Page 3 of 9

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Sat May 11, 2013 9:18 pm
by flickerfly
My daughter is under the weather so the build didn't progress, but we did buy what I think are the last few things we need. Since the J-heads can be purchased in parts from hotends.com we did that. We also opted for a heater cartridge. In a moment of mental drop-out I ended up with two steppers, That may come in handy anyway.

Today's purchase

Role Desc Qty Supplier Price S&H, Taxes Total
HotEnd Parts for a J-Head 1 Hotends.com 53.33 11.41 64.74
Ext/hotend bits NEMA 17 & Heater Cartridge 1 Ebay: rp_one_labs 19.8 3.95 23.75
ABS 1.75 ABS Pink Plastic Spool 1 Octave 31 5.5 36.5
Extruder Stepper NEMA 17 Stepper Motor 1 ATI 9.99 6.89 16.88

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Tue May 14, 2013 5:45 pm
by bwevans
How long did it take between ordering and receiving the kit from ATI?

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Tue May 14, 2013 5:56 pm
by flickerfly
bwevans wrote:How long did it take between ordering and receiving the kit from ATI?


It was real fast, as fast as anything I ordered. I ordered on May 4th and it arrived May 8th. It originated in Palatine, IL and was shipped UPS. It came in great shape. I've since ordered another item from them on Sunday. USPS says it'll be here Thursday. Smaller package so USPS makes more sense. Also, they are on the list of companies that pay royalties back to Bart for the design so that's a plus also. The only complaint I had turned out to be my own fault. It does not come with holes for the cables drilled in the makerslide, but I'm not sure how I want to do that yet anyway so I appreciate the flexibility.

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Tue May 14, 2013 6:15 pm
by bwevans
Thanks for the info, best of luck on the rest of the build!

A Bit Lost

PostPosted: Sat May 25, 2013 2:57 am
by flickerfly
Finally getting a chance to come back to this. My daughter and I played with it some and make things move a bit. I'm not sure where to find the right information on calibration and I'm not certain my power supply of choice is actually powerful enough. I feel a sharp rise in the learning curve coming on.

So, I (think) I have all the stuff. I've rigged up the X3 with the motors and the X and Y axis move around some. The Z-Axis doesn't at all though it holds its position and won't manually move easily once given a command. I have the repetier firmware and host 0.85b that I'm playing with. When I manually tell both the X and Y axis to move 1mm it seems a smaller move or awfully close to the same as the 0.1mm move. When I tell them to move 10mm, they don't move at all.

Anyway, I'm going to play with it some more tonight and try to get the endstops hooked up and verified as functional.

EDIT: Okay, I found that the travel of both the X and Y in repetier is a total of 12 (is that mm?) rather than the 200 I was expecting. I guess that probably means I have some firmware config issues.

Hmm... Maybe some math might be helpful here...

PostPosted: Wed May 29, 2013 2:17 am
by flickerfly
I found that I hadn't set anything in the microstepping options on the X3 so I pulled the drivers off and set those for 1/16 (see SD8825 Drivers Datasheet). Then I found this nifty calculator over at Prusa's website. I looked at the data sheet that ATI provided on the steppers, BOM to find the MXL belts and found the details on the pulleys.

I decided that to make this calculation I needed to enter the following info into the "Steps per millimeter - belt driven systems" calculator. Note: Repetier now has this calculator built into their software in the Tools menu. It comes out with similar, but perhaps more precise numbers. (78.8177 on the belts.)

Motor Step Angle: 1.8degree
Driver microstepping: 1/16
Belt Presets: MXL
Pulley Tooth Count: 20

Results: 78.82 steps/mm (X & Y Axis)

I went into Repetier host, hit Alt + E to open up the firmware settings and put that in for the X and Y. Suddenly, things started to move as expected over distances that are much closer to accurate. The distances aren't bang on, in fact they are actually off by a lot, but that's an improvement. I measured the X-axis with a 100mm move and it went 39mm. Clearly some work needs to be done.

Over the two short period that I had this up and running, the drivers became extremely hot; the Y driver was almost to the point of burning to the touch the first time and the next it was was the Z driver (which still doesn't move). I'm letting them cool down right now. My first instinct after cutting the power was to turn the pots down a little bit. The drivers are sometimes making lots of whiny noise in certain positions, but they were doing that before.

While I was playing with Prusa's calculators, I also tried out the "Optimal layer height for your Z axis" calculator. I found out the following using the M8 lead screw preset giving me 1.25mm/revolution. I'm not sure what this chart does for me at this point, but I'm plopping it here for later reference.

Code: Select all
Layer Height       Steps
1.0mm              160
0.5mm              80
0.3mm              48
0.15mm             24
0.1mm              16


Oh, and the "Steps per millimeter - leadscrew driven systems" calculator from the same place is relevant to the Z so here's the result of that calculation using nothing not already mentioned in this post.

Results: 2560 steps/mm (Z Axis)

Dragging this over from the SeeMeCNC EZStruder post so I have things gathered:
bdring wrote:I measured the steps per millimeter of the SeeMeCNC EzStuder at about 95.9 stes/mm

DEFAULT_AXIS_STEPS_PER_UNIT in Marlin
EXT0_STEPS_PER_MM in Repetier


NOTE: These aren't well tested and tuned values, just some calculations that will hopefully get me in the ballpark.

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Wed May 29, 2013 12:18 pm
by Gadroc
For your X/Y it looks like you have two problems. One it looks like either you have a 0.9degree steppers or you have set micro-stepping to 1/32. Second your pulley tooth count may be wrong. Looking at ATI's website looks like they are reselling reprapdiscount products now, and when I my kit from reprapdiscount I got 18 tooth pulleys not 20. I would actually count the teeth on your pulley.

2560 is accurate given 1.8degree steppers. Double check that as well.

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Wed May 29, 2013 1:19 pm
by flickerfly
Thank you Gadroc. I wouldn't have thought to actually count the teeth. The ATI Shipping List also says it has 20 teeth, but if you had 18 I better check that when I get home. I can certainly see how that would make a difference.

I'm fairly confident given another look at my stepper's data sheet that 1.8 deg/step is accurate for that stepper. I'll look at the X3 board again and make sure I didn't confuse the 1/16 jumper settings with the 1/32. That's all done with jumpers on the board right? There isn't any settings in the firmware to make in addition?

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Wed May 29, 2013 2:02 pm
by Winder
I tried calculating what the settings should be and came up with a number close to yours. Once getting to that point I started printing out this nickel test object: http://www.thingiverse.com/thing:11261

After printing I would see how well it fit and adjust my settings slightly to bring it closer in, I printed 3-4 of these total to get a nice fit. In an ideal world I would have liked to calculate the settings to get them bang on, but this was pretty easy and was effective enough for me.

Re: Flickerfly's ORD Bot Hadron Build Log

PostPosted: Wed May 29, 2013 2:15 pm
by Gadroc
As far as I can tell (a lot of the calibration Wiki's in the reprap world are outdated) it's best practice now a days to stick with the mathematically calculated values. They should be "close enough" and if not it's probably a mechanical error. This is especially true for an Ord Bot with manufactured pulleys instead of printed ones. If you are calibrating to a test object your calibration might only be accurate on objects that exact size, then when you print a larger item the errors you compensated for will then become bigger errors. In this case the math is way to far off of actual to start "fine tuning".