Sorry to reply so late, I have been sick and barely out of bed for two days
I am pretty sure that they are ALL shunt regulators.
I think you are probably right although all designs are probably not the same. I rechecked my measurements and bad/corroded connections was giving me false measurements. I'll clean and report back.
Whether thats good or bad is harder to say.
I agree that it seems an inefficient way to do it.
On the other hand, done right, it is simple and works acceptably well.
(the Honda regs for example)
The primary issue with efficiency is that dumping electrical power (into heat) that you create with the engine power can only be explained by saving production cost. A Field control configuration would make much more engineering sense but probably is a bit more expensive.
Well, yes and no ...
It absolutely would be more efficient, and better in the abstract technical sense.
But it would be much more complex mechanically. Mainly because you would now need brushes.
There are a few bikes that have field coil type alternators, but my understanding is that most use shunt regulators. Assuming most of the designers are more or less competent, that would seem to say that the advantages of shunt type regulators outweigh their disadvantages overall.
Given the uncontrolled generator input, an improved design would use a chopping series design that would effectively open and close the load from the source and regulate power that way. That might be an EMI nightmare.
There was someone on the GSR that built one (using a chopper regulator chip to drive it)
Said it worked fine.
Dunno why they are not used by OEMs. It may be EMI, maybe they have potential problems under some operating conditions, maybe because they cause stator voltages to be higher because its unloaded.
Or something else, may be just "thats the way its always been done and its worked fine till now" (Note that it DOES work fine on a lot of non GS bikes)
I would assume if there is no load on the stator legs that there is no power being generated and therefore no mechanical power lost.
The power lost is only the EXCESS power that you made and didn't actually need, the GSs don't make much excess (try to run some extra running lights and you'll find out how little)
Assume you are dumping 10 amps (thats probably too high, but we'll assume it anyway)
At 14 volts, that comes out to 140 watts. Thats somewhat less than a quarter of a horsepower.
The other more significant benefit of a modern chopping design is that it would have built in open circuit, short circuit load protection and over voltage/over current control. That would improve the susceptibility that the simple shunt regulator obviously doesn't have.
You're making WAY too many assumptions here.
Shunt regulators are inherently safe against open circuits and short circuit loads. The overvoltage is a matter of proper design (and the GS's is apparently not designed to cope with that)
But
EITHER can be done right or done wrong.
As an aside:
I have my suspicions that the problem with Suzuki's version is in the trigger part of the circuit; which in a way is not really part of the shunting circuit itself, but only drives that part of the circuit.
The "trigger" as you call it is the control so improvement based on current sense or separate voltage sense would improve the situation. Basically any error condition that can be sensed can be used to improve the control. The problem is that the shunt regulator can't really control the power generation it can only dump( shunt ) power to ground. If designed with proper power handling capability (and heat sinking) then it is probably pretty robust.
I didn't explain this part very clearly.
What I was trying to say is that the problem in the Suzuki regulator may be in one of the darlington transistors and not the SCRs (the SCRs are the part that would need the power handling capability and heat sinking)
If you look at the "darlington" type schematic, what happens as the voltage goes somewhat higher than the trigger voltage. The current into the base of TR1 starts to go up. This is multiplied by the gain of TR1 and TR2 with no current limiting anywhere. So for a very small excess voltage you get way to much current out of TR2. On a running bike this is not a problem because the voltage will never get that high ... as soon as the transistors start to conduct, the SCRs are fired and the voltage drops, thus protecting the transistors.
But jump from a running car and you let the smoke out of TR2.
However, I don't know if they simplified the schematic and just didn't show some extra resistors that would protect against this.
The generic figure from the Stator Papers does NOT correspond to the regulator internals from my OEM Suzuki manual.
I'm not sure which manuals you have but I have 4 different sources with essentially the same thing except the GS1100 uses a Darlington pair to get "crisper" on/off control than the zenor can provide and presumably better regulator/ripple control.
http://www.posplayr.100megsfree3.com/gs750/GS750_Charging_R_R_750.pdf
http://www.posplayr.100megsfree3.com/gs750/GS750_Charging_R_R_3.pdf
http://www.posplayr.100megsfree3.com/gs750/GS750_Charging_R_R_1100.pdf
I was wrong here. The schematics I have seen were all like the last one, and that was what I was refering to previously. I thought that all the GS's were like that one. I guess not ....
On the last one (the darlington schematic) your test would not work.
On the others it might or might not, depending on the resistance of the resistors and the leakage current of the diodes.
Given the detail of the schematics and the fact that they show the current flow paths and changes for the GS1100 I suspect that what they are showing is all that is in the regulator
The test you were suggesting is basically just doing half of the diode tests described in the stator papers ...
(And your indication for "it has a regulator" is their indication for "the diodes are shorted")
Assuming you don't trigger the SCRs, all you end up measuring is whatever the leakage of the trigger circuit is, which is fairly low (and unknown)(and don't forget that the bridge diodes have SOME reverse leakage too)
As you describe the test assumes that the leakage current of the rectifier diodes is much less than the control circuitry leakage current. When I first checked it it seems I had some really bad connection and so I will redo it and report back after I clean the terminals. This might not be a good test after all.
The wind is blowing (small craft advisory) and I gotta go now. \\
/
Posplayr
