#robotics Logs

Dec 26 2012

00:52 home cyberbob: cool
00:52 home cyberbob: cool
00:52 home cyberbob: sounds like an awesome thing to study
00:52 home cyberbob: I wish you the best luck..
00:53 cyberbob home: actually there are many sub fields for machine conciousness, much confused which to choose. . .
00:53 Triffid_Hunter cyberbob: cognitive science and neurology.. get into stuff like the mirror test and optical illusions and child psychology with a specific focus on developmental stages
00:54 Triffid_Hunter cyberbob: it's all well and good being up to your eyeballs in computer knowledge, but doesn't help much if all you have on the path to consciousness is vague theories rather than a detailed knowledge of human consciousness development
00:55 home mirror test?
00:55 home such as being able to identify oneself and their surroundings?
00:55 MjrTom except that humanbiology has all the laws of physics to play with and computers do not
00:55 home MjrTom: +1
00:56 Triffid_Hunter home: when animals (and small children) have never encountered a mirror before and are suddenly confronted with one, they don't initially realise that its themselves they're looking at
00:56 Triffid_Hunter home: at a certain level of consciousness, they realise that its themselves
00:57 Triffid_Hunter this is easily observed because their interactions with the mirror change from the type of interactions they have with others of their species, to more introspective actions such as grooming parts of their bodies that are difficult to see
00:57 cyberbob Triffid_Hunter: :) Actually right now I am doing a research project (some sort of demonstration for mind simulation, am academic project) but I am planning to take "machine conciousness" as field of my passion and further research
00:57 Triffid_Hunter some animals never make that leap and will treat the mirror like another creature forever
00:57 MjrTom there might be primitives that intelligence agent genomes comes with; in this case, the primitives involved might be "self" and "others"
00:57 home Triffid_Hunter: wow :D thats awesome
00:58 Triffid_Hunter like budgerigars and many other birds, who will attack the mirror until they give themselves brain damage
00:58 MjrTom this "self" primitive could be the foundational basis for learning how to articulate the world
00:58 Triffid_Hunter some birds are smart enough to make the leap, eg the african grey parrots and a few others
00:59 Triffid_Hunter dolphins and some primates can make the leap, and I suspect octopuses could too
00:59 Triffid_Hunter although deciding if an octopus has recognised itself might make things tricky
00:59 cyberbob Triffid_Hunter: +1 amazing . .
01:00 Triffid_Hunter cyberbob: optical illusions are interesting because they allow us to discover the shortcuts that our visual cortex makes in order to process the massive volumes of visual data we consume every moment
01:00 MjrTom perhaps you could devise a test where an octpus needs to use a mirror to escape some entrapment
01:00 home Octupuses are smart
01:00 Triffid_Hunter if it didn't take shortcuts, we would not experience any unusual effects when looking at any pattern
01:00 home and other niderians also I guess
01:00 MjrTom yes they are very smart
01:01 Triffid_Hunter cyberbob: so, optical illusions are fascinating if your cognitive science involves machine vision
01:01 home I am pretty interested in machine vision XD
01:01 MjrTom two facets of octopuses are very interesting; they have the 2nd most developed eye (real world sensor) and an extreme capacity to articulate the world around them - both of these things are very important for intelligence
01:02 MjrTom * two of the facets of octpuses that are...
01:02 home they are pretty curious
01:02 MjrTom "how can I articulate the world around me"
01:02 home http://www.youtube.com/watch?v=x5DyBkYKqnM
01:06 cyberbob home: :)
01:06 home lol
01:06 home did you watch it? XD
01:06 cyberbob yup, watched while thiking something else :D
01:08 cyberbob I think a machine would be able to pass the mirror test. . .(I surely can say knowledge based robots can identify themselves)
01:08 MjrTom it'd largely be about the feedback between the real world sensors, the internal modelling, and the real world articulation
01:11 cyberbob This channel is amazing, loving it . .
01:17 cyberbob http://www.conscious-robots.com/en/conscious-machines/conscious-robots/can-a-robot-pass-the-mirror.html
01:27 home lol
01:27 home cyberbob: hmm
01:27 home I need to sleep though
01:28 home interesting.. 70% chance of tjose robots passing the mirror tests..
03:02 rue_house how to read chinese:
03:02 rue_house does it look like a telephone pole?
03:02 rue_house if so, it could be a lot of things
03:03 rue_house does it look like a circle? if so, its related to a wheel of some kind
18:21 Tom_L orlok, http://www2.electronicproducts.com/Image_of_the_Day_Holiday_Ribbon_Captured_in_Space-article-IMNE_hubble_holiday_Dec2012-html.aspx
20:49 GuShH theBear: why aren't you talking here?
20:49 ShH hands theBear a copy of the new total recall m
20:50 GuShH I gave it 3 boobs up.
