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

Try to delete/rename c:\user\%username%\AppData\Local\RepetierHost\ directory before starting repetier-host.

---

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

So last night I finally got a perfect delta calibration print (this one [www.thingiverse.com]). I'm pretty psyched, but I notice my prints are still not as smooth as I would like. And by that I mean that they have a physically rough texture. I'm wondering if this is due to the jiggle of the traxxas connectors and the error in length between them, and was considering getting a set of preassembled magnetic arms. Has anyone gone this route? Have you gotten better prints from magnetic arms? Anyone tried it and regretted it?

Also, can anyone recommend bridging settings? Somewhere along the way I must have screwed mine up, as the hollow cube that printed great on my first try is no longer bridging properly.

Edited 1 time(s). Last edit at 03/01/2016 04:32PM by spiffcow.

can you attach an example of such rough texture (photo)?

Quote
spiffcow
So last night I finally got a perfect delta calibration print (this one [www.thingiverse.com]). I'm pretty psyched, but I notice my prints are still not as smooth as I would like. And by that I mean that they have a physically rough texture. I'm wondering if this is due to the jiggle of the traxxas connectors and the error in length between them, and was considering getting a set of preassembled magnetic arms. Has anyone gone this route? Have you gotten better prints from magnetic arms? Anyone tried it and regretted it?

Also, can anyone recommend bridging settings? Somewhere along the way I must have screwed mine up, as the hollow cube that printed great on my first try is no longer bridging properly.

---

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

PS: try to print something in vase mode (spiral, single wall) - is the external surface acceptable in this case?

---

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

For me slower is better for bridging. I have mine set to 20mm/s.

Quote
spiffcow
So last night I finally got a perfect delta calibration print (this one [www.thingiverse.com]). I'm pretty psyched, but I notice my prints are still not as smooth as I would like. And by that I mean that they have a physically rough texture. I'm wondering if this is due to the jiggle of the traxxas connectors and the error in length between them, and was considering getting a set of preassembled magnetic arms. Has anyone gone this route? Have you gotten better prints from magnetic arms? Anyone tried it and regretted it?

Also, can anyone recommend bridging settings? Somewhere along the way I must have screwed mine up, as the hollow cube that printed great on my first try is no longer bridging properly.

Quote
paul_delta
PS: try to print something in vase mode (spiral, single wall) - is the external surface acceptable in this case?

Here's one pic.. I'm in the middle of a long print job, but I'll try vase mode when its done. The fan in the pic isn't in vase mode, but i'm pretty happy with how it came out. The cube just didn't bridge at all though, and the tops of things are really rough if they have infill underneath. Also, the infill looks messy during prints.

EDIT

I printed a decent sized object at 0.1mm and got amazingly smooth results. So I think I'm done calibrating

Edited 1 time(s). Last edit at 03/02/2016 02:41AM by spiffcow.

It's normal. In such cases just increase the number of top layers. Some advanced slicers like Simplify3D allow to set different settings depending on layer height.
For complex object it's very handy to change the speed, infill or layer height for different part (height) of object.
For example the object has big overhang angle somewhere in the middle. To save time we can print it with 0.3mm layer but change layer height to 0.1mm for overhang part of model, etc.
Check private message with the link to S3D.

Quote
spiffcow

Quote
paul_delta
PS: try to print something in vase mode (spiral, single wall) - is the external surface acceptable in this case?

Here's one pic.. I'm in the middle of a long print job, but I'll try vase mode when its done. The fan in the pic isn't in vase mode, but i'm pretty happy with how it came out. The cube just didn't bridge at all though, and the tops of things are really rough if they have infill underneath. Also, the infill looks messy during prints.

EDIT

I printed a decent sized object at 0.1mm and got amazingly smooth results. So I think I'm done calibrating

---

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

@spiffcow: Amazing result! Looks very smooth. If you still want to reduce Traxxas joint jiggle - use springs close to the traxxas joints to pull the rods towards each other *slightly* - just enough to get rid of the jiggle, not so much as to deform the rods. That was one of the very first mods I made - though I dont know the impact. A nice set of springs is like 15 bucks (http://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20160302112653&SearchText=spring).
I dont have magnet rods yet, but I am reaaaally thinking about that mod. Wouldn't be to hard. I already have the rods - I would just need to add magnets (4 bucks) and steel balls (1 buck). I do think I'll do that.

