#homecmos IRC log for Sunday, 2011-12-11

bart416btw, somebody asked in here a few months ago why you should use a bjt as a temperature sensor instead of a regular diode07:43
bart416I think I figured it out07:43
bart416The BJT responds faster to temperature changes as well as current changes07:43
bart416The junction in a bjt forms a better diode than a diode lol07:44
reportingsjrbart416: I remember that and as soon as I read your second to last sentence (I read from bottom to top sometimes) I thought of that08:07
bart416I'm not sure why I didn't remember it either at the time08:08
bart416But was going over a schematic of a temperature sensor once more and came to me heh08:08
reportingsjrI don't think the reason is what you stated08:09
bart416it was a discrete component08:09
reportingsjrI think the reason is that it's most likely much more accurate than a thermistor and it's also much, much cheaper than a thermistor08:09
reportingsjrerr08:09
bart416It's not cause it's easier on IC scale :P08:09
reportingsjrthat last thermistor should be a ...08:09
reportingsjrfuck08:09
bart416Also, in terms of accuracy a PTAT source is better08:09
reportingsjrI can't remember the name of anything today08:09
reportingsjrthermocouple!08:09
reportingsjrwith a thermocouple you need to buy the thermocouple (fairly costly compared to transistor) and also stuff to amplify its' signal.08:10
reportingsjrand with common transistors temperature effects on them are extremely well documented08:10
bart416Actually, in this application that part of it wasn't a problem08:11
bart416It's just that it also had to be precise at cryogenic temperatures08:11
bart416semiconductor junctions will do that08:11
bart416But size was also a constraint08:13
reportingsjrhuh, I don't remember it that well. *shrug*08:16
bart416If you want precision and ease and don't care about power use, go for the 4 transistor ptat source08:17
bart416If size and power usage is an issue, use a diode or transistor08:17
superkuhHow does one use a transistor as a temperature probe? I have a temperature controller using a platinum RTD I am not exactly satisfied with.08:20
superkuhObviously I can't just switch it out, the temperature controller wouldn't be compatible. But building something else is fine.08:21
bart416Fairly easy, your transistor (like any other semiconductor) actually has an exponential curve largly dependent on the thermal voltage08:21
superkuhOh. I should've figured it wasn't linear. Still interesting though.08:23
bart416http://www.eecg.toronto.edu/~kphang/papers/2001/haft_ref.pdf08:23
bart416sort of explains it08:23
superkuhThank you.08:24
bart416ask me again in about 12 hours, then I'll be closer to my notes from a few years ago heh08:24
bart416Got a really nice reliable design08:24
horizontallywait dude i worked on something related to this recently08:24
bart416Did require a transistor array though08:25
horizontallywhy isn't the platinum RTD not working?08:25
horizontallya thermocouple / RTD is much easier to setup but more expensive than a transistor / diode temperature sensor08:25
horizontallyyou could also buy a linearized diode / transistor temperature sensor like an LM35 or (better yet) LM3608:25
bart416what's the fun in that? /P08:27
bart416You know what's sad10:04
bart416The only well documented "discrete" design is the apollo guidance computer >_>10:04
bart416fucking retards10:26
bart416master slave flip flops are NOT edge triggered10:26
bart416An edge triggered latch is something completely different10:26
bart416A master slave simply uses a delay to respond to level changes10:26
bart416I blame wikipedia for this actually10:29
soul-dhttp://www.ted.com/talks/klaus_stadlmann_the_world_s_smallest_3d_printer.html12:35
--- Mon Dec 12 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!