Thank you Ed. Good to know.
I'll keep also this one in my loop then.

SF02

I didn't want to start a new topic, so i decided to write in this one.

My initial idea was to make a sbc, and since i had arm assembly class, i wanted to try an arm based sbc, but the idea didn't get very far, and i started building my 6502 sbc, and after that the 68000 sbc, they are still unfinished, but i hope i will finish them soon (the 6502 sbc is almost finished).
Now i was thinking to give another try to the arm sbc project(after i finish this two), so i started looking for schematics and stuff. I found some open hardware schematics, like the beagle bone(a bit more complicated design). I don't really need a sbc that can only do simple stuff (i have the 6502/68000 for that), so i would want to make it able to boot some kind of Linux/Unix system. But my problem is that i don't really now where to start from, the schematics are like reading Chinese (for now)...

I get the impression they made a real effort with the beaglebone to make it simple and to use parts which are obtainable and relatively easy to use. If the schematics are complex that may just be a reflection of how things are for that kind of system. It's good that all the materials are open source (and the product itself is only EUR 70)

I'm keeping an eye on the Raspberry Pi - it should be less than half the price and hopefully will be available in the next month or so. (The first run will probably sell out very quickly.) An image of the PCB has been published but the schematics are not yet out there (and might possibly not be released). However, the SoC is neither very available nor hobby-friendly - it's package-on-package.

The Pi will run Linux at 700MHz though, much like the beaglebone, and they have similar peripherals. But the Pi only has 8 GPIO versus 65 on beaglebone - it's intended more as a learn-to-program platform than a learn-to-control platform.

I realise the Pi doesn't help serve as inspiration for your SBC plans unless they do release schematics, and even then, the SoC probably makes it not a good starting point.

Consider the Chumby Hacker Board as a source of inspiration though - that's open. It's based on a Freescale ARM at 454MHz. Meant for LCD rather than video out.

Edit to add: another open system, with a non-BGA chip: http://opencircuits.com/Linuxstamp - it's a 2-layer board, a simple system, and it looks like it's a good explanation of how things fit together. Found it from here: http://electronics.stackexchange.com/qu ... ith-an-mmu

Cheers
Ed

Thanks for the links!

Is there any way of building a sbc like that without the fancy pcb board, something cheap and that can be experimented on? I think if i put all toes ic on dip adapters, and solder them as i did on my sbc, it kinda would not work at 100MHz...
And my second question is, how do i make a sbc so that Linux can be installed, where do i find the hardware requirements?

I think there are a collection of challenges there. I'm reminded of Thomson's Rule for First-Time Telescope Makers: "It is faster to make a four-inch mirror then a six-inch mirror than to make a six-inch mirror" - not that I even own a telescope.

To build your own linux for your own SBC, I think you might be best advised to first own an SBC and rebuild the linux for that. There's a lot to know, I think, about the boot process, about configuring the kernel, getting the cross-compiler sorted out, and putting a minimal userspace in place. It might be that the Linux from scratch project would be helpful. Except, booting an ARM system is different from booting an x86 system. (I know only a little about it, because I have a Lacie Network Space system running Debian. There's a wiki which might help set the scene. "Kirkwood" is a name for a Marvell reference design, I think.)

Another challenge is that for linux you'll need at least 128M RAM (maybe 32M will be technically possible. See Tinycorelinux FAQ. Maybe even 16M - see dd-wrt supported devices)

So, you'll need a fast wide RAM bus, but then you don't want to make an advanced PCB. That's a challenge! I think making a RAM to ARM connection which works reliably at some speed - any speed - is probably the big one. Choosing the most friendly SoC and RAM will help.

(Maybe getting one of those supported routers, and building dd-wrt or a similar linux from sources, would be a good starting project? Many of those routers are really old now and worth nothing, so they should be cheap.)

Cheers
Ed

There is a major problem with pcb-s, that is the reason why my 6502 sbc is also point to point soldering. When i am doing a new project, it is almost always something that i've never done before, so it will never work on the first try. Almost always when i build something new, it does not work on its first powerup, or after a while i decide to rewire some stuff, because i found out that there are better ways of doing things... So lets say i make a schematic for a sbc, and then i order it, and i solder all the components, and then i want to make changes and try just those changes to see how they perform, i would have to order a new pcb. I can't afford to order so many pcb-s. I would need to have someone that i can consult with while i am still design phase. While i was building the 6502 sbc, it was easy, i soldered only few sockets, leaving lots of free space on the sbc, and then i would start asking questions, i would get tips, and so i would build the next segment ect... If i want to build a arm sbc, i would need some major help with it, it would be great to see some tutorials for sbc like in the 6502.org projects, with all the details, like why do you have to put this resistor here, or why does this fit here, ect... It is kinda easier to make such tutorials for 6502 sbc, because it is simple, while arm is kinda complicated. But it would be nice to find something or someone that would guide anyone willing to build a arm based sbc in the right direction, because even I don't really know where to start from, and all the stuff i learned about computers, sbc, and electronics don't help much with this kind of advanced technology. It is like walking in a new unknown world, all the devices and methods previously used are now obsolete, and the new ones lack of good, basic, and easy to find tutorials.
I can only buy a sbc and play wit the peripherals, while i cannot design from scratch the mainboard and fit a functioning os on it. You get my point...
So any kind of help with this is welcomed...