21:46 rue_shop is that why it went tits up?
22:13 theBear here ? i just got back to the land with internet, the bunder dude was spreading misinformation like it was butter, and then i got a visitor
22:13 theBear and well played rue, very nice :)
22:51 rue_shop bunder?
23:01 rue_shop I love it when a project almost puts itself togethor
23:02 rue_shop "I need a priece of flat metal... like that one on the floor..."
23:03 theBear hehe
23:11 KimLaroux haha yeah that's always nice
23:12 KimLaroux On the other hand, it sucks when a project stall because you're missing a 20¢ part
23:22 rue_shop what I'm gonna like is "hmm I need a lathe to make this part... AND THERE IS ONE RIGHT THERE!"
23:24 Jak_o_Shadows haha
23:24 KimLaroux yeah... that will be nice
23:25 rue_shop I'm working on one of the robotic arms to see how long it takes before I need to make something on the lathe
23:25 KimLaroux hahaha
23:26 KimLaroux A solution looking for a problem...
23:26 rue_shop story of my life
23:26 KimLaroux so you invented the laser?
23:26 rue_shop heh, no
23:27 KimLaroux I have a problem looking for a solution, would you have somes to spare?
23:28 rue_shop prolly
23:30 KimLaroux I've been debugging this tube/mosfet headphone amp all day
23:30 rue_shop pre tube or pre mosfet?
23:30 KimLaroux It's got a weird dynamic noise in one of the channel...
23:30 KimLaroux tube as voltage gain, then mosfet as current gain
23:30 rue_shop so tube preamp
23:31 KimLaroux yeah
23:31 KimLaroux it's weird, it's like the problem is from the tube socket
23:31 KimLaroux =/
23:31 rue_shop maybe it is
23:31 KimLaroux I swapped tubes, but it stayed in the same channel
23:31 KimLaroux how could a socket introduce a dynamic noise?
23:32 rue_shop connector
23:32 KimLaroux I just can't describe the noise...
23:32 KimLaroux it kind of sound like something is arcing
23:32 KimLaroux randomly
23:32 rue_shop what the voltage on the plate?
23:33 KimLaroux around 25
23:33 rue_shop thats low
23:34 KimLaroux yeah I know
23:34 KimLaroux it's normally around 40
23:34 KimLaroux it's a low voltage tube, 12AU7
23:35 KimLaroux but the amp was designed for a tube that pulls half the current, so using the 12AU7 in there creates a very high voltage drop over the plate resistor
23:36 KimLaroux didn't bother to change the resistor as I'm planning on replacing it with a CCS, I'm just waiting for the parts
23:36 rue_shop if I didn't have my head burried in the design of a base frame for a robot arm, I'd look up specs and do maths
23:36 rue_shop kat, would be a great person to ask
23:37 KimLaroux so, I narrowed down the problem
23:37 KimLaroux it seems to be on a single side of the tube
23:37 KimLaroux dual triode...
23:37 KimLaroux works fine if I disconnect one of them
23:37 theBear au7 is a high current tube, not a low voltage tube
23:38 theBear anything under 100v is really pushing it for them
23:38 theBear and you got a resistor for each plate right ?
23:38 rue_shop the more problems I solve the more solutions I have to problems
23:38 KimLaroux not even
23:38 theBear AND finally, it's not unusual for half a tube to die
23:38 KimLaroux nah, it's not the tubes
23:38 theBear what makes you say that ?
23:39 KimLaroux I swapped them, and the problem stays on the same channel
23:39 theBear oh, just read back, ok, you gotta understand half my working life is spent asking that question then charging $110 to swap a tube and polish a case
23:39 KimLaroux hahahah!
23:39 KimLaroux XD
23:39 theBear what kinda noise ?
23:40 KimLaroux I don't know how to describe it
23:40 theBear and seriously, whoever made a headphone amp and decided to use 40v on the plates is well, stupid :)
23:40 theBear 40v plates ALMOST sound ok for a distortion pedal with 12a?7 but not for clean stuff
23:40 mLaroux shr
23:41 KimLaroux MANY hifi designs work this way and they sound amazing
23:41 KimLaroux even final products sold for a few hundred
23:41 KimLaroux it's not like we were loading those tubes...
23:42 theBear a few hundred isn't a hifi product :)
23:42 KimLaroux we're just using them as voltage gain, then couple the signal to a high impedance input of the output stage
23:42 theBear it's all about the amount of relatively clean/linear swing they can generate at those voltages, with any load, which isn't much
23:42 KimLaroux for an headphone amp, yes
23:43 theBear err, no, few hundred is a CHEAP headphone amp, heck, a cheap chinese one
23:43 theBear a cheap chinese solidstate one
23:43 KimLaroux well, 1 V of swing into an headphone is enough to give you permanent hearing damage...
