I'm looking for an over-the-counter microprocessor to run Linux on. However I want to be able to construct the circuitry myself, so I will need a CPU that is in a package that mere mortals can work with. My assumption is that that would be a dual-inline package since that's what I've used in the past. But when I look at the Jameco and JDR catalogs I see just slow 8-bit CPUs. Can anyone point me to something more interesting, such as an system on a DIP chip type of thing?
Mere mortals can solder gull-wing chips onto boards, even when they're at fairly fine pitch. There's enough short-run PC board houses out there that you have no excuse not to lay out a board and use that.
But I agree with Sperho -- find an eval board (there's a gazillion ARM eval boards out there) that has the features you want and use that. Eval boards usually have headers full of undedicated contacts that you can use for your own circuitry. A very nice way to make a prototype, or even a one-off product is to get an evaluation board and plug it into your interface circuitry & use the stack as if it were a finished product.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
If you want to run Linux, then you don't want to develop a board. Buy a ready made single board computer (SBC). If you really want to develop a microprocessor board, then that's different.
CPUs capable of running Linux will almost certainly all be QFPs or worse. Actually QFPs are not so bad, and you can indeed solder them by hand. Oddly enough, even with a rather blunt tip iron suitable for typical "medium duty" soldering (like a 1/16" wide chisel tip).
So don't be put off by packages! I used to dread SMT, but now it's nothing.
Good day!
--
_______________________________________________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov -- NOTE: Remove "BOGUS" from email address to reply.
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
Actually I should have added something but I forgot:
I need this thing to be very rugged. The device will get banged up quite a bit and I'm a little concerned that these new 4-layer or whatever PC boards can't take shocks.
That's the other reason why I was hoping for a DIP package since their pins are thick.
Start a thread on DIP vs. SMT shock/vibration robustness.
--
_______________________________________________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov -- NOTE: Remove "BOGUS" from email address to reply.
Pins are thick because the package is large. Four layer boards are not new. Are planning to shoot this out of the cannon? How many gees does it have to endure?
You can also get the solder paste and heat gun type soldering stations for prices mere mortals can handle if you look for used stuff.
A dental pick made out of materal that solder hates like 316 stainless is a handy tool to have. You use it to break the surface tension on solder bridges.
I hope that you won't take this the wrong way, but my thoughts are that you need to learn a lot more about designing rugged electronics. Your belief that 4 layer boards and SMT parts can't take shock/vibration is unfounded.
Worst: Lead acid batteries (the second ones you buy) Cheap (sheet metal pin) sockets Mechanical switches that are hard mounted to the case Relays Cheap connectors More expensive connectors Large value ceramic capacitors Improperly installed radial leaded parts. Trim pots of any kind. Good quality PLCC sockets Machined pin sockets. SMT ICs of nearly any type :Best
Been re-working some SMD today actually. Fiddly but do-able with even just a Weller TCP iron plus fine tip.
Socket manufacturers like 3M, Augat etc do provide some sockets that take PLCC parts typically but convert them to through hole leadouts. I have my doubts that there'll be one with enough pins for a modern PC type CPU though.
Seriously, ( to the OP ) the low mass of SMD devices makes them highly resistant to shock effects.
One area where you might need to be concerned is in critical ( typically precision DC amplifer ) applications where any board stress can affect SMD resistor values. In this respect leaded parts are bettter since the lead acts as a strain relief.
Sorry, got to say you're wrong about machined sockets.
I know of various cases where either (a) vibration or (b) thermal cycling has caused DIL components to 'walk out' of machined sockets ! I'd love to see this tested on a vibration table and watch it in real time..
'Cheap' sockets are often better. They have effectively a no-return action. Augat make some wickedly good ones that grip the leads like velcro.
Well, the OP didn't say whether he just wanted to build it for the satisfaction, he has special needs that he thinks he can handle best, or some other reason. Hopefully, indeed, it is not to save money.
I hand wired an 8MHz 68000 board with SRAM. Later put the schematic into EAGLE just to gain the skill. Since the early SUN workstations were running 'nix on a 68000, it seems feasible to make Linux go there.
There was also the AT&T 3B2 unix machine, which was as I recall a 68010 machine. Presumably it ran real Unix.
A machine that I like is the LG860A. The folks I bought it from use Linux, but I am using an experimental OS
formatting link
and it runs from a CF onboard. The board has a "no moving parts" version that has a heat sink instead of a fan. It is supposed to use only about 14 Watts, and I think it is rugged enough to go into an attic in Texas (although I think one of the electrolytics in the switcher runs a little warm.) It is about half the volume of a freshman chemistry textbook. I haven't tried to break it, but it looks pretty rugged.
I have been thinking that the PIC 18F line has a lot of oompf. I haven't put it to pencil, but I have wondered if it has a much power as the original IBM PC (pre-XT, their first pc machine.)
The only main exception are the large valued ceramic capacitors. These crack easily if the PCB has a flexure shock running down it.
It is quite easy to filter out the really high frequency stuff that would be trouble to the smaller parts. Things like the on-off switch can be the biggest problem because they normally are hard mounted to the case. I think the best way to deal with this is to move the device back from the panel and use a rubber coupling.
I've done this. You must have bought some defective machined sockets. I've never seen a part pull out unless the PCB is allowed to have a large bending moment and even then the part did not come clear of the socket. It just pulled up enough that you could detect it when you pushed it back in. The fact that there was any pin vs socket motion, was enough to make us stop and redesign since such things fail over time as the gold rubs off.
I typically use 4 hours on 2X the "g" forces on a standard truck shipping profile, a small aircraft profile and 10 drops of 5 feet onto a hard surface.
Where I work, we don't allow "cheap" socket in the building. We have learned enough times that they are trouble.
In short, I stand by my order of cheap vs machined pin.
Pooh Bear wrote in news: snipped-for-privacy@hotmail.com:
Then what about the MC68000 - bloody huge dip (slow as a wet Wednesday) Dunno if you can still get them, but they made nice little tiles when I had to make a pad for a woodburner lounge heater. M
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.