The regulator rabbit hole

[Boxprod]

As one or two of you who have read my waffle so far, in my retirement quest for an “improved” but budget-friendly late-model 350 Mk3, I have completely rewired it to give a stable 12-volt system.

I’ve de-potted the red Motoplat stator and rewound it with the help and advice of Roger Lovelock (https://www.bbbikeshop.co.uk), but retained the Motoplat ignition winding, which was rewound a couple of years ago by Steve at Bradford Ignitions (now retired, I’m afraid, but Roger L could probably help).

This gives me a single-phase AC output across two wires (plus the ignition wire), and it’s more than enough for the all-LED lights and charging the 3Ah AGM battery – which is plenty for a load under 20W with all the lights on.

So the next thing is the regulator/rectifier. Being short of cash, I first tried a cheap Chinese one, as mentioned several times across this forum, and to keep an eye on things I installed a small digital voltmeter as well as a three-colour LED charging light from these kind people
https://www.improvingclassicmotorcycles.com.
Since the first Chinese knock-off charged things at over 14 volts, I got another (only a tenner on eBay), which kept things at about the recommended 13.2-ish volts.

However, I discovered that after about 20 or 25 miles, the voltage would suddenly drop to less than 12, and things would go haywire, jumping from 11.9 to 13.8 erratically. Then mysteriously it would stabilize again to 13.2-ish for another 15 miles or so before going wild once more. And so on. I feared it might be the dreaded thermal degradation which often affected the potted Motoplat stators – showing up as melted resin, apparent scorch marks and cracked insulation – but I reckoned it can’t be that because I had just rewound the stator, and there were no shorts in the all-new windings.

Suspecting the bargain-basement regulator, I spent a decent amount on an Electrex one, which did exactly the same thing. So probably not a regulator fault, but what on earth was going on?

I found out the following: all the regulators I had tried were “shunt” regulators, which prevent overcharging of the battery by simply shorting to earth any alternator output once the regulator detects a battery voltage over 13.5 or so. The alternator continues to pump out its usual maximum, which has to go to earth, which explains why regulators run quite hot, and on some bikes are notorious for it.

MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor) technology, as used in the Electrex RR122 I bought, sorts out the spikes and bumps in this process, but it doesn’t prevent the charging voltage from jumping about in the way I saw.

(It’s curious that since the passing of the old ammeter very few bikes have voltmeters, while loads of sporting cars have them, which might be because a shunt regulator and a voltmeter would scare the motorcyclist into a warranty claim.)

More research followed. AI is a wonderful modern thing, but by golly you have to dig hard to confirm that it’s not telling lies!

It turns out – and I’m sure a lot of folk know this already, but I’m not an electrical engineer, so please forgive me – that there’s another sort of regulator known as an “open” or “series” regulator. This senses the battery voltage and turns off the AC input, or in other words stops drawing current from the alternator when it doesn’t need it. This means it is kinder to the stator, which is in considerably less danger of overheating, and also causes a tiny amount less magnetic drag in the engine – although the small fraction of a single bhp released thereby won’t make any real difference.

Open regulators are commonplace on things like ATVs and adventure bikes, as well as big touring machines. These might be fitted with extra electrical kit like heating and extra lights and so on, which won’t be needed all the time but which could impose an intermittent load of 200 watts or so, and a shunt regulator would get really stressed when that load is turned off. Which is why open regulators make sense in that case.

Go looking for an open regulator and most of what you’ll find will be “replacement” units which claim to be as good as OEM, but really aren’t. They are, without exception, fakes. Cheap Chinese knock-offs just like the original ones I bought, but given a nice plastic plug to match the loom connection on your modern Japanese sportbike. If it costs less than £100 new, it’s probably not a proper open regulator made by a reputable manufacturer.

By far the most significant manufacturer is the Japanese company Shindengen, which provides OEM stuff for bike makers across the world. Their model range covers a huge variety of demands, and you can’t buy direct from them, so their website isn’t really too useful to pin down exactly what would be the right unit for my purposes. But dig further in the etherweb, and some answers do eventually emerge.