23:43 theBear that has nothing to do with what happens in the circuit before the headphones
23:43 theBear and i aint seen any headphones with THAT much efficiency
23:44 KimLaroux http://schiit.com/cart/index.php?main_page=product_info&cPath=0&products_id=2
23:44 KimLaroux made in USA
23:44 KimLaroux 350$ and that's quite an amazing product
23:44 theBear and the name is shit, no doubt for a reason
23:44 KimLaroux look at their front page
23:44 theBear but none of this is helping you
23:45 theBear front page of an american site doesn't exactly convince me of anything
23:45 theBear checked the coupling caps ?
23:45 KimLaroux You seem to have a problem with it, not me
23:45 KimLaroux yes, if I disconnect it, the noise goes away
23:46 theBear yeah, cos its been my profession for around 15 years, i know better than you :)
23:46 KimLaroux as I said, if I disconnect one side of the tube, it works just fine
23:46 theBear so is there noise at both sides of the coupling caps ? is there dc on the wrong side of them when connected ?
23:47 KimLaroux there's DC both side, and that's normal
23:47 theBear what kinda circuit is it ?
23:47 KimLaroux the other side is the mosfet bias
23:47 theBear just that stage
23:47 theBear mosfet bias on a preamp tube ? oh dear, i'm out, this is a silly circuit
23:47 theBear and you do mean coupling cap and not cathode bypass right ?
23:48 KimLaroux http://www.pmillett.com/starving.htm
23:48 KimLaroux this design is very... silly
23:48 KimLaroux lol
23:48 KimLaroux but hey I learned a shitload building it
23:49 theBear oh, there are no input coupling caps, that's err, odd
23:49 theBear wow, that's, wow...
23:49 KimLaroux yeah, it's taking for granted the source is already ac coupled
23:50 KimLaroux hahaha
23:50 KimLaroux yeah...
23:50 theBear i suppose biasing it that STUPIDLY and unnaturally high will clean things up a little at the cost of cooking your tubes, but, wow
23:50 rue_shop be nice
23:50 KimLaroux can you explain?
23:51 rue_shop or I'll make beer cease to exist
23:51 theBear and the whole tube is parallel... hmm... you get a nice half b+ voltage at the right side of the cap ?
23:51 theBear rue_shop, i am being nice again
23:51 theBear not like kim designed it
23:51 theBear wow tho, it's just, horrible
23:52 KimLaroux hahaha it's not meant to be a piece of over-engineering
23:53 KimLaroux that's not the point of the design
23:53 theBear it's not under-engineered, just horribly breaking lots of rules
23:53 KimLaroux you have to realise the original cost around 30$ to make
23:53 KimLaroux yeah, so?
23:53 KimLaroux it works
23:53 KimLaroux that's amazing =P
23:53 KimLaroux lol
23:53 theBear heh, when you put it like that :)
23:54 theBear so you got a nice 24v at the right side of that cap or what ?
23:54 KimLaroux the mosfet side?
23:54 theBear between r8 and r10
23:54 KimLaroux 17 v
23:55 KimLaroux can't go higher, or you'll blow the tube heater
23:55 KimLaroux 17 v - 4v = 13v
23:55 theBear that doesn't go anywhere near the heater, that's what is driving the gate
23:55 KimLaroux 4v being the mosfet threshold
23:55 theBear and with no signal it should be 24v
23:55 theBear perfect 220 + 220 resistive divider with a high impedance load
23:56 KimLaroux nop, look where the source of the mosfet go...
23:56 theBear what does the source have to do with it ?
23:56 KimLaroux Ah yeah, that's the original using 19J6 tubes, it's biased different for the 12AU7
23:56 theBear you still got 48v -> 220k + 220k divider -> 2k -> gate
23:57 KimLaroux in the 12au7 it's 390 K and 220 K
23:57 theBear hmmm k, so do they line up on both sides then ?
23:57 KimLaroux cause the heater is 12.6 V
23:57 KimLaroux yes, perfectly
23:57 theBear hmmmm....
23:58 theBear it just doesn't make sense... can you TRY to describe the noise/distortion
23:58 rue_shop record it!
23:58 rue_shop the internet can be used to share more than just porn!
23:58 KimLaroux yeah, it's like the noise of a tomato sauce on low simmer... with slow bubles coming to the surface and slowly blowing
23:58 theBear it can ?
23:58 KimLaroux weird
23:59 rue_shop like the sound a transistor makes when it has way too much current flowing thru it?
23:59 KimLaroux it's everything but a static noise
23:59 KimLaroux I wouldn't know this sound
23:59 rue_shop line in, record it
23:59 theBear bubbles eh ? hmmmm.... COULD be a cap issue... oh wait, that's the heater, not the , oh god, this amp is crazy
23:59 KimLaroux =D
23:59 KimLaroux he just realised hahaha
23:59 rue_shop is there a schematic I dont know about?