Folger Tech Kossel Delta 2020 Full 3D Printer Kit w/Auto-Level

As a student completely my 3 year advanced diploma for programming, and a 3D enthusiast, I for one would be highly interested in helping out with this, if I had more time.

Quote
spiffcow
Also, I'd be happy to collaborate on this project if you're interested. Having active help increases the priority of this project for my free time allotment

Quote
spiffcow
Does Repetier Host support plugins? If so I might do that.. Or submit a patch to them if they don't.

Quote
paul_delta
PS: I understand that serial communication is a piece of work... I think the best place for such calculator/calibrator- is the sub-menu in Repetier Host (right under Height map menu) that has all needed procedures to communicate with printer.

---

Owner of a QUBD Two Up
And
Folger Kossel 2020
And
Tevo Tarantula

Not a problem, Google is everyone's friend.

Quote
spiffcow
Awesome! That means minimal work to convert it!

I salute you, Captain Tim

Quote
Captain_Tim
I believe that Repetier Host does indeed support custom plugins, as This Site seems to point to them being made in C++ by my guess (would make sense to be C++ as Repetier runs on Windows, Mac, and Linux) C#, as the page says it's just a C# Library being used.

Quote
spiffcow
Does Repetier Host support plugins? If so I might do that.. Or submit a patch to them if they don't.

Quote
paul_delta
PS: I understand that serial communication is a piece of work... I think the best place for such calculator/calibrator- is the sub-menu in Repetier Host (right under Height map menu) that has all needed procedures to communicate with printer.

---

Owner of a QUBD Two Up
And
Folger Kossel 2020
And
Tevo Tarantula

That's pretty good! I had to saw my carbon rods with a pocket knife, so I'm pretty certain there's a good deal of physical error in mine.

I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points. It would be fine if you could assume the physical build is perfect, but that's an assumption that will never actually be true (at least not in my case).

BTW, your videos helped a lot in my initial build. I didn't realize that was you. Thanks!

Quote
paul_delta
Actually, I had 0.15mm using default drivers and just one shot of 6-factor dc42 calculator Here is the video [youtu.be]

Quote
spiffcow
BTW, one thing I did during this process that made a *huge* difference was switch to DRV8825 stepper drivers. They're much quieter, double your step accuracy, and cost about $12 shipped for a set of 5. I received them after I had already done several passes of calibration, and the addition of the stepper drivers allowed me to go from ~0.25mm between the highest and lowest point down to ~0.15mm between highest and lowest. Well worth it IMO.

Quote
spiffcow
That's pretty good! I had to saw my carbon rods with a pocket knife, so I'm pretty certain there's a good deal of physical error in mine.

I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points. It would be fine if you could assume the physical build is perfect, but that's an assumption that will never actually be true (at least not in my case).

Does your probe have X/Y offset from the center of effector? Slight tilt and even small offset may cause z-probe values fluctuation across the bed.
I think 0.1-0.15 is the maximum we can get for this kind of construction because of many factors like slight backlash in joints, carriage rollers, etc.

Quote
spiffcow
BTW, your videos helped a lot in my initial build. I didn't realize that was you. Thanks!

No, the only videos I published - is this one and two more with attempt to print at 120mm/s.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

Weird, it took me right past your video on to the dude who did a bunch of FolgerTech build videos.. But it still said "paul delta" as the author. To be honest I didn't watch closely -- I figured I would watch in earnest when I get home.

Quote
paul_delta

Quote
spiffcow
BTW, your videos helped a lot in my initial build. I didn't realize that was you. Thanks!

No, the only videos I published - is this one and two more with attempt to print at 120mm/s.

Nope, I stick it directly in the hot end slot when I calibrate the printer to get minimal error. I also probe at 2mm/s. My probes are repeatable down to .02mm at the margins and .01mm across most of the bed. They take an incredibly long time though. A 60 probe run (6 turns, 10 points per turn) takes about 30-40 minutes.

Quote
paul_delta

Quote
spiffcow
That's pretty good! I had to saw my carbon rods with a pocket knife, so I'm pretty certain there's a good deal of physical error in mine.

I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points. It would be fine if you could assume the physical build is perfect, but that's an assumption that will never actually be true (at least not in my case).

