I have a laptop grabbing gps coordinates and reporting them over 3g to a server, and getting a url to an image back, and displaying the image on an external outdoor lcdscreen.
Having a complete laptop isn’t optimal for a lot of reason, so im looking for new hardware and since im a .net developer gadgeteer seemed awesome, but i have some questions ofc.
Is it reasonable to do this sort of thing with gadgeteer or should i abandon my plans and just buy a bunch of laptops?
Getting a screen connected to it seems a bit tricky. atm it’s very low resolution and connected using DVI, but that will change, so i’m not sure about what resolution i will end up with and how to connect it. But is this doable in general for higher resolutions? Im assuming we will end up with some hardware between the gadgeteer and the screen, but if i have any say about it. What’s the recomended way of connecting it/communicating to it?
If this is doable, any recomendations on hardware is apreciated.
I’d steer clear of the Cerberus family as a mainboard if you’re looking at doing things graphics oriented on screens. Just stay with Cobra II, Spider, or Hydra as they have more memory to do what you need.
The tft-display interface of the hydra, is that some sort of standard for communicating with tft-screens or is that only used for communicating with gadgeteer modules?
@ Flossy - I’ve done quite a bit of work with this on a project of my own, although admittedly am having trouble with my GPRS module due to signal problems here at the house. I have gotten it working with WiFi though, so implementation shouldn’t be too different.
How big of a screen are you targeting and is this overall system going to be battery powered?
I have everything here to mock up the system for you without you having to buy everything if it’s something you’re interested in. Feel free to ping me at steve@ devonboard.com to discuss details and we’ll post progress back here (if you’d like).
The sockets are proprietary, but the electrical signals are standard. If you use breakout boards then you can connect any ttl tft lcd, as long as it’s interface is 3.3V and the resolution isn’t higher that 800x600, although you might be able to do higher resolution with hydra, g120 and spider2.
Right now the screen is a ledscreen at about 30 inches with a resolution of 144x48 i think, so for now the handling the resolution isnt a problem.
The reason for the low resolution though is that its for outdoor usage and the brightness is insane on it so it works perfectly even in bright sunlight. However, the resolution is a bit low. And snooping around a bit suggests that there is a lot going on in that market (here for ex http://www.solarview.biz/index.php?option=com_content&view=article&id=18&Itemid=10). So i absolutely dont want to end up using hardware that doesn’t scale up to the higher resolutions.
The problem now is that the total cost is $320:ish and i can get a laptop with built in 3G for less than 300. and it isnt really a problem if i need 1, but in the long run ill need a few hundred of them if everything goes as planned, so by then the price is a factor.
Also ease of installation, maintenance etc is important by then and a laptop offers none of that as i have noticed in my earlier tests (freaking screensaver and power saving kicking in on a linuxclient, windows update started on a windows client etc, i have had all types of errors known to man so far =).
Still, im a bit unsure of what to do =/… Can you get someone to manufacture a board with specific features/specs? im assuming 1 will be expensive as hell but it could be worth it for many. I absolutely want it programmable so i can put some intelligence on the client aswell even though the server does most of the work now. And if so, how does one go about to do such a thing?
you can always get a custom board made with specific components on it. How much it would cost depends on what you really need, and what you can use to start (and what skills you have available to you). GHI do custom design work in their consulting service.
Gadgeteer is intended to be a rapid prototype technology, so once you prove you can make it work with what items you can get/expect for the long term, then you can build custom boards based on exactly what you want.
So we’ve given you some thoughts here. How about we actually try to “assemble” something that would work? Here’s my first attempt and thought process behind decisions.
Cobra II (ECO, or WiFi, you decide). This gives you a G120 based device, the current “cream of the crop” module. It also means that you can use the Cobra II as the basis for designing your own device, it has freely available design files. It is a Premium device so you get GHI’s additional SDK features, which includes USB Host.
3g. This is going to be tricky in my view. I don’t know how you’ll figure that bit out. Others with comments??
GPS. Any serial device will work suitably, and features like update rate and reception indoors are more likely to be deciding factors that you’ll want to control.
Screen. If you’re sure you’re going to a TFT screen, then go for the TE35 as your proof of concept. This will allow you to have a decent resolution and show things working - although it’s not going to be as good as a screen designed for outdoor use. Then, once you decide on an actual screen, you can figure out how to “adapt” it into a set of Gadgeteer sockets, like the TE35 does (check it’s schematic for details)
I suspect that the device with the display is a seperate entity from the GPS/GSM device, right?
For the GPS/GSM/Camera device you can maybe look at a Cerberus with a serial camera. I don’t know how big the images are that comes from the camera though, so ram might be a problem, but if you can work around that then it is a lot cheaper. You can either store the image locally in some flash storage, then send it via GPRS, of you can send it via GPRS as it comes in from the camera.
For the display device I would also recommend the G120/Cobra2. Or…
Or (GASP!) maybe a raspberrypi? Especially if you want to connect a 30" display.
[quote]i would also add long time availability as requirement, what if one of the few hundreds is broken?
the pic you get back has the same resolution as your screen or do you need a kind of translation between both resolutions?[/quote]
Absolutely, also some atleat half-easy way of switching the hardware if it isn’t working properly, but that feels like the easy bits compared to the rest. I’ve built the server aswell so it will deliver images suitable for the screen even if we end up using different resolution screens, so no processing is needed for that other than perhaps decoding the imagedata. and possibly some transition between them, but thats easy enough as long as we aren’t very high resolution.
You sure have =), my main questions was if it is this doable with gadgeteer, and it seems so. Also if it works as planned, is there some way to “productify” it, and that also seems solvable.
I have been looking at raspberry pi aswell, but got the recomendation not to use it. But now i have to check out windows embedded aswell (that i didnt even know about until it was mentioned by andre.m) before i order anything.
you have suggested to use the GHI GPS module, and then later on you also suggested to use hydra as the mainboard,
but on the GPS module page, it specifically said that it won’t work with hydra. Am i missing something?
If I recall correctly then the GPS will not get a GPS lock while connected to the Hydra. It seems to be a problem with the Hydra contaminating the power supply with noise, or the GPS being over sensitive to that noise.