#milkymist IRC log for Saturday, 2012-08-11

lekernelwhat program can I use to compute the electric field generated by multiple electrodes with a given geometry and at given potentials?07:51
mumptaicst microwave studio?08:26
mumptaior is it just "an oddly shaped capacitor at DC"?08:27
lekerneljust static field08:47
mumptaithat's way simpler, but right now i'm not a aware of a software for that (that won't also do the complex stuff and comes with that cost)08:49
mumptaiwhat are you trying to solve?08:52
lekernelquickfield looks pretty good, if only it weren't windows and evaluation-type license08:52
lekernela pretty simple problem in fact :) four electrodes arranged in a square08:54
lekernelwith two potential differences for X and Y08:55
mumptaisome years ago i had a 3d frequency domain solver, that should be able to do that08:55
lekernelthe end purpose is to be able to generate an arbitrary and constant 2D electric field inside the square by varying those two potentials08:55
lekernelI want to simulate it to see how much the electrodes for one axis will mess up the others, and what I can do to prevent this (eg space the electrodes etc.)08:56
mumptaireduce the poles to infinitesimal small charges and calculate the superposition of the filed of all four at any place you are interested in08:56
mumptaiand be a bit carefull that the reduction of the pole size stays valid08:58
lekernelhow can an infinitesimal small charge generate a constant potential (inside the electrode)? this doesn't sound like a good model...08:59
mumptaiit isn't09:01
mumptaiits just simple09:01
lekerneltoo simple I think :)09:01
mumptaibut it hold up as long as you are far away from it09:01
lekernelI'm not. the very purpose of this simulation is to see what happens close to the electrodes09:02
mumptaiif the electrodes are big compared to their distance, it won't work too well09:02
mumptaiokay, than you need to work with the integrals covering the electrodes surfaces09:03
lekernelthis is holding the two vertical electrodes at half potential11:45
lekernelif I don't... there's a very bad "shielding" effect http://imgur.com/La0Ue11:47
mumptaibut could let the side plates just float12:06
azonenberglekernel: unrelated question, are you familiar with NoC architectures at all?12:06
lekernelmumptai: which requires transformers in the drive electronics ...12:10
lekernelazonenberg: a little bit12:10
azonenberglekernel: Would you be interested in commenting on a very early draft of a paper i'm writing?12:11
azonenbergRight now i'm basically proposing what i want to make, most of the code isn't actually implemented12:11
lekerneljust ask... :)12:11
mumptaior high-value resistors12:11
mumptai(comparable to the resistor ladder in a pmt)12:13
azonenberglekernel: see PM12:14
larschow does a NoC work? Instead of a register read or write you send a packet and at some point get an answer back in another packet?14:25
lekernelusually yes14:26
lekernelthe problem, of course, is that every transfer has massive latency14:47
larscwell I suppose that's one of the interesting reasearch topics in that area, how to keep the latency reasonable low14:58
--- Sun Aug 12 201200:00

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