Could these be SCRs? I don't understand this circuit!

[MaxSimmonds]

Hey Guys,


I reverse engineering this circuit:

https://imgur.com/a/jemT1vB

It looks like these could be SCRs, unfortunately I kinda destroyed the packages whilst I was depotting the PCB:

and I can't find any datasheet for what I think says "IS125 03F USA". I don't recognise the logo either. ​


it's a voltage regulator rectifier, and I was thinking that the random 3 legged package might be an SCR, but I can't seem to get an LT spice sim working. I can do it if I assume they are mosfets (see attached) but I just don't think you would use them in this application, the current through them would be too high for an SCR. This is from a motorbike, so the alternators on them can be fairly powerful (300W for example). This is a single phase motor bike, hence the single phase rectification.

Any ideas if this could even be an SCR, or is it more likely a FET?

[T3sl4co1l]

So this is a rectifier between alternator winding (single phase) and battery / system DC power?

I suppose they could be SCRs.  Shunt the alternator winding when over voltage?  Crude, but I suppose it'll work.  I would guess the SCRs should be rated 50V 10A or more, how much exactly is needed depends on alternator rating and maximum module temperature.

MOSFETs would dissipate many times more power than the same SCR circuit.

Tim

[CaptDon]

Although it seems very odd, I had a Suzuki TT-500 Titan that used SCR's to short out the alternator to control the charge rate. The alternator winding I guess was cooled by the incoming air/fuel/oil mixture since it was a two-stoke engine and the alternator stator was internal to the engine with permanent magnets providing the excitation. I have also seen SCR's in series with the single phase alternator stators. They are wired in a way that they normally trigger and a transistor and zener arrangement hold off the trigger when the battery is at full charge voltage. This whole el-cheapo arrangement really only works because there are no filter capacitors and even after the bridge rectifier the voltage still drops to zero at zero crossing of the stator A.C. output which allows the SCR's to operate on a 'per half cycle' decision. The permanent magnets and self inductance as engine speed increases help limit the total fault current in the SCR-Shorting scheme.

[CaptDon]

Although it seems very odd, I had a Suzuki TT-500 Titan that used SCR's to short out the alternator to control the charge rate. The alternator winding I guess was cooled by the incoming air/fuel/oil mixture since it was a two-stoke engine and the alternator stator was internal to the engine with permanent magnets providing the excitation. I have also seen SCR's in series with the single phase alternator stators. They are wired in a way that they normally trigger and a transistor and zener arrangement hold off the trigger when the battery is at full charge voltage. This whole el-cheapo arrangement really only works because there are no filter capacitors and even after the bridge rectifier the voltage still drops to zero at zero crossing of the stator A.C. output which allows the SCR's to operate on a 'per half cycle' decision. The permanent magnets and self inductance as engine speed increases help limit the total fault current in the SCR-Shorting scheme.

[T3sl4co1l]

Although it seems very odd, I had a Suzuki TT-500 Titan that used SCR's to short out the alternator to control the charge rate. The alternator winding I guess was cooled by the incoming air/fuel/oil mixture since it was a two-stoke engine and the alternator stator was internal to the engine with permanent magnets providing the excitation. I have also seen SCR's in series with the single phase alternator stators. They are wired in a way that they normally trigger and a transistor and zener arrangement hold off the trigger when the battery is at full charge voltage. This whole el-cheapo arrangement really only works because there are no filter capacitors and even after the bridge rectifier the voltage still drops to zero at zero crossing of the stator A.C. output which allows the SCR's to operate on a 'per half cycle' decision. The permanent magnets and self inductance as engine speed increases help limit the total fault current in the SCR-Shorting scheme.

That should be something like this:
https://www.homemade-circuits.com/scr-voltage-regulator-circuit/

Voltage drop can be reduced by using SCRs for the two positive FWB diodes, and regular ones for the negative.  (In that case, two R+zener networks will be needed, one for each SCR.  Except at low voltages -- peak input less than Vkg(max) -- where the gates can be wired together.)

In the shunting circuit, a single SCR can be used, after the FWB, only if there is a second series diode into the cap -- otherwise the SCR would crowbar the cap, a non-starter.  And obviously, this would increase the voltage drop another step, so has a similar problem as the above two cases.  Using two SCRs, as the module apparently did, seems like the better option.

Tim

[floobydust]

TO-92 isn't going to be handling the power here. There should be bigger SCRs, sometimes raw dies attached to the case.
If only we had a make/model, 6V or 12V, scooter how many CC's, or better pictures etc.

I've never seen any MOSFET designs, everything is stuck in the 1970's with the circuits simply copied.
Shunt regulators in small engines are prevalent and terrible for wasting excess energy as heat, but more efficient with lower voltage drop. I guess extra stator windings are too much cost to generate the extra volt or two.

[T3sl4co1l]

I've never seen any MOSFET designs, everything is stuck in the 1970's with the circuits simply copied.
Shunt regulators in small engines are prevalent and terrible for wasting excess energy as heat, but more efficient with lower voltage drop. I guess extra stator windings are too much cost to generate the extra volt or two.

It's kinda funny, because SCRs are so stupendously cheap (not much more than diodes of the same rating?), and electrical efficiency is such a negligible factor in the overall design.

