Software

[Da_Blitz]

Well as some of you may have noticed there hasent been much work as off late.

this is mainly due to the software issue, ie what pcb tools will i use. well i had a look at the "pcb" program on sf.net and it seems it now includes an autorouter, i dont know how good it is but ill give it a try. its still lacking features i "want" but i belive at thius point in time that i could do the project with the software in its current state

so basically it looks like gEDA and pcb forthe pcb and scehmatic software, this is probbelly a better thing than what i wanted to use as most peopel who are want to take a look wouldnt need to buy expensive software, but those people probelly wouldnt mind the pictures

i will give it a spin in the next couple of weeks, i have some free time coming up again so i think its time for the feature freze and some diagarms, at the least i will have the power supply up which should be "fault tollirent" unlike whatever they put in my Z , it should also be able to exceed 0.5A on the usb socket, not recomeneded of course but those who need a bit of surge power to jump start a usb drive will apreciate this.

i will try and design it for 0.7A @ 5v normally and 2A surge (basically using caps for the surge, if you dont need this feature then drop the caps but considering how they are less than a cent anyway i dont see the point)

if i get a chance i will knock up some of the pelimanary usb to iMX3 stuff and some of the support circitry (ie clocks and passives) as these always need to be double checked (you always forget somthing). this is somthing that a non technical person can do, just look at the page numbers i provide as refrence and double check the wiring looks the same in thier examples asit is in the pics i provide, and check the svalues are the same except where stated.

i will have to look at a couple of IO things as well, did anyone want a rs232 level serial port (12v to -12v) or ttl (5v to 0v), is an SPI header somthing that people want (ie add user sensors) and do you want i2c, if so a dedicated connection (ie no devices on the bus) or a shared one (cheaper for me, you only have to be careful that when you add a device you dont use an exsisting device)

some other features others may not have cconsidered, besides the gyroscope/accereometer would you like features like a compus? or teamprature?, i was thinking that a temp sensor on the cpu and an ambient one would be nice, easy to do with I2C. perhaps one near the battery and power supply would be advisable as well.

well thoughts and input people, this week is out but next week the holidays begin (much to my relief)

[speculatrix]

serial these days is less of an issue, as long as there's enough power to drive a usb to serial adaptor

an external I2C or SPI might be useful; tho' SDIO is also based on SPI so an SDIO card which does general purpose I/O would be an attractive proposition. What sort of interface is the car OBD-II?

a compass + accelerometer coupled with GPS allows dead reckoning when GPS coverage lost, so is quite an interesting combination.

[Da_Blitz]

i belive most cars use the CAN bus, this chip dosent have any hardware to talk to it however and its non trivial to implement

[Duncanla]

DB, how far along are you with the design of the electronics?

I'm not asking to rush you I was just wondering where I can help out. Have you worked out any schematics yet? I am willing to assist in this process by helping to create any symbols or footprints that you may need for gEDA. I have no experience creating these in gEDA but I am picking it up relatively quickly. If you need help with this, we will need a list of exactly what chips we will be using, etc.

Also, you said you have ideas for how the design of the case and layout of the electronics will go, it would be nice to get some quick sketches up so we have an idea of what you would like to work towards. I think it would also attract more people to help if they could see where this project is headed.

I would also be interested in working on qemu emulation for the hardware. I have lots of programming knowledge but zero knowledge about qemu, but I think it is something I could work out. This would mostly be beneficial to the software development for the PDA.

I was thinking of a way that qemu could be used to aid in hardware design and I think it would be interesting to modify qemu to work almost like a circuit analysis program. We could program the behavior of the signals of each chip in qemu and then import a netlist from gEDA to define the connections between the chips. This would be sort of a combined circuit analysis / hardware emulation program. It would, however, be tedious to program the behavior of each signal especially with the ability of the i.mx31 to change the functions of its pins but I think it could be done with some work and would end up being very beneficial to the development process and would greatly lower the costs of having to have multiple prototype PCBs made.

