Aveling wrote:
A dead battery (master switch left on for days) is a completely different matter because it will have little or no residual voltage to resist the inrush current when you connect jump leads from another battery that can source 1000’s of amps into a what is a dead short.
I’ve never seen a deep discharge lead acid battery act as a short! The behaviour I’ve seen from every deep discharged battery I’ve seen is that the charge current at the start is minuscule, then it builds up over the first five or ten minutes of charging. The battery may or may not hold a charge afterwards. I’ve certainly never seen one behave anything like a short circuit.
I just cannot see how one would be able to regulate the output of an alternator that has permanent magnets for the rotating field. You can only dump surplus voltage into heat sinks when only partial output is used. Those types of AC generators are widely found in motorcycles, push bike generators, mopeds etc. . There is some maximum voltage at high revs in this design and the rest is taken care of by the heat sink via the “regulator” .
Vic
BC permanent mag alternator
Regulators for the Rotax (which has no field coil either) seem to contain a bridge rectifier with two diodes and two thyristors, the output voltage determined by phase controlling the thyristors.
vic wrote:
I just cannot see how one would be able to regulate the output of an alternator that has permanent magnets for the rotating field. You can only dump surplus voltage into heat sinks when only partial output is used.
Regulation of a PM generator is typically done with pulse width modulation, and is commonplace.
A separate issue with PM generators for aircraft use is that if the PWM regulator fails, you can have a situation in which you’re pumping energy in an uncontrolled manner into a fire. It’s not fail safe.
My first exposure to PM generators was with those British bikes 
After a while I came across the Ducati regulators that are also used by Rotax, and somebody I knew explained PWM to me… I’m not an EE. Now I’m dealing with 25,000 times the power output
those Ducati regulators are very well known in the Rotax clubs, less so for reliability. A friend of mine had a Ducati 860 GT from new, still has, and the wet battery was cooked by the regulator with a real mess on the chrome of the silencers. He was really mad at that . . .
Vicvic wrote:
I just cannot see how one would be able to regulate the output of an alternator that has permanent magnets for the rotating field. You can only dump surplus voltage into heat sinks when only partial output is used
“Dumping surplus voltage into heatsinks” is only necessary for a linear regulator. Switch mode regulators are vastly more efficient and often have very low losses, for an example you are using right at this minute – your PC power supply (well, the detail will be different, but the same operating principles). For example, the 1kW supply on my PC will take an input voltage from 100 to 250 volts, and has relatively small heatsinks (there are some losses, the conversion is never 100% efficient, but the losses are nothing compared to the kinds of losses you would have with a linear regulator). Imagine the size of heatsinks needed for a 1kW linear regulator that could take an input voltage range of 100-250 volts and output 3.3, 5 and 12 volts! They’d be bigger than my entire PC. Yet the 1kW power supply in my PC has a tiny little heatsink.
Modern switch mode regulators are very good and come in all shapes and sizes on a convenient IC, generally the designer just needs to add the inductor and feedback. I built a switch mode regulator for a solar panel that would output 6 volts, off a panel whose output voltage could rise to 20 volts or so. I didn’t even use a specialist SMPS IC, it was based on the humble 555 timer IC and even that was almost 90% efficient (and required no heatsink). By comparison, a linear regulator wouldn’t even be 30% efficient and would require a substantial heatsink.
I just cannot see how one would be able to regulate the output of an alternator that has permanent magnets for the rotating field
The output voltage will be proportional to RPM, which is a bit useless. That is what the classic bike dynamos do.
The only “proper” way is to have a switch-mode regulator on the output.
The open circuit voltage of a bicycle dynamo is proportional to RPM, but with a load attached they act as reasonably good constant current devices – at least the older ones did. The laminations of the core are made deliberately thick so that eddy currents increase and as RPM increases the dynamo becomes less efficient. On one of the high-end models they achieve much the same by having a high inductance winding so there’s a large flattish spot on the power/rpm curve that covers all reasonable speeds without wasting power. You still did need two Zener diodes just to protect the bulb from really fast downhills but most of the regulation came from the construction of the dynamo itself.
A while back I tried to work on a switching regulator for a dynamo, but worked out that at normal speeds it was more efficient just to drive the bulb directly – even with Schottky diodes the losses in the rectifier at low or cruising speeds outweighed the gains. A switching regulator only made sense if you had a high-voltage dynamo where the voltage drop became negligible. The problem then became that you had to produce enough power at low speeds e.g. 3mph – but at 60mph the regulator had to put up with several hundred volts. At least one of the first dynamos to use a switching regulator for efficiency (i.e. so it could use an efficient core with thin laminations and avoid wasting energy) was a bit notorious for self-destructing at high speed – burning out the regulator, not just the bulb. People didn’t enjoy plunging downhill at 50mph in the pitch black.
Arguably bike dynamos are quite impressive devices – an aircraft alternator has to provide a fairly constant output from 500-3000 rpm (factor of 6) but a bike dynamo has to produce a reasonably constant output at 3-60mph (factor of 20).
