Morgan Build, Ann Arbor MI (US)

[knarfishness]

Hey Folks,

Made some good progress this last week putting together the frame for my Morgan. I’m converting my spare Prusa Mendel v2, so my hunt for parts has been pretty minimal. (Still need to come up with 6805 bearings).

I’m using aluminum pipe instead of the recommended brass (or copper) as it was a bit cheaper in my area. Most of my “SAE Conversion” has been done with a Dremel tool, so far as the tube fitting is concerned.

Anywho, I’ll update with questions/progress pics as I go:

https://www.dropbox.com/s/ogxjhhsc326waz4/2013-07-10%2001.34.25.jpg

-Frank

[Quentin Harley]

Good progress! Let me know how those Alu pipes work.

[Quentin Harley]

6805 is a bearing commonly used in bicycles, and if you go to your local bike shop you will find it as hub bearings. Get the cheapest ones – the ceramics are great, but too expensive for the use!

[knarfishness]

Aah, bicycles! That should make them easy to source.

I’m curious if you’ve experienced any differences in the bowden tube setup between 1.75mm and 3mm filament. I’ve got both hot end options available.

[Quentin Harley]

Only used 3mm on Morgan so far. To all accounts 1.75 works better for bowden, but I have 3mm down to an art now.

Next machine will be 1.75. Building it right now…

[knarfishness]

Updates!

https://www.dropbox.com/s/1znrf9ikl48bxu8/2013-07-15%2022.24.39.jpg

I’ve received my 6805s, so I figured it was time to go ahead and prep the frame.

PVC is painted black now, as well as the oak frame stained a nice reddish color to class it up a bit.

One thing I did notice during assembly is that the top/bottom mounts for the smooth rod seem to be a bit more narrow than the distance between the bearings that hold the print bed. For now I’ve just cut the parts to get the fit right, and then I can expand the mounts at the top/bottom if need be. Any idea why that is?

Thanks for looking 🙂

-Frank

[Quentin Harley]

Looking good. What was the distance on the brackets? And on the Z-bed bracket? I might have made a typo in Openscad when I recently fixed up all the parts.

[knarfishness]

I’ll double check with the calipers when I get back to my workshop, but I can’t imagine they’re more than a centimeter off on either side (2cm overall missing width on the top/bottom brackets).

[Quentin Harley]

On Morgan master it should be 175mm between the rods. Di you go for 8mm or 12mm rods? 12mm is 195mm between the rods.

Morgan lite is 160mm.

[knarfishness]

8mm rods and bearing were what I had left over from the Prusa V2. I can double check my logs to make sure I printed an entire 8mm set, but the large bracket is definitely the smaller one.

Although looking at those numbers, it looks like the discrepancy that I’m seeing…more double checking required. 🙂

[knarfishness]

After some checking, it seems that the version of the 8mm top/bottom z brackets came out too short. They can’t be the 12’s, just judging by the holes. That being said, if you’ve got a level, those pieces work just as well as 3 separate parts.

Little more progress this week (it’s been busy), but the drive system is in place and ready to be belted up. Time to start messing with Marlin and the PrintrBoard.

[video src="https://www.dropbox.com/s/v8cz9m2ivkvuq87/2013-07-18_22%3A54%3A057.mp4" /]

[Quentin Harley]

Really starting to come together!

[knarfishness]

Well, I’ve got the first powered movements done. (Almost wrecked it ><)

Anywho, I’m at the point where I need to mount endstops, but I can’t see clearly in the videos you have up where you have them placed. Any tips?

Thanks!

-Frank

https://www.dropbox.com/s/co78cj83rcp5lw5/20130723_181902.jpg

[Quentin Harley]

Hope it helps…

[Quentin Harley]

I found the problem…

The Z-mount bracket was the wrong type. I omitted to upload the 8mm version of the bracket.

So sorry! The files are now corrected on the Git repository.

[knarfishness]

Aah, gotcha. Sticking with my split pieces for now, as the frame is already together, but if I do a teardown after first tests I’ll get them replaced.

I took a look at the endstop placement picture you sent me as well. I see the placement (but no endstops?) in the pic, and I have my mechanical endstops attached there.

A few questions moving forward:

1. What on the wheel triggers the endstop? My best idea was to use thumbtacks/nails pushed into the wheels to push up against it.