One problem to overcome would be to emulate the signals as the pass through different components between chips. However, one solution to this may be to use spice. We would then have a mix of qemu and spice in order to emulate this device.

I am quite interested in helping to get this project off the ground because I am planning on designing a portable media player that would be very similar to this device, except with some modifications (ata hard drive, no gsm, more sound/video output, etc.)

Let me know if I can help, because I know you are busy.


-Duncan

[Da_Blitz]

most of the schematic is in my head still, at the moment learning geda is my biggest priority ever scince i learnt that it does have autorouting

i am not 100% sure what you meant by the qemu, if you meant connecting the chips to a qemu processor through a bus or writing drivers for each chip to confirm that the drivers on the virtual cpu work

ethier way it looked like to much work for too little gain

feel free to contact me through alternative means if you want, also would you mind filling out teh details on your media player?

[Duncanla]

Sorry about the cryptic description of my ideas with qemu, I wrote it in a hurry.

I thought it would be nice if we could use qemu to emulate the PP. We could use qemu to emulate each chip but then import a netlist from gEDA to tell qemu how the chips are connected together. This would give us greater hardware emulation for the hardware design and would also allow to take advantage of the imx31's ability to hook everything up in several different ways. That way we could just hook the chips together in gEDA, export that to a netlist, and load it into qemu and have it emulate what we just drew up. Then as the hardware design changes, or people need to adapt the design to their own needs, they can simply change the schematic, get the netlist, and start emulating their new design.

Sorry, I'm in a hurry tonight too with school and all so I hope that makes sense. I mentioned spice before but I quickly realized that would be overboard and much harder to implement. If it still doesn't make sense, then it is probably too complicated and/or unnecessary to our needs as a hobby project. As I think about how all this would be accomplished I agree that the benefits would probably not be enough to warrant all of the work that it would take, however it may prove to be useful if enough of the people who design the PP actually use it.

I'm running out of time so I'll leave it at that. You guys sound like you have much more of this kind of design experience than I do so I'll leave it up to your wisdom.

Real quick, for the media player I'm thinking that I will use much of the hardware you guys have been talking about. Some of the features I'm thinking about:

ATA 1.8" hard drive
WLAN
External USB
Large touch screen [the larger one you guys were talking about (5" ?) or maybe slightly larger, I haven't looked into it much and don't know much about them]
Video Out
High quality sound (haven't looked into that much either)
Bluetooth (maybe, so I can carry around one of those folding keyboards if I need it)

Basically I still want it to be like a pda but much more media based. Really, it will have much of the hardware in common with the PP (power management, wlan, bluetooth, usb). Actually the new iPod touch is close to what I want except the hard drive is too small, and after having an iPod, I really don't want my media to be locked down like that. Not to mention with an open hardware and software design I could add any sort of functionality I decide to it. I do understand that this won't be as small as any commercial device but I don't expect to be able to reach those sizes on my budget. I'm more interested in functionality really.

[speculatrix]

check out the Archos series, they have some really nice hardware. However, whilst many are linux based, I think that they need to be 'cracked' open in s/w before you can add 3rd party apps.

[Duncanla]

Wow, the archos 605 seems to be almost just what I want. I wouldn't hesitate for a second if there was open software for it. I really want something that I can hack around with and it would also be nice to have an open OS running on it so that I could play any media types that I want.

[Da_Blitz]

it sounds like your best off contacting me directtly

the qemu stuff sounds alright but writing saftware to emulate the hardware would be more work than neessacry, i would rather spend more cash on hardware and push out several prototypes

as for your needs, it almost looks like a subset of what i had in mind  i had planned to attach a CF card off of the ATA interface but with the way that SD cards are going it might be even better for me to drop that and go SD for the root FS. ethier way it wouldnt be too hard to have both

just PM me and i will send you more details about how to contact me