3D Printing Newbie With a Fisher Delta

Hi Guys,

New to the forum and to the world of 3D printing in general. I was lucky enough to receive a Fisher 1.0 for Christmas, and have successfully built it and commissioned it... And gone through the robot printing right of passage.

However, I have had a few issues since that initial success. Full disclosure -I have had to edit the config.g file to allow the printer to maintain a static IP address, as I could never get the web interface to load by using the printer name as the UR -and trying to work out what the DHCP server in the router had dolled out to it each time was a pain. That's the only change to the config I've done, everything else is as it was out of the box.

What I've noticed since the first initially good print:

1)Homing the axis seems to be hit and miss. Some times it does it as you'd expect - up the towers, triggers the limit backs of a fraction. But it occasionally crashes into the the switch and would appear to keep driving the belt, as I can here a clicking noise which I assume is the belt slipping on the timing cog.
2)AJAX web comms drop outs happen from time to time.
3)The build tak surface which was great the first day of printing seems to have lost it's grip. The last couple of print attempts have not adhered, and I end up with a rats nest of extruded filament, usually stuck to the effector and the brass nut.

I've tried to search for answers on here, but I'm getting lost in the lingo and the G code fragments posted up. Any help would be gratefully appreciated.

Thanks

Jason

[www.greatoldone.co.uk]
[www.youtube.com]

1. Assuming it is just one tower that keeps moving, it sounds like the carriage is not triggering the microswitch reliably. You may need to adjust the microswitch mounting or bend the lever arm a little.

2. Which firmware version are you using? There have been many improvements since the Fisher was first shipped. The latest version is 1.17a.

3. I am not an expert when it comes to BuildTak, but there is a Google group for BuildTak support at [groups.google.com]. You can buy replacement BuildTak sheets, or replace it with something else.

HTH David

---

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 agree with DC42 check your micro switches are triggering reliably. Some people had problems with the crimp connections and soldered the micro switch wires.

Prints not sticking:

1)Print head not at the correct height:

Could be linked to your micro switch problem. Your start point keeps changing. so the print head is not at the correct height.
Do more than one bed alignment, check the result at the end of each run, they should tend to zero ( or very small value) eg:
run 1, start 4.999 end 0.000,
run 2 start 2.555 end 0.000,
run 3 start 0.400 end 0.000,
run 4 start 0.0200 end 0.000.
now print.
If this isn't happening
check micro switch connections.


2) what is the room temperature? if you don't have a heated bed, you may have difficulty if the temperature is less than 20C.
Make sure the printer is not in a draft.
Clean the build tack with washing up liquid to remove grease and dirt.

Print head not at the correct height:
Check the bed mounting brackets aren't cracked,
Check the connections to bed probes are tight
Pulleys not slipping on the motors

Hi guys,

I'll check the microswitches. I do know that the mounting slots for them aren't in line with each other, as per the pics on the build instructions. Mine are offset a fraction so when the switch is mounted it is at a slight angle, exposing more of the switch actuator to the carriage.

I upgraded to 1.17a last night, so I'm know on that firmware, and v1.1.3 of the web interface. I rehomed after doing so, did an auto calibrate and got good prints of the robot and rings again.

After that, I decided to try printing something that wasn't on the sd card. Oh boy. What a mess. I downloaded an object from thingiverse, and using the fisher ini file in slic3r generated the g-code for it. But it didn't print properly at all. The first layer appeared to be done much faster than the sample objects, and I don't understand why... the ini file has the first layers at 20% of the speed of subsequent layers. As it's going so fast, the filliment gets dragged all over the platform and ends up in a wad around the extruder.

Two things what temperature are you printing at ? Using Sli3r, Filament settings , first layer = 209, other layers = 206. (Good start point for PLA)

Print settings: loops (minimum) = 1, Minimum extrusion length = 20mm (if the item is small 40mm is better, ensure that plastic is flowing before printing the object.

Did you level the bed before the second print?

hi all,

Yes, I homed and auto calibrated before attempting to print.

My fillament settings where set to zero - I've changed them to the values suggested. They are currently set to:
dia 1.73, ext multiplier 0.95, extruder 1st layer at 209c, subsequent 206.

My print settings had no loops, so it's now set to this:
loops 1, dist from obj, 5mm, skirt height 1, min ext. length 40mm, brim width 0

None of this has any effect on the print. It still ends up a mess on the nozzle.

Have you sorted out the micro switches?

