Good. God. / Computer Science / Forums

Forums

4hv.org :: Forums :: Computer Science

« Previous topic | Next topic »

Good. God.

previous 1 2 3 4 next

Move Thread

LAN_403

Bjørn

Wed Mar 14 2012, 12:16PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

I think RISC OS can map pages of other tasks to the same address to get them out of the way of the running task (all tasks think they run at the same address) but none of the versions I have will protect anything.

Ash Small

Wed Mar 14 2012, 09:36PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Am I correct in assuming that, with 8 GPIO pins, I can have 256 output 'states'?

Steve Conner

Wed Mar 14 2012, 11:09PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, and you can usually also set them to be inputs, which is a third possibility.

You can also blow them with excessive voltage and then you don't have 8 any more.

On modern hardware they may be 3.3V and not 5V tolerant.

The I2C bus can be used to add extra (slow) GPIO, using chips such as the PCF8574. This is quite old and clunky now, but newer chips may be available. I used three PCF8574s to add 24 extra outputs to my processor.

Ash Small

Thu Mar 15 2012, 05:32PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Steve Conner wrote ...

Yes, and you can usually also set them to be inputs, which is a third possibility.

You can also blow them with excessive voltage and then you don't have 8 any more.

Yes.........

I was thinking something along the lines of using the GPIO pins to 'drive' another chip, which sampled data from numerous other sources, and controlled other processes. The sampled data could then be fed back, possibly via another port?

simple, and reasonably fast, maybe?

(could presumably also be achieved using the USB ports, but that might well use up a lot more processor power?)

EDIT:

Steve Conner wrote ...

On modern hardware they may be 3.3V and not 5V tolerant.

I think I read that they connect direct to the processor, and run at 1.8v, but I'd need to re-read it to be certain.

Bjørn

Thu Mar 15 2012, 06:19PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

With 700 000 000 instructions per second (ignoring the 24 GFLOP GPU that is harder to use) there is no need to worry about computing power, only about bad code.

You can get as many input and output lines as you want using shift registers like the 74HC595 on the SPI port, leaving the GPIO free.

The voltage levels are 3.3 V in and out.

How to access the hardware under Linux (it might even work):

Steve Conner

Thu Mar 15 2012, 06:26PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, the shift register trick is commonly used, and will work if the SPI port does.

The Linux code looks quite reasonable. I guess the mmap() function with the MAP_FIXED option is what makes the MMU do your bidding.

Conundrum

Thu Mar 15 2012, 06:34PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4061

Nice, shame they didn't put onboard serial.
At least a USB to serial IC which is standard, then us PIC/AVR folk could program them directly from the RPi.

-A

Bjørn

Thu Mar 15 2012, 10:32PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

Yes, the shift register trick is commonly used, and will work if the SPI port does.

Even if they confused the lines to the SPI it will be possible to bitbang it at double or triple digit MHz.

That said this is a very bad choice for a microcontroller in most cases. It is a general purpose computer more powerful than the 7th PC I ever owned. It only makes sense as a microcontroller if you need to control very little that requires huge amounts of computing power and or a high resolution display.

Steve Conner

Fri Mar 16 2012, 08:01AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, it's a sexy solution looking for a problem.

My first experience of hardware hacking was with an Amstrad word processor. I got the joystick interface, and it came with a manual giving the IO port address if you wanted to use it in your own programs. I had great fun connecting random things to the joystick inputs and writing BASIC programs to interpret the results. My favourite was the real life Pong table using a mini hovercraft as the puck. Of course the games got a bit rowdy and it got smashed.

But anyway, I hope the RPi is capable of giving today's kids fun experiences like that. Formative ones too, in my case I have been making a living connecting random things to computers ever since.

Ash Small

Fri Mar 16 2012, 12:00PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

BjÃ¸rn wrote ...

That said this is a very bad choice for a microcontroller in most cases. It is a general purpose computer more powerful than the 7th PC I ever owned. It only makes sense as a microcontroller if you need to control very little that requires huge amounts of computing power and or a high resolution display.

Yep, my reasons for wanting to use it as a microcontroller are firstly, the MKS vacuum equipment I'm planning to use is/was designed to be controlled using very primitive RS232 technology, and secondly, the price (Â£20).

I may also use it to provide the signals for controlling the HV transformer gate drivers and magnetron drive circuits as well, presumably via buffer/slave chips.

(no huge amount of processing power required here

)

EDIT: I've not written much code since the days of the ZX80 and Jupiter ACE, though. (except HTML)

previous 1 2 3 4 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob