Unsolved Mystery. Weird Z0 behavior around build perimeter.

[Jimustanguitar]

I would be real curious to watch the dial indicator as head travels in a large circular pattern, instead of just probing at fixed points. First on the boro print surface, then on that piece of Mic-6 tooling plate I have.

Scott is gonna loan me a rolling ball dial indicator tonight so that we can test this.

[626Pilot]

I don't think *Solved* belongs in the title of this thread until there's a solution confirmed to work for more than one person.

[bvandiepenbos]

I don't think *Solved* belongs in the title of this thread until there's a solution confirmed to work for more than one person.

I agree

[RegB]

I'm glad that the folks at SMC were able to fix Jimustangguitar's printer but I live in Arizona so it would be a bit of drive me . I really don't see how spacers could be the fix for this issue since unless the combination of the spacers and the clips are deforming the glass at points 123 to bring them into virtual flatness with ABC. Am I missing something?

ABC = 12am (z-axis), 4am and 8am
123 = 2am, 6am and 10am

Pre change to Carriage Horizontal Offset:
When I read the post that SMC had a fix involving the Carriage Horizontal Offset value (raising it from 35 to 37.5) I came home and tried it out first thing. Before changing the Carriage Horizontal Offset I did the 4 point calibration out of the manual and had it dialed in perfectly on my Rostock V1 unmodified system. I printed a 100 mm radius disc and as expected it was perfect at ABC positions and not adhering due to extruder height at positions 123 being to high.

Post Carriage Horizontal Offset change:
I changed the the Carriage Horizontal Offset value from 35 to 37.5 and tested the height at x,y,x = 0,0,0 where changing the Carriage Horizontal Offset value had the effect of pegging the paper to the plate. I tested at x,y,z=0,90,0 (A) and found the paper pegged extremely hard. I had to raise the z value by 0.6mm to release the paper so changing the Carriage Horizontal Offset had the effect of virtually increasing the max length values and making the system extremely convex.

I went through the standard 4 point z height calibration again until it was perfectly dialed in - which really took some time because of how convex the system was. Upon completing the calibration I printed my 100mm radius disc again with no change to the results that I had before changing the Carriage Horizontal Offset value prints - points 123 were still printing high and ABC were right on.

By the way, I ended up changing the PRINTER_RADIUS value to 201.1 from 198.7 to flatten out the convex that the Carriage Horizontal Offset change introduced.

Conclusion
Bottom line is this weird Z0 behavior does not have a solution yet. My system's effect print diameter is degraded to well within the imaginary triangle for any quality prints

Q: Isn't the Carriage Horizontal Offset value change for Max V2 ONLY ?
i.e. a BAD thing to do on Max V1s.

[WayneMcCleve]

Q: Isn't the Carriage Horizontal Offset value change for Max V2 ONLY ?
i.e. a BAD thing to do on Max V1s.

You may be correct RegB. I didn't read the supposed fix of changing the Carriage Horizontal Offset (CHO) value from 35 to 37.5 was only for V2. If you are correct then there are two reasons for the "*Solved*" to be removed 1) not a fix for us with a V1 and 2) stating that the CHO value change is the fix and then tweaking someone's base plate heater bed with shims until it works is not a fix.

PLEASE SEEMECNC FOLKS, REMOVE THE "*SOLVED*" AND CONTINUE TO LOOK INTO A REPEATABLE SOLUTION TO THIS PROBLEM. ALSO, PLEASE EITHER TAKE DOWN OR ADD COMMENTS TO YOUR YOUTUBE VIDEO WHERE CHANGING THE CHO IS STILL BEING OFFERED UP AS A SOLUTION. See my post on Aug 20th to this thread.... CHO change did not improve or fix the problem for my V1

[Jimustanguitar]

I put the solved in the title (the OP owns the title, not SeeMe moderators). I can take it down if it makes you feel better, but the issue that I started this thread for IS resolved on my machine.

If someone has tried the shimming fix that I used and it has not fixed their issue, they have yet to post about it. Your move, internet!



Carriage offset should be the same between the V1 and V2. The design of the part changed, but not the functional dimensions.

[WayneMcCleve]

I understand Jimustanguitar that for you the issue was solved and I am glad for you.... truly. Thank you for starting the thread in the first place. This information / discussions has been very enlightening. I thought I was going crazy and had spent many hours trying to calibrate my system and now I know that I am not alone and this issues seems to be a combination of an inherent problem in Delta type 3D printers but that only partial explains the issue. The degree to which some of our systems are experiencing this issue I believe to be either a Rostock design or configuration problem and we really need the experts at SeeMeCNC to not dismiss this issue.

Your fix required the guys at SeeMeCNC to physically work on your system after you had changed Carriage Horizontal Offset value. They added shims etc.. My point is that the Carriage Horizontal Offset value change alone is not the fix but that is what SeeMeCNC is communicating to the community. On the SeeMeCNC's official YouTube channel they have a video explaining that this issues has finally been solved but it says nothing about shims being part of the fix. Also, there have been more than one of us in this thread asking for explanation how shims could be a fix without warping the plate to achieve a "virtual" flatness.

