The perennial "how to charge a sealed lead acid battery realllly cheaply and in a vaguely OK manner" question. More of what I discussed/asked a month or so ago. If I meander on about this it may prove useful to others and may prompt some useful input. I'm not going to spec this in fine detail as the general requirement is clear enough. Short description: Aim is to float a 12v battery at 13.7 volt nominal at about zero cost. Rate of charge when not at float voltage is not too too critical and can be adjusted by simple things like eg series resistor and (gasp) limiting heat sinking on any regulator used to force it into limiting / shutdown above certain power levels. Also scale answers down by 2/3 for an 8V battery. Input voltage is a vaguely constantish voltage which may be set within reason with a say +/- 0.5 volt margin. battery has capacity in the 3Ah - 7 AH range. C/3 max charge rate is tolerable but as low as /10 - C/20 for 3/7 AH battery probably acceptable. Typical bottom end Asian SLA's spec sheets allow up to C/3 charge. An 'obvious' solution is simply a 13V7 voltage regulator with suitably low reverse leakage (doesn't drain battery), stable-ish output voltage and temperature characteristics that vaguely match what the battery wants. ST make such - their PB137, 13V7 regulator. (Interesting name - Pb = lead, 137 -> 13V7), now you'll remember it :-) ). 2 V ish headroom needed for a 16V+ supply. Googling doesn't suggest wide availability. Findchips (20 distributors) lists only Mouser at $0.50/1000. I've little doubt that others do similar at a similar price (as opposed to the largish range of 'proper" SLA charger chips at a too dear $USN. The LM340/LM7805 or LM317 with scaling resistors that one would otherwise be tempted to use is cheaper BUT have nasty problems with good accuracy without adjustment . EG the LM317 adjustment current is 50 uA typical and 100 uA max so even with a stiffish reference divider the variation in set point voltage at 12v can vary by about half a volt depending on initial adjustment current. This is far too much too ignore for SLA float purposes. Also, back feed down reference chain from battery is far too high too be tolerable so a series output diode is needed, detracting from regulation accuracy. PB137 is rated at 10 uA draw AFAIR. An LM358, precisionish cheapish reference, a high side PFET (or pnp) and a small bipolar driver, add R's to taste, would do better and probably cost about the same. I can feel a processor based solution coming on - might as well keep the processor busy. Vref will be calibrated during setup of the main system and can then be accurate enoughish for this purpose. Processor is AVR ATmega but all the above is processor independent (and almost unrelated). Processor solution has advantage of allowing boost charge to 14V4 ish and then drop back to float. Thoughts? Russell McMahon -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist