Analysis of an AC voltage reglator/charging circuit

[canadaboy25]

Hi all,  I'm looking for some help analyzing this circuit since I'm having a bit of trouble wrapping my head around it.

The device is a voltage regulator for a snowmobile.  The engine has a generator winding underneath the flywheel which generates an AC voltage with frequency proportional to the speed at which the engine is turning.  I'm not sure how many magnetic poles the flywheel has, but the engine speed should be in the range of 1000 - 8000 rpm.

The regulator serves two purposes:
1 - Regulate the AC voltage to approximately 14VAC to prevent the voltage from destroying the incandescent light bulbs when the engine speed is increased.
2 - To rectify the AC voltage and charge a 12V lead acid battery.  The only purpose of the battery is to run the electric starter motor used to start the engine.  All light bulbs are powered directly from the AC voltage.

The regulator had stopped charging the battery and would intermittently stop regulating the AC voltage, causing all of the light bulb filaments to blow (very frustrating).  I disassembled the regulator (it was potted in a rock-hard potting compound which only softened up with a hot air station) and reverse engineered the schematic, which is attached to this post.  The diodes labeled "Power Diode" are large diodes with no markings that are heatsinked to the metal enclosure of the regulator.  All 3 PNP transistors are MPSA56.

The nodes labeled "R", "G", and "Y" are the three wire connections to the regulator (red, green, and yellow).  The red wire connects to the positive terminal of the 12V battery.  The ground reference of the schematic is connected to the negative terminal of the 12V battery.  The yellow and green wires are connected to the AC generator in the engine.

Clearly, when "Y" is positive and "G" is negative, current will flow through D9 and D6, charging the battery on that half cycle.  However, to charge the battery on the other half cycle, current would have to flow through D8 and SCR1.

I'm having trouble wrapping my head around the gate control circuits for the two triacs and would appreciate if someone could give me an explanation of how this circuit works.

Thanks!

[Kim Christensen]

Looks like D5, Q2, Q3, & SCR1 are for regulation of the battery voltage.
When the battery voltage is low, Q2 is off which means Q3 is on which fires SCR1.
When the battery voltage gets too high, D5 starts to conduct, which turns Q2 on, which turns Q3 off, and thus SCR1 doesn't fire as much each cycle. (D7 + D5 voltage drops, plus the divider formed by R1 & R7||R11 determine the regulation voltage)

Now, the AC limiting looks a bit more bizarre. But it looks like this is accomplished with D1, D3 , D4, Q1 and SCR2.
When the peak to peak AC gets too high, D1 conducts, turning Q1 on which triggers SCR2 which shorts out the AC winding (weird) and clamps the voltage for the rest of the 1/2 cycle. Either D3 or D4 is conducting depending on the polarity of the AC cycle at the time.
 

[canadaboy25]

Looks like D5, Q2, Q3, & SCR1 are for regulation of the battery voltage.
When the battery voltage is low, Q2 is off which means Q3 is on which fires SCR1.
When the battery voltage gets too high, D5 starts to conduct, which turns Q2 on, which turns Q3 off, and thus SCR1 doesn't fire as much each cycle. (D7 + D5 voltage drops, plus the divider formed by R1 & R7||R11 determine the regulation voltage)

Now, the AC limiting looks a bit more bizarre. But it looks like this is accomplished with D1, D3 , D4, Q1 and SCR2.
When the peak to peak AC gets too high, D1 conducts, turning Q1 on which triggers SCR2 which shorts out the AC winding (weird) and clamps the voltage for the rest of the 1/2 cycle. Either D3 or D4 is conducting depending on the polarity of the AC cycle at the time.

Thanks for the explanation!

So essentially, the charging of the battery can only ever be "half" turned off because the other half cycle will always charge the battery through D6 and D9.  Lead acid batteries are tolerant to a low continuous charge current even when full, so I suppose this makes sense.

I'm glad you find the AC regulation circuit strange, because that is the part I was most confused about.  It appeared to me that SCR2 would short the windings which I didn't think made much sense, and figured I was missing something.  If this is in fact how it operates, then when G is positive and Y is negative, SCR2 will fire sometime before the peak of the half cycle, assuming that the AC voltage is above the limit.  This seems fine (sort of), but how would the negative half of the cycle be regulated?  At the zero crossing point, SCR2 should unlatch and then it will be reverse biased when Y is positive and G is negative, and there will be no regulation?

Also, since it seems that the circuit relies on the 12VDC rail for a reference through D1, if there was no battery connected to limit the voltage, the AC regulation would not function as well?

[canadaboy25]

I decided to throw together a quick LTSpice simulation to see if it would show anything interesting.  With the AC voltage source set to 100V and a series resistance of 200 ohms, and a DC source of 15V with 10 ohms resistance, the results are attached.

At the start of the simulation, SCR2 never conducts.  However the AC voltage is regulated to ~20V amplitude.  I suppose they are relying on the battery to limit the AC voltage through the charging diodes.

You can see the current through SCR1 (light blue) every other half cycle as expected.  Once C1 charges up to a certain voltage (dark blue, scaled to be more visible), SCR2 fires and clamps the voltage.  This only happens on one half of the cycle as I suspected.


