so lets say I have this crazy idea to drive a shift register clock input with a crystal
but I have no idea where to start
Pointers?
Well a crystal is but a two terminal passive component. It will only generate clock pulses if it's part of a oscillator circuit. In a AVR processor chip the crystal is wired to pins that then connect to internal circuitry to cause the crystal to oscillate and generate clock pulses.
So if you wished to clock a shift register, you require a clock oscillator. One solution is to use a clock oscillator module that generates clock pulse. It has to be purchased with the clock speed you desire and the voltage level you require for the shift register. These modules use three active pins, ground, operating voltage input and clock pulse output. Here are examples of crystal oscillator modules:
Lefty
well poop, guess I should go back to digikey and make another order
(I used to like jameco, but their site is impossible to find anything on and they will jack me 9 bucks for shipping and take 14 days to show up)
I assume the pinout on these things are the same, since I only saw 1 datasheet with them on it?
I assume the pinout on these things are the same, since I only saw 1 datasheet with them on it?
Well the ones in a 14 pin DIP package are pretty standard, pin 7 ground, pin 14 Vcc, and output on pin 8 I think, with pin 1 being a no connection. (yes 14 pin dip package, put just four pins brought out!)however it's always best to get a datasheet.
Lefty
heh soon as I thought about it the darn things show up in the mail (with some other stuff) now they sit here mocking me
:-[
-
You can probably make a crystal oscillator using a couple of CMOS gates. For example:
Digital circuits page -
Some chips (other than microcontrollers) contain a built-in oscillator. I'm thinking of the xxxx4060, which has a crystal oscillator and 14 stages of clock-dividing flipflops, but I doubt whether it's the only one.
yea I have been looking at various schematics such as
but there is a few holdups
- its a 25.175mhz crystal and my scope is 20mhz
- it requires quite a bit of fooling around to do it right
- I am trying to keep this as simple as possible, with the fewest amount of components
guess it would have been better to do some more research
poking around on citizencrystal.com (which is who made the crystal) I am starting to get some basic idea
least its progress
I am curious to know what you are going to do with this.
Any data you send to it has to be synchronise to the shift register clock. This seems like more trouble than it is worth and takes more processor cycles than simply supplying the clock from the processor.
ghetto (vga) character generator experiment, Ill be dividing the clock by 8 which should give me a ~3.1 mhz pulse which (i am pretty sure) the arduino can pick up on to sync with the register
which probably sounds really hair brained after I bought the wrong crystal part lol
but the crystal oscillator is on order and will be here over the weekend so I dont have to fool with it
Its possible to build a crystal oscillator using an odd number of inverters chained together - for highish frequencies 1 might do, lower frequencies might need 3 or 5. Also probably need load capacitors too, and some experimentation (sometimes a crystal will run at a harmonic).
So if you've a spare invertor (or NOR/NAND) gate hanging around...
BTW a 20MHz oscilloscope will see 26MHz signal, it will be attenuated somewhat.
and some experimentation (sometimes a crystal will run at a harmonic).
yea its already going to be a headache project, I know it can be done but I would probably end up spending 10x as long trying to get that part running right then the whole deal
but thx
What are you thinking about trying, Osgeld? External clocking multiple ATMegas? True random number generator? Clock signal injector? S&Gs?
ghetto (vga) character generator experiment,
Well that link takes me back, it is the sort of thing I used in the 70s to produce writing on a TV set for decoding radio teleprinter signals. This was before micro processors arrived on the scene.
should give me a ~3.1 mhz pulse which (i am pretty sure) the arduino can pick up on to sync with the register
Do the maths, you are not going to do this even with assembler, this gives you roughly 5 clock cycles to detect an edge read some memory and write out data to the shift register.
well I was a bit smarter the second time around and also got an oscillator that will work for rs170 (for a much slower dot clock)
so my sit back and relax attitude towards this project, after nearly burning myself out on the led matrix project, has provided a interesting side effect
although YES i am having quite a "fun time" even with timing out monochrome qvga to something the arduino can "sync" with, fooling around I managed to create an odd pixel clock that works OK with text on a overlay to a 640x480 vga signal, and would be usable for a RS170 with a pixel clock that allows 100 pixels per line ( * undetermined x computer generated lines)
in theory :-/
the idea is a 25.175mhz clock dumps into a f /16 divider made out of 2 74x74 dual d type flip flops, which is providing a 1.9MHz clock
or 0.515 microseconds per clock cycle, RS170's active video time is 51.5 microseconds giving 100pixels per line
this is pushing the arduino quite hard, giving data look up, counting, and sync but I have some cards up my sleeve, including "dual mcu" willingness IF necessary, and a brand new old stock from a radio shack that is a throwback from 1987 (god bless you!), 4164, 64Kbit (8kbyte) dynamic ram chip
the ram is important for me to learn considering I am doing a 100% restore on my 1985 Apple //c (rom2) and the thing has 128KB of it, along with its pretty easy to upgrade in the future outside of my nostalgia as parallel ram has not changed much
Why all this trouble?
cause I want to learn more hardware!
if I wanted crap on a screen I have a pic32 that can do 640x480 monochrome in software no problem
even all the division and counting done by the logic chips and somewhat by the arduino could be done in maybe even a gal setup, which I have a tube of and a prom burner!
but I want to do it at the lowest level I can handle at this point for experience sake 
using tools that we all have access to, for if nothing else, learning
the ram is important for me to learn considering I am doing a 100% restore on my 1985 Apple //c (rom2) and the thing has 128KB of it, along with its pretty easy to upgrade in the future outside of my nostalgia as parallel ram has not changed much
You're restoring an Apple //c? What happened to it that caused it to stop working? I find it curious; I have a couple of //gs boxes in my shop that work fine (well, last time I played with them, which was a while back).
Now my Altair - that's a beast of a different color; that thing's going to be an absolute bear to restore when I get-round-to-it. As far as I can tell, it sat under a pile o' junk and has more dirt in it and such; ugh. Getting the power-supply up and running (without exploding - it likely needs new caps) will be the first job...
Good luck on your project; it sounds like a fun a thing to try a hand at. I know when I get around to working on my Altair, I might try building a TV typewriter for an interface to it (I have a copy of the "TV Typewriter Cookbook" - so it shouldn't be too difficult, outside of sourcing the components).
have no clue what happened to it, it will occasionally attempt to boot leading me to beleive its mostly rotten capacitors
I bought it from a mac collector a little while back for a song and a dance, I never personally owned one when I was a kid we had a IIe, but I always liked the looks of the IIc better + its small which is nice when in storage (its just barley larger than a typical laptop)
I bought it from a mac collector a little while back for a song and a dance, I never personally owned one when I was a kid we had a IIe, but I always liked the looks of the IIc better + its small which is nice when in storage (its just barley larger than a typical laptop)
Yeah, I had a friend who had one back in high school; seemed dinky compared to my TRS-80 Color Computer (not as small as a Timex Sinclair 1000, though). Of course, on the Color Computer end, we had the MC-10...
Its too bad you can't find any of the old school display chips; many of them would be nearly perfect for the Arduino (parallel addressing and data bus aside, of course).