2. Where is the home location supposed to be (in relation to the print bed)?

3. Home-X and Home-Y seem to operate both motors simultaneously, a feature of the SCARA firmware I’m assuming. However, it seems that once homed, the already homed wheel will start to turn _again_ blowing past the endstop and causing it to go a bit crazy. Is one always supposed to home before the other?

Thanks a ton for the assistance, I can’t wait to get printing with this thing. 🙂

-Frank

[Quentin Harley]

The small magnets fits in the hole in the wheel closest to the homing position. The exact position is not critical, because calibration of the end stop homing offset will take care of that – it only needs to be outside the normal operational range.

I am going to make a small video to make it a bit more clear.

[knarfishness]

Aah, I’ve never used the magnetic endstop setup before, I’ll have to look into that. There isn’t a firmware setting in marlin that specifies different homing patterns for mechanical and hall endstops, is there?

Looking forward to more videos.

[Quentin Harley]

😉

[knarfishness]

Looks like it’s still marked as private. Thanks for taking the time to video it!

[Quentin Harley]

Try it now?

[jade39339]

Great video, that helped me alot to understand what was going on!

[knarfishness]

Perfect, gotta love youtube.

I’ll have to hunt around for a few hall effect sensors this weekend in order to replace the mechanical ones that I currently have. Still not sure why the mechanical ones would act differently, but that’s a mystery for later.

What I’m curious about now is that the bed homes to the bottom and not the top. I’m still using the threaded rod setup for Z and not the leadscrew, and it seems to me like it would take quite a while for it to come all the way up to start a print.

Thanks again.

-Frank

[Quentin Harley]

It doers take a while, but remember that the steps per mm would be greater with threaded rod. You should also use lower microstepping, perhaps 1/4 or 1/8 in order to prevent absurdly high step rates.

Do not look for a leadscrew yet though… there might be a upgrade in the future that could optionally remove the need for a lead screw, but a lead screw might be slightly more reliable. So if you have a lead screw already, you need not worry that you wasted your money.

[Quentin Harley]

PS: On the mechanical switches, you can make it work if you can make the switch region larger. It should not be easy for it to miss the landing spot.

In order to save myself a lot of stress, I first simulated the homing routine by hand, monitoring the limit sensors using the gcode M119. This way you can double check that the machine is indeed switching the limits.

Also remember that they should be wired as “MAX” limits, not min if the hardware has a distinction.

[knarfishness]

Alrighty, more updates!

After discovering that I had X and Y reversed (*facepalm*), I’ve got homing working with my shiny new hall effect sensors. Way cooler than mechanical endstops.

I tried a few dry runs earlier, but it seems like it’s still trying to print right on top of the shaft, so I think that some offset adjustment is in my future.

-F

[Quentin Harley]

Good. I think it is time for the calibration video.

[florent]

Can I ask something that might be a silly question ? Surely it’s because my english is poor, but I can’t find a difference, looking on google, between a threaded rod and a lead screw. Do they differ just by the level of accuracy ? Or is there more than this ?

[jade39339]

Threaded rods are not as I understand it uniformly accurate with regards to the pitch of the screw where as lead screws more or less guarantee uniform pitch, diameter, etc… within a given tolerance range. That’s why they cost much more than a threaded rod.

[florent]

OK ! Thanks a lot !

[Quentin Harley]

That is correct. Threaded rods were designed to fasten stuff together. Is was never meant to be accurate.

[Quentin Harley]

Right. I took the video, but this one I will have to edit together since I used screencapture and my phone to demo the calibration.

Give a me a day or two…

[Fig]

Hi Frank, good job!
I’ve never used magnetic endstops either, I have no idea what type and where can I buy them…Could you give me a website link of the supplier in order I can buy them directly.
And is there any calibration of Marlin firmware(Quentin’s armlevel) after mounting them on machine?
Thanks

[Quentin Harley]

Hi Fig,

There is a calibration routine, but homing will work even without calibration. I have shot some video a few hours ago, and need to edit it together.

I am also going to make a calibration flowchart to make it very clear.

Cheers,
Q

[Fig]

Thank you Quentin!
Awaiting for your nice video.
I’ve assembled arms, drive wheels and frame, but I don’t know where to buy the hall effect sensors. I searched them on ebay, but that seems different from your video shows… So, could you give me a hall effect sensor supplier’s link?

