Saturday, August 3, 2013
Been on hold for the Texas summer - so hard to get out in that garage when it is 105 F out there... but I did recently acquire my very own drill press so things are moving along better.
This is a total spam post - selling the phone that started it all:
eBay listing for word.a.bot phone...
It was a good ride.
Wednesday, May 22, 2013
Well after a few long evenings in the garage redesigning the Z axis and working through various problems I finally got back into CAMBAM to generate some 2D tests - I'm a very pleased with the results thus far. Precision seems to pretty good without much tweaking and all axisis are functioning properly!
Sunday, April 21, 2013
X, Y .... Z?
Been a while, but I'm finally getting motivated again to keep the project moving - had to take care of a few "honey-dos" prior to moving forward.
Found a older computer (c.2004, Celeron 2.6ghz put a clean copy of XP) to push the GCode - will probably do the designing one of my non-garage machines since this is just barely fast enough that I don't take my sledge hammer to it.
Got the power supply (Goodwill, thank you!) all wired up to the GRBL shield, flashed Arduino with 0.8 GRBL and have the X & Y axisis running great. No binding or hang-ups. Only issue thus far is with the anti-backlash springs, which I don't think are quite big enough. They are getting hung-up on the treads, so I left them out for now.
Put some good time in last night with the Z axis - the more I work with this particular design, the more I realize that it is extremely easy for the small errors to add up into a machine that may not have the accuracy I'd like, especially when I don't have a lot of precision tools to build. A learning experience before moving on to bigger and better.
Saturday, February 16, 2013
Progress Check
Have the X axis all wrapped up and the Y axis all set as well - just working on the carriage supports, which, are a royal pain to get all lined up and the holes accurate. Also, having some trouble finding all the necessary hardware (Homedepot doesn't seem to know that M3 screws are actually produced).
Very excited to actually get a motor mounted (this week?)
Very excited to actually get a motor mounted (this week?)
Saturday, February 9, 2013
Wednesday, February 6, 2013
Word-a-bot is Getting Some Upgrades
With some new (to me) NEMA 17 steppers, some updated GRBL and a few ideas from MYDIYCNC.COM, Wordabot is about to get some vital upgrades. More to come!
Sunday, November 4, 2012
The Fun Part
Been a while, but unfortunately building Wordament playing robots is not my full time job... unlike some people...
So, the goal now is the match as many 3+ letter words as possible in a 2 minute window. Of course, the more letters the word has, the better the chance that you'll get some good points out of it - similar to Scrabble. Obviously 2 levels of touches won't get us many words, but as we progress, the combinations get exponentially greater. Level 2 - 84 touches, Level 3 - 408 touches, 4 letter words - 1,764. Even Excel, backed by an i5 has a hard time now matching up 6,714 rows against a 5 letter dictionary base of almost 9,000 5 letter words. That is a lot of number crunching!
Using my master 'possible touches' map I came up with a simple (to start with) formula in excel to try to find the next title in the series:
=IFERROR(IF(VLOOKUP($N3,$B$4:$J$19,O$1,FALSE)=0,"",IF(VLOOKUP($N3,$B$4:$J$19,O$1,FALSE)=$M3,"",VLOOKUP($N3,$B$4:$J$19,O$1,FALSE))),"")
Let me break that down a bit.
In this case cell 'N3' would be the last title that I was sitting on and the array 'B4:J19' would be the possible touches map. Now, lookup the last title I was on to get the next possible value. 'O1' would be a dynamic indicator of which column of the possible touches array that I was trying to lookup on. In this case, directly East which would be title number 3, however, this is checking for a zero, in which case, it returns a blank. The next vlookup checks to see if a title has already been used, if so, returns a blank, then the final vlookup actually returns tile 3.
This is the basis for determining all the potential values - once you have a completed set for a number of letters, those become the paths to the next letter, in which the formula above is applied adding and extra check to make sure that you don't pull in tiles that have already been used.
Possible 7 letter words:
=IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=0,"",IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=$Q3,"",IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=$R3,"",IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=$S3,"",IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=$T3,"",IF(VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)=$U3,"",VLOOKUP($V3,$F$4:$N$19,Y$1,FALSE)))))))
Looks like a mess when it is pasted down here, but the only difference is the check to make sure nothing is being used twice.
Once I have all those values, I concatenate the results and boom - I have my pathways.
On to dictionary matching techniques! Then GCode production similar to those AMAZING CNC milling machines and 3D printers! Stay tuned!!
Subscribe to:
Posts (Atom)
.jpg)
.jpg)