Does your probe have X/Y offset from the center of effector? Slight tilt and even small offset may cause z-probe values fluctuation across the bed.
I think 0.1-0.15 is the maximum we can get for this kind of construction because of many factors like slight backlash in joints, carriage rollers, etc.

Quote
spiffcow
I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points.

There is NO LIMIT to the number of data points you can use in the web calculator, if you are patient enough to wait for the computation to finish.

Edited 1 time(s). Last edit at 02/16/2016 02:14PM 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].

Quote
dc42

Quote
spiffcow
I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points.

There is NO LIMIT to the number of data points you can use in the web calculator, if you are patient enough to wait for the computation to finish.

Apologies again, you are correct. I was looking at the directions in step 4, but did not try just setting a higher value.

Is there a significant delay with more points? My app is just a port of your code to C#, and calculation of 100 probe points is nearly instantaneous.

Quote
spiffcow

Quote
dc42

Quote
spiffcow
I was able to get a very small error in the dc42 calculator, but it fell apart once I added more data points. I once got it down to .05mm in the dc42 calculator, but when I actually ran G33 there were other points with a lot higher error. I think that's the big weakness in the dc42 web calculator, in that it just doesn't allow enough data points.

There is NO LIMIT to the number of data points you can use in the web calculator, if you are patient enough to wait for the computation to finish.

Apologies again, you are correct. I was looking at the directions in step 4, but did not try just setting a higher value.

Is there a significant delay with more points? My app is just a port of your code to C#, and calculation of 100 probe points is nearly instantaneous.

Also, I notice there is a copyright notice now that I didn't notice before: "These calculators are COPYRIGHT Escher Technologies Ltd. Please don't steal them!" Does your previous statement (that the code is GPL) hold true? If not I will take down my app. Likewise if you personally would prefer that I remove it until I re-implement the calculations on my own. I mean no offense by deriving from your work.

I'll take a look at converting to a plugin this weekend and keep you informed. I might also look into converting their bed height calculator to use better points for delta printers (the whole grid-based height map doesn't really correspond to where you typically find errors for delta printers after all).

Quote
Captain_Tim
As a student completely my 3 year advanced diploma for programming, and a 3D enthusiast, I for one would be highly interested in helping out with this, if I had more time.

Quote
spiffcow
Also, I'd be happy to collaborate on this project if you're interested. Having active help increases the priority of this project for my free time allotment

Quote
spiffcow
Does Repetier Host support plugins? If so I might do that.. Or submit a patch to them if they don't.

Quote
paul_delta
PS: I understand that serial communication is a piece of work... I think the best place for such calculator/calibrator- is the sub-menu in Repetier Host (right under Height map menu) that has all needed procedures to communicate with printer.

Do keep us informed. I for one am interested to see this type of plugin implemented, and it would be a life saver for a lot of people using Delta printers. After the basic plugin is made, perhaps auto-corrections to the EEPROMs could be the next feature. So when they start to probe the bed, it will take the height error and do the math, then ask the user if they'd like to "change the EEPROM values to the following: (List the new EEPROM Values)". That would really streamline it.

Quote
spiffcow
I'll take a look at converting to a plugin this weekend and keep you informed. I might also look into converting their bed height calculator to use better points for delta printers (the whole grid-based height map doesn't really correspond to where you typically find errors for delta printers after all).

---

Owner of a QUBD Two Up
And
Folger Kossel 2020
And
Tevo Tarantula

Quote
spiffcow
I'll take a look at converting to a plugin this weekend and keep you informed. I might also look into converting their bed height calculator to use better points for delta printers (the whole grid-based height map doesn't really correspond to where you typically find errors for delta printers after all).

It sounds promising!

PS: I agree that the grid system doesn't work well and set of circles will work best for delta.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

Yep, that would be pretty awesome! You could just set it on auto-calibrate and stop it once it's as close as you want it. It would also allow testing targeted points for a settings change (e.g. test an angle correction based on the expected output of the previous data points to determine if it corrected the problem).

I'm getting kinda psyched about this project.

