On Thu, Jan 18, 2018 at 5:07 PM, Louis Pearson desttinghimgame@gmail.com wrote:

On a slightly different topic, how much would it take for you to design a eoma68 compliant netbook housing?

it'd be about the same amount of time as the laptop (about 18 months) as there is almost nothing in common that can be utilised between the two, with the possible exception of some of the libraries used in the 3D CAD design.

specific components that would need to be researched are:

* keyboard and securing a guaranteed supply of the same * trackpad and securing a guaranteed supply of the same * battery and securing a guaranteed supply of the same

each of these - on their own - will require contacting a hundred to two hundred separate suppliers, each of whom will demand, "who are you, are you real, are you one of our competitors fishing for information, are you a waste of our time, are you going to order 50,000 units because otherwise we're wasting the factory's time" and so on.

it's *extremely* complex basically and NOTHING that was done in the 15in laptop can be re-used. at all.

I was interested in the laptop, but it was (and still is) outside of my budget. Do you think you could make a netbook housing that is more budget friendly, or are there too many unknowns for you to say?

having done several now i know pretty much exactly how long it will take, that's not the main problem. the main problem is, you have to find a sponsor willing to pay my time to do it, and at the end of all that, you then have to have a crowd-funding campaign where a minimum of 250 preferably 1000 preferably 10,000 people *actually want one*.

l.

Not to mention that, at least in terms of hardware, there's very little that's standard about laptops, ever -- the display protocol, sort of maybe, and the drives, and that's about it. There;'s a reason those machines tend to go together and come apart like a jigsaw puzzle without the box!

Forgive a top-post, please, Luke - I'm on my phone.

Coreboot, IIRC, is a replacement for BIOS/UEFI. So if you have the original system's motherboard intact - in which case you cannot drop in the chip you want to drop in - you can replace the contents of what is essentially the boot ROM chip with coreboot. That's as far as that goes...

In case you do not understand what BIOS and UEFI are, read on...

When a computer is first turned on, the CPU automatically copies the contents of the BIOS (or UEFI) into RAM, which is called 'shadowing' the ROM. It then jumps to a specific address, hard-coded into the CPU, to start execution of part of those instructions.

For the record, historical processors typically started either at 0x0000, assuming a 16b address bus, or at 0xFFFF. The 8086 and 8088 did something different, and I forget now what that address was, but it was in the middle somewhere, IIRC.

*ahem*

The code executed at boot is enough to test how much RAM is present and functional, and to bring up various parts and pieces of the system so that it can function cohesively and coherently. Hard drive interfaces and accessing. Some sort of display function for output. Keyboard and mouse interfaces. *Et cetera*.

Once this is complete, and a limited 'sanity test' (POST, the Power On Self Test) is executed, the BIOS (UEFI) code loads the OS bootloader into RAM and begins executing that - whether it's GRUB or NTLDR, that is the part where the OS begins to take over. The bootloader pulls up the kernel and whatever init program is present, and away you go.

...all that to say that coreboot basically can't help you here, because all of that is what coreboot duplicates, and that's *all* it does.

Sorry for the long yarn of explication, but I wanted to be thorough. My hand hurts now, though, so "here endeth the lesson", as my mother often says :)

On Jan 18, 2018 9:35 PM, "Bill Kontos" vkontogpls@gmail.com wrote:

On Thu, Jan 18, 2018 at 6:58 PM, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:

...

no.

it is a vast amount of work. the LCD has to be researched (if its datasheet is even available). a conversion circuit has to be designed and manufactuered.... and before that it is necesssary to work out if there is room for it.

the keyboard hsa to be reverse-engineered

the trackpad has to be reverse-engineered

the connectors have to be researched (heights, sizes), PCB heights measured.... or you have to cut holes in the casework to get the PCB to fit.

the battery has to be researched and reverse-engineered, paying attention to safety as you could set fire to it if you get it wrong.

What if the laptop in question has coreboot support?

---

arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk

BTW -- not *everything* is nearly as complex as Luke would have you believe. Close, but not quite, and the two doozies more than make up for the easier bits...

Keyboards are invariably a passive switch matrix (look it up if you don't know -- you should, it's worth your time) and not that hard to reverse-engineer if you have a couple to burn through with a multimeter and hook probes. It's a matter of an afternoon or two. Touchpads are almost always either USB or PS/2 and not some alien bull**** (relatively speaking) like I2C or SPI unless they're bullt into a desktop keyboard (I have seen this, particularly with wireless keyboard/mouse combos). You probably can find a datasheet for the 'pad's controller chip and that will give up the info on how it connects to the world.

It's the screen and battery interfaces that are going to get you and get you good, time-wise and effort-wise.

LCDs are *very very very* rarely VGA (a *very* few *very* old laptops did that) and are often either LVDS or (if in a nearly brand-new machine) eDP. Once in a particularly blue moon, you'll potentially find one that runs parallel RGBTTL as its interface (my mother had a cheap tablet with an SVGA [800x600] screen that was RGBTTL.) The problem with LVDS is that the spec excludes a standard pinout and leaves that to the manufacturers -- so you've got to basically get in there with an old-fashioned oscilloscope and poke around until you have an understanding of which pin does what -- not a simple task when there's forty or fifty pins on a connector! ...oh, and some will be logic-level (typically 5v but sometimes 3.3v or even 1.8v) and some will *ahem* not be logic-level but rather analog or differential signals, which is why a logic analyzer won't do you here -- you'll blow its input circuits sky high. (...which you *really* don't want because logic analyzers are expensive... at least, the *useful* ones are.) RGBTTL is easier -- you can use a logic analyzer for that, if the datasheet doesn't have the pinout (which it basically always does) -- but, again, there's not really a standard there as far as pinout is concerned, and you have a fiddly surface-mount connector with remarkably tiny pin pitch to deal with as well -- after all, if you can actually solder to a flexible PCB (which is what those cables always are) without ruining it, you have soldering skills of near-mythical level and I have a few projects for you to help me with :P ).

