... I must admit Ive been too chicken to try this calibration on the DC42 fork yet but will later today.

I agree the 78 for the S calibration could be a lot simpler, I read it as being the distance between the small ridge near the corner of the piece and the centre of the screw ... I measuerd that as 45mm - but im unsure if this is correct but am trying again tonight.

James

I fixed a couple of bugs relating to M556 a few months ago, so I think you can safely use M556 with my fork.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Yep I am currently printing the calibration piece again with axis compensation. Works nicely so far! I used the 45mm setting as advised.

Hi,
I got similar problems with my Huxley Duo.
I have to print the calibration piece, I would like to know which firmware are you using.

I started changing values for homing but then I read you start working on dc42 firmware and I don't understand which is the best to use..

Thanks in advance for your support
Davide

I think you need to be a bit more precise with the description of your problems. What I did:

• Use the dc42 firmware
• Use 45 mm for the length of the calibration piece instead of the 55 mm

Read here about that dc42 firmware.

Hi Teleman and everyone contributing to this thread: just wanted to say thanks for the advice, it was very useful when I was going through similar issues during my build and commissioning my Huxley Duo.

I have one question though, when I do the bed compensation what I get seem to be to be rather large differences in the readings for the four corners. An example from the last time I did the compensation:
"Bed equation fits points [30.0, 5.0, 0.054] [30.0, 124.0, -0.517] [134.0, 124.0, -1.156] [134.0, 5.0, -0.434]"
Is this normal for the Huxley and is it able to deal with it or should I look into it? I have only printed a few things, but they seem on the whole to be ok (I am using the dc42 fork by the way)

Davide, not sure if this will come a bit late, but this is what I did:

- I initially commissioned the printer with firmware 0.78c as this was the one that was already installed in the Duet, and I was having issues finding the way to update the firmware, so I decided to ignore the update strong recommendation from the instructions temporarily.
- Printed the calibration pieces just fine with 0.78c, in fact I was pleasantly surprised the inaccuracies in the 90deg angles were not too high.
- Printer would go crazy with the M556 command. I decided to change to the dc42 fork (1.04f at that time) as the official RRP one seemed to be giving people some issues as Teleman reported.
- Everything seems to be working well, except a few minor issues such as the one I report above with the bed compensation accuracy.

Oh, one last thing, I agree with Teleman and do not use the papers in the corners, not necessary. Just make sure you do your z-probe calibration with the kapton tape only as the probe reads slightly different values.

Will