Welcome! Log In Create A New Profile

Advanced

Thinking about failed printing....

Posted by Traveltrousers 
Thinking about failed printing....
November 12, 2013 06:22PM
Hi guys

So at the moment, we have 3 ways to print.
1. Eject the SD card, load a file, put it back, start the file.
Great reliability, but what a pain....
2. Upload to the SD card via pronterface...
Reliable, but really, really slow for large files.
3. Print from Proterface/Cura.
Quick, easy, but a major pain when the job fails.

Why can't we combine 2 & 3?
A small slice of g-code is sent to the SD card, from where it is run, followed by the rest of the slices.
If the program disconnects, it just reconnects and sends the next slice...
Quick and reliable, and convenient...

What does someone who knows Python think?? :p
Re: Thinking about failed printing....
November 12, 2013 06:44PM
How slow is "really, really slow for large files"? What do you call large?

It currently takes my bed about 12 minutes to reach 135C. If we are talking around this kind of ballpark then it should be relatively easy to set the bed target temp to first layer temp, then start the upload of the file to the SD card, then kick off the job when done. Someone should make this an option in OctoPi...
Re: Thinking about failed printing....
November 12, 2013 07:11PM
I just sent a small, 430kb 28 layer file to the SD card and it took three minutes. That's slow!

If I transferred it by hand it would be half done by now, and I don't have to return to the printer to start it off.

I'm printing PLA, so I start the heating before the move to 0 and it hit 190 degrees 20 seconds after that, waits for 60 seconds (while my bed gets to 90) and then starts.

I have a slight worry I will damage the mini SD slot with over use...
Re: Thinking about failed printing....
November 13, 2013 02:46AM
What you propose would probably require modification of the firmware. It would need to assess the availability of g-code and safely pause the print at completion of the current layer (assuming it has all the g-code for that layer).

Printing from SD card is safer in that respect because you know the printer has all the g-code it requires to complete the print. Any form of connected print either normal or what you propose is subject to disaster in the event that the g-code stream is terminated before the print is completed.

Regards,
Neil Darlow


I try to write with consideration for all nationalities. Please let me know if something is unclear.
Printing with Mendel90 from fedora 25 using Cura, FreeCAD, MeshLab, OpenSCAD, Skeinforge and Slic3r tools.
Re: Thinking about failed printing....
November 13, 2013 04:04AM
I find that if I have the PC next to the printer and connected to the same mains extension and don't run anything too stressful on the PC, disable sleep and automatic updates then it is 100% reliable over USB for normal prints. For things with lots of tiny line segments, e.g. yoda, I need to use SD card to get the bandwidth.


[www.hydraraptor.blogspot.com]
Re: Thinking about failed printing....
November 13, 2013 05:02AM
@nophead: I agree that with those provisions you might be able to print successfully over a USB connection but there is always the concern of a PC crash etc. With SD card printing you do remove a whole set of variables.


I try to write with consideration for all nationalities. Please let me know if something is unclear.
Printing with Mendel90 from fedora 25 using Cura, FreeCAD, MeshLab, OpenSCAD, Skeinforge and Slic3r tools.
Re: Thinking about failed printing....
November 13, 2013 10:43AM
It's mostly been reliable, but I do use my computer for some intensive tasks sometimes, but then this morning (and yesterday) I was away from the computer and it stopped halfway through...

I borrowed* a raspberry pi from a friend, and just put octoprint on it so we'll see how reliable that is...

* I'm 'fixing' for him :p
Re: Thinking about failed printing....
November 13, 2013 05:46PM
Quote
nophead
I find that if I have the PC next to the printer and connected to the same mains extension and don't run anything too stressful on the PC, disable sleep and automatic updates then it is 100% reliable over USB for normal prints. For things with lots of tiny line segments, e.g. yoda, I need to use SD card to get the bandwidth.

This is what I am doing too and found it has been working well so far (I also disabled screen-saver, sleep, etc.). I have a disability and have not figured out how I could move the card back and forth from PC to printer. This option seems out for me unless I ask someone to do it for me.

BUT, here is my question. I have an UP! printer and after it slices it sends the print to the printer. Sending to the printer doesn't take long. Is it because the USB cable for the UP! Is not a mini cable like we use with the Mendel90? Or is it something in the Melzi that is slowing things down and thus a limitation therein?

I sliced a squirrel from Thingiverse at .2mm using Slic3r and it took maybe two hours to upload to the Melzi. It would have taken a few minutes with the UP! I would totally love to see a much faster transfer for RepRap.
Re: Thinking about failed printing....
November 13, 2013 06:23PM
Gary, you can use an "sd card extension cable" if you have trouble reaching the back. They are ~$10.
Re: Thinking about failed printing....
November 13, 2013 06:54PM
USB can transfer much faster but Repraps use a USB to serial converter rather than a direct connection to the ATMega, which is makes it slow, especially on Windows. You can make it a bit faster by reducing the serial port latency in device manager.


[www.hydraraptor.blogspot.com]
Re: Thinking about failed printing....
November 14, 2013 01:26AM
Quote
RyanMark
Gary, you can use an "sd card extension cable" if you have trouble reaching the back. They are ~$10.

Cool. Thank you! I looked that up and it solves 50% of my problem. I'll work on the other 50%. Best regards ..
Re: Thinking about failed printing....
November 14, 2013 01:46AM
Quote
nophead
USB can transfer much faster but Repraps use a USB to serial converter rather than a direct connection to the ATMega, which is makes it slow, especially on Windows. You can make it a bit faster by reducing the serial port latency in device manager.

Thank you, I will learn how to do this.
Re: Thinking about failed printing....
November 14, 2013 08:18AM
Is it possible to tap directly into the serial port on the Melzi board?

With a 256mb Octoprinter I need 1 USB port for the wifi and then there isn't a printer connection, unless I buy a USB hub and by that time I may as well have just bought the 512Mb pi....

The octoprinter is pretty good, no fails so far. Not quite as responsive or easy as pronterface, but pretty close....
Re: Thinking about failed printing....
November 14, 2013 09:37AM
There is a second serial port available on the expansion port so theoretically you could change the firmware to use that.


[www.hydraraptor.blogspot.com]
Re: Thinking about failed printing....
November 14, 2013 11:12AM
That word 'theoretically' sounds like a challenge!! :p

After a bit of googling, I found this which looks like it should provide a suitably cheap interface between the pi serial pins and.... the TX/RX/VCC/GND pins on JP16 of the melzi board?
Re: Thinking about failed printing....
November 14, 2013 11:28AM
The PI also has some direct UART connections so you could just make a direct connection as long as you have a solid ground and keep the wires short.


[www.hydraraptor.blogspot.com]
Re: Thinking about failed printing....
November 19, 2013 08:36AM
Keep us up to date on this. I'm moving in your direction, but much slower winking smiley Managed to get OctoPi setup but still juggling work commitments to really get everything setup good
Sorry, only registered users may post in this forum.

Click here to login