Back
[04:25:38] <Jartza> oh yeah
[04:25:55] <Jartza> vga + uart works. screen clearing works without losing uart data.
[04:26:00] <Jartza> next, control sequence support
[04:26:09] <Jartza> I'm planning to support subset of ANSI escape codes :)
[04:26:35] <Jartza> https://www.dropbox.com/s/i8b0a8e2019y058/1439024513.mp4?dl=0
[04:44:58] <LeoNerd> Jartza: Bonus points if you can implement DECDWL / DECDHL
[05:00:14] <Jartza> :)
[05:00:21] <Jartza> LeoNerd: I'm not sure if I want to :D
[05:01:08] <Jartza> theoretically possible, but I think the characters are already double-width and double-height all the time :)
[05:01:13] <Jartza> with 32x14 characters
[05:12:59] <LeoNerd> Mmmm
[05:13:00] <LeoNerd> True
[05:13:18] <Jartza> ahh. fixed tiny glitches showing on screen during clear. much better now.
[05:15:16] <Jartza> this is getting bloated, already 880 bytes of binary
[05:15:22] <Jartza> plus 2kB of font data, of course
[05:16:06] <LeoNerd> So, just over a quarter of a tiny85
[05:16:44] <Jartza> yea
[05:17:20] <Jartza> well, actually binary and font is already a bit over 1/3rd
[05:17:31] <Jartza> 2928 bytes total
[05:18:01] <Jartza> I guess the ANSI-support will bloat the code quite a lot still
[05:19:51] <Jartza> I had to clear the screen in chunks of 64 bytes so I didn't break the hsync
[05:39:36] <LeoNerd> Morning RikusW
[05:39:51] <RikusW> Morning LeoNerd
[05:39:55] <LeoNerd> I've just been playing with current-sense amps. Got a lovely capture here of the current draw out of an intelligent LED
[05:41:01] <LeoNerd> http://home.leonerd.org.uk/local/screenie/TS1102-vs-ADM4073.png
[05:41:34] <LeoNerd> The white (ref) curve is the TS1102.. 50kHz bandwidth thing.. horribly slow. The blue is my shiney new ADM4073, which claims 1.8MHz
[05:43:32] <LeoNerd> I feel totally justified in spending £4 for that chip now
[05:47:29] <RikusW> interesting
[05:48:21] <LeoNerd> I've turned it into a tiny little current shunt probe for my 'scope. It's literally a board with just that chip + a shunt resistor of 0.1ohm on. then connectors for power + scope output.
[05:48:40] <LeoNerd> Really simple. :) Also a lot cheaper than the ~$1k things people like Tektronix want to sell me
[05:49:32] <RikusW> the white doesn't look like the blue at all..
[05:50:00] <LeoNerd> It's what you'd get if you took the blue line and smoothed it out with a capacitor
[05:50:10] <LeoNerd> It just doesn't cope with that sharp peak
[05:50:24] * RikusW got a tek RM561A scope (vacuum tube type)
[05:50:31] <RikusW> its ancient...
[05:51:19] <RikusW> and it still works
[05:52:57] <RikusW> http://www.barrytech.com/tektronix/vintage/tekrm561a.html
[06:04:07] <LeoNerd> Mmm.. proper oldschool there ;) I have a Rigol DS1054Z
[06:04:19] <LeoNerd> (which took the above screenshot)
[06:16:35] <RikusW> I don't use it a lot, but its handy to have :)
[06:16:51] <RikusW> I did get a new probe for it though
[11:44:38] <Jartza> wtf
[11:44:45] <Jartza> why is attiny sram full of crap when I power it up
[11:45:04] <Jartza> I thought the memory was all zero in boot
[13:48:58] <hypermagic> hello my friends
[13:49:28] <hypermagic> Jartza, actually sram is volatile only if power supply voltage goes below the data retention voltage level
[13:50:03] <hypermagic> also, one does not assume memory to be empty before use, unless you use calloc/zero it out
[14:56:17] <Casper> Jartza: the ram on avr is not zeroed up at boot, so will be at whatever that the noise caused it to be at power up. A reset will also not reinitialise the ram, but will "zero" the registers (but maybe not all)
[15:35:25] <Jartza> Casper: well yeah, the memory seems to also corrupt in reset. at least on this attiny :)
[17:05:25] <Jan-> hihi
[17:08:09] <Jan-> this may be a stupid question, but can I (ab)use the spi peripheral to send data out to normal nothing-special shift registers?
