The ODOTS test bed - an example of approximate perfection

HAHA - something that works! (maybe)

So - an exciting day, all individual components have been tested individually. The RFID unit has successfully read and written from both MIFARE 1k and 4k cards, the clock has kept time and had its time set, the buzzer has buzzed and the Microcontroller has sat though it all tolerating my antics. Special mention to the 200mm Telescope which donated one of its off-cuts to the testbed as a back board.

The test bed is really about getting everything together in one place so that it can be worked on as a piece. As mentioned earlier all individual components have been tested to the project from here on out is really about putting it all together, using the clock to generate data to be stored on the card, using the buzzer (and soon to come LED) to indicate to the user operational behaviour and finally building in the magic serial to computer interface that will allow each ODOTS checkpoint device to have its time re-calibrated, its function defined and its ID set.

At this point in the project I have a month and a half (if you squint) before my self imposed deadline which is good as if I can get major functionality done this week I will have a month of poking around, testing (and breaking) before a test-with-public at the end of September and a write up with tutorials by the end of the year. This probably means that it is a good time to look at how the project is tracking against its original objectives. For reference the requirements stated were:

1. Have 100 competitors on a 20 checkpoint course for under £500.
2. Ensure top level designs and software are straightforward and accessible.
3. From a performance standpoint checkpoint registrations should ideally take <20ms at a range of under 0.5m, while competitors should have the ability to record as many checkpoints as possible. Orienteers, being used to current timing systems, will also require the checkpoint registration times to also be recorded to around centi-second accuracy over a course that could take 24 hours!
4. From a safety standpoint it would also be helpful if checkpoints remember who has visited them as this can help with search and rescue (if needed).
5. From an operational standpoint the entire system should still work if it has been submerged for a while in a nice wet bog.

The engineer requires that I address these in the simplest way possible, point by point ("Yes" may be a bit too brief).

1. Cost analysis (a very exciting spread sheet) currently predicts ~£18 per Checkpoint unit (for the proof of concept) and <£1 per competitor recording card. 20x£18 + 100x£1 ≈ £500, so this is probably an achievable goal so far.
2. Arduino libraries are brilliant at keeping you from getting too bogged down in detail, which is generally when the unreadable spaghetti appears. So far the code has been written with the utmost effort to keep relevant variables and constants centralised and accessible. The top level Arduino sketch is currently looking like psuedo-code (because it almost is) but there should be no reason for this to break down. Source code has been split down into device specific headers, or general ODOTS functionality. The proof of concept will be built with through-hole, and widely available, components making it easily replicatable by someone with little specialised tooling or expertise.
3. Very straightforward functional requirements, these will all be met. Choosing the MIFARE standard gives access to 1-4kbytes of storage (100s of controls) with an RF interface that can generally conclude an exchange within 10-20ms at close range. Engimancy with the clock will allow millisecond precision of timing for the same unit and close to that between units, laziness more than anything else limits the available time frame, the clock has records of the current century if needed!
4. This is the only objective that may be bent, memory is at a premium in the AtMega chips, unfortunately this feature may have to wait until the system is made slightly more professional.
5. Lunchboxes and a contactless interface are acting fairly well to keep the entire design pretty easy to keep waterproof. 

Other than final integration big tasks looming on the horizon are dealing with the power situation (especially for the RTC, I am thinking a coin cell hiding behind a diode...) and actually developing my ideas about the enclosure. Also trying to make sure that any user is able to handle the built product intuitively will be good to do. Making an interface for setting Card and Checkpoint IDs as well as changing Checkpoint unit function which is simple as possible with an eye for starting to get it to run on minimalist extra hardware will be a must at some point.

An updated logo, I realised that the antenna is too good
a shape to not include!