Folger Tech Kossel Unboxing and First Impressions (Pic Heavy) (Page 7 of 8)

try adding a washer between the wheel and the plastic carriage to see if that helps with the drag you are feeling.
I found that mine runs much more smoothly with a washer in there - gives just a smidgen more space between frame & carriage. It also helps with wheel alignment.

I have the Sintron carriages on order, might even arrive today if I am a bit lucky. Now that this has become an experiment might as well try a few options that might come in handy when I build a larger Kossel.
I also asked Folger Tech about their opinion and advise but so far all I got was a request for pictures of the assemblies. I guess this is still a somewhat new industry and so one cannot expect perfection, but the whole reason why I went with a kit instead of a scratch build is that I wanted to avoid some of the trial-and-error pitfalls. I know from experience how much (well, and money) these cost and paying for someones experience that has been put into a kit is usually well worth it.
We'll have to see how this one plays out after I solve the slider issue but I am not sure I made the right choice with this kit.

Speaking about instructions...
I was looking ahead to see what I can work on that doesn't in some way or another relate to the sliders and thought I would work a bit on the electronics.
I unzipped the firmware package and found 2 directories containing what appears to be the same version 1.0.0 of Marlin. One is called "KosselRevB" and the other "KosselRevBPE". No readme and also no PDF for the firmware install in this revision. I have the PDF of Rev. A but that would still not explain why the two directories and which one to go with.
Are you guys actually using Marlin or is Repetier the way to go for the Kossel?

I ran a file compare between those two folders. The firmware is identical. There are two lines different in the Configuration.h, the PE version is setting INVERT_E0_DIR to false for a geared extruder drive and it also sets MANUAL_Z_HOME_POS to 1120 versus 315 in the non-PE version.
Does this ring a bell with someone here as to what this PE might mean with these settings in mind?

So the carriages and the Delrin wheels arrived today. The Delrin wheels are twice as wide as they have two ball bearings and, of course, I didn't have any M5 bolts with 35 mm length lying around. So I could not try them yet but, man, these wheels are really high quality! I will get those bolts just to see how they roll along the 2020. I might want to use for my XXXL Kossel.

Now the Sintron carriage kit really surprised me. The quality of the printed parts is simply excellent and once I put on together for testing I was blown away.

The carriage slides onto the 2020 with zero force (something I cannot do with the FT carriages) and you actually have to tighten the tension screw quite a bit to reduce the play and have the carriage follow the rail precisely. But once that was done the carriage was sliding up and down with almost no force and was actually rolling down the rail by the force of its own weight. Hallelujah! That's what I was dreaming of!
Here's the video that proves it: https://plus.google.com/photos/10428891 … 4870373058

I would say this is sufficient :-). The Attachment block for the diagonal rods is 0.8mm narrower than on the FT ones, but I think I can tweak that by putting washers underneath the ball mounts.
I am definitely back in business and now I would only need time to finish the build. Apparently the weekend passed and I am back on the treadmill though.
"heartless" definitely saved the day with his hint towards the Sintron carriages. I will touch base with FT and see if maybe they have a problem with their supplier as the carriages seem too narrow for the 2020.

glad to hear you got decent carriages from Sintron - the ones I got were a mess, but the wheels are great - about the only redeeming factor out of that whole kit...
I re-printed the carriages in ABS for use on my FT kossel (basically reprinted everything in ABS, as I am not fond of PLA for printer parts)

I can also sympathize with the CS thing - they are a bit on the slow side - but they do take the comments and suggestions into consideration for future updates/upgrades (whether they respond specifically to you, or not). It is my understanding that there are only 1, or maybe 2, people that handle all of the email... they are a pretty small company... but yeah, even a simple "we will look into this" kind of response would be nice.

Why would you assume the print head moving towards the build platform to home is wrong?
The Z axis moving opposite your controls is also not uncommon but can be changed to your liking.

I get what your thinking, that would make for a very cumbersome start for each print.  The nozzle would have to travel to the farthest point away from the build platform to home.  Then once it knows where it's at travel the full length of Z again just to start printing?  Certainly not how I'd do it.  Perhaps there's endstops on both ends of the travel (which is not that uncommon) and there's nothing wrong with using a sensor as an endstop on one end for repeatability where it's needed.  Then using a cheaper mechanical switch as an endstop where it really doesn't matter.