[17:09:46] <Jan-> I mean, it seems like it oughta work
[17:09:53] * Jan- peruses datasheet *headscratch*
[17:12:29] <ali1234> Jan-: yes, absolutely
[17:12:43] <ali1234> SPI is just a shift register anyway
[17:12:52] <Jan-> thought that made sense
[17:12:57] <Jan-> thanks :)
[17:13:03] * Jan- is a beginner, cant'yatell
[17:13:31] <ali1234> also you might want to look at the MSPI port, which lets you use the serial UART as SPI
[17:13:59] <ali1234> the benefit of doing that is that the UART has a FIFO, where as the normal SPI peripheral does not
[17:14:14] <Jan-> um, er
[17:14:21] <ali1234> that's handy if you're synthesizing arbitrary bitstreams with tight timing requirements
[17:14:21] <Jan-> oh, you mean, more pre-sending storage
[17:14:26] <Jan-> right
[17:14:27] <ali1234> right, yes
[17:14:36] <Jan-> *nodnod*
[17:14:49] <Jan-> It's just lighting 7-segment displays.
[17:15:16] <Jan-> I'm mainly interested in reducing the cpu workload, rather than timing it very tight. The actual tight timing is on the strobe line of the shift registers.
[17:15:29] <ali1234> you know you can just buy a chip for exactly that right?
[17:15:43] <Jan-> You can, but this needs to be camera-safe. So, I need it not to be strobed.
[17:16:22] <Jan-> therefore brute force approach. one 8 bit shift reg per 7-seg (plus decimal point)
[17:16:28] <ali1234> fair enough
[17:17:05] <Jan-> it does work :)
[17:18:32] <Jan-> also it's super easy to get the parts.
[17:18:53] <ali1234> MAX7219 isn't exactly hard to get
[17:19:16] <Jan-> never used 'em
[17:20:23] <ali1234> max scan rate is 1300 Hz though
[17:20:30] <ali1234> so probably not camera safe at all
[17:21:13] <Jan-> no.
[17:21:24] <Jan-> I mean often it's not the end of the world
[17:21:42] <Jan-> but when you can do it right.
[17:22:25] <Jan-> the other thing is that it's very quick to update from a cpu point of view. and you can flick one strobe line and the whole display updates, like *plink*
[17:23:33] <ali1234> yeah, you can chain the shift registers too so you just need SPI and one GPIO for the strobe
[17:23:38] <Jan-> that's the plan
[17:23:44] * Jan- has a plan!
[17:29:02] <Jan-> I'm not sure how fast I can update a load of 595s with 64 bits of data. Fast, I guess.
[17:29:36] <ali1234> faster than you can see for sure
[17:30:26] <twnqx> but faster than a cam can see...
[17:31:01] <ali1234> not an issue when you're not multiplexing the segments
[17:34:00] <Jan-> I guess the current drain's gonna be quite high.
[17:34:15] <Jan-> still
[17:34:23] <Jan-> there's always the special technique called Off Mode.
[17:34:55] <ali1234> shift registers might not be able to supply enough current for your LEDs
[17:35:21] <Jan-> I have ULN-2803s
[17:35:22] <LeoNerd> There's some special version of a 595 with high-current drivers
[17:36:14] <ali1234> that's gonna be a huge board if you make it using DIPs with those external drivers :)
[17:37:49] <Jan-> yeah - although they're gonna be like 1.5 inch displays
[17:37:56] <Jan-> and there's eight digits, so y'know
[17:38:00] <Jan-> it's pretty big anyway