Regular rubber bands and paper clip do the trick )

Quote
docpayce
@spiffcow: Amazing result! Looks very smooth. If you still want to reduce Traxxas joint jiggle - use springs close to the traxxas joints to pull the rods towards each other *slightly* - just enough to get rid of the jiggle, not so much as to deform the rods. That was one of the very first mods I made - though I dont know the impact. A nice set of springs is like 15 bucks (http://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20160302112653&SearchText=spring).
I dont have magnet rods yet, but I am reaaaally thinking about that mod. Wouldn't be to hard. I already have the rods - I would just need to add magnets (4 bucks) and steel balls (1 buck). I do think I'll do that.

---

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

So a few days ago I asked you guys for help diagnosing a print and adjusting my settings, it fixed some of the issues, like the stringy artifacts, some blobs, and a "droopy" chin. One of the issues was lines in the layers. This issue is still persisting. Below is a link to an album consisting of 2 images. Both images are of 3 bulbasaur prints, one's with flash, the other isn't. Most of the problems are on the forehead and bulb part of the prints.

Bulbasaur Trio

Starting from the left:

Far left is the original print. While it had it's own problems, it had the best forehead and bulb. It still had some layer gaps, but it wasn't pressing.

Middle one is with adjusted speeds, temperature, and extruder modifier set to 1. This fixed all the other problems but the layer issues and seem to have made them a bit worse.

Right one is with an extruder modifier of 1.075. It seems to be the worse one for layer gaps as it has them on the side, across the face, and even one or two on the legs.

What could cause these layer gaps?

---

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

If your extrusion with is calibrated - there are another possible sources of this problem : temperature, speed and layer height.
Layers tend to shift on complex surfaces with big angle - mostly because of filament contraction. Try to reduce the speed first and then layer height.
From my experience it always takes the same time to get good quality print:
- for big layer height you have to reduce speed
- for good speed you have to reduce layer height

Quote
Captain_Tim
What could cause these layer gaps?

---

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

PS: also you may try to increase extrusion with. It should increase contact surface and help layers to stick.

Edited 4 time(s). Last edit at 03/02/2016 06:08PM 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)

Fun fact! The revision of Marlin that FoglerTech has in their file store isn't actually based on a full-spec Marlin 1.0.x, and so DC42's site will help to calculate the smooth rod length, but not the endstop offset, since the M665 and M666 commands are literally not in the code base of that repo (okay, DC42's site will technically still calculate it, but you won't be able to do anything with them without moving to newer firmware).

Anyhow, after finally fixing some issues with my hot end and getting as much mechanically tuned as I can, I've moved onto some mid-scale prints that print wonderfully for the first 1.5-2 cms, and them go crap (notably, a 50% scale verison of [www.thingiverse.com]). I finally diagnosed that at about that point, height wise, the head seems to be dragging over the previous layers for a couple slices, getting out of whack, until the origin is out of sync and then it starts printing normally again just shifted off center, producing some unintended real-world glitch art.

My best bet right now is that the motors / stepper boards are overheating and skipping steps, so I'd love to know what others are using for Vref if using a standard FT RevB kit. However, my other question is, if I am in fact getting steps skipped, wouldn't I expect to see it in artifacts in all three dimensions, and not just a compression of the Z-dimension?

Edited 1 time(s). Last edit at 03/04/2016 12:51PM by foryst.

Quote
foryst
Fun fact! The revision of Marlin that FoglerTech has in their file store isn't actually based on a full-spec Marlin 1.0.x, and so DC42's site will help to calculate the smooth rod length, but not the endstop offset, since the M665 and M666 commands are literally not in the code base of that repo (okay, DC42's site will technically still calculate it, but you won't be able to do anything with them without moving to newer firmware).

Anyhow, after finally fixing some issues with my hot end and getting as much mechanically tuned as I can, I've moved onto some mid-scale prints that print wonderfully for the first 1.5-2 cms, and them go crap (notably, a 50% scale verison of [www.thingiverse.com]). I finally diagnosed that at about that point, height wise, the head seems to be dragging over the previous layers for a couple slices, getting out of whack, until the origin is out of sync and then it starts printing normally again just shifted off center, producing some unintended real-world glitch art.