If you home and calibrate the bed then print the reprap robot everything is fine?
If you use a model from slic3r you have problems?

Also make sure that after you auto calibrate, you do not home the printer again before you start printing. So don't use a G28 command in your slicer start gcode unless there is a G32 after it.

Most delta printers are designed so that the carriages run in to the homing switches at the top of their travel. This gives a very reproducible trigger position. The Fisher gains a little extra build height by having the carriages run past the homing switches instead. This doesn't sound as good to me from the point of view of getting a reproducible trigger height. But I don't have a Fisher, so I can't be sure. Small changes in trigger height will be handled by auto calibration, as long as you don't home the printer after auto calibration.

Btw in response to a request from a Fisher owner, I added a Z probe recovery time to the M558 command in firmware release 1.17, to save you from having to use G1 and G4 commands between the G30 commands.

---

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
ramblejeff
Have you sorted out the micro switches?

If you home and calibrate the bed then print the reprap robot everything is fine?
If you use a model from slic3r you have problems?

Jeff - yes, and yes. Gcode from the ad card image works perfectly. My model, sliced with the fisher ini file doesn't.

@DC42 - that all sounds marvellous, even if I have no idea (as yet) what that means.

I think I know whats happening here.
The robot is already sliced and the printer settings are embedded in the robot.g file you send to the printer. You can see these settings if you open up the robot.g file in a text editor (NOT Ms Word!) go to the end of the file then scroll up to the line: (or use find)
;avoid_crossing_perimeters = 0

this is the start of the printer settings and continue to the end of the file. These settings are the same as you see under the 'Print Settings', 'Filament Settings' and 'Printer Settings' tabs in Slic3r. If you open your file you will see these settings are missing and/or different to the robot.g file. I suspect you haven't loaded the config bundle fisher.ini into Slic3r and corrected some of the values.

Download the Profiles bundle from RepRap RepRap Printer Bundle
Unzip the bundle you will see a file called Fisher.ini. In Slic3r goto File>Load Config Bundle. In the file dialog that opens navigate to and click on the Fisher.ini file, click on Open. If you click on the Filament Settings' and 'Printer Settings' tabs you will see some default values have been loaded.

Go to the Printer Settings tab
click on General, Click on Bed Shape: Set select Circular and diameter 150 mm.
click on Custom G-Code, if the start code box contains

G28 ; home all axes -------> this is what dc42 is talking about. G28 homes the head, which deletes the auto bed compensation, so the print head is at the wrong height!
G1 Z5 F5000 ; lift nozzle

[The values below might not be the best but they work!!]
delete whatever is in the start code box and insert

M83
T0
M116

in the end G-Code box delete what there and insert

G91
G1 Z40 F4000
G90
G1 X0 Y0
T99

Click on Extruder 1, Change the nozzle diameter to 0.4 mm and save the changes.

Go to Filament Tab. Set:

Diameter to 1.75 mm (or measure your filament and put in the actual diameter).
First layer to 209, other layers 206
Bed temperatures should be zero ( if you have a heated bed set it to 60)

save the settings

Go to Print Settings Tab. Set:

Layer height = 0.2
First layer height = 0.35

Save the settings.
Slice your model save it to the printer, home the head, Auto bed compensation then print.

What happened???

Edited 1 time(s). Last edit at 01/09/2017 06:55AM by ramblejeff.

All of my settings are now as per what you suggested. see attachments:








Still no dice. Workflow was:

- re-slice object with settings as per above.
- upload g-code.
- re-home, and re auto delta calibrate.
- print object
- hit e-stop as it's all turned into a rats nest.
- sigh

rinse, repeat.

I've checked the microswitches - leads are all attached, and I have continuity from the switch to the board. I've also bent the levers out more, so they are now almost touching the guide rods.

I did a line by line comparison of my sliced file and the one for the robot. There are several differences between them:

My file: end_gcode = G91\nG1 Z40 F4000\nG90\nG1 X0 Y0\nT99\n\n
Robot: end_gcode = G91\nG1 Z40 F4000\nG90\nG1 X0 Y0\n\n

My file: retract_length = 3.2
Robot: retract_length = 3.5

My file: start_gcode = M83\nT0\nM116\n\n
Robot: start_gcode = M83\nT0\nM116\nG1 X0 Y0 Z4 F4000\nG30\nG1 Z[first_layer_height] F4000\n

And to add insult to injury, I can churn robots out till the cows come home. Look here's another one:



It must be slicer settings. I've checked all the carriages, switches, belt tension etc etc.