Just to pick up on a much earlier comment:

The second of Sam's figures is from a link which is now defunct, so I did some searching and found this link which contains a figure of 36 Dhrystone for a 1MHz 6502, but also contains some rather higher scores, depending on the surrounding system and the compiler used. cc65 looks good at 52 Dhrystone /MHz, 40% better than others:



Thanks to Groepaz for doing the work!

Cheers
Ed

Well, indeed, all my comments about getting some system and experimenting with building linux from source are about addressing the novelties of getting a toolchain to work, building a kernel, getting the right device drivers, arranging for the right services to run at boot time and so on.

For the electrical and system engineering problems, it seems you need to build something. If you can find the stories of the progress of existing systems as they were designed and debugged that might help. If you can keep your system minimal and simple that must help. If you make your own PCBs that will be cheaper, but it's another learning curve.

Of course, if you ditch the idea of linux, and go for a multitasking forth, that's so much smaller you can use on-chip RAM and the system engineering is a lot simpler. Maybe some other minimal OS would do the same.

I agree that having advice and support is very valuable. Are there hackerspaces in or near your university? Could you even start one?

Cheers
Ed

If there are, i can't find them, the problem is that everyone who knows a thing or two about this kind of stuff will probably try to do something with software, or with peripherals, but not a sbc from scratch. I asked lots of questions to my professors, but it seems they started to abandon hardware design, and nowadays hey tend to do only hdl, and software projects...

I can see why - even with breadboards, you may have problems reproducing results reliably across a whole class of students. And fitting in multi-layer PCB design and manufacture, and soldering surface mount, would be quite difficult.

Speaking of surface mount, have you found a suitable ARM SoC which isn't BGA, or have you confidence you can deal with BGA?

BGA might be a problem, i only have a soldering iron with, but i do have a small tip, so i could solder smd components. Once i managed to solder a smd ic(my first ic) with a 100W soldering gun.

Here's a good resource:
http://electronics.stackexchange.com/qu ... re-not-bga
(by which I mean the site is worth searching(*), as well as it being a useful post)

Also, note that uClinux is a linux for systems without an MMU. That opens things up a bit.

Cheers
Ed

(*) searching or indeed using, by posting a well-worded question.
(*) I've just updated the previous post cataloguing open systems, having found a relevant recommendation from a post on http://electronics.stackexchange.com

It would be good if i could move this topic in a topic in which we could attempt to find out more about arm sbc, i would only need Garthwilsons aprooval.

I think we covered this pretty well here, and I posted my position and several agreed. What I probably should have been more clear about is that Mike did not give me permission to do just anything I want, but to help him keep the forum running the way he wants it. I don't want to betray his trust. Whatever he decides is what goes. When there was something major like starting the programmable-logic section, I did check with him first. I can ask again about a non-65 forum section, but it seems like we've already been through this. It's probably easy to start a new forum on Yahoo Groups, and you can put links to it here to get more exposure if you like.

I don't see any problem with constructive rabbit trails here and there as long as there is some relevance. (A superbowl topic under "General Discussions" does not qualify! The section subtitle does say, "Let's talk about anything related to the 6502 microprocessor.") Constructive rabbit trails can include discussion about other processors as long as there's something there we can apply to 65-family processors and programming and support. Posts about computer history, logic with tubes, core memory, etc. are acceptable since they are involved in how we later arrived at 65-family computers. The lines separating the appropriate from the inappropriate are fuzzy though, and I hope no one will put me in the awkward situation of having to be the enemy of anyone other than the spammers that I try hard to protect the forum from.

Ok, so until i find some other place to discuss this, i just write here.

BigEd posted a link http://opencircuits.com/Linuxstamp and this seems to be a good starting point. The device does not have much stuff on it, and the schematics fits on one page. I could use my 6502 sbc as a terminal(since it has a crtc) so i can have my pc free from that task(to be able to do other things). Lets say for start i would want to make a clone that would actually work.

Now i have 3 problems:
1. I don't know where to find the ic for a fair price without toes scyrocketing shipping prices, maybe the Unicorn electronics might help.
2. I don't have the fancy pcb, i could find a way to get a normal single layer pcb. I think if i put all the ic on adapter sockets and wire all point to point with ribbon wires it would spell disaster.
3. I also need a JTAG.

As for the questions i have only two:
1. Where does the 180MHz clock come from?
2. Why the power supply has lots of components?
3. Why all modern computers have coils in the power supply?