Replace roms with flash? Increase RAM?

Many of our hpcs have removeable roms, essentially for upgrading to a newer OS.

Would it make sense to remove the rom chips from the daughterboard and substitute compatible flash chips? Ok, we need to find pin compatible flash memory. That doesn't seem to hard,
Then we need a way, preferably through serial/usb to reload the flash chips with the OS.

I can certainly think this would be ideal for the MP world where many are considering/looking for rom updates for their devices. If there are guys like Simon out there willing to expend the time we could have custom flash images with everyday programs like vxutil and cecommander in flash..
This might not be a bad idea for people conidering linux as well.

Perhaps it would be easier to make new daughterboards with blank flash (provided there's a tool to flash it) and why not increase the ram with pincompatible larger capacity chips?

some pipe dreaming but I see a lot of cheap consumer devices with newer OSes flashed in and added ram.

edit: I understY

Edited by wallythacker 2006-01-20 2:12 AM

From my limited digital electronics knowledge... wouldn't you need the Flash to match the read speed of the original ROM too? I have no idea if that's even a problem tho...

Also, can you elaborate on the benefits? The only thing I can sorta understand is Linux, because there are so many, erm, geeks working on it.

But for the benefit of HPC, I don't think you can even access any extra data you stuff into the replaced "ROM", unless you rebuild the OS.

OK it's almost midnight and I have a weird feeling that I might have said something stupid in the above

Yes, it's possible and people on the Jornada linux mailing list have been looking at it. There's no reason why we couldn't increase RAM/flash significantly, but the effort involved in creating the design then tooling up for production of a very limited run of complex SMD devices means the outcome is expensive...

I will chase the leader on the mailing list to see if progress has been made.

BTW, people are replacing the flash on their Zaurii with larger capacity chips

Yeah, I heard about the Jornada one. And they're only planning to do this for Linux on ARM Jornadas (7xx and possibly 820). Like I said I see no way for this to benefit CE tho. Even with the population of Jornada, they still haven't come up with enough people willing to invest in it. Perhaps it's cheaper to get a large CF or even a Microdrive just to run Linux.

Yeah I can see all those Linux, ahem, geeks wanting more space for their Zaurii Question tho, can they screw play with the Zaurus OS just like any Linux?

Anyway I think replacing a Flash with a larger Flash is one thing, replacing system ROM with Flash would be quite another (in terms of software).

Oh and I forgot one thing. You can't possibly write to the "ROM" thru the system, so you need to create a seperate module to write to it. Like a board that you can plug the "daugther board" into, and has a connection to parallel port (USB would be too complicated) so you can flash the chip with a PC or something.

Aha, so there is some work going on. Good. It makes sense to me. Having a large flash would allow, among other things, a nice safe storage area or a place to install essential programs. I like what Mullenger has done with the SL4 in providing a fair sized flash file store. So in effect it's similar to enhancing the OS.

I would *think* there would be a fair market for say, 128/256mb ram boards for 720's and/or 64/128mb flash rom modules.

The MP series seems a likely market also because of the swappability of the roms between the various models. I'd be very interested in large flash and ram modules for my MPs.

edit: Why couldn't we write the flash rom through the system ala ppc? I admit it would take some serious skills but with some of the incredible things I've seen done so far I wouldn't be surprised.

Edited by wallythacker 2006-01-20 2:25 PM

Because it's not just a flash, it's a flash in place of the ROM. Even with a PPC you cannot replace a ROM with flash and expect to be able to write to it.

I think you are right. The mechanism for writing to the flash rom needs to be in place too. Think of how a CDR or CDRW *cannot* be written onto if you put it into a CDROM drive.

Well, bypassing the direct writing to flash for the moment, there's no doubt all hpc2k users would benefit from larger ram chips. hpc2k is capable of addressing 256mb (in 32mb maX chunks) , more than adequate IMO to bring our devices to pretty modern specs.

Back to the flash, it wouldn't be the first time external writers were needed to write flash and it won't be the last time. Again, using Mullengers 4.2 image, Simon managed to add goodies the to image as well as fix many bugs, so I don't buy the fact that extras like vxutil/gsfinder/portrait couldn't be added to a hpc2k flash based image. Or a way couldn't be found to utilize the spare flash space for persistent storage like Mullenger has done.

