My CoreXYU Printer project

Quote
o_lampe
What's written on the 6-pin IC at the top left?
It might be a PIC controller, triggering the two bigger MOSFet chips at the bottom.
Ever tried to run the pump with 12V?

I do believe it did not have any text on it. I’ll have another look when I get back home.
I have not tried to run it on anything other than 24V, what would you expect 12V would to? Run slower? I could give it a try.

Took another look at that IC, there are some text there but I can not read it. I have a 500x usb scope on the way, might be able to read it arrives in a week or two...

Quote
lars.arvidson
...I have not tried to run it on anything other than 24V, what would you expect 12V would to? Run slower? I could give it a try.

I expect the levers to swing softer and make shorter moves. Question is, what will the valves do? They are passive and usually open/close firmly only when there is enough airflow.
A variable input voltage would show you the sweet spot. That's why I suggested fast PWM ( with variable duty cycle ) and a huge capacitor in parallel. That would be similar to a buck converter.

Tried adjusting the voltage and it worked pretty good down to 12V! Was a lot less noice at 12V and it did still pump some air. The question is how mush is needed.

Airpump test

How big capacitance do you think is needed (or how would I go about to calculate it)?

Edited 1 time(s). Last edit at 08/12/2017 10:02AM by lars.arvidson.

There is a capacitor at the input of the PCB. With a higher PWM frequency it will probably be big enough. ( Hoping it is a low resistance type "low ESR")
It's hard to tell, how much ripple voltage is left without an oscilloscope.

You already have quite a large capacitor in parallel (the big one in the picture). What you need is a resistor or even better an inductor between the Duet and the pump. Best would be an inductor of at least 100uH used with a high PWM frequency e.g. 50 or 100kHz. You must add a flyback diode if your Duet PCB is not the 1.02 revision. If you use a resistor, try a few tens of ohms.

Edited 1 time(s). Last edit at 08/13/2017 03:07AM by dc42.

---

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].

Thanks guys! I do believe I managed to get it to react to PWM. I had to connect it directly to the duet. My guess is that the mosfet board can’t switch fast enough for higher frequencies.
I only had 0.25W resistors so could only use 5ohm without exceeding the limit (too much). There was not a big difference from 100% to 50% but when down to 10-5% there was a pretty big difference. Maybe it will work better with bigger resistor or a inductor.
Would this one work? I guess this have to low power rating?
I’m sorry to say my electronics skills are not the best…

Cable management/wiring will the bane of my project... One of the PT100 hotend temp had been acting up a bit lately. When I took a closer look it turns out that one of the wire was broken...

Quote
lars.arvidson
Thanks guys! I do believe I managed to get it to react to PWM. I had to connect it directly to the duet. My guess is that the mosfet board can’t switch fast enough for higher frequencies.
I only had 0.25W resistors so could only use 5ohm without exceeding the limit (too much). There was not a big difference from 100% to 50% but when down to 10-5% there was a pretty big difference. Maybe it will work better with bigger resistor or a inductor.
Would this one work? I guess this have to low power rating?
I’m sorry to say my electronics skills are not the best…

That inductor would probably do the job, but this one [www.banggood.com] would be better.

---

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].

I been printing some dual color froggs lately (thingiverse).
I have had problems with (among other things) stringing and zits. I have gotten one new flexible drive shaft for the Nimble (still waiting for the other one) and both breach blocks replaced but the big difference was change of slicer settings. The left frog in the picture below is with old settings and right with new settings (same hardware setup).

Hires picture here.

Firmware version is 1.19RC6 (slightly modified), slicer cura 2.7-beta with new reprap firmware option selected. Base print speed 80mm/s for both frogs. The right one I changed retraction from 2 to 3mm, combing from no-skin to off and turned on coasting. I also decreased temperature from 210 to 200 (I have done this before and it did not help much with stringing but reduced the oozing).
I’m pretty happy with the result but still room for improvement… Might try a dual color benchy again with these settings…

Edited 2 time(s). Last edit at 08/15/2017 06:49AM by lars.arvidson.

Benchy turned out ok!