My best bet right now is that the motors / stepper boards are overheating and skipping steps, so I'd love to know what others are using for Vref if using a standard FT RevB kit. However, my other question is, if I am in fact getting steps skipped, wouldn't I expect to see it in artifacts in all three dimensions, and not just a compression of the Z-dimension?

IIRC the movement motors are 0.4 amp. I replaced my drivers with a set of DRV8825 and targetted 0.2V for the movement steppers (hard to get it that low without it dropping off completely..) and 0.7 for the extruder. I think the calculations are different for the Polulu drivers though. I'm getting amazing results now, so I highly recommend it.

Also, if you're changing your angles based on dc42's calculator, make sure and print a delta calibration object to verify the angles. The calculator only has the heights to go off of, where an actual calibration object can tell you exactly your printer *thinks* your angles are. If you're off by more than 0.5 degrees it's a good indicator something is physically wrong with your build to the point that software correction won't fix it. Trust me on this, I've learned it the hard way.

Edited 1 time(s). Last edit at 03/04/2016 01:27PM by spiffcow.

Quote
foryst
My best bet right now is that the motors / stepper boards are overheating and skipping steps, so I'd love to know what others are using for Vref if using a standard FT RevB kit. However, my other question is, if I am in fact getting steps skipped, wouldn't I expect to see it in artifacts in all three dimensions, and not just a compression of the Z-dimension?

I went as far as to look at the resitors on the driver boards with a microscope to check the values. *If* you received the same drivers (I would bet), the exact voltages are 0.32 V for the small NEMA17's (gives 0.4 amps) and 0.64ish for the big extruder one (0.8A - can go to 1.2A if you want to).

Note that the print you are aiming for is reaaaally hard to do, especially at 50% scale. The small overhanging parts tend to stick up a bit (if not cooled VERY efficiently) and get stuck on the printhead moving by on the next layer. (... ... that was a little bit hard to understand I fear. Did you get me?)
First thing I would do is get an easier model and check whether you run into problems there, too.

You have some photo of the print?

Also, if you had missed steps ... you would HEAR that. Very distinct and fear-inducing noise.

Another possibility (not likely tho): Are all GT2 belts at approximately the same tension?

Edited 1 time(s). Last edit at 03/04/2016 04:18PM by docpayce.

Try to reduce travel acceleration and printing speed. This model is travel-intensive and any misconfiguration will affect the print more and more with every next layer. Check your extrusion as well - overextrusion and high speed may result slight wrapping so dragging nozzle will accumulate positioning errors.

Quote
foryst
Anyhow, after finally fixing some issues with my hot end and getting as much mechanically tuned as I can, I've moved onto some mid-scale prints that print wonderfully for the first 1.5-2 cms, and them go crap (notably, a 50% scale verison of [www.thingiverse.com]). I finally diagnosed that at about that point, height wise, the head seems to be dragging over the previous layers for a couple slices, getting out of whack, until the origin is out of sync and then it starts printing normally again just shifted off center, producing some unintended real-world glitch art.

My best bet right now is that the motors / stepper boards are overheating and skipping steps, so I'd love to know what others are using for Vref if using a standard FT RevB kit. However, my other question is, if I am in fact getting steps skipped, wouldn't I expect to see it in artifacts in all three dimensions, and not just a compression of the Z-dimension?

Edited 2 time(s). Last edit at 03/04/2016 04:32PM 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
docpayce
I went as far as to look at the resitors on the driver boards with a microscope to check the values. *If* you received the same drivers (I would bet), the exact voltages are 0.32 V for the small NEMA17's (gives 0.4 amps) and 0.64ish for the big extruder one (0.8A - can go to 1.2A if you want to).

Note that the print you are aiming for is reaaaally hard to do, especially at 50% scale. The small overhanging parts tend to stick up a bit (if not cooled VERY efficiently) and get stuck on the printhead moving by on the next layer. (... ... that was a little bit hard to understand I fear. Did you get me?)
First thing I would do is get an easier model and check whether you run into problems there, too.

You have some photo of the print?

Also, if you had missed steps ... you would HEAR that. Very distinct and fear-inducing noise.

Another possibility (not likely tho): Are all GT2 belts at approximately the same tension?

