Back
[00:55:48] <anton02> are there filters that can filter by decay rate instead of just frequency?
[01:26:39] <Thrashbarg> anton02: filter what by decay rate
[01:27:57] <anton02> the signal
[01:28:06] <Thrashbarg> what component of the signal
[01:28:16] <anton02> amplitude
[01:28:25] <Thrashbarg> so you want a voltage controlled amplifier?
[01:28:46] <anton02> no
[01:29:22] <Thrashbarg> I'm still not clear on what you're trying to do
[01:29:45] <anton02> say you have a decaying sinusoid
[01:30:00] <anton02> it decays by 1/3rd every 2 microseconds
[01:30:04] <Casper> signal -> carrier -> AM -> low pass
[01:30:19] <Casper> -> ADC
[01:31:16] <Thrashbarg> anton02: go on... if you already have that decaying sine wave do you want to measure its amplitude?
[01:31:57] <anton02> i guess that would be part of it. but then end result is that i want to measure the decay rate
[01:32:11] <Thrashbarg> okay so you want an envelope follower?
[01:32:31] <anton02> i want to disregard everything else like a filter would disregard all other frequencies than the frequency of intereest
[01:32:43] <Thrashbarg> what frequencies are we talking about?
[01:33:35] <anton02> 100hz
[01:34:23] <Thrashbarg> so you could probably send your raw signal to your ADC and do the filter and detection in software
[01:35:24] <Thrashbarg> then again the idea is obtainable in the analog domain too :P
[01:35:50] <anton02> so in other words its pretty hard
[01:36:01] <anton02> on a low end micro
[01:37:01] <anton02> i was hoping there'd be a certain filter config that would work
[01:37:27] <Casper> you want a DSP, not a filter
[01:39:11] <Thrashbarg> can be done with an ATMega I think but it's "good, fast, cheap, choose two"
[01:50:03] <Casper> you forgot "easy"
[01:54:13] <Thrashbarg> that'd probably come under 'cheap'
[01:54:16] <veek> easy isn't as important nowadays :p
[01:54:19] <Thrashbarg> heh
[01:54:34] <Thrashbarg> it's a business saying, cheap covers labor costs
[02:05:36] <Casper> 2:40am...
[02:05:41] <Casper> and I'm hungry... grrrr
[02:06:16] <Casper> Wondering what to cook... because it's late...
[04:23:22] <sabor> veek: have you found out how LDS works meanwhile?
[04:23:37] <veek> sabor, hey yep
[04:23:42] <sabor> kk
[04:23:52] <veek> bad book :p thanks though :)
[04:24:03] <veek> well good book but spotty
[04:24:22] <sabor> hehe
[04:24:45] <sabor> the avr instruction set manual is also a good thing to read
[04:24:58] <sabor> explains a lot of this low level stuff
[07:48:32] <LeoNerd> Hrm.. struggling (and failing) to find any AVR DMX receiver code online that actually pays attention to the timing of the break and mark-after-break conditions
[07:48:49] <LeoNerd> All of them just ignore it, meaning they'll get very upset by tiny little break conditions if the timing down the line is just slightly off
[07:49:27] <LeoNerd> Currently I'm not quite sure how it's possible to implement.. maybe you'd have to use PCINT on the RXD pin and count delays using a timer on that?
[07:56:28] <LeoNerd> Wow, even Mat Hertel's one doesn't :(
[07:59:10] <LeoNerd> In summary: everyone on the internet is wrong
[08:22:56] <Tom_itx> hah
[09:14:42] <LeoNerd> Well, this is awkward. The only way I can think of to do break/make-after-break timing analysis on incoming DMX signal involves keeping the UART RX turned off during break/mab and use PCINT + a timer on the RX pin, and if the timing is OK then enable the RX afterwards. Problem is, the RX pin is in the same PCINT bank as the SPI port's SS
[09:14:48] <LeoNerd> which I'm also PCINT'ing
[19:44:30] <LeoNerd> Ugh, oopsie. Turns out that DC/DC converter isolators tend to have minimum load requirements. So my 1W (== 200mA at 5V) converter would want to see a minimum of 20mA
[19:44:40] <LeoNerd> MAX481s tend not to draw even close to 20mA. Ever. ;)
[19:46:09] <LeoNerd> Even the smallest of the range, the 250mW (50mA) would want to see 5mA, which even then is ten times more than the MAX481 draws
[19:46:43] <cehteh> time for some load resistor :d
[19:47:21] <cehteh> or add more leds to the project .. blue led for power on haha
[19:47:41] <LeoNerd> Hah.. I'm gonna need a lot of LEDs to get to a minimum load of 20mA
[19:54:34] <Casper> or use a better one
[19:57:50] <LeoNerd> Are there any you know of..? Something that can go down to microamps of load?
[20:03:50] <carabia> probably not microamps, or they'll prolly be hella expensive
[20:04:09] <carabia> should get down to single digits in microamps anyway...
[20:04:39] <carabia> but unless it's battery-powered, who cares. It's not like a dozen milliamps saves the fucking world.
[20:05:20] <LeoNerd> Huh? No I mean, minimum load
[20:05:29] <carabia> Yes, and?
[20:05:38] <LeoNerd> Most of the DC isolators I find all want 10% load or they basically don't regulate the output any more
[20:05:52] <LeoNerd> If you draw only a few microamps from the socalled 5V output, it could be anything up to about 10V
[20:05:53] <carabia> I know... Did you not understand what I said?
[20:06:12] <LeoNerd> I'm wondering what you mean about being battery powered
[20:06:44] <carabia> well, bleed resistors or running 20 millis worth of leds can be a bit meh, if your device will be battery-powered, eventually
[20:06:51] <carabia> but if not, who the hell cares.
[20:07:14] <Casper> LeoNerd: I wonder, is a bit more voltage possible?
[20:07:20] <Casper> if so use a shunt regulator
[20:07:36] <Casper> so when the load increase it will basically disconnect itself
[20:07:45] <LeoNerd> Hm.. I suppose I could just stick a tiny 7805 on the other side
[20:07:54] <LeoNerd> or an LDO version anyway
[20:08:01] <Casper> 5.6V zener?
[20:08:41] <carabia> Seems like I got myself in a pickle anyway. I randomly got this lcd without a nv charge pump. Amazing.
[20:08:42] <LeoNerd> Oh hmm... like that
[20:08:55] <LeoNerd> That stands a chance of working
[20:09:21] <cehteh> LeoNerd: do you need to stepup or stepdown?
[20:10:16] <cehteh> for that little load, why dont you just use a linear regulator if it is regulating down
[20:11:08] <cehteh> err forget it .. isolator :D
[22:54:28] <carabia> Heh, figured it out. I'll just use a spare max232 to generate the contrast bias voltage for this lcd. =D
[22:55:30] <carabia> I saved a few $$ (...) and bought a bunch of these, note to self that not all lcd's contain charge pumps.
[22:57:21] <carabia> I'm only getting enough contrast for a nice viewing angle at around -6v