Quote
Captain_Tim
Do keep us informed. I for one am interested to see this type of plugin implemented, and it would be a life saver for a lot of people using Delta printers. After the basic plugin is made, perhaps auto-corrections to the EEPROMs could be the next feature. So when they start to probe the bed, it will take the height error and do the math, then ask the user if they'd like to "change the EEPROM values to the following: (List the new EEPROM Values)". That would really streamline it.

Quote
spiffcow
I'll take a look at converting to a plugin this weekend and keep you informed. I might also look into converting their bed height calculator to use better points for delta printers (the whole grid-based height map doesn't really correspond to where you typically find errors for delta printers after all).

A lot of people stay away from Delta printers because of the calibration aspect of them. This could open the door to a lot of people who don't want to deal with the often time, long and stressful action of calibrating a delta. Then delta printers could be more appealing to new enthusiasts then regular Cartesian printers (let's be honest, deltas look a lot cooler anyways).

Quote
spiffcow
Yep, that would be pretty awesome! You could just set it on auto-calibrate and stop it once it's as close as you want it. It would also allow testing targeted points for a settings change (e.g. test an angle correction based on the expected output of the previous data points to determine if it corrected the problem).

I'm getting kinda psyched about this project.

Quote
Captain_Tim
Do keep us informed. I for one am interested to see this type of plugin implemented, and it would be a life saver for a lot of people using Delta printers. After the basic plugin is made, perhaps auto-corrections to the EEPROMs could be the next feature. So when they start to probe the bed, it will take the height error and do the math, then ask the user if they'd like to "change the EEPROM values to the following: (List the new EEPROM Values)". That would really streamline it.

Quote
spiffcow
I'll take a look at converting to a plugin this weekend and keep you informed. I might also look into converting their bed height calculator to use better points for delta printers (the whole grid-based height map doesn't really correspond to where you typically find errors for delta printers after all).

---

Owner of a QUBD Two Up
And
Folger Kossel 2020
And
Tevo Tarantula

Quote
spiffcow
[Is there a significant delay with more points? My app is just a port of your code to C#, and calculation of 100 probe points is nearly instantaneous.

AFAIR the time taken to do the maths is proportional to the number of factors you are adjusting and the square of the number of points.

Quote
spiffcow
Also, I notice there is a copyright notice now that I didn't notice before: "These calculators are COPYRIGHT Escher Technologies Ltd. Please don't steal them!" Does your previous statement (that the code is GPL) hold true? If not I will take down my app. Likewise if you personally would prefer that I remove it until I re-implement the calculations on my own. I mean no offense by deriving from your work.

I am happy for you to use and adapt the code if you acknowledge me as the source of the original.

---

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
spiffcow
Also, I notice there is a copyright notice now that I didn't notice before: "These calculators are COPYRIGHT Escher Technologies Ltd. Please don't steal them!" Does your previous statement (that the code is GPL) hold true? If not I will take down my app. Likewise if you personally would prefer that I remove it until I re-implement the calculations on my own. I mean no offense by deriving from your work.

I am happy for you to use and adapt the code if you acknowledge me as the source of the original.

Thank you. I do mention your calculator in the source code, but I will add a reference to the readme file as well

Quote
Sven.L

Nebbian:
I am just not sure about anything. One the one side, Paul tells me that the probe has a cable going to the RAMPS and another cable goes to RAMPS from the switch. On the other hand, my manual says I need to connect the orange-red-brown cable (from the probe) to the white-red-black cable (with the cut-off white wire, from the switch). But that also means, that there is no wire at all going to the RAMPS. STRAAANGE!!! And there is no additional slot for any connection to the RAMPS. Could it be, that the manual of Folger Tech tells me wrong things? HEEEEEELLLLPP!!!!!
[attachment 72667 Photo7499b.jpg]

This is, how it is now, based on the Folger Tech instructions:
[attachment 72668 Photo7497.jpg] [attachment 72669 Photo7499.jpg]

I really don't understand what's missing here. You have the switch, which has a white, black and red wire. This is plugged into the orange, red and brown cable. The other end of that orange, red and brown cable should be plugged into the RAMPS, in the Z Min position.

As far as I can tell you're not missing any parts.