Thanks for the driver readings. I'm pretty sure that that cements that I'm at least running hot, in addition to possible other things. As for the print, it certainly is a more complex one than some of the others I've done. My most successful/complex print was another reduced scale one of these (https://www.thingiverse.com/thing:1087797) which printed mostly accurately, in addition to a number of progressively better calibration cubes.

I haven't heard any crazy noises that would sound like missed steps, but I will admit one of my belts is noticeably tighter than the others, from some post-build tweaks involving better leveling of the top triangle. If the motor tuning and further tuning of extrusion amount as suggested by paul_delta doesn't work, I'll try loosening that belt by a tooth or two.

Quote
foryst

Quote
docpayce
I went as far as to look at the resitors on the driver boards with a microscope to check the values. *If* you received the same drivers (I would bet), the exact voltages are 0.32 V for the small NEMA17's (gives 0.4 amps) and 0.64ish for the big extruder one (0.8A - can go to 1.2A if you want to).

Note that the print you are aiming for is reaaaally hard to do, especially at 50% scale. The small overhanging parts tend to stick up a bit (if not cooled VERY efficiently) and get stuck on the printhead moving by on the next layer. (... ... that was a little bit hard to understand I fear. Did you get me?)
First thing I would do is get an easier model and check whether you run into problems there, too.

You have some photo of the print?

Also, if you had missed steps ... you would HEAR that. Very distinct and fear-inducing noise.

Another possibility (not likely tho): Are all GT2 belts at approximately the same tension?

Thanks for the driver readings. I'm pretty sure that that cements that I'm at least running hot, in addition to possible other things. As for the print, it certainly is a more complex one than some of the others I've done. My most successful/complex print was another reduced scale one of these (https://www.thingiverse.com/thing:1087797) which printed mostly accurately, in addition to a number of progressively better calibration cubes.

I haven't heard any crazy noises that would sound like missed steps, but I will admit one of my belts is noticeably tighter than the others, from some post-build tweaks involving better leveling of the top triangle. If the motor tuning and further tuning of extrusion amount as suggested by paul_delta doesn't work, I'll try loosening that belt by a tooth or two.

Watch the output log for something like "delta movement error" or "bad checksum". If you see that there's a good chance that you need to check your settings and/or you're getting missed steps.

Another thing to check for -- make sure your pulleys are all secure. I had a sideways printing issue early on that turned out to be an issue with one of the pulleys slipping ever so slightly.

Z-probe offset. I printed paul_deltas magnet z probe carriage for the rev. B kossel. I tried different settings in the following configuration.h file to input the new offsets without success, just alot of aborted bed crashes of my new sensor mount. The current and original restored settings are below.

#define Z_PROBE_OFFSET {0,0-.6,0} // (0,14, -6.5,0) {24.6,-21,0,0} {24.6,-21,0,0} X, Y, Z, E distance between hotend nozzle and deployed bed leveling probe. (0,19-8,0)

So, am I even on the right track?

On a side note to paul_delta your magnetic couplings are a god send , in that they are almost idiot proof. They gave way and protected my mechanicals on each of my aborted autoleveling attempts!

What firmware? You have to adjust your z-probe height - it will be around 100mm, depending on how close your probe to the hotend tip. If it doesn't work for configuration.h - check if your Eeprom is ON. I am not very familiar with Marlin but I guess it has some settings in Eeprom for z-probe that overrides your configuration.h. In Repetier it's "Z-probe height" option.

Quote
Gingerbreadman
Z-probe offset. I printed paul_deltas magnet z probe carriage for the rev. B kossel. I tried different settings in the following configuration.h file to input the new offsets without success, just alot of aborted bed crashes of my new sensor mount. The current and original restored settings are below.

#define Z_PROBE_OFFSET {0,0-.6,0} // (0,14, -6.5,0) {24.6,-21,0,0} {24.6,-21,0,0} X, Y, Z, E distance between hotend nozzle and deployed bed leveling probe. (0,19-8,0)

So, am I even on the right track?

On a side note to paul_delta your magnetic couplings are a god send , in that they are almost idiot proof. They gave way and protected my mechanicals on each of my aborted autoleveling attempts!

---

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

PS: if you will figure it out - let me know, I'll add your settings example to Thingiverse description.

---

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

Thanks for the quick reply. I've been uploading the folgertech configuration.h file to repetier host each time print. I have never flashed my eeprom thinking to take baby steps first, especially since I don't know what I'm doing.