[knarfishness]

I snagged my hall effect sensors on Amazon for a pretty decent price:

They work well with those tiny ‘thumbtack’ fridge magnets.

[knarfishness]

Hooray for Friday free-time, I now have a build log, with pics!

http://dieselpoweredrobotics.com/?page_id=184

I’ll link this from the first post as well, so people see it. Hopefully after calibration I can get the finishing shots.

[Quentin Harley]

Get your arms calibrated…

Morgan arm calibration guide

[knarfishness]

Working my way through the flowchart, but I haven’t had too much time lately. I’ll let you know how it goes when I get done.

Thanks again.

[Fig]

Hi Frank,
I bought hall effect sensors on Amazon, the type is same as yours. But they didn’t work fine, I linked them to RAMPS by wires directly, was I wrong? Do sensors need another electronic board?
Fig

[Quentin Harley]

Hi Fig,

Two things to remember when using the hall sensors:

[Fig]

Thanks Quentin,
Do I need to buy sensors with boards like this? (http://www.ebay.com/itm/3-x-Hall-Endstops-react-magnetism-3D-printer-reprap-prusa-mendel-etc-/150872387181?pt=LH_DefaultDomain_2&hash=item2320b1ee6d)
Or just A3144E sensor is OK?

[Quentin Harley]

The sensor should work as is. I use the updated sensor in my machine though… A1101E

Mine switches at 100G, whereas the 3144 is not clear… The datasheet does not give proper switching values for it.

I would recommend getting A1104e hall effects. They are very cheap, and work a charm.

[knarfishness]

Hrmm, just started messing with Morgan calibration again, and my board isn’t responding to M360, M361, M362. G28 to home still works fine. Any idea what’s up?

[Quentin Harley]

All movement in disabled until you hit home, but I assume you have done that first.

You are running the latest version of qharley/Marlin armlevel branch, right? What does the firmware report when you connect to Morgan in Pronterface?

[knarfishness]

Yeah, the homing works fine, it’s just the second time I went to run this calibration, it doesn’t respond to M360/1/2.

It’s reporting fine via the console as well, as the SCARA version of Marlin.

[Quentin Harley]

Sounds weird.

Which electronics are you using again?

I just performed the whole calibration routine again after I had to take apart the Daughter of Morgan.

[Quentin Harley]

Grab a fresh copy of the Morgan firmware (armlevel branch) and reprogram your board. When startin up it should state :
Author: (qharley, Morgan SCARA armlevel)
somewhere in the startup text on the gcode terminal. If not, you might have the incorrect branch. If you have the right branch. you may have set something up incorrectly, and you need to reset to factory defaults using M502.

Hope it helps.

[Quentin Harley]

How is Morgan doing?

[knarfishness]

I’ll be verifying this afternoon that the branch that I’m on is correct, as well as resetting the board if I’ve got to. Even though the calibration looked close last time, my dry-run print was still trying to print itself on top of the top wooden frame, and not the bed.

Once everything is working properly, I’ll need to more permanently secure my belts to the drivewheels, and then it’s print time!

[Quentin Harley]

Post your boot splash settings here before you start. You can always come back to it, or help me understand what is going on.

[knarfishness]

< 6:00:37 PM: FIRMWARE_NAME:Marlin V1; Sprinter/grbl mashup for gen6 FIRMWARE_URL:https://www.morgan3dp.com PROTOCOL_VERSION:1.0 MACHINE_TYPE:Morgan EXTRUDER_COUNT:1

That’s my splash message. Trying again today, M360 is once again working. Just had to hit the reset button on the PrintrBoard.

I’ll let you know how calibration goes.

[Quentin Harley]

For fastening your belts you could also look at the very nice belt clamps angrychisel made for his Morgan derivative

I will be printing a couple, and should have something similar on Morgan git soon.

[angrychisel]

Oh Quentin please don’t tell anyone about my over engineered belt clamps. I plan to make something simpler in the future, but I needed something quick to tighten the belts. I’m a tad bit embarrassed by them. I promise I will develop a simpler and easier to adjust system when I get a chance. I actually thought about incorporating the dovetail into the wheel in the future, that way you would only need to print 2 extra pieces. The belt wedge and the belt tightener.

[Author]

Posts