One of the major flaws with the LibreTag system in its current state is that if the power is turned off mid game all the variables containing the players health and remaining ammo are lost. This can either be an annoyance or a way to cheat, either way it needed to be fixed.
I toyed with several ideas like saving all variables to the built in EEPROM every 10 seconds or so, but I was worried that after a while I may wear the EEPROM out. There is also the disadvantage that since the save is happening every 10 seconds people could still cheat in that time window. I finally came across the idea of getting the Picaxe chip to detect when the power is turned off and then save all the variables to EEPROM. By using a large capacitor (2000uf) on the 5v supply line, the Picaxe gets just under a second of runtime before the voltage drops too low.
I initially tried lower capacity capacitors but it was only at about 1500uf that the picaxe managed to stay on long enough to detect the power loss. I created a simple IF statement, once the input voltage dropped below about 5.5v (I’m using a 5v LDO regulator) it went into a loop sending a test character out the serial connection to the pc till the power completely dies. During testing what I saw was really strange since at 1000uf nothing was received, but when using a 1500uf capacitor, about 20 characters were received. I would have expected that when using a 1000uf to at least get a couple of characters received, I am guessing this is to do with the logarithmic nature of a capacitor.
Anyway like I said with a 2000uf capacitor the Picaxe stays on for long enough to save the health, ammo and number of remaining clips to EEPROM. With a few more modifications to the code, the gun now reads these values back out at power on and hence the game state is preserved. The only thing now is that I have to re-flash the gun once the health or ammo is depleted, which means I better get working on the Admin box to remotely configure the guns…..
Since I first built a working circuit for the LibreTag project, I always found that testing indoors proved to be without any need of aiming. My walls are white and so much IR is generated by the IR LED,that you can almost point the gun in any direction and score a hit.
I have been thinking of ways to be able to scale the power output of the guns down depending on where you intend to use them. My best idea so far was to use another FET to switch the IR LED either through a ~100 ohm resistor or straight to ground, to give a x1 or x1/10th power selection. However the problem with this is that it would never give you a chance to say scale from a large room to a small room.
Tonight I had a brain wave. After glancing over the data sheet for the 38khz IR receivers I noticed that they have a fairly high drop off as the frequency diverges from 38khz. I wondered if I could use this to my advantage by changing the PWM frequency of the transmitting IR LED.
I am happy to report it works wonders. I still need to implement a menu option to make the “strength” selectable, but with the freqency set to 50khz the sensitivity of the receiver, as per the graph, drops down to about 20%. It makes the gun usable indoors, with no sign of any reflections. It is maybe a little too under powered at 50khz, but this can be tweaked.
Work has been progressing rather slow recently. I’ve been very busy and just haven’t had the time to do some serious work on the guns. Coupled with the shorter days and awful weather we have been having in the UK the last few weeks, means I havn’t had a chance to go outside to do wood work / painting.
However I have more or less finished one gun now, with another 2 not far behind.
As you can see I’ve painted them with a large red side, this is so they comply to the VCRA (Violent Crime Reduction Act). It sad that we have to do such things, but better safe than sorry. I have also been experimenting with different types of sights and have attached a prototype one to the gun. It seems fairly accurate, I was easily able to aim perfectly from over 100ft. I will be testing some other types of sights in the coming weeks and hope to post a article comparing them.
The last two pictures are of the sensor. It is mounted in a Tic Tac box, which seems to be a pretty good fit. There will eventually be 3 sensors per gun, but I need some more people to eat Tic Tacs first.
Code wise, not much has changed. I have tidied up some of the code and made a few corrections to the parity encoding/decoding which has reduced the number of corrupt virtual bullets. It seems good enough now, with only one being received every 30-60 shots.
After what seems like too many hours, I have finally assembled the first Libre tag gun.
The exposed circuit at the top of the gun is a sensor, I need to put a RJ45 socket on the gun so that the sensor can be attached properly, but that is a minor task. The only other task is to mount a IR LED and lens into the lens assembly and mount it on the gun.