Sorry, can't resist to post this picture

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

OK Sven L., read my lips

1. You have a blue servo that drops the probe (black mechanical clicking switch) to the working position when you need it to calibrate the printer.
2. You need to connect this servo to the RAMPS - directly to the 3-pin connector near the RAMPS reset switch, using extension cable because the servo wire is too short to reach RAMPS.
3. You have a mechanical switch/probe (black) that you need to connect to the RAMPS z-min ground (-) and signal (S) pins. You can do it using 2 or 3 pin extension cable. If you use 3 pin cable, you need only 2 wires, so in this case you just don't use the white one and solder black and red to the mechanical switch. So, for this EXTRA extension cable you need to cut one end (opposite to the end that you will connect to RAMPS pins) and solder black and red wires of this end to the probe (mechanical black switch).

That's it.

Edited 3 time(s). Last edit at 02/16/2016 08:50PM by paul_delta.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

Having problems when I run g29.
The head moves as though it is scanning the bed, in the x and y but steps up incrementally on the z until it hits the endstops at home. At that point I have to kill switch it or it will jam against the end stops.
I'm using repetier host v1.6.0, folgertech kossel rev b with the inductive probe, and the folgertech configuration rev.b file. I've set bed height and the z probe register correctly (lights up and shows in the log) . The g28 code works well, and the printer responded correctly when I ran a test calibration spiral print, though that was a sad print to say the least.
I'm sure I have wired or miscoded something since I am thrashing in the dark here and would like to preemptively apologize if the solution to this problem was offered earlier in this thread . My sad excuse for a 3d vocabulary hinders my search and comprehension of what little I have found and "understood"...thanks

PS: as far as I can see from your picture - you destroyed extension cable because you cut wrong end.

Edited 1 time(s). Last edit at 02/16/2016 08:57PM by paul_delta.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

Hey!

Just a short 'hello' from my side. I got my Kossel 2020 this christmas and got it working reliably...ish now. Happy to be of help if I can.
btw (kinda commercial break ): If you don't know what to do with your spare endstop, use it as a z probe: [www.thingiverse.com]

And a question:
In the discussion on dc42's / spiffcow's calculators and whether to use 6 or 7 factor calibration, paul pointed out that this will throw off x/y accuracy. Is this really the case? I changed delta diag rod length from 238 mm to 242 mm (both time doing dc42's 6-factor, 10 point calibration) and can't see a significant difference, thou with the 242 mm print-outs should be 97% the dimensions. Will the radius correction just cancel out any changes to the rod length? In that case, isn't 7-factor always to be chosen more or less?

BR!

Edited 1 time(s). Last edit at 02/17/2016 03:26AM by docpayce.

I had this same problem. Try uncommenting this line in your configuration.h settings: #define DISABLE_MIN_ENDSTOPS

Quote
Gingerbreadman
Having problems when I run g29.
The head moves as though it is scanning the bed, in the x and y but steps up incrementally on the z until it hits the endstops at home. At that point I have to kill switch it or it will jam against the end stops.
I'm using repetier host v1.6.0, folgertech kossel rev b with the inductive probe, and the folgertech configuration rev.b file. I've set bed height and the z probe register correctly (lights up and shows in the log) . The g28 code works well, and the printer responded correctly when I ran a test calibration spiral print, though that was a sad print to say the least.
I'm sure I have wired or miscoded something since I am thrashing in the dark here and would like to preemptively apologize if the solution to this problem was offered earlier in this thread . My sad excuse for a 3d vocabulary hinders my search and comprehension of what little I have found and "understood"...thanks

I'm having a problem where my print space is being limited somewhere to a square about 155-ish mm (~6in) and looks something like this (it's supposed to be printing a circle):



Can someone please tell me where this is happening? I have my print space (bed shape) set to circular in slicer with a diameter of 200mm, so this shouldn't be happening as far as I can tell... Is there something in the firmware limiting this?

YES



P.S.: If you tell us what Firmware you are using, we might be able to help you... :-)

Quote
Legrand
I'm having a problem where my print space is being limited somewhere to a square about 155-ish mm (~6in) and looks something like this (it's supposed to be printing a circle):

...

Can someone please tell me where this is happening? I have my print space (bed shape) set to circular in slicer with a diameter of 200mm, so this shouldn't be happening as far as I can tell... Is there something in the firmware limiting this?

Thanks again, Nebbian and Paul_Delta.