Hirez link

I find my cable management a bit lacking as I have stated before. The cables to the hotend can easily get to mush tension (start pulling on the connectors or wear against the carriage) or get too loose (get stuck on the frame or the other carriage).
I have been looking at using something like a cat6 cable. They have 8 conductors, are cheap and readily available. Endstops, fan, pt100 temperature sensor and z-probe should be no problem using this kind of cable. Heater might be a problem and might have to use other cables or would at least need a few of the pairs in a cat6 cable...?
The carriage with the probe currently need a lot of conductors. 2 for fan (24V), 2 for endstop microswitch, 4 for pt100 and 5 for BLTouch sensor (2 of the 5 are ground).
That’s 8 (4+2+2) conductors for the carriage without z-probe and 13 (4+2+2+5) for the other.
I wonder if one could use same ground for fan, endswitch and probe? If so I could get away with 10 (4+1+1+3+1) conductors, still two conductors too much… Maybe if I use two cables, use three pairs for heater and one pair for something that is not sensitive to interference (maybe the endstop or fan?) and if I can use a common ground the rest would fit in the other cable.
I got a few keystone cat6 jack. They are pretty big, don’t know how I would be able to fit two of them on my small carriages… Maybe I could make it work if I do a carriage design that integrates the extruder (replace the nimble shell and only use the gears).

I needed a longer cable for my PanelDue and thought a cat6 cable would be perfect for this. The duet3d wiki states the resistance per conductor should not exceed 0.1 ohm. I have seen dc loop resistance figures from 10-20 ohm for 100m. For “non loop” 1m that would be 0.05-0.1 ohm. I guess there might be some resistance in connectors too. Anyway, I tested with a 3m cable which will probably put me over 0.1 ohm but everything appears to works.

See the bukito . It uses flat cables for cable management and is very neat. U can use multiple parallel lines for higher amps

I wouldn't put the fan on the same ground as the endstop and thermistor. Fans generate noise, the last thing you want on the thermistor input.

---

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

The thermistor needs a separate ground, because on the Duet it connects to VSSA (analog ground) instead of normal ground. In theory the fan and probe could share a ground, but if that ground wire breaks then 24V will be fed to the probe via the fan, which isn't good. Also, using a common ground for the fan stops you switching the fan on and off, because the Duet switches the ground aside of the fan (like other electronics does).

You could use a common ground for the endstop switch and the BLTouch signal ground.

---

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].

Quote
ekaggrat
See the bukito . It uses flat cables for cable management and is very neat. U can use multiple parallel lines for higher amps

Thanks for the suggestion. I do however feel a flat cable is not the best match for a printer where the hotend moves both in x and y. There’s also the flexible driveshaft and the air tube running to the hotend.

Quote
the_digital_dentist
I wouldn't put the fan on the same ground as the endstop and thermistor. Fans generate noise, the last thing you want on the thermistor input.

Quote
dc42
The thermistor needs a separate ground, because on the Duet it connects to VSSA (analog ground) instead of normal ground. In theory the fan and probe could share a ground, but if that ground wire breaks then 24V will be fed to the probe via the fan, which isn't good. Also, using a common ground for the fan stops you switching the fan on and off, because the Duet switches the ground aside of the fan (like other electronics does).

You could use a common ground for the endstop switch and the BLTouch signal ground.

Thanks for the information! I have a PT100 with 4 conductors so no normal thermistor. If two pairs is enough for the heater I can put fan and endstop on the other two pairs in that cable. From the other cable I'll use two pairs for the PT100 and the other two pairs for the BLTouch. Will probably need 2.5m cables.
I think I'll have to do some testing to see if 2 pairs are enough for the heater...

If you are using 24V power then I think 2 pairs should be enough for the heater.

---

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].

Printed a few stringing tests. There’s quite a difference between the extruder with the original cable and the one with the new cable. With same settings, same filament brand and type (not same coluor), the extruder with old cable produce a lot of hair fine stringing/fluff while the other prints with only minimal stringing. The last cable should arrive in a few days.

Edited 1 time(s). Last edit at 08/20/2017 09:40AM by lars.arvidson.

Printed a pretty big spiral vase [www.myminifactory.com] . Its about 380mm high. I intended to print it at 0.2mm layer height and 80mm/s but forgot that Cura spiral mode (not surprisingly) use outer wall speed which is defaulted to half of the print speed unless set explicitly so it printed at 40mm/s. Took a bit extra time to complete

Original picture

From a distance, it looks good but if you light it from above and take a closer look, the white filament really shows of the artefacts produced by the printer.

Original picture

I do believe that a lot of the artefacts both the z-banding and the ripples in x and y are (at least partly) due to slack in the Igus bushings I use (iglidur G, sleeve bearing). They are designed for a press fit but recovering a press fitted bushing for the next iteration of the design can be hard so I have not done that tight a fit for them. This leaves a bit of slack. If a put a finger on the carriage I can feel little bit of shaking/resonance that I believe is due to the bushings not gliding perfectly on the shafts. It also shows up a bit in a mesh probe grid where you see difference in height depending on what direction the carriage with the probe is moving before probing. The probe is not mounted dead centre and when the slack in the bushings will allow the carriage and the probe to tilt slightly and the distance to the bed will depend on what direction the carriage was moving.