It's certainly worth thinking about. If we had some fanatics (think how the linux crowd has done so much for so many platforms) with intensive electronics skills all that I'm suggesting is within reach.

wally,
While I do agree that increasing RAM is mostly beneficial... to be honest it might not be worth the cost. IMO the 728's 64MB of RAM is just about the right size for HPC devices. Yes I was excited about the 128MB RAM in my Smartbook when I first got it, because it's just like anything else in personal computing, the higher the performance the cooler the device However I quickly found out that I really do not need all that memory. And sometimes I do wish my battery doesn't drain 5% every night just to keep all those 128MB alive.

I know many people like to install a boat lot of programs on their PDAs, but not many of them actually use all that software frequently, even among the biggest geeks. Installing less-frequently-used programs on memory card is more efficient than getting more RAM.

About the Mullenger release... nothing was added to an existing image. The whole image was built from scratch. This is not an easy task, and I think being a former Siemens employee would have to do with how Mullenger did it. Just one more time about the file storage... the hardware was flash to begin with, not a ROM chip replace with flash.

Many great things can be done, it's only a matter of who can do it. I am sure HP can easily release a CE5 upgrade for 72x, or NEC can make that MP1000. But that doesn't necessarily mean someone else can do the same

Well, IIRC IPv6 wasn't in cenet 4.1 initially.

Are you saying Simon completely wrote his image from scratch? If you are that supports my position that someone else could rewrite a 720/MP image from the get go.

If Simon used the 4.1 image as his starting point my same position applies. Someone else could certainly mod a hpc2k image to add new functionality.

IMO, The writing to flash chips is a moot point. Creating a flash writer capable of holding the 720/mp flash boards seems pretty easy by comparison. It's just a matter of the right lines being connected up, then blast in the new image.

I run low on raM more often than I like with 32mb, especially when I have multiple browsers open and am downloading moderate size files. I dislike having to juggle the memory allocation because of the "file cannot be written to cache" error. I'm confident 128mb would reduce the problem and 256mb would eliminate it. I wouldn't be using the ram for program installation. I still believe programs belong on persistent storage and ram is workspace.

I think increasing ram would be the easiest task of all. Find pin and power compatible chips with more capacity than the originals and a few minutes with a SMT soldering station and the job is done. There's no need to develop a new ramboard that I can see. Perhaps the new larger capacity ram would be even more power frugal than the originals.

support for ram icrease: http://www.pocketpctechs.com/main~unit~HP_Compaq_Jornada_720-300~area~upgrades-ram~action~updateunit.htm

They appear to just dismount and remount higher capacity chips like I said.

it wasnt too hard to do to "rewrite" the image. just install PB 4.2 (platform builder) and now what drivers are needed etc etc even i could do it if i knew the right drivers (from PB again, of course, it is just a matter of selecting them)

problem: no available PB for HPC2000. there was PB 3.0 but even if you acquire it somehow, you'll lose pocket excel/access, and the standard outlook PIM suite. just like cenet. (if you don't mind it then great - i don't )

without this, i don't see how you could change the image. anyone?



you'll see 64 mb is enough reallocation? running out of ram? never again!

I agree increasing RAM is not that difficult. However I also said that it was not worth the cost.

Just look at that 64MB upgrade... boy oh boy, I don't want to predict the price point of the 128MB version. I don't even want to think about 256MB.

BTW, 720 owners who missed out on the Rocketman deal will prolly regret it

On the Rocketman deal, we both agree. I have to find out what's up with my forthcoming 720.

It's interesting to note that the price difference in the chips to go to 128mb is pretty minimal for a vendor, but I guess they have done their r&d for 64mb and they're gonna stick to it. I can't find it now but another vendor offers 128mb for a 720 for $200 IIRC. Still pretty steep. But it can be done if we have the money or dedication.

I wish I had access to SMT equipment. I'd be experimenting with ram like a madman.

edit. Monica got me thinking. If Simon can build a 4.2 image using PB why couldn't someone (ex hp guru) with the knowledge build a 4.x image for the 720s? Or any other hpc for that matter?

I think we're just scratching the surface of what can be done with our older hpcs.

Edited by wallythacker 2006-01-22 11:17 PM