Hello,
i'm training on arduino and i would like to develop an automatic stopwatch with sensor for small races (30-100m). So i have my ultrasonic sensor which works fine.
The first sensor is not a problem because it is on the same place as the arduino.
Now i belive that if i just use a 100m cable (cat5 for instance?), it will not works due to the length...
What is the common solution to this problem?
I am sorry i am neither good in english neither in basic electronic...
Thanks for help.
Long wires pose three types of problems:
A CAT5 cable has 4 twisted pairs. The twisting is because it minimizes interference, especially with a balaced connection. Its hard to tell beforehand if you run into any problems but i imagine it is worth trying. If it doesn't work you have the option og using an arduino at each end and have them communicate
For 100m of CAT5, you may only need to "condition" the signal appropriately at the Arduino end. There are many types of ultrasonic sensors out there with different types of outputs (pwm, analog, digital, transistor).
What sensor are you using and what is the I/O voltage level of your Arduino board?
Done something similar, but used a laser diode and a LDR to make a photo-interrupter.
Seem to recall a ztx 510 and bc108c transistors after the LDR for gain and drive for a relay.
Could get a dual op-amp package, configure cct for amp and comparator, instead of transistors and relays(was a long time ago!).
Could cheat a little and have the arduino make the starting Bang and begin the count there.
This would negate the need for a starter gate and the long wire run, just stand at the finish line with arduino, big speaker and your finishing gate.
Edit:
Actually, leave the arduino at the finishing gate with a long wire to the start with just a simple monetary push to make button for the "GO".
Again only a single gate solution.
Then you could get fancy and use a mic to analogue input, set a high threshold to detect the starter pistol and that start the count, again single gate system for end.
Hi man
My suggestion: Don't use any sensor directly conected to long cables. You need a second equipent to capture and transmit the distant sensor data, let say, another arduino unit.
After that you need to transmit it and for that theres a lot of ways:
1 - RS485 that can reach 1000m
2 - Wireless radio like zigbee
3 - I2C bus with a P82B96 IC
Regards
Since most ultrasonic sensors have low frequency output (this model is 40Hz) the attenuation of the signal through 100m CAT5 would be negligable.
So, essentially, the proper way to do this is to use an RS-422 transceiver chip on each end, one pair of the CAT5 being used for this balanced signal and another pair to provide power to the remote sensor. If you are running the system from 9 or 12V, then you should send that voltage down the cable and use a separate regulator at the sensor end to derive your 5V. And in fact, you can parallel the other two pairs in the cable with your power lines to reduce resistance so that you have three wires for the positive supply and three wires for ground.
All that said, the sensor itself may well have sufficient drive capability (it is principally the capacitance of the cable that is the concern) to drive the line directly and the only thing needed on the Arduino end would be diodes to clamp voltage spikes outside of its 5V supply rails.
What you do not need, is a second Arduino as of course you would still have to figure out the communication between the two.
What is the common solution to this problem?
Go wireless.
Or use RS485 buffer on each end to send/receive.
or use a "current loop" in the cat5 cable but it will require a smal current generator with transistor and zener diode + supply voltage like 12 or more volt (depend of the cable lenght/resistance) I've already done such kind of current source diagram for a friend who wants to use a switch over a long cable lenght.
going wireless will add a lag in the measured time.
but if the question is more "how to send a power supply over this lenght" the answer is higher voltage. this is why the power over lan is a 48v system.
you increase the voltage and use "point of load" power supply for the connected equipement.
increasing the voltage allows less current over the wires and decrease the copper impedance influence.
same as the electric power network in the world (high voltage with big lenght and local low voltage distribution with transformer)
Hi, the RS485 idea is the best and most reliable system.
I have been involved in a timing system that was 3km long with timing points at 500m intervals, the system used rs-485 between posts and performed as expected, the long runs being compensated by the RS-485 setup.
Tom..... 
complex solution using MAX485 line drivers compared to a simple current loop to give a start signal! no?.
except if he want to transmit asynchronous informations between two boards
A pair of cheap RF modules would probably be cheaper than the cat5 of that length. Heck you could probably even afford another UNO to drive it and still be quids in.
and add lag to measure the time. except if the transmission (read encryption + decoding) delay is exactly the same each time.
Genesis92:
and add lag to measure the time. except if the transmission (read encryption + decoding) delay is exactly the same each time.
You're probably only talking micro seconds.
more a milliseconds I think and not constant so you can't make an offset.
even working at 115.2Kbds (and such radio modules are rare) with only one 8 bits word that give 70uS
only to "talk" to the module to send data. except if you have an already configured one with "transmit enable" logic line but you have the coding/decoding time in the emitter/receiver modules and it takes some delay to achieve this function.
On 100m distance 1mS even 100uS can make the choice between the winner and the looser.
Genesis92:
more a milliseconds I think and not constant so you can't make an offset.
On 100m distance 1mS even 100uS can make the choice between the winner and the looser.
Perhaps we should arrange a race of RF vs Cat5. By the time you allow for the capacitance of 100m of twisted pair, I think it should be a pretty even match.
this is why I was talking about current source and it's not the same application price.
he can use standard cheap cable not CAT5
I doubt very much that the OP is expecting to express times with any more than 1/100th of a second accuracy. Any latency associated with RF will be imperceptible.
100uS can make the choice between the winner and the looser.
Not if they're both being timed with the SAME few microsecond latency.
yes this is what I've said ecept if they have a constant transmit time and using an offset time but i'm pretty sure the codind/decoding latency will not be constant.
i'm working in wireless (224MHz/433MHz/bluetooth/wifi/RFID) laser barcode scanners since 15 years, it's my regular job 
so I give my opinion
Bluetooth and wifi 
No wonder you're talking about latency
Straight RF, different story