I hope to get some Igus filament (I180-PF) soon that I plan to use to print tighter fitting bushings.

Quote
lars.arvidson
Printed a pretty big spiral vase [www.myminifactory.com] . Its about 380mm high. I intended to print it at 0.2mm layer height and 80mm/s but forgot that Cura spiral mode (not surprisingly) use outer wall speed which is defaulted to half of the print speed unless set explicitly so it printed at 40mm/s. Took a bit extra time to complete

You could have turned the speed slider in DWC up to 200% to put it back up to 80mm/sec. But 80mm/sec is pushing the limits for good quality perimeters on most printers.

---

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].

Quote
dc42

Quote
lars.arvidson
Printed a pretty big spiral vase [www.myminifactory.com] . Its about 380mm high. I intended to print it at 0.2mm layer height and 80mm/s but forgot that Cura spiral mode (not surprisingly) use outer wall speed which is defaulted to half of the print speed unless set explicitly so it printed at 40mm/s. Took a bit extra time to complete

You could have turned the speed slider in DWC up to 200% to put it back up to 80mm/sec. But 80mm/sec is pushing the limits for good quality perimeters on most printers.

True. I might do the same print again and change the speed on the fly to see how it affects the artefacts. My guess is that 60mm/s should be doable...

Printed same vase as yesterday. First bit at 100% speed (40mm/s), second at 150% (60mm/s), third at 200% (80mm/s) fourth at 250% (100mm/s) and last bit at 300% (120mm/s). Changed the temp from 200 to 210 on the last bit. At 120mm/s the printer did some irregular popping sounds (hard to describe) I could not quite locate…
Not that big difference in the artefacts between the different speeds. The ripples in x/y might stretch proportionally with the speed. Some of the z-bands look a bit more pronounced at higher speeds.



Edited 1 time(s). Last edit at 08/21/2017 03:43PM by lars.arvidson.

Quote
o_lampe
What's written on the 6-pin IC at the top left?

G235

Hi Lars,

I came across your thread while researching large scale printers. I'm currently running a home brewed large delta printer but my current printing requirements have grown beyond my printable volume. I've been fond of CoreXY kinematics and my current situation is the perfect reason to build a new printer. Your build checks off all my boxes for features in my next printer. Would you consider sharing your BoM and designs?

Regards,

Allan

Hi Allan,

Sorry for the late reply, I have not worked mush on this project last months, partly because of the amount of changes I want to try and the amount of work it will require that I don’t know where to start.

I would be more that happy to share the stuff I have with you (and anyone else that’s interested) but I would not recommend you build this printer in its current state. It is a work in progress and there is so much that could be improved! Currently I’m pondering switching to belts for Z instead of leadscrew to see if I could get rid of z-wobble. I would also want to try out linear rails to increase precision and get the travel speeds I would want. I would like to try four rods for gantry and route the belts closer to the center (in z) of the carriage to minimize the skew. I would want to support all the pulleys both from top and bottom. Basically, a total redesign

There are some core-xy designs that are much more mature and complete than this one on this forum that you might want to look at. I know that digital dentist has well documented design that seems to work well and has a big build volume. I’m sure there are others as well…

Quote
dc42
You already have quite a large capacitor in parallel (the big one in the picture). What you need is a resistor or even better an inductor between the Duet and the pump. Best would be an inductor of at least 100uH used with a high PWM frequency e.g. 50 or 100kHz. You must add a flyback diode if your Duet PCB is not the 1.02 revision. If you use a resistor, try a few tens of ohms.

Finally got time to try out the air-pump with the inductors I ordered. I tried a 100uH and a 330uH and both worked well at PWM frequency of 100k. The 330uH gave more change in airflow when going from 50% to 100% and was close to off at 10%.

I’m thinking of trying belts instead of leadscrew for z-axis to see if I can eliminate z-wobble. I’m going to use three independent steppers for the bed levelling feature of the Duet. Currently I use 17HS19-2004S1 for the lead screws. Would it be possible use this stepper with 16 or 20 tooth pulley (no gears) to drive the belt? Feels like the resolution would be a bit too low? Without microstepping: 2 * 20 / 360 * 1.8 = 0.2mm/step. 16 tooth pulley and 1/16 microstepping: 2 * 16 / 360 * 1.8 / 16 = 0,01. 0,01 sounds good but don’t know if using microstepping is a good idea?

