Back
[01:42:06] <rue_bed> no, but today I was watching a high performance servo video where the author commented that PD was desirable overpid
[01:42:16] <rue_bed> it was for a walking robot
[01:47:22] <Casper> rue_bed: do you know where I can get some good servo that could be powered 24/7 without issue?
[02:14:54] <rue_bed> hobby or 'real'?
[02:17:03] <Casper> whatever is reliable and not too expensive
[02:17:14] <Casper> will go in an air exchanger/heat recovery
[02:17:18] <Casper> to valve up the air...
[08:11:30] <Levitator> Greetings, sweet homies.
[08:11:53] <Levitator> I have just discovered that someone left an AVR in my Arduino. I have seen beyond the pale.
[08:15:43] <Levitator> Wows. Printf is kind of expensive. An entire 2kB. Otehnoes.
[08:16:48] <Levitator> MrFahrenheit:
https://www.youtube.com/watch?v=-BfmKUW0ZZ8
[08:17:21] <MrFahrenheit> that's the one
[08:18:12] <MrFahrenheit> although I've not heard this version before
[08:18:41] <Levitator> I don't think it's that different. Some extra guitar, I guess.
[08:19:01] <Levitator> Can never have enough Brian May, though.
[09:20:25] <Levitator> I really wish NetBeans would listen to me when I tell it where to read headers for Code Assistance.
[09:22:49] <Levitator> I have an intense dislike of silent failure.
[09:31:35] <twnqx> i think the size of printf comes largely from float handling, and there's a version that doesn't support floats that's smaller
[09:31:45] <Casper> Levitator: now, if you find printf expensive, add true float support and sin()
[09:40:29] <Levitator> Oh, incidentally, how do you add float support, -lm?
[09:44:25] <Levitator> Also, is there any C++ standard library support at all, or is it all C?
[09:45:09] <twnqx> basic float support is there by default, math support comes with -lm indeed
[09:45:22] <Levitator> Neat.
[09:46:07] <twnqx> no idea about c++, sadly. i know that a C++ compiler exists
[09:46:20] <Casper> twnqx: since when? last time I had to add float support myself...
[09:46:22] <Levitator> Works great so far.
[09:46:39] <twnqx> Casper: no idea, but a find /usr/avr shows me tons of libm
[09:46:41] <Casper> was quite hard to debug actually, because it didn't threw any warning
[09:46:50] <Levitator> Just haven't had any luck including standard C++ headers. C++ app code works fine as long as it doesn't reference the standard library.
[09:47:15] <twnqx> Levitator: i have libstdc++ on my system, so i guess not
[09:47:35] <twnqx> given that it seems impossible to get sub-MB c++ binaries on x86 i'd say they don't exist, though
[09:47:45] <Levitator> Would probably bloat the daylights out of the code, but was wondering if there was partial support.
[09:52:27] <Levitator> This is so much fun. I'm so disappointed in myself for not having played with MCUs sooner.
[09:59:39] <twnqx> heh
[09:59:48] <twnqx> better late than never :)
[10:12:36] <yids> hello
[10:12:55] <yids> anyone got experience with
http://munacl.cryptojedi.org/atmega.shtml ?
[11:31:32] <_ami_> is there any board for GPIO pins based on avr? probably GPIO pins over usb?
[11:32:12] <LeoNerd> ?
[11:34:34] <_ami_> LeoNerd:
http://nanorivertech.com/viperboard.html -- close to this.
[11:34:55] <LeoNerd> Ah.. similar to a bus pirate
[11:35:00] <LeoNerd> Quite a few devices in that space
[11:36:55] <LeoNerd> I've been pondering my own attempt at it. Probably a 32U4. Will ignore the bus pirate's text-based interface, so binary only.
[11:38:08] <_ami_> LeoNerd: i was thinking abt supporting GPIOs over IP through linux driver.
[11:38:22] <LeoNerd> OK. that's a quite different thing
[11:39:28] <_ami_> LeoNerd: do you think its helpful if implemented? as you know more abt hw than me.
[11:40:03] <LeoNerd> ?
[11:40:20] <_ami_> consider a ntwrk where you have a rpi2 and few wifi sensors (esp8266, other PCs, rpi2).
[11:41:05] <_ami_> now the sensors GPIOs (esp8266, other PCs, rpi2) will be shown as GPIO pins in your main server which is rpi2.
[11:42:10] <_ami_> so you basically write userspace code on rpi2 and deal with those pins locally.. and internally linux kernel or some other mechanism takes care of forwarding GPIO commands to respective sensors
[11:42:28] <LeoNerd> Eh..
[11:42:34] <LeoNerd> You probably wanted plan9
[11:42:42] <_ami_> plan9?
[11:43:08] * _ami_ googles
[12:02:32] <_ami_> LeoNerd: is it plan9 same? i don't think so. also it seems pretty outdated. 80s stuffs
[12:14:11] <sebboh> hia folks.
[12:14:50] <sebboh> I have no idea what I'm doing. FYI. :) Question: an atmega8 has an internal clk but I can also provide an external one, right?
[12:19:14] <sebboh> how slow can an external clock be before the chip fails to operate? I mean, 1hz, one per second... will all the values in the chip stay in place for that long wait between clk pulses? Sorry if 'pulse' isn't the right word. :)
[12:20:28] <LeoNerd> I think AVR core is all static logic, so in principle I don't think there's a lower bound to how slow the clock can be
[12:20:40] <LeoNerd> I know sometimes people run them off 32kHz watch crystals for example
[12:23:18] <sebboh> Has anybody made a computer in which the clock doesn't run when the thing is idle? As in... you know how a textbook 8-bit adder will ripple across starting from where the clk hit to when it is 'done'? In principle is there any reason why a whole computer couldn't operate in that manner? So the CLK is ... the enter key?
[14:13:15] <Levitator> Is there really no way to get C++ exception support?
[14:16:10] <Levitator> Because I've come up with a really dirty way to fake it.
[14:17:15] <Levitator> Make up a macro that wraps pseudo-throwing functions and templatizes them so that the default implementation fails a static assert, and the valid one is dependent an an argument which is the output of another macro which wraps calls with an exception test.
[14:17:44] <Levitator> So that way, there's static checking that you are trapping all throwing functions, and they can unwind the stack using ordinary return statements emitted by the second wrapper macro.
[14:29:16] <Levitator> sebboh: How many thousands of times do you want to hit the Enter key to complete an operation?
[18:34:12] <Lambda_Aurigae> sebboh, yes, an avr will run dc to max frequency...in other words, 1 hz or slower to its max frequency...and beyond.
[18:34:36] <Lambda_Aurigae> I've used a 555 timer in 1-shot mode to do single stepping through program on an atmega32...
[18:34:44] <LeoNerd> Hah
[18:34:56] <LeoNerd> single stepping by means of a clock pin :)
[18:35:09] <Lambda_Aurigae> yup.
[18:35:34] <Lambda_Aurigae> and a second 555 in free running mode to generate a slow continuous clock pulse train.
[18:35:59] <LeoNerd> Speaking of 555s, have you seen the TimerBlox chips?
[18:36:05] <Lambda_Aurigae> no.
[18:37:14] <Lambda_Aurigae> cute.
[18:37:42] <Lambda_Aurigae> all smt packaging though.
[18:47:24] <carabia> Lambda_Aurigae: you got a call from the 80's earlier, they said they have a brand spanking new shipment of dips for you fresh from the oven!
[18:49:46] <Lambda_Aurigae> sweet
[18:50:03] <Lambda_Aurigae> will go well with my 70s wire wrap.