#robotics Logs

Oct 03 2014

00:05 rue_house you will see what a fool your being
00:08 e_house whispers to everyone, for an unsigned 3 bit input, the 4 bit signed value is equal to the unsigned
00:11 marvimias rue_house, well, each bit Tn depends on S and all of U0..Un, so I'm not clear what "table" you have in mind
00:11 rue_house so, unlike the univeristy kid with the large logic circuit, just use wires
00:12 rue_house a table is an x-y layout of numbers
00:12 rue_house it relates one thing shows along the x axis to another thing shown along the y axis
00:12 marvimias hmm, I think I might have put the sense upside down on the U0..U(n-1) bits
00:12 theBear you're = you are
00:13 rue_house tables can be divided into columns
00:13 marvimias well, you have only 15 possible permutations
00:13 rue_house theBear, dude, I'm tired, can you show marvimias how to make a table of all the bits of a 3 bit unsigned number?
00:13 eBear whispers "pssswwswswswss - pass it
00:13 marvimias but they're spread across 4 dimensions
00:14 rue_house http://en.wikipedia.org/wiki/Two%27s_complement
00:14 rue_house there is a good example of a table in this page
00:14 marvimias rue_house, pack all the U bits in a normal integer, and then T=S?-U:U, and read the bits out of T
00:14 rue_house they have a column showing all the possable combinations of 3 bits
00:15 rue_house and the tranlation to signed and unsigned values
00:15 theBear nup, i'll do homework if i go to college, till then i know and that's enough
00:15 marvimias almost all CPU's do 2's complement; for those that don't, you can say T=S?~U+1:U
00:15 marvimias assuming T & S are both unsigned storage in C
00:16 rue_house your design was to convert 3 bit unsigned to 4 bit signed
00:16 rue_house all values of the 3 bit number are positive
00:16 rue_house which is a property of them being unsigned
00:17 rue_house the last bit of a 4 bit 2s complement number is the sign bit
00:17 rue_house if the sign bit is set, the value is negitive
00:17 marvimias I said, use an unsigned type in C; but interpret the bottom 4 bits as the signed value
00:17 rue_house if the sign bit is clear, the value is positive
00:18 marvimias otherwise what sort of 4-dimensional lookup table were you planning to use?
00:18 rue_house so the possable values for a 3 bit unsigned value are 0, 1, 2, 3, 4, 5, 6, 7
00:19 rue_house the unisgned 3 bits of the number 7 are 111
00:19 rue_house the signed 4 bit of the number 7 is 0111
00:20 rue_house if a person were to draw a table, comparing all possable values of a unsigned 3 bit number to the equiv 4 bit unsigned values, they would find that all the bits match with the highest of the 4 bit value always being 0
00:21 rue_house marvimias, ):P
00:21 marvimias yes. so we don't change them when the sign bit is 0
00:22 rue_house if your not good with tables, you will never understand state machines
00:24 marvimias and when the sign bit *is* 1, we invert all those where at least one lower-order bit is 1
00:24 marvimias problem solved
00:24 marvimias to negate a 2's complement number, you take the 1's complement (bitwise inversion), and then add 1
00:24 marvimias I was assuming you already knew that
00:24 rue_house for a 3 bit unisgned value going in, the resulting 4 bit output will *NEVER* have the high bit set (the sign bit)
00:24 marvimias that means you get the new sign bit for free out of T3 if you provide U3=0
00:26 marvimias what is your S bit for then??
00:26 rue_house I'm not sure if marvimias understands the question he asked...
00:27 rue_house I was asked what the relation from a 3 bit unsigned number to a 4 bit signed number was
00:27 rue_house the answer is they are equal, because of the limits of the input value
00:28 rue_house if I didn't know it in my head, writing out a table would show it clearly
00:31 rue_house I would probably write it like this...