As you can see I do have a heated bed to fit, but I didn't want to complicate matters putting that on whilst I had print issues.

What I would do now is open up the G-code file for the failed print delete everything from ;avoid_crossing_perimeters = 0 then copy everything from below ;avoid_crossing_perimeters = 0 in the robot.g file and paste it into the Failed print G-Code. Then try and print.

I've attached two files that I have printed on my fisher. maybe you can see where your settings are different. Check the top of the file as well everything up to the G1 codes, are comments or used to set the printer. ( The code G1 tells the head where to move).

For a list and explanation of the G-codes see here:G-Code

if you want to try them, I will could export my slic3r settings for you to try.

Hi,

Your ini file would be great... As you can see, the robots are thanking the maker as they've finally seen something that isn't them or a ring being printed!



One thing I did note, from your g-code files is that you're using slic3r version 1.1.7. I'm using 1.2.9... Is there a reason for this (known issues in later versions)? I just downloaded the latest version from the slic3r website.

BTW - I tried what you suggested, cutting and shutting the code from robots into my object. It was a successful as the other attempts. ;-)

I'm using version S;ic3r 1.2.9 as well that was the version when I started this 3D printing thing.

My config bundle contains lots of files, for printer use fisher, I haven't tested fisher 2 yet. The filament settings are for excel fil PLA, they are a good starting point, just select your colour. In the print settings Ignore Fisher2 i haven't tested them. The ones starting with 'quality' are the ones to use.

If its not the bottom of the .g file that is wrong it must be the code up to the first G1 code.

Good luck

Thank you.

I may have stumbled across the issue I was having though. It would appear that the object I downloaded from thigiverse that I was trying to print is built in such a way that the bottom layer is very small. As soon as those bits are complete, it goes full speed and is attempting to extrude in the air.

I downloaded another object (a dice), sliced it using the ini file I've been attempting to use and got a decent-ish print. The corners are a little saggy on one side, but all in all a damn sight better than what I had been achieving.

That sounds about right. try the same object with support see if it will print. The saggy corner can be caused by a number of things, try rotating the object 45 deg around the vertical axis. It could be a draught or parts of the bed being a different temperatures causing the filament to cool at different speeds.

Now you are getting prints off you should calibrate the printer, important is the extruder, 10mm is 10mm of filament, x,y,z 30mm is 30mm, then the print bed. I know the fisher has bed alignment but if you calibrate the bed the alignment is much quicker.

Good Printing!

On my Fisher with Slic3r 1.31.6 I have to add a Z offset to make sure the print starts at the right height, the offset is different between PLA and ABS. For PLA I use 0.15 and for ABS -0.3 although the gcode appears to limit this to 0.05mm.

BTW The new cubic infill in 1.31 is awsome.

---

DC42 Kossel 330mm x 2meters
My Thingiverse Creations

OK - calibration. I assuming that there are various offsets I can set within slicer to bring parts into dimensional accuracy? Is there a how to anywhere someone can point me to? And should I do this now, or fit the heated bed and then do it (my gut tells me after as the bed is bound to calibrate differently)

BTW, here's the dice I printed last night (on the left), and the one I've just printed using slic3r 1.31.6 from the Prusa website, with the ini I had cobbled together (not tried yours yet, Jeff). The cubic infill is indeed awesome. And it has no saggy sides, as I turned the part 45deg on the platter a suggested.



Dimensionally, I believe this is supposed to be 20mm on a side, and actually measures 19.8/9ish on the x & y axis, and 20.4 on the z axis.

Oh dear.

I've just attached the heated bed to my printer, so I could start the calibration process with it on the fisher. Assembling it and putting it in was simple. If only the PSU was fit for purpose. I received a new one with the bed, 19v 6.3A to replace the original 19v 3.4amp PSU that came with the fisher kit.

It didn't come with another mains cable, so I used the original's one - plugged it in, ready to go. Turned round to upload the new heated bed file, and there's a loud pop, the unmistakable smell of an electrical fire and when i look at the PSU there's smoke poring from the PSU mains cable socket and from around the strain relief gland of the DC output lead.

Needless to say it was unplugged at speed. I've contacted eMaker (the outfit I bought it from), and am awaiting their response.

Can't do better than these instructions:
dc42 Delta Calibration

[duet3d.com]

both are by dc42 who helped develop the Duet and the software on your Fisher so if he doesn't know...