I would advise you to use the FT firmware provided on their cloud drive for both printers as it is already closely preconfigured. It just needs a few tweaks based on your build decisions.

The normal process is as follows. Start Pronterface or host of your choice. Send a G28 command through the terminal and the system should move all three axis to the top of the towers and stop when they hit the endstops. Then move down a bit and back up slower just till they stop. NOTE: MAKE SURE YOU HAVE ADJUSTED ALL MOTORS TO THE PROPER CURRENT LIMIT. THEY WILL GET HOT ENOUGH TO DEFORM THE ABS CORNER PIECES IF THEY ARE SET TOO HIGH!!

You then have to measure two physical values and put them into the firmware. The first is the distance from the center of mounting screw on the effector to the center of the mounting screw for each slide on the carbon fiber rod. These should be very close to the same for all three towers. If they are very close then you can just average. If one is more than a few mm off then you need to cut the longer rods down to the size of the shorter one, or buy new rods and start over unless your rods have length adjusters that where not glued in place.

Once you get that value it needs to be placed in the firmware under the following line.

Define# DELTA_DIAGONAL_ROD 230 //mm (215) should be a value around this range..

You next need to measure the distance from the center of the nozzle to each tower face. This value also should be very close. If it is not then you need to adjust the height of the endstop switch for the tower that is off.

Once you get that averaged value it needs to be placed in firmware under this line.

Define# DELTA_SMOOTH_ROD_OFFSET 178 //mm (137) again this could be any number around this range depending or your build choices.

Once you set these values upload the firmware to the printer. Once it reboots, send another G28 home command and again all axis should move to the top as they did before.

Now type the following command to send G1 Z50 F4000 and this should move the effector down to within 50mm of the bed or where it actually thinks 50mm should be.

The firmware should have a starting value of 253.9mm for the Z height. Using a piece of paper repeat the last command with lower values including decimal values once you get very close to the bed until you can just feel the paper dragging between the nozzle and bed. You can even use negative values if you need to go passed zero to get the nozzle to touch the paper.

When you get to the point where the paper is dragging then record the value that you last put in. You will subtract this number from the original start value of 253.9 or add the number if it was a negative value. You then take the result and put it in the following line of the firmware. you may also find that as you get very close you may have to home and then put in the very small value. When I was setting up mine I found if I sent a value of like 4.25 and then 4.23 that there was no change. But when I send a G28 between them then I could feel more tension on the paper.

Define# MANUAL_Z_HOME_POS 253.9 <---this value actually represents the exact distance from the bed surface to when the endstops at the top are triggered or the max travel of the Z axis.

As to Wards question there are no max or lower endstops. It instead counts steps from the upper endstops till it reaches this soft value of whatever was set for Define# MANUAL_Z_HOME_POS 253.9 and it knows that is where the zero or start position should be.

It is just a more complex way to define your Z max but since all math for model building relies on this point it is important it is precise.

Beyond this there is also the auto homing switch offset you will need to set ad the flatness value as well will need to be set to compensate for the bed not being totally flat and the arch that the arms move in as they move across the bed from center to edge. I am not going to get into that here as it should be outlined in your manual or configuration guide on the FT cloud drive.

Heartless also linked some very good calibration guides she found on the web in her build thread that had ore information on calibrating the flatness.

Thanks for verifying that the G28 should move towards the end stops, Carl. I have this figured out now and have the latest Marlin 1.1-RC2 running on my modified Kossel with reasonable positioning. The auto bed leveling shows a matrix that is indicative of slightly concave print bed which is also slightly tilted. I will remove, or at least reduce, the tilt tomorrow and might try a glass plate to see if that is a more flat surface than the aluminum plate but I am not sure how I can make the glass visible to the inductive probe yet.
Other than that it seems to work fine in dry runs. I might feed it some filament tomorrow if I have enough time after work and can level the bed.

ok, luckily I did copy and save that guide as a word document for future reference.

also attaching a reference pic for the various measurements & where they are taken from.