The battery invariably runs SMBus (short for System Management Bus) for communications, which is an I2C variant -- not too bad to deal with -- except that, like LVDS, there's no standard pinout, and oh by the way you need a charging circuit as well as knowing the commands to send and receive over SMBus to make it charge, discharge, and read out its level. Hint: the industry standard name for the controller chip in the battery is 'fuel guage', oddly enough -- you'll likely need to find its datasheet (good luck!) to get the commands, unless you really want to blackbox a battery that can literally burn your house down if you don't handle it with kid gloves. (Seriously, look up lithium battery fires on YouTube. If that stuff doesn't scare the p*ss out of you, either you're one serious pyromaniac or you have no bladder.)

If you can handle all of that, you can do *exactly* what you're looking for. Personally, I'd rather figure out a way to basically make a portable desktop from readily available components (which I've done three times now, actually) and make do with that.

--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Fri, Jan 19, 2018 at 2:59 PM, Bill Kontos vkontogpls@gmail.com wrote:

On Fri, Jan 19, 2018 at 4:57 AM, Christopher Havel laserhawk64@gmail.com wrote:

...

Forgive a top-post, please, Luke - I'm on my phone.

Coreboot, IIRC, is a replacement for BIOS/UEFI. So if you have the original system's motherboard intact - in which case you cannot drop in the chip you want to drop in - you can replace the contents of what is essentially the boot ROM chip with coreboot. That's as far as that goes...

Ok so it doesn't really do that much on it's own. Got it, thanks. I was thinking more along the lines of old thinkpads which have a decent amount of reverse engineering done to them. Even the x230 has the keyboard layout reverse engineered by a sysadmin who wanted to soehorn the old 7 row on it.

it's just not worth it and drives up the price of the second-hand thinkpads, and pisses everyone off who is selling them to libre conscious people as they can't make a business case for even bothering to buy and convert them to libreboot.

l.

On Mon, Jan 22, 2018 at 8:50 PM, Bill Kontos vkontogpls@gmail.com wrote:

On Fri, Jan 19, 2018 at 5:04 PM, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:

...

---

it's just not worth it and drives up the price of the second-hand thinkpads, and pisses everyone off who is selling them to libre conscious people as they can't make a business case for even bothering to buy and convert them to libreboot.

So buying a thinkpad and converting it to an ethical device is pissing off those who buy thinkpads and convert them to ethical devices?

yyyup! the effort by the vera-apparatus team caused a *lot* of problems when they started buying up used thinkpad laptops to convert over. the supply is extremely limited.

I'm not really into retro computers as these people seem to be but from my understanding they are very negative to each other as if we don't have enough to fight against already.

that would be one way of looking at it.

... or... it could be the case that it is hopelessly unrealistic and just not worth the time and effort - in fact due to the amount of time and efffort it takes to do the disassembly and conversion, in combination with the fact that second-hand machines are firstly distributed world-wide so must be shipped to one location for disasseembly and second they're in short supply _anyway_, it could be viewed as being highly environmentally IRRESPONSIBLE to spend significant energy resources on converting such products.

in other words: when you add up the amount of time and effort proposed to be spent, and convert it to an actual dollar amount, i estimate that it would come to an amount that would EASILY fund the development of an entirely new type of computer.

one that can be designed to be repaired, upgraded, respect software freedom and not end up in landfill.

... .yeh?

l.

On Tue, Jan 23, 2018 at 1:26 AM, zap calmstorm@posteo.de wrote:

...

in other words: when you add up the amount of time and effort proposed to be spent, and convert it to an actual dollar amount, i estimate that it would come to an amount that would EASILY fund the development of an entirely new type of computer.

one that can be designed to be repaired, upgraded, respect software freedom and not end up in landfill.

... .yeh?

You are correct, and I wish I had realized this a lot sooner. My bad...

yyehh i've been down this evaluation path a number of times now on this list, with different groups of people at different times. it... kinda puts a dampener on peoples' enthusiasm for doing home-grown "hackaday" style projects... but... hackaday projects are for people to learn (and teach other people) electronics. this project is *specifically* about reducing *world-wide* e-waste on a *massive* scale by making desirable long-term upgradeable computing appliances, thus keeping stuff out of landfill as long as possible... and that *has* to be done not by disassembling pre-existing deeply flawed "Designed for Obsolescence and Manufacture" products but by going *right* back to the very source of the problem.

totally different approach that's really hard for some people to understand or accept, the scale is about a hundred thousand times larger than they're able to get their minds around.

l.

On Mon, Jan 22, 2018 at 11:21 PM, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:

in other words: when you add up the amount of time and effort proposed to be spent, and convert it to an actual dollar amount, i estimate that it would come to an amount that would EASILY fund the development of an entirely new type of computer.

one that can be designed to be repaired, upgraded, respect software freedom and not end up in landfill.

I think you misunderstand the use case here. It's a personal project. I would totally do it myself if I had the skills. Besides I don't see why minifree & the likes are entitled to the entirety of the used thinkpad supply. It's not like their fight isn't a losing one anyway.

--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Fri, Jan 19, 2018 at 2:35 AM, Bill Kontos vkontogpls@gmail.com wrote:

What if the laptop in question has coreboot support?

that helps identify some of the components and peripherals... but all of those peripherals wtith the exception of any "management" ICs actually built-in to say the battery will be discarded, won't they?

so coreboot will have all these options which #define ICs... that are going to be chucked out into landfill.

l.