My thought in asking for the *Solved* to be removed is that your fix doesn't match what SeeMeCNC's is saying the fix is and they haven't explained the "why" for shims being a fix for your particular case. To me, without an explanation it is not a repeatable solution and more akin to duct-tape and barbwire. Also, with the *Solved* on this thread I feel it gets dropped from their radar and many of us are still needing the experts at SeeMeCNC to continue to be engaged in finding a robust solution.

I guess one of us could start another thread on this topic but then we loose all the good discussion, background and continuity around this issue. Thanks for considering

[RegB]

From a "process" POV I really LIKE the way Jimustanguitar set things up.
Basically a two step process 1) Set the end stops to match the base 2) set the snowflake and glass plate to be parallel to the base.

It certainly has appeal and seems to be the "right" way to do things.

Well... assuming that the towers are all equally spaced, are facing the right way and are at right angles to the PLANE of the bed.
Not sure about my confidence in all that, other than that the melamine parts are laser cut, but in the standard build process there aren't any checks to ensure that tolerances haven't built up...

The only other "Con" I have with this process is that once done the stand-offs/spacers all have to go back to their designated stations if/when the snowflake ever has to come off and the paper shims all have to be tracked.
Not sure if they will shrink from drying over time, but I think I would try to find something a bit more stable wrt humidity.
Piano technicians know about these things, but I don't think they take out the paper shims and stack them for re-use at the same place - BICBW.

I haven't done this (Yet), though I probably will.

[Jimustanguitar]

Looking forward to hearing your results. I'll post my GCode when I get home.

Feel free to PM me if you have any questions about the specific process that I followed.

[RegB]

Something (actually a LOT of things) happened last week.
I was GOING TO print something out Wednesday morning, so I turned the printer on and manually set the temps to the usual 225 and 90 for ABS.
Then the LOTS OF THINGS got in the way of me doing that print; personal, family, business, etc.
I finally got around to it last night, i.e. the printer had been sitting there preheating itself for 4 1/2 days.

Surprise, SURPRISE !!!
I got the best prints evverrrr...

So, I am left wondering if the whole rig reached some thermal stability that I just haven't seen before {puzzled}.
That whole base enclosure probably takes a lot longer to warm up than the print bed itself, which I recently timed at ~13 minutes from 24C to 90C (V6 Onyx).

I am NOT printing anywhere near the print perimeter, but I am printing multiple parts and they are coming out VERY WELL.

[geneb]

Maybe a little heat soak did you good.

g.

[Eaglezsoar]

Maybe a little heat soak did you good.

g.

Perhaps a day in the sauna would complete the job for Jim.

[Jimustanguitar]

Maybe a little heat soak did you good.

g.

Perhaps a day in the sauna would complete the job for Jim.

I'm in a really good place with my printer, I probably ran it for 10 hours this weekend... It went well enough that I feel the urge to take it apart again for upgrades I'm upgrading my Onyx to a newer version. It will coincidentally be interesting to see if the problem follows the bed itself, or if it's part of the machine's frame. Stay tuned

[RegB]

Heat soak or drying out of the (wood fiber based) melamine.

I think my printer may have approached the conditions seen at SeeMeCNC and by some of you heavy usage folk, i.e. the "always on" condition.
MAYBE a few problems go away when a steady thermal state is reached - which might mean that SeeMeCNC doesn't get the problems that some of us evening and week-end printsters get.

Things are going so well that I want to print a whole lot of things NOW, or I would volunteer for a 3 day shutdown followed by an attempt to print after only a 15 minute warm up.

I can shut it down tonight, leave it alone tomorrow and try something Weds morning.
Those "math gears" might be a half decent test and amuse a few family members - yeah, I'll try those.

I hope this isn't going to be like a piano that has to wait for MONTHS to acclimate to the home before tuning - then tuned several times a year )-;

[Jimustanguitar]

I hope this isn't going to be like a piano that has to wait for MONTHS to acclimate to the home before tuning - then tuned several times a year )-;

It's not a maintenance trap, but it is responsible to do a "pre-flight" check so that you catch things early. I'd rather spend 10 minutes tightening a screw than waiting 10 hours for a print to fail at 99%.

[Flateric]

Something (actually a LOT of things) happened last week.
I was GOING TO print something out Wednesday morning, so I turned the printer on and manually set the temps to the usual 225 and 90 for ABS.
Then the LOTS OF THINGS got in the way of me doing that print; personal, family, business, etc.
I finally got around to it last night, i.e. the printer had been sitting there preheating itself for 4 1/2 days.

Surprise, SURPRISE !!!
I got the best prints evverrrr...