I have been looking at something like 17HS15-1684S-HG20, a 10:1 or 20:1 geared stepper. They do have 8mm shaft with a key-way. I don’t know where to find pulleys with key-ways, anyone know a good source? Might not be commonly used with GT2? Might need to switch to some other kind of belt/pulleys? Could probably use bigger pulleys with geared stepper and steel reinforced belts.

Maybe I could get away with printed gears and use the steppers I already have?

Any input would be appreciated!

How heavy is the bed, support, and anticipated maximum print?

Before I added the worm drive to UMMD, I had a 184 oz-in NEMA-23 stepper and 20 tooth pulley lifting the 3.5 kg mass of the bed and support. I didn't test it to see how much additional weight it could manage.

https://vimeo.com/196123930

If you put a 20:1 drive on a single NEMA-17 motor (they go up to about 100 oz-in torque), it should have enough torque to lift your bed and print without adding additional motors, unless the bed is really huge. I wouldn't bother with using a 0.9 degree stepper. The gear box will give you more than enough resolution. It's hard to say if a 20:1 planetary gearbox will prevent bed-drop on power off - that will depend on the weight being lifted, and the motor's detent torque. I went with a worm gear mechanism because I had more confidence that it was going to prevent bed drop.

While you could go with 3 motors and gearboxes and drive them independently to autolevel every time you power up, as I recall, your printer's construction looked pretty solid. You may not need to adjust the leveling at all after initial setup. If that's true, adding two extra motors with gearboxes is a lot of unnecessary expense. OTOH, if you use a single motor, you have to assemble a mechanism to drive 3 belts to do the lifting. So it comes down to your level of confidence in the construction of the frame, what you can shoehorn into the machine, the depth of your pockets, and personal preference.

A couple more bits of info that may be relevant... I'm running UMMD's Z axis motor at 1A (so probably getting no more than 100 oz-in torque) and it doesn't complain about lifting 7+ kg at 20 mm/sec. UMMD hasn't required any bed leveling adjustment since I last took the bed out of the machine to make some mods, even after traveling in my car a few times. The worm drive/motor I used was about $100 via ebay and the long keyed shaft was about another $30.

---

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

I haven't built a printer with a bed that moves in the Z direction, but I can see the following potential issues:

1. Z resolution. Delta printers commonly use 16 or 20 tooth pulleys for the tower motors, so in terms of Z resolution, it's possible if the friction isn't too great. For the best print quality I recommend using 0.9deg motors on deltas, not the 1.8deg motors that you have. If a single 1/16 microstep causes the effector (or in your case, the bed) to move - hence the low-friction requirement - then 16 tooth pulleys give you a resolution of 10um using 1.8deg motors, or 5um using 0.9deg motors. 20-tooth pulleys give you 12.5um and 6.25um which are not only larger but also more difficult values to work with.

2. Lifting force. Delta printers only need to lift an effector, which is normally light in weight. You will need to lift the mass of the bed. The required torque needed per motor is:

required_torque = (bed_mass * g)/(pulley_radius * number_of_motors)

The radius of a 16 tooth GT2 pulley is (2 * 16)/(2 * pi) which is close to 5mm.

To ensure good microstepping positioning, the required torque should be no more than about 5% of the motor's holding torque (perhaps 10% at a push) - and that is after you have derated the published holding torque by 0.8 to allow for the fact that the motors will get too hot if you run them above about 80% of rated current, and by a further 0.71 to allow for using microstepping.

3. The detent torque of the motors won't be sufficient to prevent the bed falling under its own weight when you reduce the current. So you will need to make alternative arrangements.

So I think you will need some sort of reduction mechanism. Given that, I am inclined to agree with digital_dentist that one motor with a gearbox and belt drive to 3 leadscrews is probably the simplest solution, even if you do already have the necessary spare motors drivers on your DueX5. The only disadvantage I can see is that you will need bed levelling screws, assuming that you can't get a fine enough adjustment by loosening pulley grub screws and rotating the pulleys on their shafts. Another possibility is 1 or 3 stepper motors with integrated planetary gearboxes.

Whatever gearbox(es) you use must have low backlash. Personally I wouldn't want more than 10um of backlash in the Z axis, which implies that the backlash in an output shaft driving a 16 tooth pulley should be no more than about 0.1 degree. That's a tall order. The weight of the bed may help to reduce the backlash in practice.

In your position, I think I would try to remove the leadscrew wobble, by using 2 or more linear guides and avoiding over-constraining the leadscrews. I recently managed to avoid leadscrew wobble on my SCARA printer by replacing the leadscrew and nut along with careful adjustment of the leadscrew bearing position. But perhaps you have already tried this.

Edited 1 time(s). Last edit at 10/22/2017 12:01PM by dc42.

---

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].