On 31 Oct 2011 at 9:47, David Riley wrote: We're approaching the time of year when my field of vision is polluted with those damned flickering LED outdoor Christmas tree light displays. I must have some sort of visual impairment because I can pick out the flicker from a mile away--literally. It reminds me of a nest of ants crawling all over the display. It seems that the usual method of implmentation (in the US, anyway) is to use a string of LEDs to self-rectify the AC line current, resulting in a pronouced 60Hz flicker. I picked up a couple of strings of said lights last year (GE brand) and hooked them to a fullwave bridge rectifier, with a single large (220 ?F filter cap). The flickering ceased, but the quailty of light is rather harsh, particularly when compared side-by-side with an incandescent string. l suspect it won't be long before consumer and legislative pressures means that our Christmas seasons will be uniformly polluted by hordes of ants dancing over our lighting displays. Moral: Even a 60Hz multiplex rate is too slow for some people's vision. --Chuck

On Mon, Oct 31, 2011 at 9:29 AM, Chuck Guzis <cclist at sydex.com> wrote: It might be an interesting project to take a few strings of these lights and build something like a Life display. You would be the envy of all of your retro-computing neighbors. http://www.deepdarc.com/2010/11/27/hacking-christmas-lights/

Den 31/10/2011 kl. 17.29 skrev "Chuck Guzis" <cclist at sydex.com>: Chuck why so Dark :-) lighten up! Here in the EU we have 50 hz.... And i agree. I see it to, and hate it to. Regards Benjamin

Is that when every pin can act as either an 'anode' or 'cathode' driver, so you can ewffectively have an LED between any pin and any other pin? -tony

I think that's required to avoid other LEDs turning on, isn't it. If a pin has to be either high or low, then (without giving it much thought) can you do any better than conventional multiplexing? -tony

Sure, I meant if you added no external logic. If you regard those 6 oupuyts as row/column drives to a conventional LED matrix, yhe bes you can do is to split them in hald and have 3 rows and 3 columns, for a total of 9 LEDs. I think that for this sort of circut, given n outputs : n even : (N^2)/4 (or (n/2)^2) n odd : (int(n/2) * (int(n/2)+1)))/4 It beats normal multipelxed LED matrix (not adding any external logic). But for charlieplexing to work, and not haev other LEDs turn on, you need indivdualy-controllable 3-state outputs -- that is any of the pins must be able to be high, low, or high-impedance, independant of all the others. -tony

Yes, but I am no so foolish as to assuem that all thw world (or even all of digital electronics) involves microcontrollers. Understnadign how a tirck like this works and what it requries (and why) means you can use it, or a modified verison of it, in other situations -tony

Yeah...as a kid, I had a calculator that used ten-electrode neon tubes ("Nixies") for display, and, while I don't remember all that much of the driving circuitry, I think it multiplexed. (It was somewhat broken, and - hardly surprisingly in view of my skill and knowledge at the time - the more I tried to fix it up, the brokener it got. I wonder what became of it.) /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

On Oct 31, 2011, at 6:17 PM, Tony Duell wrote: Most of the time, these days, that's all it's used for; I've seen some app notes where, if you're really clever, you can drive 4 LEDs per pin (plus 1 extra) without any additional external electronics. Very nice for either pin-constrained micros or large sets of LEDs. For example, most cheap clock radios have one 4-pin DIP running the whole show (and it's pretty much been the same chip since the '80s). It uses the 60/50 Hz zero crossing on the AC line to keep time, which gives it very precise time (Laurent Hammond, inventor of the AC synchronized motor which ran both home clocks and the Hammond organ, took advantage of this and now the power companies must make sure that the power cycles average 60/50 Hz over the day to a few ppm, if anecdotes inform correctly) and also gives it a handy time base for multiplexing half the LED segments. Normally, as mentioned earlier, a 60 Hz refresh frequency might cause a headache, but I imagine there are holdup caps on the LEDs so they're not quite so obnoxious. - Dave

On Nov 1, 2011, at 4:29 PM, Tony Duell wrote: Yeah, I corrected that later. Silly mistake. :-) May be true; the few cheap clock radios I've disassembled to try to fix (because their alarm noises were particularly effective and they were no longer made, not because I was too cheap to pay $9 for a new one) certainly all had the same chip (or generic functional equivalents) driving them. They were all multiplexed drivers, and the datasheets went into great detail about how to wire up 7-segment displays properly. Admittedly, it's a statistically invalid sample (about 3, if memory serves). Perhaps I'm misremembering the size of the DIP; it could have been a 28-pin. Those clocks are long-dead (they turned out to have dead ICs of one type or another, and since they were all long out of production, replacements were unavailable), so I can't look them up again. - Dave

On Tue, Nov 1, 2011 at 9:35 PM, Mouse <mouse at rodents-montreal.org> wrote: This is a very common arrangement of one-chip inexpensive LED clocks in the US. I did some clock hacking a while back because I wanted a 24-hr clock at the South Pole and those are hard to find from US sources (not so in NZ or Oz, but those are 220V - the 50Hz vs 60Hz is often a jumper option, but the voltage is usually a fixed-winding transformer, not so easily hacked). Nice. My hack was to take a larger clock (3" digits) implemented as a bunch of 805-sized SMT LEDs on a PCB and populate the missing LEDs (onto bare pads) to allow the 10s-of-hours digit to render 2 as well as 1 (cheap American clocks can only render blank or 1). The clock chip won't try to render a leading 0 when in 24-hr mode, so only six of the seven segments were on copper on the board. I got the 805 LEDs from a pile of dead Quantum IDE hard drives - the size and color of the activity lights were a match for the bare dice and flying wire LEDs installed on the clock at the factory. Yep. I've hacked 3-4 types of cheap alarm clocks and the 12/24 mode pin is usually right at one end. Some clock PCBs even have solder pads to make an easy jumper. The only issue I've seen is the cheapest clocks only have 2 segments populated in the 10s-hours place. If you made your own digits, then you wouldn't run into that. As you should be - reverse engineering, repurposing, extending functionality... hacking all. -ethan