Up! Mini 2 + Duet electronics! (Linear rails and real hotbed!)

[bjorn]

Bed is all glued up; used some "704" allegedly thermally conductive silicone like stuff that should be good to 250C to bond the heater wire to the heatsink, and covered it all with regular red high temperature silicone, added a layer of ceramic wool for insulation and covered with a silicone impregnated sheet of fiberglass sandwiched onto some aluminium foil.

Phone died while measuring, but it got to 97C in just under 6 minutes in the end. Acceptable. Added a 120C bitmetal thermostat and a 150C thermal fuse and a bond to PE for some safety measures. Just need to attach the new PT100 sensor and wire in the external solid state relay.

[bjorn]

Bed hooked up to Duet with an cheap 2 channel SSR module, allows adittion of enclosure heater from left over heater wire.

2019-03-26 17_52_21-data_image_svg+xml;base64,PD94bWwgdmVyc2lvbj0iMS4wIiBlbmNvZGluZz0iVVRGLTgiIHN0YW.png (401.28KiB)Viewed 19981 times

[tiertini]

Hey @bjorn, any news on this very interesting project you started?

[bjorn]

Not yet, the belt and idle pulley for the axis that moves the bed broke, and while my quick fix let it get back to work the damage to the belt was so severe it completely failed after a while. And its a weird GT1.5 fixed length belt with matching idle pulley + some other proprietary bits that broke.

Tiertime finally got back to my support ticket, but claims the printers warranty is up after only 8 months, and my local vendor seems to have vanished completely.

I'm not shelling out $66 + DHL shipping for the belt and other parts that broke. Going away for easter, but I'll try to publish the latest config and .stl files for mounting etc when I get back.

[bjorn]

With the printer well and truly hosed, I moved my attention the the CNC router, with duet electronics and some general TLC I got it running, and if we ignore the odd apprentice marks and wafer thin sections that should have some meat to them I got the pockets for the N35UH magnets milled out, and re-surfaced the section for the thermal cutoffs. nice and flat. (maybe I could even machine the other surface flat flat flat flat after stiffening up the CNC router a little!)

Tapped the ends to accept delta style ball joints so the bed can pivot without strain. All that remains is to saffrifice one of the busted flex boards and make some holes to insert some counter sunk screws from under the flex board, up to the new heat bed and grub screw a thumb wheel for adjustments. The screw will then screw into the ball joint pulling the bed closer, with springs providing siffness. It'll cost me 20mm or so Z height, but again no irreversible changes made to the original hardware (except the already busted flex board).

Magnets are due to arrive any day, and i need to rewire the thermal cut offs and make a small bracket to screw them down as well as the thermal adhesive, but all in all the bed is just about all dressed up with nowhere to go.

[bjorn]

Its been a long journey, but its over now.

Payment provider accepted my claim for all the defects and issued a full refund. Printer is now restored to original condition and awaiting vendor to collect it. I might edit some of the posts and add some bits and pieces over time, but all in all, this thread is dead.

[arhi]

For me it was Smoothieboard or Duet - and the Duet community just seems to be in the open source spirit and the Smoothieboard borderline arogant.

:D :D :D heard this so many times ... on the other hand, I prefer the smoothie attitude that they don't allow crap into their code and comparing smoothieware code with marlin or duet you see a difference between professional embedded systems developers (smoothie) and a total hack made by ppl that never before did anything embedded :D :D :D :D ...

both choices of mcu's are terrible, but imo duet is worse :( ..

anyhow I have a "real" question here, not just babble ... WHY o why did you staid with PT100 sensors on the UP when you replaced the whole thing?! I understand when ppl (like me) change only the CPU module and use existing electronics, you keep the PT100 so you don't need to rewire everything and on-board pt100 amp will not properly work for some NTC, but since you replaced the whole motherboard too why not just replace the PT100 on the extruder with a standard NTC?! same for the bed :)

[bjorn]

I decided to see what was possible with as little modifications to the original parts as possible. Straight up replacing a part is easy to reverse and in the event of problems the printer could be put back into stock configuration with ease for warranty repairs despite the rather intrusive upgrade.

So while I could absolutely replace the smd thermistor on the bed and in the extruder, PT100 has its advantages when it comes to accuracy (at least compared to the cheapest NTC parts), and they're also better suited for higher temperatures (300+). Havent tried the Up extruder past ABS, but I fully intend to print polycarbonate and nylon so keeping PT100 support seemed like the way to go.

As far as smoothie go, i do have a smoothieboard, that runs the LPC port of GRBL...

[masrodjie]

Decided to double check the pin out for the print head PCB i've seen here before. Seems the Mini 2 has a slightly different layout. But as mentioned in the heated bed thread, I don't think the wiring can handle the move to 24V, so I need to use the original power supply once I get the heat running.

Edit: Will upload new sketch, this one has a flaw.

Top 3-pin is fan.
Left 4-pin is heater/PT100
Middle 2-pin is LED
Bottom 4-pin is extruder motor.
IMG_20190303_165336831.jpg
IMG_20190303_165347624.jpg

I can't see the image

[masrodjie]

Realized the motor pin out and pin spacing were the same so they plug straight in, end stops needed an adapter but quick work.

Need to print som parts to hold test indicator to verify steps/mm and stuff, but so far so good.

