I’m running two delta printers side by side. One has ramps 1.4 with Marlin and the other is an MKS Sbase board with the most recent release of smoothieware. When I generate a file through Simplify3D and print it on the Marlin board, the completion time is relatively close. Usually a little early but within 5% of the estimated time of print. Using the same G-Code file on my smoothie printer, the print takes twice as long. Both print out properly calibrated sizes.
There is a jumper on the MKS board for 32microsteps vs 16microsteps. I have it set to 32 and I’m wondering if there is a timing issue or a variable I need to set to increase the pulse rate?
I have two more tests to do before I can confirm this is caused by a problem with board timing.
move the jumper to the 16 steps and run the print again. (I’ll have to change the steps per mm in the config)
I have a proper smoothie board V1.0a. I can set up that board for the delta printer and test the timing. (I’ve been saving this board for a better project)
Any thoughts on what else could be causing my printer to run at half speed? Does anyone know of a variable that would alter the board timing to reflect 32steps per mm?
This is probably just acceleration and junction deviation ( jerk in Marlin ) parameters being different, resulting in different speed curves, and because speed is time times distance, difference in print times.
Acceleration range was tested from 300mm/s^2 to 1000mm/s^2. It didn’t appear to improve the speed.
I also tested the 32microsteps vs 16microsteps. There was no difference in speed.
I did not test Junction Deviation. It is currently set to 0.05. I will need to read more on the function and recommended range in order to test.
The 0.5 and 4000 worked to speed the machine up thank you. I did a little reading and some testing.
Based on my testing if you’re coming from marlin, a quick estimate for junction deviation is dividing the max jerk by 60. This is a bit odd because the reading I’ve done indicates the max jerk value is in mm/s where the junction deviation value is in mm. If this were a simple unit conversion problem, multiplying the max jerk by 60 would cancel the seconds unit and give mm as a result. But we know it’s more involved than that, based on the reading on the GRBL method of calculating jerk, the values for junction deviation are not a simple unit conversion. However…… dividing max jerk by 60 is a simple rule of thumb that gets us in the ballpark for the smoothieware junction deviation value. It was some interesting reading, thank you for pointing me in the right direction.
Approximate junction deviation when converting a machine from Marlin to smoothie.
(DEFAULT_XYJERK ) / (60) = (junction_deviation)
If: DEFAULT_XYJERK = 20
then: 20/60=0.33333333 = starting point for the junction deviation value
therefore: junction_deviation = 0.33