Fighting with computers

Computers are not always friendly.

Monday, November 25, 2013

You get what you pay for ...

Last weekend I have hosted the fourth edition of a workshop where people build their own Prusa i3 3D printer. Trying to learn from past mistakes I attempted to cover all the basis, but unexpected things always come up. 

But what got me really angry was the poor quality of one of the boards we used. Once I realize that boards were like that, I double-checked each one, tracing back with the schematics the different pins involved in any faulty behavior observed. Re-soldering those pins did solve the issues discovered, but it was a painful and time-consuming operation, on hardware you already paid for and expected some QA on. (Well ... don't). 

This is the letter I sent to the manufacturer.


In my last workshop I have used 10 of your RAMPS 1.4 boards. In a previous communication I reported 3 out of 10 boards showing some fault. Because of that, this time I checked all the boards in advance. What I saw was a mess: 7 out of ten showed some malfunction. I used several hours fiddling with them till I've got all sorted out. All problems were due to faulty solder work (whether it is the solder material or the work, I do not know).

What's even worse, three of these boards crap on me when people connected them during the workshop (not sure they were the three that worked ok for me during the initial test) forcing me to have to fix them on the fly. (Mechanical stress during insertion might have trigger the fault as these were already tested by me).

I find all the above totally unacceptable, and while I am not planning to use your hardware in the near future, I wanted you to know why, before I mention it to other people.

Maybe you should consider to add a new line of product of "Known to work" or "fully tested" RAMPS boards. I cannot understand how such an awful statistics are possible unless boards are not tested at all. (By testing I mean some real test. Just putting a sticker on them does not apply).

Kind regards,

Miguel Sánchez
Valencia, Spain

Thursday, November 14, 2013

Mini UP! 3D printer, first impressions

A friend of mine recently bought a Mini UP! 3D printer from PP3DP. He was looking for a ready to use 3D printer within a budget and that model was my first advice.

Once I have seen and used the printer it meets the needs of my friend and it has impressed me on how easy it was to get the first print. However, it came almost without any instructions so many people may have a difficult time using it until they figure out that an online manual has to be available somewhere.

Print quality using defaults is very good and so is printing speed. Not so bright is the way heated bed works in this printer: no temperature sensor is there, so you have no idea how hot it is.

Large parts will warp badly unless bed is hot enough and well-leveled. Once you patiently wait for at least 15 minutes for the bed to heat up and level the bed, the adhesion is pretty good and any problems we had before were gone. Still, I do not like the idea of not having an accurate idea about the bed temperature.

On the software side, both Windows and OSX are supported, but the latter is not working nicely in Maverick (program will beach-ball for every print, sometimes print is interrupted before the end, sometimes not). Windows version seems to behave better, but only after you install the drivers, that can be found somewhere in the folder you installed the program (but oddly enough are not installed when you install the software).

Still, this small printer delivers solid performance in a small package and I like it. The light included in the hotend makes it easy to follow the printing process without opening the cabinet nor risking to cool down the heated bed.

You can build your own printer from a kit more cheaply, but it is a lot of work. This printer will save you some work and frustration (but it will take away some of the pride of doing it on your own).

Thursday, November 07, 2013

Programming Teensylu/Printrboard from Linux

I have been testing these two boards as a posible replacement for RAMPS in my 3D printers. Nothing wrong with RAMPS, but these two boards are simpler, cheaper and smaller. But they share a problem: they are not easy to program from Arduino IDE.

After a long dance I was able to do in Windows, and I am almost positive same can be done in Linux, but I am kind of tired of trying out things.

The bottom line is that because these boards are based on ATMega1286 they lack of a USB-to-serial chip as many Arduino boards do (except Leonardo).  So that means the usual bootloader and protocol are not a choice here.

Long ago, Paul Stoffregen developed an Arduino-like board using similar chips and built all the required add-ons so they could be programmed from Arduino IDE. These tools can be adapted to be used in the AT90USB1286-based boards with LUFA CDC bootloader and they work ok in Windows (once you manage to get every detail just right).

I was not so lucky in Linux and after some small research I realized that avrdude was used in Windows and that protocol involved was avr109.

So, mostly as I reminder to myself, I am using this line of code for uploading compiled code to my Teensylu board in Linux:

sudo ./avrdude -p at90usb1286 -c avr109 -P /dev/ttyACM0 -C ./avrdude.conf -U flash:w:Marlin.cpp.hex

I am using averdude version that comes with Arduino 1.0.5, not sure if it does matter or not.