If you search the fisher forum for calibration you should find the answers to any questions you have and browsing the reprap.org site will provide loads of info.

Thanks Jeff.

I wish I'd left well enough alone. It was printing. Now it isn't, the heat bed PSU is toast, the machine doesn't reliably home anymore. I knew it wasn't going to be easy getting in5o 3D printing... hmmm.

3D printing is full of highs and lows Check the SD Card? you might reload the firmware. eMaker might have some suggestions. There is a steep learning curve but keep going the results are well worth it.

OK - my new PSU has arrived, and it works - to a certain extent. It doesn't emit smoke, so thats a good thing.

The problem I have now is that the bed will heat, but the moment the G Code turns on the extruder as well, the printer dies (no power - extruder fan stops spinning, lose contact with the web server). Resetting by pulling the DC plug re-inserting doesn't work. cycling the AC does though... So I assume something is tripping in the PSU when it starts to draw more current.

Has anyone else had this problem? I don't believe the hardware on the printer is misconfigured or connected wrong - unless ive got the polarity wrong on the bed. But I have the red positive into the positive terminal. And even so, it's just a big resistive element, so it wouldn't matter on how the power flows, right?




What should I be looking for?

I've spoken to eMaker (Jean) and they told me to look at the M301 commands to adjust the power to the bed and the extruder. They pointed me at their git-hub, with their sample config.g for values.

I've been sat here for ages trying to work this out. The values are the same as theirs. My config lines:

; Thermistors
M305 P0 R4700 T100000
M305 P1 R4700

; Tool definitions

; *** BED ***
M301 H0 P5 S0.7
; *** BED ***

M563 P0 D0 H1 ; Define tool 0
G10 P0 S210 R-273 ; Set tool 0 operating and standby temperatures\

; *** EXTRUDER ***
M301 H1 P12 I0.4 D80 W180 S0.6 B300 ; Set tool 0 PID parameters

; *** EXTRUDER ***

M92 E310 ; Set extruder steps per mm

Ive spent several hours playing with the S parameter as they suggested, and nothing works. As soon as the bed gets to a set temp, and the extruder is turned on, the printer turns off as the PSU trips, and an AC power cycles is required. Here's the PSU rating plate if it's any help.



Does anyone have any ideas I can try tonight before I ring them again in the morning???

@GreatOldOne

I measured Fisher's power consumption to build a UPS.
In idle Fisher takes 6 W to 7 W of power. During the heat up of hotend, it takes 60 W and then stays at 40 W for the whole printing session.





I suggest you start printing while turning off the heated bed, then set the temperature of heated bed at 60 C in printing.
If Fisher stops, you have a under rated power supply or faulty heated bed.

My understanding is most cheap power supplies provide 70% to 80% of the max power.

Edited 2 time(s). Last edit at 01/20/2017 10:43AM by janpenguin.

Thanks Jan

I've been speaking with Jean at eMaker, where the bed & Supply was sourced from. He asked me to measure the resistance of bothe the extruder and bed heaters - which where 5.7Ohms and 3.3Ohms respectively. Here's the reply:

"Both of your heaters have rather a low resistance. The cartridge heater in the nozzle isn't far out, but the bed resistance is definitely too low. With your current setup, the nozzle is drawing 63.3W (spec is 60W), and your heatbed is drawing over 109W. The power supply is rated for 120W.

I'll check our stocks and send you a replacement."

So that's why the PSU is tripping. I'm no waiting for the replacement(s) to turn up.

Hi,

My new power supply that came with heated bed kit trips too.
I have not contacted eMaker about the problem but it sounds just the same. I control the heaters manually and found that if I enable the bed and the nozzle heater it will sometimes be ok, but if I try to home at the same time it trips.
I have adjusted my workflow to get around the problem. Firstly I enable the bed at 60 and clean it while it warms up, then enable the nozzle at 210, clean any remnants from that as it melts then home all, calibrate and then print. This is all with just PLA but I do now get consistent repeatable prints than stick well to the bed, I normally turn off the bed after 3 or 4 layers, it stays warm well anyway.

I was planning on making my own PSU or possibly paralleling this one with the old one via some sort of load splitting circuit if needed.

Please keep us updated with any news on replacement parts as this is very annoying if I forget to do everything in the right order.

Steve.

Will do, but as of today I've not heard anything back from eMaker after the original reply; I've sent a reminder to them asking for an ETA