PC104 on a mini race car

[%X]

I think you will find they already have that bit covered (check their list and diagram).

That they are going to try and seek someone from the Computer Building is probably the best they can hope for, so long as they ask for people with Embedded Systems Experience).

--
********************************************************************
Paul E. Bennett...............
Forth based HIDECS Consultancy
Mob: +44 (0)7811-639972
Tel: +44 (0)1235-510979
Going Forth Safely ..... EBA. www.electric-boat-association.org.uk..
********************************************************************

e t

It doesn't appear to include a motor controller. Rather the pieces needed to make one. I also think the Freescale device isn't complete enough for their needs.

Robert

[%X]

To those who have arrived late in the thread, I suggest you read back to the OP's original post. You are asking some of the same questions that have already been answered.

It is obvious the team have no real plan in place which is why they are struggling but random suggestions of software are not really going to help. They need on-the-spot assistance and hopefully they will find it in the Computer Engineering department if the college runs Embedded Systems Courses.

So far we do not fully appreciate their time-scale but it sounds like it is fairly short. The car has to be big enough for the PC104 system to be seen as a viable central controller with enough battery power to move it.

We are all going to struggle to help them at this remote distance so I hope they find the people they need closer to their location.

--
********************************************************************
Paul E. Bennett...............
Forth based HIDECS Consultancy
Mob: +44 (0)7811-639972
Tel: +44 (0)1235-510979
Going Forth Safely ..... EBA. www.electric-boat-association.org.uk..
********************************************************************

I also failed to see any sign of a real motor controller - just points where their "controller" will connect to their drivers and sensors, and a quaint belief that a PC/104 board is a motor controller without the slightest idea how to implement it.

The MC3PHAC is a cheap and simple device that is easy to use. I agree that it doesn't have all the features that such a car would need. But a controller that would do everything they want would cost more than they can afford, and be more complex to use. With the MC3PHAC, they have at least some chance of getting the car moving.

I've tried to give a little advice (I wouldn't say /random/ suggestions...). But I thoroughly agree with your points here - the best help has to come from people who can see what they've got and talk to the team rather than from halfway around the world.

I think you're working at way too low a level. Before you even start to look at bits and pieces or flowcharts, you have to have a proper *project* plan in place, if you don't you'll just spin your wheels and fail to deliver anything. Unfortunately, IME academic institutions are often woefully deficient in that department, so you may want to find a friend or relly that works managing technical projects, who can help you through the process. You need to define an *achievable* objective, which may well be a lower target than you originally envisaged if you've lost time and are struggling with understanding. Then you need to identify the resources you need, which will include gaining familiarity with the technology. Then you need to break the work down into clear stages, with outcomes that you can manage and measure, and assign the tasks.

Given what you've described so far, I'd expect that setting up the platform and maybe prototyping some basic interfaces such as the motor drives may be as far as you can get. That's cool if it's reality, be up front with your supervisor and have them involved in the process. If s/he is dissatisfied with that, then go higher.

If you really do need to accelerate the process, then the correct approach is to engage the appropriate knowledgeable resources and review the project plan. Flogging yourself or others won't work.

I'd like a dollar for every uni (= college) project that I've seen go this way. The cause is almost always incompetent and negligent leadership by the institution.

Darth,

This looks like a very cool project. It is great that you are helping out -- it appears that it is not your "baby," but nonetheless you are offering assistance in your field of expertise.

Your Universty computer department will probably point you in the correct direction for resources in the embedded area -- keep us posted on the progress. Henry Ford once said that he did not necesarily know about a field of expertise -- but he knew how to locate a person that had it in a few minutes ... No worries.

Stay cool.

Don

building

Cheers guys and thanks Don for your encouragement.

Some good news. I dropped by the lab where all the electrical parts are been worked on. The electrical engineer showed me some of his latest work. At the moment he is using xPC Target to run run the motors. A lot of the system has been designed in Matlab.

The stack of boards include a PC104, Power Supply board for PC104, 2 x I/O boards (I think for all signals) and 2 x PWM boards. I mentioned to him what people have posted here about one board not enough to run both motors and he said that's why there is 2 PWM boards (9ch each).

He said once he has finished all the modeling, he will put all matlab files into once master file and change the format of it (somehow) so it can be sent to the PC104. He was confident of this as he said he has sent small programs to the PC104 and they seemed fine.

They've also purchased a 1GB CF card that inserts to the side of the PC104 board. There is one problem with this. In order to make it bootable he made a floppy disk image onto the card and as a result the 1GB card now only reads 1.44MB. Using a laptop HDD doesn't seem like a good idea because of its moving parts and the vibrations from the car. Looking into a USB atm.

So far, positives looks to be no need for embedded work if transferring the Matlab file to PC104 works fine.

--------------------------------------- Posted through

I doubt that they'll find an undergraduate with the experience needed to jump right in.

To the OP, you're better looking for a postgrad doing robotics research or a lecturer (they're all profs over there aren't they?) who's running a robotics team.

They will have most of the background you need and should be able to advise on the best approach.

If this is a final year project then finishing it probably isn't critical for a good result (at least this was the case when I was studying) as long as it's written up properly. Make sure you've got your bits completed properly.

Nial

I have no experience with Matlab but I suspect there's a _big_ difference between running it on a lab PC and an embedded Celeron.

This guy sounds like he knows enough to be dangerous, I don't know what's involved but the CF should be able to be properly formatted.

Keep us informed of progress!

Nial.

That doesn't make sense. CF should behave exactly like HDD, but slower. I have many systems booting off 1G or 2G CF. Even if FDD image is needed. If could be a virtual drive image within the ext2(or

3) file system. Alternatively, a 1MB partition 1 and a 0.99G partition 2. Can you ask him to explain the need for the FDD format?

Yup. The person might have thought to make it like a bootable CD/DVD where the boot image loads drivers to access the rest. But that shouldn't be necessary here.

George

Just wondering how this project turned out.