Straight line curves down at top of the sheet (solved)

:grin: Gotta download the files, slow link…

1 Like

I’m not certain where you’re going with this, and I’m assuming you’re talking about the feedrate for the G01 gcode as used in your cut file, yes?

in v0.68, CNC_Functions.h, it is constrained in line 416:

CNC_Functions.h:416: feedrate = constrain(feedrate, 1, 635); //constrain the maximum feedrate, 25ips = 635 mmps

In v0.78, CNC_Functions.h line 39:

CNC_Functions.h:39:#define MAXFEED 900 //The maximum allowable feedrate in mm/min

There has been much water under the bridge since those days, and in v1.08 it is calculated in three places:

GCode.cpp:630: sys.feedrate = constrain(sys.feedrate, 1, sysSettings.maxFeed); //constrain the maximum feedrate, 35ipm = 900 mmpm
GCode.cpp:688: sys.feedrate = constrain(sys.feedrate, 1, sysSettings.maxFeed); //constrain the maximum feedrate, 35ipm = 900 mmpm
Motion.cpp:100: MMPerMin = constrain(MMPerMin, 1, sysSettings.maxFeed); //constrain the maximum feedrate, 35ipm = 900 mmpm

Note that the calculations in v1.08n are based on a value that is sent from GC as $15 (look for $15 in log.txt to see what yours comes as) but if never changed programmatically defaults to 1000 in the firmware Settings.cpp line 78 (this also happens when an EEPROM clear was done, or the firmware says “Restoring defaults”):

Settings.cpp:78: sysSettings.maxFeed = 1000; // int maxFeed;

To change it in EEPROM without editing the firmware, you could make a macro $15=900 and send that. I don’t believe it’s available as a setting in GC directly.

4 Likes


1 Like

I see, well v0.68 limited the feed rate 635, v0.76 limited it to 900.
On my own rig, using $15=900 made the horizontal cuts near the top of the sheet good, where using the EEPROM default of 1000 gave me cuts similar to the problem in your pictures.
Maybe try using the ‘$15=900’ macro to see if that makes a difference, back it down to 850 or 800 or so to find the speed limit on your setup?

3 Likes

I wonder if we should consider a ‘speed limit’ setting in the GC settings, and/or maybe an optional speed limit calibration step? Some investigation into the effects on the speed limit of surface angle, friction, bit size/rate/sharpness and probably other things as well would be needed…

1 Like

Without reducing the weight, I currently need to go down to F200 to get something that looks like straight. Is there a current limiting or shield protection in the software?
Edit: I’m a bit surprised that the weight of the sled did not make it in the settings. Would have expected that with calculating chain sag. My guess is that its is calculated somehow.

The paths I can follow to get more info are:

  • measure voltage at motor (hopefully during the week)
  • try versions that I’ve seen cut straight at F635 and F1000
  • exchange power supply, but if I can’t use this Power Supply - alternatives I’ll wait a month or more.
  • I have 2 v1 motor-shields. Are they still supported by the software? Or would I even have to go back to them trying v0.68?
  • turning that 1 line into several segments

No, the only limits are the feed rate ones.

It is not directly calculated, but is incorporated into the chainSagCorrection value in an inverse-squared manner. Also incorporated into this value are aspects such as the chain weight, bed lean angle, as well as a variety of scaling parameters, so directly pulling the weight from this value would either require that other data be input or guessed.

I do think it gives a close-enough relationship to sled weight however that it could be used to create some sort of maximum speed setting based on sled weight though.

3 Likes

there is a speed limit, we also have code to detect that we are exceeding the
max speed the machine is able to do (will be in next week’s update)

1 Like

That would be the software feed rate,

and this would be the physical/electrical limitations of the equipment. The software doesn’t look to current limiting or shield protection.
Another variable to check along with voltage on the motor would be temperature of the H-bridges. They don’t have protection, and could have reduced performance at high temperatures, which would also show up as reduced voltage at the board end of the motor cables.

1 Like

i’m pretty sure we don’t have temp monitoring of the drivers in the hardware

You’re right, it would take some form of remote sensing.

Given the kind of singularity of this cuts, don’t bother to much. Working on it.

Thank you all for all the support, suggestions, clues, hints and inspirations.
This amazing combined community effort put a list for me together to test, that pushes me in the background for a moment. Don’t think I will not be reading!

What I really would love to read is, a combined task force of min. 10 Maslows testing 1.08 thoroughly from all sides.
Starting as if you just opened the kit.
By the time the new kits arrive at their destinations, the support from our side will get everyone from unpacking to cutting.

It was suggested to open a post, but sadly my current powers do not allow me to create a spark strong enough to ignite. Volunteers?

1 Like

you have shown a weakness in the system, and figuring out what’s happening will
make everyone’s machines better.

1 Like

Only 2 things that I’ve changed solved the issue.

  • I replaced the 5A power supply with a 10A one.
  • I tested 1.09

And the WINNER is (for now until proven otherwise) @Robert!
With a new power supply and 1.09 (please join us on the mission Let’s make v1.09 the most stable release yet) the cut seem to be straight. A restricted feed rate might have contributed, there is more to find out.
That I’m able to continue testing in the ‘Heavy Weight’ category is good news.
Thank you all.

7 Likes

Based on this outcome; for those outside the US who need to source their own power supply anyway is it worth going straight for a 12v DC 10A over the recommended 5A to be on the safe side?

Cheers!

I would vote for YES. The price should not be much different I hope and it’s good to have a solid reserve.

There is also a chance that a future motor shield will have stronger chips, allowing for heavy load :relaxed:

1 Like

What is the risk of burning out the motor shield by providing it up to 10A of current? Would it be prudent to go to active cooling (i.e., heatsink w/fan) if you up the available current?

I failed to provide the proof it was the power supply.
The version was changed from 1.07 to 1.09 that decreased the max feedrate from 1000 to 750 and perhaps some other changes. But with 1.07 I had to go down to feed 200 to get a straight line.
All speculation till now.

Active cooling is never a bad idea. As far as I know we are running the shield within, but close to upper limits (with a heavy sled and/or short vertical hight, top of the sheet, motors are pulling against each other)

Under normal operation the motors and shield should not pull more A then needed. If a chip shortens, the amount of magic smoke might increase. I like to make my power supplies ~30% stronger then I calculated. I don’t remember why, perhaps to compensate for errors in the math or from old transistors those days, that had the best performance around 70%.

Currently running this china 12V/10A without magic smoke.



Kindly ignore the Bedouin style mounting :grin:
I live in a country under (not so active) development, plus the window AC is blowing sideways over the electronics.

4 Likes