trashing a computer?

I am attempting to make a fairly complex robot, with a.i and all. any recommendations for destroying a computer for usable parts? also, how does one get the software and hardware to interact so that the damned thing doesn't have to be connected to a computer all the time?

Reply to
J. T. Laurie
Loading thread data ...

hack saw and dremel tool... soldering iron and solder sucker.

take basic electronics (analog), advanced electronics (digital), and microprocessor courses at your local community college.

if you don't want it connected to a computer, you'll need at least a CPU on the robot, and a wireless data link.

Rich

Reply to
aiiadict

Reply to
J. T. Laurie

Use a hot air gun on the MOBO to get off the components you want. Also, I don't think there's a lot of useful stuff on modern computers, but 486's had lots of logic circuits you could use, plus connectors and so on.

If you want AI, well good luck. I'm happy with Artificial Stupidity. That's about the level of my robot work. But you'll need a computer of some sort c>>> I am attempting to make a fairly complex robot, with a.i and all. any

____________________________________________________ "I like to be organised. A place for everything. And everything all over the place."

Reply to
Tim Polmear

Hi JT,

your computer has drives (floppy, hardisks, CD or even DVD) and the motors can be stripped out and applied as robot motors. Even you fan motor could be converted if you can find a way to mount a gearbox on it.

Beware though that the motors that you will get are stepper motors and they are not very good for robotics if not used in feedback loops. This means that you will have to find a way to measure the motor position by either a small potentiometer or an optical encoder. The second solution leads to better accuracies and less noise.

The problem with stepper motors is that they are very good for constant load applications and they can drift if applied in variable load applications. Robots are complex mechanism and this emans that their load is always changing. Chances to loose track a stepper motor is very likely.

On the drives, some mechanical parts could be reused.

I will soon lead students to bult simple robots with bits and peices from different computers and also VHS players.

Have fun and good luck,

LHR

Reply to
laurentien

JT,

For the most part, computers aren't worth stripping for parts. You're not going to be able to use any of the chips on a modern motherboard in anything you hand solder one-off. As far as using the whole computer to control a robot, ATX computer systems are really only appropriate for large (say, 50+ lb) robots with huge batteries that require high computing power.

Computer PSUs can be reused if you need cheap 3.3/5/12v dc power, but you're probably better off with a *real* lab power supply, even a

1970s model off ebay.

The motors in disk drives are pcb-integrated synchronous pancake motors and are basically useless for robotic purposes. Floppy drives have one tiny stepper that is so cheap it doesn't even have actual bearings.

Ink jet printers are usually good for a few stepper motors -- or dc motors with encoders on a few models -- and a bunch of little gears and mechanical parts. When you take these apart, keep the mechanical subassemblies intact if you can -- gears that you know fit together are infinitely more useful than a bag of random ones. Small laser printers often have one somewhat larger stepper motor. The laser itself is invariable an infra-red band diode.

Junked UPSs are a great source of free SLA batteries -- especially since it's likely that such batteries have been maintained perfectly (remember, they are *always* connected to a digitally controlled bulk/ float charger) and cycled very few times, albeit violently. They also have several big mosfets. (You'll want to replace the solid-block heatsinks.)

Old keyboards have an 8049 mcu, which if you're really motivated you might actually be able to use. USB keyboards, not so much. Mechanical mice have a pair of quadrature encoders.

Don't forget what your time is worth. The other day I found a junked UPS (I live in nyc, I walk by a lot of other people's trash.) and got a pair of batteries worth maybe $30 each new for just the exercise of carrying them home. If you're spending half an hour ripping up a keyboard to get a $1 chip, however, you may want to consider if your engineering skills are really worth less than minimum wage.

Also, you cannot disconnect a robot from its computer. It stops being a robot if you do that. If this is all news to you, you may want to consider building a simple robot before the complex one.

-chris.

Reply to
e c kern

Reply to
J. T. Laurie

Maybe you are asking the wrong question. Why go all to that trouble?

From my limited experience as a student working on a second robot project...

Why not use a microcontroller... like a good old Microchip

formatting link
PIC. Or there are 16- and 32-bit ARM models around if you want that much grunt, but starting off with an 8-bit micro is a good way to start. There are several books on PIC robotics around, although they focus on BASIC programming. I have basically used what I have learned at college, and lots and lots of r e s e a r c h.

If you go with a PIC, there are cheap C compilers around less than $200, and some student versions for free from Hi-tech and Microchip. I bought a compiler produced by CCS from a regular here - Don at dontronics. The reason why I made that choice was that there are many libraries and functions available. Also check out mikroelektronika

formatting link
there are many people that use the forums there, and there are free compilers for trial use. Their compiler has many functions and libraries.

Microcontrollers are often more suited for robotics projects because of the many hardware peripherals available on-board the chip. A microcontroller is a SCC - single chip computer with IO, memory, and communications, etc all on one chip.And it's a lot easier to prototype and solder, power, hook up etc etc.

First you need to do a lot of homework and planning before you charge in

- the research is critical, or you will waste time, effort, and money and maybe save a poor old PC...

Even if you want some serious computation, you can get the micro to transmit the data to a PC wirelessly to crunch numbers if you like.

If this is your first project, maybe start with a simple design.

As a student at a technical college, I take my teacher's advice and develop simple systems first, and then link and develop these and build your skills level.

Cheers

:-]

Dale

J. T. Laurie wrote:

Reply to
Dale T Stewart

PolyTech Forum 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.