X/Blue (moves bed in/out): 110step/mm, Backwards, Active high/NC limit switch at Xmin.
Y/Red (moves extruder left/right): 110step/mm, Backwards, Active high/NC switch at Ymin
Z/Yellow (moves bed up/down): 110step/mm, Backwards, Active high/NC switch at Zmax


It can home, and move to soft limits. Next step work out what voltage fans, leds, beds, heaters etc are getting and wire it up.
Worth noting that Trinamic drivers will not run these silently at 12v (maybe can be tweaked, but 12v was louder than original hardware). When running off 24V they are virutally silent, will upload video comparing homing sequence soon(tm)

Hotend Heater (presume 19v/measured 8.2ohm*)
Hotend Fan (rated 5v 0.2A)
Hotend LED 5v
Bed Heater (presume 19v/measured 14.1 ohm@20C, 18.5ohm@75C .**)
Case LED 19v
Case FAN (rated 24v 0.3A)
Motors: 400 steps, 25ohm, 30uH, 0.4A (inductance measured, current best guess)

* Same as below more or less. 24V means 70W instead of 44W to heater.
** Feeding it 24v will yield 1.7A/40W but too much for the FPC cable. Can go up to 20V if you use the two conductors for the z-height. 4*0.5A = 2A. Derate to 70% and get 1.4A/28W which is still twice the original power. Might work, but if reverting to soldering then adding some nice thick gauge silicone wire is better. Still need to determine where the drop from 19V to 14.4V occurs on my printer to make a recommendation for supply voltage if not making any changes.

Bed FPC pinout: (maybe reversed)
1+2 Z height probe
3+4 PT 100
5-7 heater +
8-10 heater -

FPC front panel pinout (useless after upgrade)
gnd
gnd
gnd
nc
txd
rxd
nc
5v
5v
5v

FPC the other one pinout:
LED 19v
LED -
Door sensor 1
Door sensor 2/Switch NC
Door sensor 3
Switch NO
Switch COM
LED R-
LED RB+ 5v
LED B-

One thing I'll give them credit for, proper bonding. Main board has a 1x5 header with all leads going to the metal chassis. Which I can only assume means the external power brick isn't actually isolated but ground referenced. Interesting choice!

Thats about as far as I'll get until the PT100 board and the FPC adapters arrive. Doable to print mounting adapter to go between the old board mounting holes and the Duet mounting. But it will be a challenge to get power and ethernet in through the unmodified rear panel, only option seems to be cables and grommets through existing holes and external connectors or even permanent fixed length leads.

Can you share the STL/STEP for the board mounting?

[bjorn]

Need to make an adjustment and do a test print, but I'll hopefully get around to it soon ish. Picture had a flaw, but will redo it as well

[vedika]

Thanks for the update..

[bjorn]

Can you share the STL/STEP for the board mounting?

https://a360.co/2wpTgSG

That is AS-IS; and its not very stiff, but it was the quick and easy method that would be small enough to print on the Up Mini 2. However the full fusion project is available if you want to make changes. I was planing on trying to use the screw holes for the electronics cover as additional fasteners, but not sure I'll get around to it.

[arhi]

I decided to see what was possible with as little modifications to the original parts as possible. Straight up replacing a part is easy to reverse and in the event of problems the printer could be put back into stock configuration with ease for warranty repairs despite the rather intrusive upgrade.

:D

So while I could absolutely replace the smd thermistor on the bed and in the extruder, PT100 has its advantages when it comes to accuracy (at least compared to the cheapest NTC parts)

the smd one on the pcb is fairly good but the one in the metal casing is pos :( ... I opened few (one original that came with upplus2 and one that I got from e3d) and inside the metal case is a pt100 sensor in 1/8W resistor case insulated inside fiber/silicone shroud so the accuracy is terrible (there's 2 layers of insulation from pt100 to the outside case)
[/quote]

Havent tried the Up extruder past ABS, but I fully intend to print polycarbonate and nylon so keeping PT100 support seemed like the way to go.

I print PC using upplus2 without issues (actually had lot of issues as bed could not get hot enough so had to add better heating element and external power supply for the bed so that PC don't warp like crazy, also had to put upplus2 in enclosure etc etc.. but the extruder itself handles PETG, PC, Nylon ok)

[bjorn]

the smd one on the pcb is fairly good but the one in the metal casing is pos ... I opened few (one original that came with upplus2 and one that I got from e3d) and inside the metal case is a pt100 sensor in 1/8W resistor case insulated inside fiber/silicone shroud so the accuracy is terrible (there's 2 layers of insulation from pt100 to the outside case)

I honestly don't even know why I'm surprised. On the other hand the extruder has been flawless* through quite a few spools of ABS so far so I had no reason to suspect it be sub par.

*) one jam after idk, 5kg maybe, replaced nozzle and back to printing.

[arhi]

the temp precision of the heated element is overrated. precisely controlling temperature of the plastic on the mouth of the nozzle, that would maybe produce better results (tbh I tried, I made a TC on the tip of the nozzle by electrowelding wire to the tip, and no difference even with +-0.1C on the mouth) but since most filaments have 50C + range where they are printable...

[masrodjie]

Are you sure the step per mm for XYZ is 110? It's using GT1.5 and 20 teeth so it should 106.67