Things get more interesting when I set the DC voltage source to slowly ramp up, simulating the battery taking a charge. (2nd waveform)

At first, SCR1 is conducting, charging the battery on both halves of the cycle.  After the battery charges to a certain point, SCR1 no longer fires and the battery is only being charged on one half of the cycle.  During this time, the negative half of the AC wave is not regulated and reaches -90V.  This seems to quickly charge C1 and shortly after, causes SCR2 to start firing and clamping the unregulated negative half of the cycle.

So it would seem that they are relying on the battery to regulate the AC voltage while it is charging, and once it is full, they use it to regulate one half of the cycle, while the other half is shunted by SCR2.
Of course, my results may not be overly accurate because my values for the series resistances of the sources, and frequency of the AC source are just guesses.  I suppose the couple unregulated cycles that make it through are not enough to blow the filaments in the bulbs?  Seems like a very strange design.

[SeanB]

Yes it regulates the negative cycle, with the battery being used as a clamp for the positive cycle, and the SCR1 side dropping charge current for the battery when it reaches 14V4 to a lower value, and then SCR2 is there to clamp the negative side, shorting out the coil, as that will limit current due to the core saturating hard as the current rises, limiting the current flowing in it. Not the best, but has an advantage in being cheap and easy to implement, plus they do not need to have a charger in for the versions with pull start, simply have the lamps on all the time, and put in the AC regulator, which is basically SCR2 side wrapped with a bridge rectifier, to do the same. You can replace with a lot of universal regulator units as well, some are better, some are worse, but all operate on the same rough principles.

Most common faults are the diodes and SCR units cracking from the thermal cycling, and going intermittently open circuit, which as you found means high AC voltage on a half cycle, depending on the diode or SCR that breaks. Normally the generator is the same number of poles as the number of cylinders, times 2, simply because the generator uses the same magnets as the ignition circuit, so for a single cylinder 2 stroke it is 2 pole, for a 2 piston 4 pole, as they normally are wound on the same rotor as the ignition system.

[Terry Bites]

So if the engine stops you get plunged into darkness- scary.

[canadaboy25]

Yes it regulates the negative cycle, with the battery being used as a clamp for the positive cycle, and the SCR1 side dropping charge current for the battery when it reaches 14V4 to a lower value, and then SCR2 is there to clamp the negative side, shorting out the coil, as that will limit current due to the core saturating hard as the current rises, limiting the current flowing in it. Not the best, but has an advantage in being cheap and easy to implement, plus they do not need to have a charger in for the versions with pull start, simply have the lamps on all the time, and put in the AC regulator, which is basically SCR2 side wrapped with a bridge rectifier, to do the same. You can replace with a lot of universal regulator units as well, some are better, some are worse, but all operate on the same rough principles.

Thanks for verifying my results.  Interesting you mention the pull-start version.  I have an identical snowmobile, only difference being it does not have the electric start or 12V battery.  The lights are turned on and off with a switch just like the electric version.  The regulator on the pull-start version is just a single wire device which bolts to the frame of the machine.  I have one of these regulators that have failed as well.  I may tear it apart just to see how they have implemented that one.

I did replace the regulator on the electric start model with a new aftermarket one.  However, I found that the bulbs flickered really badly at idle when the regulator was connected.  When the regulator was disconnected at idle (the voltage was low enough not to blow the bulbs) the flickering stopped.  I thought the new regulator may be faulty, but after understanding how it works, it seems that the flickering is a natural result of the SCR firing and clamping the voltage.

Most common faults are the diodes and SCR units cracking from the thermal cycling, and going intermittently open circuit, which as you found means high AC voltage on a half cycle, depending on the diode or SCR that breaks. Normally the generator is the same number of poles as the number of cylinders, times 2, simply because the generator uses the same magnets as the ignition circuit, so for a single cylinder 2 stroke it is 2 pole, for a 2 piston 4 pole, as they normally are wound on the same rotor as the ignition system.

While disassembling the regulator, I found that the cathode of D8 was corroded and no longer soldered to the metal plate.  I could not find any other obvious faults inside it.  Perhaps one of the components is cracked like you mentioned, however, all test good when removed.  After replacing the regulator, the battery now charges, but after going for a ride, I found that my taillight bulb had blown.  At some point during the ride, the main fuse that connects the 12V battery to the system had vibrated out of its holder.  This obviously would have caused the AC voltage to go unregulated, and blow the bulb.  I will have to check and secure all of the wiring to the battery, I had no idea that it would be critical to the AC regulation side of the system.

The engine is a two cylinder, so it would likely have 4 poles on the flywheel.

So if the engine stops you get plunged into darkness- scary.

Yes, the only time the lights work is when the engine is running.  Very common on snowmobiles since most of them do not have electric start or a battery.  I imagine the ones with the electric start also operate off the AC voltage just to minimize the differences between the two systems.

[floobydust]

These PM generator regulators are antique 1970's designs that have not really been improved. Popular on small engines in mopeds, tractors, ultralights etc.
Many of them are shunt regulators - they short the stator out when voltage reaches the setpoint. Of course this is a waste of engine power. chinese GY6 150cc is a "modern" example.
Here are schematics of other ones that do not short out the stator. Kubota, Rotax, Ducati etc.  They are all using SCR's doing active rectification, gated-on when the voltage is below setpoint. An extra output for the CHARGE lamp.
A good starting point to understanding OP's oddball circuit.
more in this thread: https://www.eevblog.com/forum/projects/ac-generatoralternator-charger-rs1105-replacement/