On Tuesday 14 September 2004 13:12, Rajesh Rajani wrote:
Don't be so hard on them :-). The issue is direction. There are a lot of good ideas out there, but they lack direction. It comes upon the more experienced people like you to guide and encourage people with ideas.
I used to guide college projects, but found that too much spoon feeding was involved, as very few had an aptitude for engineering and problem solving in particular. It resulted in my doing the thinking and problem solving (otherwise the students would have not been able to complete the project). Also increasing bussiness pressure made me stop.
The onus is also on the new people to not take this mailing list for granted and expect easy ready-to-eat answers all the time. Google is your friend.
If people are interested we could flag off an embedded project on either ARM which is the most popular and easily available processor or Power PC which has great performance, or ye ol X86 (Via C2/3) which has a little of everything + familarity. There are others available too like 683xxxx, Mips, Sh, roll-ur-own Sparc on xilinx / altera. This involves a lot of HARD WORK which our grrreat engineers find wholly unpalatable. Why to rack ones brains when one can cut and past from EFY /Elektor sundry 25 yr old gizmos like wireless network (1 infrared reciever, 1 ir led, 1 555). If u thought it was 802.11 perish the thought, most havent heard of such complex decimals "in the syllabus". And microprocessors=8085 menta kit (it should be labelled dementia kit).
I am not in this space so I may be off here, but I think that the issue here will be availability of the components, multi layered boards, etc.
Procuring samples is fairly straight forward. Multi layer boards are the main cost.
Designing custom hardware and getting a board built will be very time consuming
3 to 4 weeks. 4layers.
unless you are working in some company or know people providing such service. It may be a better idea to try and acquire some SBC and try to port Linux on it.
download. connect serial cable. run flash script. boring and uneducative to say the least. In fact most sbc come with a cd and linux preinstalled.
Then getting development tools like Emulators/BDM/JTAG debuggers is going to be another issue unless you already have something at hand.
GNU toolkits have all the jtag tools. gdb works on em.
The point being, that doing embedded software on custom hardware involves a lot of supporting material and this is the main deterrent to its growth as a hobby.
It is a pity that even with tools like GNU/Linux and mediums like the internet and email our level of education and awareness has become so pathetic. Jeez here i go again like an old foggy.
I am personally interested in robotics. This would involve mechanics, electronics, embeded hardware and software and tonnes of fun. It would also provide self education to those involved / seeking.
I am interested in it too. Unfortunately, the mechanical and the electronic design is way beyond my skills. Software is something that I can do though. I have been thinking of implemented something like http://protowrxs.com/nolen/robotics/LTBot/default.asp or http://www.evolution.com/er1/what_overview.masn
http://oap.sourceforge.net/ http://www.robo-etf.org/ http://www.orocos.org/
using standard hardware rather than creating a custom board for it. Should be simpler. Controller cards are available or can be designed for stepper motor control, etc. Cost of the project components is an issue.
Off the shelf components are atleast 30 times more expensive. Except for the system controller which can be an EPIAM2 mobo available locally, all circuits are quite simple. However The EPIAM2 is a power hog compared to ARM and would require a hefty power pack. or be powered from mains.
The cost of time is an investment in knowledge so it can be discounted at the moment.
Tell me if you want to do something with this.
We will require a few more guys.