For a decade and more I used four Sanyo N110 NiCd cells to power the receiver and servos. The energy used for an F5B flight I measured at 25mAhr for a two servo wing (Surprise 11) and 40 mAhr for a four servo wing (Avionik B06). So the 110 mAhr was plenty. I tried equivalent weight (30gms) NiMH with bad results. The voltage fell to dangerous levels at high (2.5A) currents, especially in cold conditions.
New NiCds became difficult to get, then in addition I had space problems in the B06. So I decided to jump to LiPo cells with a regulator. 2S LiPo packs intended for indoor flight can supply 5A with ease, and weigh 20gms. The only problem is that they need a regulator to reduce the voltage from 8.4 V to something acceptable to the radio. Switching regulators are on offer to do this, but I feared them as in the Avionik they will be hard up against the receiver and interference would be a problem. Linear regulators are available and used by some, eg Joe Mouris (Flight power) and David Hines (Thunder Tiger). But they are big and heavy and I could not see how to fit them into the Avionik. So I decided to build my own.
Commercial linear regulators are big because the designers build them for continuous operation. For F5B this is not necessary. The currents can indeed be 2.5A or more, but the total energy throughput is only max 50mAhrs in 15 minutes, an average of 200mA, generating very little heat. So a small PC board is all that is required. The one I used is 38mm x 19mm as shown below.
The board is double sided, but the other side is left blank to help with heat dissipation. The position of the tracks is determined by placing the regulator on the board and drawing round it with a scriber. The tracks are cut out using a grinding tool, I use a small tool sold to draw images in crystal glasses (from wife), but I think a Proxxon or Dremel would be fine. The next picture shows the components in place.
The regulator is the LM 1084 see http://www.national.com/ds/LM/LM1084.pdf R1 is 120ohm and R2 is 390ohm. This gives about 5.3V which seems fine. At 2.5A this delivers a full 5V at the receiver connection. There is space alongside the 120ohm on the circuit board to add another resistor in parallel which will increase the voltage, see the formula on the datasheet above. Below is a regulator wired up, in fact the fourth one just after testing and before kapton taping. Note that in this one there is a 2.2k resistor in parallel with the 120 Ohm one as the initial voltage was only 5.2V (resistor tolerance) and the servos on my Avionik B08 seem a bit slow, so this resistor raised the voltage to 5.4V. I am not a fan of high radio voltage for fear of brush wear in the servos.
Note that the battery is removeable for charging. In an earlier implementation I followed my NiCd practice of installing the battery permanently and having a separate charge wire, see below:
This worked fine until the model was left switched on for several days. This happened in the past with NiCd but the batteries recovered fine. Not so the LiPo. After several tries I got them to take charge again and get up to the proper voltage, but on testing at 2.5A the output voltage of one cell was way down. This required a solder job to replace. The plug and socket on the removeable battery don't seem to drop much voltage, they are used in the manufacturer's standard radio installation after all. Here they are upstream of the regulator so any voltage drop is not seen by the receiver. I just have to guard against the Mike Seale experience in Prague I think, of the battery disconnecting after a vigourous launch.
George. I'd forgotten that. It was actually Neuhardenberg (1998) and installing the flight battery must have pushed the receiver away from the receiver battery so it was only held loosely in place. The launch then disconnected it. Luckily the motor cut and the model landed reasonably softly. No damage but a zero for that round and then I had o fly more carefully for the rest of the competition. I now either hard-wire the battery to the switch or tape the connector so it can't come apart. Mike