Marlin firmware’s Configuration.h, last attempt update, I changed from the folger tech default settings in "Z_probe offset" to {0,0,-69.4,0} based on info from a m114 command. When I run a g29 it attempts to plow through the bed.

Quote
Gingerbreadman
Z-probe offset. I printed paul_deltas magnet z probe carriage for the rev. B kossel. I tried different settings in the following configuration.h file to input the new offsets without success, just alot of aborted bed crashes of my new sensor mount. The current and original restored settings are below.

#define Z_PROBE_OFFSET {0,0-.6,0} // (0,14, -6.5,0) {24.6,-21,0,0} {24.6,-21,0,0} X, Y, Z, E distance between hotend nozzle and deployed bed leveling probe. (0,19-8,0)

So, am I even on the right track?

On a side note to paul_delta your magnetic couplings are a god send , in that they are almost idiot proof. They gave way and protected my mechanicals on each of my aborted autoleveling attempts!

There is something is wrong with your settings, as I said, check your Eeprom, make sure that your Z max and z-probe offset from the hotend tip is correct.
Check if your z-probe returns correct condition when triggered.

Quote
Gingerbreadman
Marlin firmware’s Configuration.h, last attempt update, I changed from the folger tech default settings in "Z_probe offset" to {0,0,-69.4,0} based on info from a m114 command. When I run a g29 it attempts to plow through the bed.

Quote
Gingerbreadman
Z-probe offset. I printed paul_deltas magnet z probe carriage for the rev. B kossel. I tried different settings in the following configuration.h file to input the new offsets without success, just alot of aborted bed crashes of my new sensor mount. The current and original restored settings are below.

#define Z_PROBE_OFFSET {0,0-.6,0} // (0,14, -6.5,0) {24.6,-21,0,0} {24.6,-21,0,0} X, Y, Z, E distance between hotend nozzle and deployed bed leveling probe. (0,19-8,0)

So, am I even on the right track?

On a side note to paul_delta your magnetic couplings are a god send , in that they are almost idiot proof. They gave way and protected my mechanicals on each of my aborted autoleveling attempts!

---

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

PS: looks like you are using obsolete version of marlin (FT version?). If you want to stick with Marlin - I'd suggest to download fresh version from [github.com]. Or better switch to Repetier...

---

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

Can i simple migrate my actual marlin configuration.h to reptier or need to see something?

@Gingerbreadman: -69.4 seems to low for pauls z probe. It should be in a range of -100 ... -130 mm. Home (G28), lower the z-probe manually until it switches on and note the z-position. Thats your offset you need to put in. Also always use G28 before G29 (see documentation: G29
Hope that helps. :/ And do consider switching over to Repetier.

@foryst: Leveling the top triangle is *NOT* as important as you might think. The position of that triangle and the endstops is easily compensated by endstop offsets in Repetier and Marlin. Especially if you use tools such as dc42's (Escher) delta calculator.
I suggest to loosen the top triangle at one tower, pull it upwards quite strongly to tension the belt and fasten the triangle-to-tower-beam screw while holding that tension. Do that on all three towers. The resulting "tone" of the tensioned belts should be a low to medium bass tone - not to tight, not loose. It is not as important that all belts hit exactly the same tone (though certainly being a plus), as GT2 belts are very stiff and will give you nice 80 steps/mm for a reasonable tensioning range.
If your belt is to loose, you will have huge issues with backlash, which will eff up a LOT. Especially in Deltas.

Edited 1 time(s). Last edit at 03/05/2016 06:20PM by docpayce.

Quote
PrimeraRata
Can i simple migrate my actual marlin configuration.h to reptier or need to see something?

Basically yes, you use the same values to a large part. You need to configure your configuration.h from scratch though. Which is easy given the tools Repetier has on-board:
http://www.repetier.com/firmware/v092/

Edited 1 time(s). Last edit at 03/05/2016 06:19PM by docpayce.

No need to configure it from scratch - use my configuration that you can find in the footer of this message. Use it as is (the whole archive) or just take the configuration.h file, upload it to repetier configurator [www.repetier.com] and adjust for your needs.

Edited 2 time(s). Last edit at 03/05/2016 08:03PM 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)

I downloaded your config and reading i found #define DELTA_DIAGONAL_ROD 230 // mm/ My diagonal rod is 240 mm, what more i need to change?