00:32 rue_house 3 but unsigned in | value | 4 bit unsigned out
00:32 rue_house 0 0 0 0 0 0 0 0
00:32 rue_house 0 0 1 1 0 0 0 1
00:32 rue_house 0 1 0 2 0 0 1 0
00:32 rue_house 0 1 1 3 0 0 1 1
00:32 rue_house 1 0 0 4 0 1 0 0
00:32 rue_house 1 0 1 5 0 1 0 1
00:32 rue_house 1 1 0 6 0 1 1 0
00:32 rue_house 1 2 1 7 0 1 1 1
00:32 rue_house 'its ok Bender, there is no such thing as 2'
00:44 rue_house marvimias, if you ask someone for help with a problem, sometimes you have to open yourself to how THEY solve it.
01:02 rue_house ok, I need to work out how to delete an item from this char***
04:36 Tom_itx rue_house thinks in canadian
11:23 root66 Hi, I'd like to make something like this http://vimeo.com/93923609 but I am not sure what kind of motors these are or what size I would need to hold something like a drum stick. Any advice?
11:34 Hyratel root66, you need a servo and a custom bracket
11:39 root66 I guess what I am asking is what kind of servo, in order to quickly hit and release the drum with a reliable amount of latency. I imagine the really cheap ones wouldn't cut it.
11:40 root66 SG90's for example
11:50 Hyratel oh, for fast strikes?
11:51 Hyratel for that you need a voice coil
11:51 Hyratel or a motor and a return spring
11:51 Hyratel you need to have as little drag as possible
11:52 Hyratel also the shorter you can make the throw, the better, it'll have a snappier response
12:02 root66 Hyratel: I was thinking about rigging something up with a rubber band as the return spring and an SG90 holding something lighter than a drumstick. I just wonder if the claim of 60 degrees in 10ms is accurate or if it will vary slightly from motor to motor
12:03 Hyratel it will. that's why you have to calibrate
12:04 root66 ok, thanks for the input
20:49 rue_shop2 bingo
20:50 rue_shop2 the reactor got displaced and isn't coupled to the exchanger properly
20:50 rue_shop2 that expains why I couldn't get any proper heat out of it yesterday
20:50 Tom_itx easier than you thought ehh?
20:51 Tom_itx are you using this exclusively or as a supliment heat source?
21:01 rue_shop2 its a primary suppliment
21:01 rue_shop2 if I have to turn the gas on, i can
21:01 Tom_itx are you using chips or logs?
21:02 Tom_itx you need a conveyor you can feed a tree into it
21:33 rue_shop2 chunks
21:33 rue_shop2 from scrap pile at landfill
21:34 RifRaf tried osh park for pcbs?
21:35 Tom_itx yup
21:35 Tom_itx US mfg
21:35 Tom_itx only catch is the purple mask
21:35 RifRaf just got some, yep came pretty quick with free post
21:35 RifRaf i made some purple pcbs once :) so is kinda familiar
21:36 Tom_itx hackvanna is another one
21:36 Tom_itx he's an aussie like you
21:36 Tom_itx was commuting to china
21:36 RifRaf i have found a bridge though between 2 pads on the soic20
21:37 RifRaf trying to see if its meant to be that way though, might be in the design
21:38 rue_shop2 I'm getting the same output I did last night from it at half way to going than I did yesterday at full bore
21:39 Tom_itx don't work up a sweat
21:39 rue_shop2 theBear, RifRaf do you know if mp3 players will dial down the quality for slow processors?
21:40 rue_shop2 I upgarded the jukebox, and, dispite the difference of the missing white noise from its fans, I'm SURE there is more high and low end in the music
21:41 RifRaf hrm, did not think they would? wouldn't converting to lower quality use more processing than just playing a file?
21:43 RifRaf the bridge on this pcb is between the 5v and 3.3v input for the power in, pretty sure that should not be shorted
21:44 RifRaf looks fine on here https://644db4de3505c40a0444-327723bce298e3ff5813fb42baeefbaa.ssl.cf1.rackcdn.com/uploads/project/bottom_image/xaUeX7dY/i.png
21:48 RifRaf well it is bridged in the eagle design, so in that case there is nothing wrong with the pcb's, they look great
21:56 rue_shop2 hahah boiler delta is 14c! I couldn't get past 6 yesterday!