Rover Wakey Design Notes P. McGill * The 5 V supply powers only the PC-104 stack at 6 W or 1.2 A. We'll make this a 3 A supply for future capacity, and use the design from the SES Wakey. * The 12 V supply powers only the two serial relay boards. Each board draws 35 mA in standby and 157 mA with all relays closed. * All other 5 V and 12 V loads are exposed to seawater and are supplied from additional isolated power supplies; Wakey doesn't need to provide this power. * We must galvanically isolate the Device Enable line from the acoustic modem from Rover Wakey's circuit, since the enable line goes outside the sphere into seawater. A 4N35 optocoupler can be used, but the modem side must be powered by a 3 V lithium battery since the Device Enable line is just an open-collector transistor with no power source. * If we want at least 4 mA to flow in the 4N35 LED (yielding a CTR of about 100), we need (3 V - 1.3 V - 0.3 V) / 4 mA = 350 ohm. The closest standard value is 348 ohm. With this value the board (hardware v1.3.0, S/N 1) will wake from sleep at 25 degC with a little as 1.1 V driving the wakeup circuit. * If we want 50 uA to flow in the 4N35 output transistor when it's on, we need a collector resistance of (5 V - 0.3 V) / 50 uA = 94 kOhm. 100 k is close enough. This is enough current to pull the 4N35 collector down to about 0.3 V. * Reverse polarity protection can be provided by a fuse on the input feeding a reverse-polarity diode. If the power is connected backwards, the diode will conduct and blow the fuse. The circuitry will still see -0.3 V (Schottky) or -0.6 V (silicon) for a brief time until the fuse blows. The LM2936 max ratings state up to -50 V is okay, but the TPS5430 and TPS54260 have a max rating of only -0.3 V, so a Schottky diode will have to be used. The reverse leakage current spec of the diode must be only a couple of microamps since this is a constant load on the input power. An STPS5H100 Schottky diode is rated for 100 V PIV, 5 A If, and 3.5 uA Ir at 25 degC. * The LM2936 regulator currently used has a quiescent current (Iq) of 15 uA. This could be replaced with a TPS7A16, which has an Iq of 5 uA. * Rover Wakey Specifications Input Power: 20 to 34 VDC Reverse polarity will blow fuse but save board Output Power: 5 VDC at 3 A, fused at 2 A 12 VDC at 2 A, two outputs each fused at 1 A Communications: RS232, 9600 bps, 8, N, 1 Modem Wakeup Battery 1.5 to 7 V Input draws 4 mA at 3 VDC Sleep Time 1 to 65535 sec (18.2 hrs) Sleep Current 30 uA * Improvements to v1.2.0 SES Wakey firmware needed for next release: 1) A set serial number of, say, 255 should return "not set". 2) Enhance command parser so parameter can be entered on same line as command. 3) Add logic to wake up when activity is seen on Modem Device Enable. Anomalies seen with hardware v1.3.0 * S/N 1 board U6 LM2936 regulator failed, outputting 8 V. This happened as ammeter was being put in power feed to monitor current, but no excessive voltages should have been applied to board. Replaced U6 and board seems to be fine, so PIC survived overvoltage. * S/N 1 board with v1.2.0 firmware drew about 4 mA in sleep mode after a power cycle, but could not duplicate failure on subsequent sleeps or power cycles. Now drawing normal sleep current of 30 uA.