A switching regulator for example, would be a fine addition for electrical efficiency, and plenty reliable given today's parts and ratings.  But what is efficiency, anyway?  Ultimately it's fuel savings, but so what, the engine is a pitiful 10 or 20% or whatever, electrical amounts to a tiny percentage of that, and there are far, far lower fruit to pick -- like just upgrading from 2 to 4 stroke, fuel injection, various emissions controls (better fuel burn / less unburned waste), and I mean, just plain old tuning, or even driving habits for that matter.

So, from a system design standpoint, it really doesn't many any sense, and yeah sure, just keep cooking the winding, who cares.

Tim

[MaxSimmonds]

Hey thanks everyone!

@floobydust , it was off a royal enfield 500cc bullet. 2007. I checked the service manual, and it only showed the following (incorrect) diagram:

https://www.manualslib.com/manual/818784/Royal-Enfield-Bullet-Res-500.html?page=180#manual << source

https://imgur.com/a/nzIP4tU

I'm afraid I can't give you a better image of the ICs, these are all I have:

https://imgur.com/a/yyNssSx

None of which are any good.

My question, though, is (assuming these SCRs which I think is a good assumption) how would you design it with only 1 zenner and 1 resistor to shunt the AC signal? When I try to do it, it'll shunt one half, but allow the other through, this is when I have both shunts on each AC input. I've never really worked with SCRs before, only FETs/BJTs. It feels like it should be simple, but I get confused with the resetting of the SRC

Any help (even better, an LTspice sim showing the SRC in action shunting AC signals!) would be very much appreciated!

Max

[floobydust]

OK I looked at your pictures and the SCR's were a TO-220 package part with rivets to the case, that got broken off. Now this makes sense. I would say the SCR's are house numbered IS125 USA, sensitive-gate and KAG pinout, isolated tab. It looks like an AC shunt regulator. The PC board has provisions for more circuitry, this is the simplest, cheapest design possible and I'm not a fan of Sanwha electrolytic capacitors in hot places. They don't even seem to have a control transistor. It's like a British PC board the Koreans/chinese did their magic on.
You can use the usual Rotax circuit or did you want to repair this? It's main goal is to be cheap and easy to manufacture. chinese quad/atv/motorcycles/mopeds use the same module.
https://www.eevblog.com/forum/projects/ac-generatoralternator-charger-rs1105-replacement/msg1256470/#msg1256470
From what I can tell, if the system voltage is too high, the zener conducts and activates that SCR which latches on for that 1/2 cycle, shorting out the stator and wasting its output as heat. At zero-cross the SCR turns off, and then the other 1/2 cycle it repeats with the other polarity SCR.

Apparently Lucas came up with the shunt regulator in the 1960's using a big stud-mount zener diode and here we are 60 years later. I'd seen small engines using shunt regs that made it into ultralights and too much cringe.
It must be that MecE's are lazy and just copy old EE tech, or the pennies count and no one other than Shindengen has done any improvements like use mosfets.

[CaptDon]

My 1984 Evinrude 3 cylinder 70hp outboard had ONLY a bridge rectifier, no regulator!!! The battery voltage would climb to nearly 16vdc. Often I would keep the running lights on with the motor running in the daytime to hold the voltage down to around 13vdc to avoid damage to my other marine electronics.

[MaxSimmonds]

Thanks @floobydust!

So I did a sim of the circuit I did a daveCAD of:

https://imgur.com/a/GlSzc2O

I have connected it as you said, KAG, but it doesn't make sense (and nor does it work!). The gate would be grounded. Of course, it doesn't work.

How would I fix this? I don't want to repair this regulator, rather, it's for my dad and he want's a smaller version of it (it was really big with that cap). I'm thinking of redesigning it, but it bothers me that I don't understand how something this simple works!

[floobydust]

First comes the schematic and parts ID, then comes the simulation...    I will look at that later. How many amps/watts is it rated?
I would say the bulk of the size is from the heatsink, large through-hole parts, and because these are heaters in a way. Nobody has made a modern version of these paleolithic regulators for the masses. I imagine with Schottky diodes or active rectification it would be much better.
Rough estimate is about 3W per amp and on the highway that is a lot of heat, for the stator as well.

[T3sl4co1l]

K and G are swapped.  Doesn't make sense any other way.  Not sure about typical SCR pinouts, how many there are, but a left-right flip isn't uncommon elsewhere e.g. BCE vs. ECB in Japanese transistor families.

Tim

[MaxSimmonds]

Hmmm, that doesn't work either!

https://imgur.com/a/tY8ZOJj

I've attached the ltspice file and zipped lib, maybe someone can show me where i'm wrong?

[T3sl4co1l]

R2 is too big.

Tim

[MaxSimmonds]

Oh man now I feel stupid, I put that there as I thought the output (for some stupid reason) needed to be isolated, and I thought the sim wouldn't run without a ground, well how stupid of me!

Thank you very much!

[floobydust]

The SCR's might be flipped, not sure how the unit goes together but the TO-220 pieces don't line up. It still might be a custom part, an isolated tab SCR I thought is uncommon, unlike triacs. Realized the circuit has nothing to limit zener+gate drive current? the 510R resistor is G-K. Sometimes these modified, cheaper deals don't even work.
It's just weird to have Korean capacitor, USA on the SCR, and it's a British PCB and motorcycle.