So, I am left wondering if the whole rig reached some thermal stability that I just haven't seen before {puzzled}.
That whole base enclosure probably takes a lot longer to warm up than the print bed itself, which I recently timed at ~13 minutes from 24C to 90C (V6 Onyx).

I am NOT printing anywhere near the print perimeter, but I am printing multiple parts and they are coming out VERY WELL.

I have to seriously wonder how much this may be the case. I run my printer essentially 24/7. Literally for the last few months on end. My prints are fabulous, I do a precheck calibration before most prints.

But it seems to me that way back when I didn't print constantly I had far more issues. hmm, something to consider. I wonder if the rmax was built entirely of aluminium (ok yes expensive) if it would be an entirely different animal.

(Eric runs off to check his personal stock levels of aluminium flat stock....)

[626Pilot]

I wonder if the rmax was built entirely of aluminium (ok yes expensive) if it would be an entirely different animal.

Not terribly expensive. The usual way to build a Rostock is an aluminum T-slot triangle on top, and another one or two at the bottom, and to fit the triangles to plastic or laser-cut melamine pieces that have just enough clearance to accept them without any wobble. I think you can get a Rostock Mini like that for under 700 bucks, including electronics. In a kit, of course. Frankly I like that better than the melamine/lexan plates that SeeMe uses because their plates don't fit the T-slot as tight as other designs. It allows slop that the others don't. Maybe SeeMe will use tighter fitting plates on future designs.

[Moeparker]

I am having this exact issue. I changed the Carriage Horizontal Offset to 37.5 two months ago and between the towers I still had gaps, at the A-B-C points as per the pictures I've seen from Jim.

[img]http://i.imgur.com/a2aKMAJ.png[/img]

I didn't know shims were needed. Tonight I will take off the bed and start the shimming process. I've done the 4 point calibration and it's spot on at X-Y-Z-Center, but then at A-B-C it's a mess between the towers when I try to print there. I've been stuck in that triangle for a while now.

Before I tear into the printer, do I understand this right? Firmware set to 37.5 for Carriage Horizontal Offset value, and then shim up the snowflake at 1-2-3-4-5-6 as needed to remove the gap I get at points A-B-C, between the towers? Won't that change X-Y-Z as well? Maybe I'm just confused and overthinking this.

I don't have any of the hardware to do the dial indicator measurement method. If that's something I should do does anyone have links to where I could get the hardware for that? Thanks a million.

[Jimustanguitar]

So.... In the end I bailed on points ABC all together. There is no way to directly adjust those points, so I concluded that it is fruitless to make adjustments in one place and to take measurements in another.

Make your adjustments AND take your measurements at the same 6 points on the snowflake. That's what worked for me.



Looking forward to hearing about your findings! You may be the first person to confirm that this process helps.

[Moeparker]

Started the process today. My #1 and #2 spots were lowest, about 0.22 and 0.32 lower than the highest of the 6 points. I'm in the process of making shims now. AAA magazine cover is 0.1mm thick, magazine page is 0.05mm and tin foil is 0.01mm. Making my list of how thick each shim needs to be is like reading a quest from a video game. "Collect 3 mag cover shims and 2 tin foil shims for screw hole #2".

[Jimustanguitar]

Keep in mind that perfection is an endless pursuit. You'll have to decide if you want +-.002, 3, 4 or whatever... Know what your goal is and how close you should get before you stop and decide to be happy

[Moeparker]

No joy. I went through those steps, I have points 1-6 all within 0.05 of each other, and I still have very significant Z0 gaps (1-2mm) between the towers. At the towers I'm flush, but between it's the same thing, huge print ruining gaps.

What does kind of work, 3 layers of tape between the towers. It is not the solution I want, but we have orders to fill and we have to get this working again. I've rotated the glass around, still have the same gaps. I'm convinced it is somewhere in the math, how the printer sees the world. I want to try a few more things in firmware but it'll be after I get some prints out the door.

[bot]

I haven't read this entire thread, but it sounds familiar.

I had similar issues with a home built rostock running marlin about a year ago. I seem to recall my solution lied in tweaking the physical endstops in connection with the smooth_rod_offset variable. I was able to flatten the bed within reasonable spec... I don't remember exactly the steps I took, but I seem to recall it had to do with the interplay between the towers that moved the most distance to bring the carriage to the "in between towers" points. I think I would adjust the end stops so they seemed out of spec, but this caused the bed plane to reveal a warp in the "in line with towers" plane. Which I would then adjust with the smooth rod offset.

I could be forgetting poorly, though, I recently had a doozy of a head injury.

[Moeparker]

Thanks for the info. I'm going to compile my data and send it in to support, see what they can make of it. I hope they see a simple mistake I made and it's fixed in two shakes of a lamb's tail.

I did notice that in the newest version of the Rostock Max V1 firmware the carriage horizontal variable is 38.4, not 37.5. Interesting.

[Steve123]

So is shimming the answer? or is this still an ongoing problem?