First, my own language is German, that is why even despite of all the reading of your answers I am doing, not everything I can do catch - many many words in it that I don't understand. Google Translate does offer translations (I use it often for your postings), but the arriving sense is often distorted. I simply cannot capture all meanings and hints in your postings. It is like reading a book in another language: yes, it can be read, and you think you understood most, but still, not everything of the contents is captured as easily as if the book was in your own language. Fault on my side. Just be patient with me, despite of that.

When seeing the photos inside my building manual and the text that instructed me to cut the white wire as it is not needed, I did so. The manual also said that the white-red-black wire needs to be connected to the orange-red-brown wire from the probe, which I did consequently, too, of course. It does not speak about more than one white-red-black wire. That was before posting inside this forum, at that time I just referred to the manual alone, not yet to any Wiki.

Reading then all your answers, my understanding is, that it seems I am short by one cable: another white-red-black cable. The cut white-red-black wire from the switch goes, as I am understanding now, to the RAMPS board. And the un-cut white-red-black wire from the orange-red-brown probe wire goes also to the RAMPS board. Which is much more logical to me than the result based on the building manual.

If my understanding is correct, I am short of one (un-cut) white-red-black wire. Even worse, I went to many shops but had to fund out that I cannot purchase such things here in Jamaica. But I would be willing to pay with PayPal to someone, who can prepare a simple letter envelope and send such a cable to me. To afford 14 to 50 US$ for shipping, it feels not so right, too, if the value of the cable is only 2 US$ and if a regular letter for 2 or maximum 5 US$ can do the job. In my private life, I do seed swaps with many people all around the world, and thus I know the prices of regular letters from the USA or Europe or Asia to Jamaica. And such a cable is not thaaat sensitive, it can be sent inside a regular letter.

(Side Question: is there any German building manual that is also accurate?)

Edited 2 time(s). Last edit at 02/17/2016 08:12AM by Sven.L.

Sorry. Should have been more detailed.

I'm using the FT marlin firmware.

Quote
Sir_Death
YES



P.S.: If you tell us what Firmware you are using, we might be able to help you... :-)

Quote
Legrand
I'm having a problem where my print space is being limited somewhere to a square about 155-ish mm (~6in) and looks something like this (it's supposed to be printing a circle):

...

Can someone please tell me where this is happening? I have my print space (bed shape) set to circular in slicer with a diameter of 200mm, so this shouldn't be happening as far as I can tell... Is there something in the firmware limiting this?

Quote
Sven.L
If my understanding is correct, I am short of one (un-cut) white-red-black wire.

No, you still need at least two of them because you killed your extension with wrong cut (you cut wrong side).

Quote
Sven.L
To afford 14 to 50 US$ for shipping, it feels not so right, too, if the value of the cable is only 2 US$ and if a regular letter for 2 or maximum 5 US$ can do the job. In my private life, I do seed swaps with many people all around the world, and thus I know the prices of regular letters from the USA or Europe or Asia to Jamaica. And such a cable is not thaaat sensitive, it can be sent inside a regular letter.

You are too obsessed with these wires. Buy them on eBay from China. It's $0.71 per 100cm cable with free shipping [www.ebay.com].

For now, while waiting for your wires from China, you can just skip this step and calibrate your printer manually.

Edited 3 time(s). Last edit at 02/17/2016 09:07AM by paul_delta.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

PS: Even more, this kind of probe will create more problems than regular manual calibration. This probe arm design is poor and will slip on servo gear. I'd suggest to redesign the arm to get something like ⌜-shape mount. This way it will be possible to have this black switch right under the nozzle - to avoid offset errors.

Edited 3 time(s). Last edit at 02/17/2016 09:21AM by paul_delta.

---

My [www.thingiverse.com] stuff, my configuration of Repetier 0.92.6: [forums.reprap.org] (use Arduino 1.6.5 to compile)

Quote
Legrand
Sorry. Should have been more detailed.

I'm using the FT marlin firmware.

I do not know FT Marlin, but I assume this function similar in all Marlin Versions.

search for line like this in configuration.h:

#define DELTA_PRINTABLE_RADIUS 100

The value at the end of the line is what you have to change...

Edited 1 time(s). Last edit at 02/17/2016 10:26AM by Sir_Death.