8-Bit CPU

Hi guys,

So I'm building my own Delta from scratch. It's coming along really well, I've spent a fair chunk of money on extrusion, power supply, bearings, motors, etc. At the moment I can't afford a Smoothie board or any ARM based controllers, they seem really hard to find in Australia.

I was wondering how bad an Atmega (Arduino) controller is? I have a spare Sainsmart controller ( Basically a glorified ramps all in one PCB ) Link here

I was hoping I can use this? Well, I'm sure I can, but how bad are they? Will the print speed be limited? Will stuff just come out ugly?

Thanks for any info!! Also any other controller board suggestions would be really good.

Edited 1 time(s). Last edit at 07/27/2016 08:31AM by aaroncm.

It's certainly possible to drive a delta printer using an 8-bit CPU,.but you are likely to face additional obstacles such as:

- Difficulty in supporting a graphical LCD becuse of insufficient CPU power to calculate delta movement at the same time
- Limited travel and/or print speed, especially if you want to use x32 or greater microstepping or 0.9deg/step motors for better resolution
- Lack of a a good auto calibration facility

The 32-bit boards provide additional benefits compared to most 8-bit boards. Have you looked at the Duet 0.8.5 and Duet WiFi? They have IMO the best delta printer firmware of any board (including least squares auto calibration, and segmentation-free delta movement), and with the recent fall in the £ they are excellent value for non-UK buyers.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Thanks for the reply

I'd love one of those Duet boards but it'll be a loooong time before I can afford one.

Do you know of any other boards or clones that are a little cheaper? Do you think it's at all possible to get an > 8-Bit controller for around $50AUD?

Thanks!

I'm running 8 bit controllers on both my deltas, and haven't run into the issues that DC42 mentions.

On my big delta, it's running 0.9 degree steppers, with 32 bit microstepping, and I have it set to 300 mm/s travel speed. I don't know if it gets that fast, I haven't measured it, but it sure seems fast enough. Quality is fine. Speed for me is more limited by the properties of the plastic and mechanics, rather than capabilities of my firmware. I usually run at around 40-50mm/s print speed, as I like quality instead of speed.

I'm quite happy with the 4 line LCD panel as a control source. It works quite well for me, and I can't justify the added expense of both a duet and a graphical interface board when what I've got is ticking all the boxes for me.

I would recommend Repetier instead of Marlin, however.

Edited 1 time(s). Last edit at 07/27/2016 11:14PM by nebbian.

I run an 8 bit board (arduino) and I had problems with its speed when delta segments per second (DSPS) was set to 200. But 200 is too big value. It leads to maximum segmentation error of 0.0012 mm which is unnecessary low for step size of 0.0125 mm. I did not have any speed problems after I changed DSPS to 90. 90 segments per second leads to maximum segmentation error of 0.006 mm, still precise enough. Even setting DSPS to 70 leads to segmentation error of only 0.01 mm. In other words, DSPS is often set to a too big value by default which unnecessarily loads your controller CPU.

All these values are for print speed of 120 mm/s, diagonal rod length 256 mm. The machine does not have any LCD. If you plan to use lower print speed by a factor of K then you can use even smaller values for DSPS (by the same factor of K). Now, some people may tell you that segmentation error of 0.01 at 120 mm/s is too big. I do not agree. Almost all of those people will use belts and in most cases their dynamic errors from belt elasticity will be much worse than 0.01 mm. Well, maybe not if they use small accelerations. But it does not make sense to use print speed of 120 and small accelerations. With high print speed and slow acceleration your print time will not be much shorter anyways (especially for models with short infill lines). Moreover it leads to uneven extrusion at the beginning/ends of paths. Pressure management and flying extruder or flex drive may help here though.

Thanks for the info.

I'll definitely be sticking with my Mega2560 then. I'll update on how it goes once I get it working.

Thanks again!

Quote
hercek
I run an 8 bit board (arduino) and I had problems with its speed when delta segments per second (DSPS) was set to 200. But 200 is too big value. It leads to maximum segmentation error of 0.0012 mm which is unnecessary low for step size of 0.0125 mm. I did not have any speed problems after I changed DSPS to 90. 90 segments per second leads to maximum segmentation error of 0.006 mm, still precise enough. Even setting DSPS to 70 leads to segmentation error of only 0.01 mm. In other words, DSPS is often set to a too big value by default which unnecessarily loads your controller CPU.

All these values are for print speed of 120 mm/s, diagonal rod length 256 mm. The machine does not have any LCD. If you plan to use lower print speed by a factor of K then you can use even smaller values for DSPS (by the same factor of K). Now, some people may tell you that segmentation error of 0.01 at 120 mm/s is too big. I do not agree. Almost all of those people will use belts and in most cases their dynamic errors from belt elasticity will be much worse than 0.01 mm. Well, maybe not if they use small accelerations. But it does not make sense to use print speed of 120 and small accelerations. With high print speed and slow acceleration your print time will not be much shorter anyways (especially for models with short infill lines). Moreover it leads to uneven extrusion at the beginning/ends of paths. Pressure management and flying extruder or flex drive may help here though.

Thanks hercek!

This is excellent insight. I wondered often why DSPS 200 and DSPS 100 produces identical printing quality.