Schematic of voltage regulator/rectifier?

Bike is a 1982 KZ750H2. Does anyone have a schematic of the circuit INSIDE the regulator/rectifier box? I could reverse engineer it if I could get past the epoxy sealer without damaging the circuit, but that's not possible. Or were these circuits a trade secret?

It's a pretty simple circuit, trade secret or not, but could be made many different ways.

It just an averaging circuit with a threshold for firing SCR's. I don't think they include a temperature compensation because most people don't ride in super cold temps.

Here's one regulator I had working, but never developed it because it would have cost more than Oregon unit and/or the ebay ones. It could probably be simplified further and would definitely need a lot more real-world testing.
home.comcast.net/~loudgpz/GPZweb/RegRec/GPZvRegMagnetField.html

The rectifier portion is standard. You can save power on the rectifier if you use germanium or selenium instead of silicon. Here's a simple silicon rectifier.
home.comcast.net/~loudgpz/GPZweb/RegRec/...3phaseRectifier.html

@ Loud...

Other then power dissipation (split current) across the diode banks, any other reason you jumper across to the other AC input for each FWB rectifier AC phase input?

OMR

polkat wrote:

Bike is a 1982 KZ750H2. Does anyone have a schematic of the circuit INSIDE the regulator/rectifier box? I could reverse engineer it if I could get past the epoxy sealer without damaging the circuit, but that's not possible. Or were these circuits a trade secret?

I wouldn't try it just because the rectifiers probably need to handle about 100V and 100 Amps. Those would be monsters to try to find. The SCR shunt probably has to handle maybe 50 Amps or so with at least 200V breakdown. It's not technically sophisticated but monster size devices.

Another problem with auto electrics is they are subject to high peak voltages that happen quickly but still degrade the semiconductors over time. That's the reason so many rectifier/regs turn up dead for no aparrent reason.

loudhvx wrote:

It just an averaging circuit with a threshold for firing SCR's. I don't think they include a temperature compensation because most people don't ride in super cold temps.

Not sure about bikes, but the units we designed for Mopar and GM definitely required temp compensation per their specs.

loudhvx wrote:

It's a pretty simple circuit, trade secret or not, but could be made many different ways.
home.comcast.net/~loudgpz/GPZweb/RegRec/GPZvRegMagnetField.html

You might be able to improve temp tracking by moving D1 above R1 and then adding a resistor above the emitter of Q1 to balance that current mirror better (?) The two emitter tail resistors coul be small value (like ten Ohms) because all they need to do is provide a little degeneration to balance the diode against the VBE of the PNP transistor.

Yeah but if u did that,the MMCIC could have a premature discharge leading to a weak angle of dangle,resulting in overloaded flux capacitors,cloudy turn signal fluid,and a VERY dissatisfied customer.Humpf.BJ

bountyhunter wrote:

loudhvx wrote:

It just an averaging circuit with a threshold for firing SCR's. I don't think they include a temperature compensation because most people don't ride in super cold temps.

Not sure about bikes, but the units we designed for Mopar and GM definitely required temp compensation per their specs.

The final voltage on the KZ's seem to be more affected by the voltage losses in the detection circuit wiring. (brown wires on the KZ's). Bikes just seem to be more crude in general so I don't think they put a lot of sophistication in it. They just kept the final voltage lower to reduce battery boil. This just means the battery may take longer to charge and may not necessarily charge to 100%. That's probably why bike batteries don't last as long as car batteries. I see car systems go up to and sometimes over 15v, but the KZ's with a good system generally doesn't go much above 14.5v.

Anyway, the one I built acted identically to the factory unit when I tested it on the bike.

bountyhunter wrote:

loudhvx wrote:

It's a pretty simple circuit, trade secret or not, but could be made many different ways.
home.comcast.net/~loudgpz/GPZweb/RegRec/GPZvRegMagnetField.html

You might be able to improve temp tracking by moving D1 above R1 and then adding a resistor above the emitter of Q1 to balance that current mirror better (?) The two emitter tail resistors coul be small value (like ten Ohms) because all they need to do is provide a little degeneration to balance the diode against the VBE of the PNP transistor.

Thanks,
Yeah, as I said, I didn't really want to develop it much further due to cost.

bountyhunter wrote:

polkat wrote:

Bike is a 1982 KZ750H2. Does anyone have a schematic of the circuit INSIDE the regulator/rectifier box? I could reverse engineer it if I could get past the epoxy sealer without damaging the circuit, but that's not possible. Or were these circuits a trade secret?

I wouldn't try it just because the rectifiers probably need to handle about 100V and 100 Amps. Those would be monsters to try to find. The SCR shunt probably has to handle maybe 50 Amps or so with at least 200V breakdown. It's not technically sophisticated but monster size devices.

Another problem with auto electrics is they are subject to high peak voltages that happen quickly but still degrade the semiconductors over time. That's the reason so many rectifier/regs turn up dead for no aparrent reason.

The rectifiers I used were just the simple bridges in the link I posted. It seemed to handle the load and shunting just fine without much heat. Each bridge is rated at 20 amps, so considering the three phases it's will provide more current than the main fuse can handle. The only issue would be the negative side diodes during shunting. I don't recall if I've measured that. I do know these have been used on several bikes successfully, even in shunt-type systems.

The SCR's I have listed worked fine even without heatsinking. They didn't get super hot, if I recall, and they were only about the size of a to-220. I probably did run the bike at higher RPMs with the lights off to stress-test the SCR's, but only for 20 minutes or so. They were, however, the components that pushed the price up too high. They had a 300amp surge rating. I would definitely recommend a big heatsink for them, though.

I'm not sure if the factory units have capacitors in them, but I suspect that is the part that fails first during normal use (after 20 or 30 years). That's because the most common failure mode I've seen on the reg/rec combos is that they fail in such a way that they seem to work fine at idle and lower RPMs, but as the RPMs go up, the voltage gets shunted at lower and lower voltages. Eventually the regulator shunts continuously at a sustained RPM. I think over time, this eventually burns out the SCRs.

This confuses people because they do their voltage testing while the bike is parked and motor running. Everything seems fine until they go on a long highway run, then the battery is dead at the end of the ride. The way to find this failure is to have some sort of voltage detector on the bike all the time.

Old Man Rock wrote:

@ Loud...

Other then power dissipation (split current) across the diode banks, any other reason you jumper across to the other AC input for each FWB rectifier AC phase input?

OMR

Just to split the load among more diodes. It doesn't get split evenly due to tolerances in the diodes, but adding more still helps take some abuse.

hocbj23 wrote:

Yeah but if u did that,the MMCIC could have a premature discharge leading to a weak angle of dangle,resulting in overloaded flux capacitors,cloudy turn signal fluid,and a VERY dissatisfied customer.Humpf.BJ

Couldn't have said it better myself.