/diy/ - Do It Yourself

>>1783562
>5-color pangolin sinew
uma delicia

>>1783744
>assuming exist somewhere
found the econ major
>dedicated watchdog timer IC for a 2 day timeout
if they do exist, they're either
1. designed for one particular socket on one particular device type for one particular buyer, not adjustable, and wouldn't be worth listing in the catalog, or
2. secretly a microcontroller
Panasonic once had a long period timer AN6780, with a built-in divider chain, but neither they (nor Matsushita semiconductor) are around anymore. why not? MCUs won and working engineers don't have time for that shit

>>1783790
>can knock out a weird waveform generator in 5 minutes
if that works for you, ok
>not having a CPU, programmer, and skeleton projects ready to go
that's 200 frustrating hours that you should have gotten past already, so that you could have written the code in 2 minutes
>main()
>{
> const int period = 172800, on_time = 1;
> for(;;) {
> PORTB |= PIN0;
> DDRB |= PIN0;
> delay_ms((period - on_time) * 1000);
> PORTB &= ~PIN0;
> delay_ms(on_time * 1000);
> }
>}
> // avr-gcc -o pulse.elf pulse.c
> // avrdude -U flash:w:pulse.elf:e -U (fuses as required)
et viola, PB0 sinks tens of mA every 2 days for one second, ±2% at standard voltage and temperature. no close-tolerance components required (but you could add a crystal very easily for ±0.005% accuracy over the full commercial temperature range)

>>1783807
>unless you want to do babby projects forever
I... I think they do, anon

>>1687971
did you look inside the tape to see if the edges were starting to ruffle? have you cleaned the head with 99% IPA and a Kimwipe? (don't use cotton swabs directly on the head, the fibers will fuck your shit up)
get a little mirror on a stick so you can see what's going on down there. Harbor Freight has Chicagoland locations right?
>measure belts
place belt on flat surface, hold two pencils upright, separate until they pull the belt into a narrow oblong with as little stretching of the belt as possible, mark, measure
if you can't successfully google up a "suitable" (forget about exact) replacement belt, another trick you could try is heading to your local auto parts store and picking up some P.B. Blaster, which is sometimes used to rejuvenate automotive belts
mechanical alignments should be in service manual territory

>>1688054
>milliseconds
jeez grandpa
the STM32F030 would be much more than plenty. it's got a core that can do big math at a suitable speed (no hw multiply, sorreh), a 12-bit ADC that can be automated (some STM32F have two or more), and most importantly its timers are fire
you may be able to do this with a 555 and a few more components. need more constraints. need more input

>>1688102
>t.hipster
many magnificent things have been done even within those limitations. open-loop gain and negative feedback absolve many sins. however, it did tend to cost quite a bit and cooperated heavily with the macro-scale physical domain, using odd materials and components that are hard to get and use today thanks to Western deindustrialization among other reasons, and isn't *as* practical anymore for that reason
physics also don't help. the bulk of any interesting analog or digital circuits that you can get to work on a plugboard can be done in software with much less fuss and alignment. analog is really only interesting at speeds where traces need to be treated as transmission lines

>>1687999
show tits and coat hanger walker

>>1646068
>muh DAC
there are highly integrated USB DACs that you could incorporate right on the board
most USB device stacks also include sample code to bridge a USB audio class function to an I2S DAC, or make it trivially easy to build one. USB also allows for a device to offer multiple functions to the host, so you could add in one or two 3.3V-level async serial ports or whatever, just update endpoints in the code and touch up the descriptors to match