My alternator doesn’t give a massive output because I don’t need it, so I could use a regulator from the smaller end of the modern spectrum. After a while I found that a 2019-ish Honda CB125F was fitted with a single-phase open regulator built by Shindengen with the model name SH879AA. Of course they’re difficult to find new, and the internet keeps trying to sell you the fake one no matter what you do, so how about a used one from a motorcycle breaker?

That led me to a business called James Sherlock Ltd (trading on eBay), and I bought one in very good condition for all of £19. Then came the puzzle of what to connect to which pin.

I said earlier that you have to work to get the truth from AI, and I learned this the hard way when I researched the wiring convention for this particular regulator. After a lot of AI digging, I was fairly confident that the 5 pins should be connected in a given pattern. The two AC wires into the top two pins, earth to the bottom left, switched positive to the bottom middle, and permanent positive to the bottom right. But when I connected the regulator in this way, my onboard voltmeter showed a drop to 10.9 so I hastily disconnected it and tested it with a multimeter – which revealed I had fried a diode, rendering the regulator useless.

Back to the internet, and kick AI hard, and eventually it admits that the wiring convention should be as follows: AC to the top two, but battery positive to the bottom left, nothing to the middle, earth to the bottom right. Then, when I try to confirm its verdict by asking the same question in a slightly different way, it says exactly the opposite for the +ve and -ve pins, and no matter how forensically I question Google’s AI search engine, it cannot confirm which way to connect the thing. Every time I ask it, it says one thing in full AI mode, and then the opposite in AI overview. And this is going to destroy jobs and take over the world? Judging by this evidence, it’s going to short itself out before it gets anywhere. Here’s hoping.

The next search was for how to test the regulator with a multimeter to see which pin is which – and that’s not too hard. Although, of course, the AI idiot isn’t terribly clear about the process, not seeming to understand that absolute accuracy is vital when dealing with the physical world rather than the virtual one.

Every one of my AI searches gave answers that proved to be completely wrong. When pressed to give an answer that was based on real data, it simply collapsed: “Data not available”.

It’s entirely possible that the specification of the SH879AA isn’t available, because Shindengen are said to have suffered badly in the Fukushima tsunami, but that was in 2011, and these regulators have been fitted up to the present day, I think. But the factory only sells to motorcycle manufacturers, so they may guard their secrets carefully.

Whatever, James Sherlock sold me another unit, and more by luck than judgement I managed to wire it up in the correct pattern – which if anyone’s tempted, is like this:
• Top LEFT – Battery positive
• Top RIGHT – EARTH negative
• Bottom LEFT – AC input
• Bottom MIDDLE – Switched positive
• Bottom RIGHT – AC input
That’s when you’re looking into the wiring socket on the side of the unit.

Shindngen SH879AA pin mapping.jpeg

s-l1600.jpg

With the regulator wired in what I believe to be the right way, I held my breath and started the old bike. Voltmeter at 14.1, dropping to 13.8 with the lights on. Phew.

So why aren’t bikes given open regulators in the first place?

My initial thought was unit cost. Shunt regulators are much cheaper, and that could explain it. No manufacturer would spend £100 rather than £10. But the components in the circuitry are pretty similar, so maybe it’s that old Catch 22. Open regulators are more expensive because fewer are made, while shunt regulators are cheap because loads are made? Not convinced.

Here are a couple of interesting (to me...) articles on Grant Tiller’s website. I know not everybody agrees, but I found them very informative as I went on my quest.
https://britspark.com/regulator-rectifier
https://britspark.com/regulator-rectifiers-alternators#Open-Type-also-known-as-Series-Type-SCR-Regulator-Rectifier

Best wishes to all! And PLEASE don’t believe anything of a technical nature that comes from AI,

Pete.

[Jordan]

Nice to know that someone else spends too much time trying to improve Italian electrics.
For what it's worth, here's my thoughts:
Some early bikes fitted with alternators had no regulators as standard, relying on balancing the output with the load so not wise to change headlight wattage too much. A double pole light switch put more charging coils into circuit as required.
Shunt type regulator of the most basic type is a single component zener diode which was used very successfully on British bikes from 1960s.
It can also work on our 12v-converted Ducati singles with a simple wiring change. Trouble is, that zener diode isn't easily available now.
But combined rectifier-regulators are, fitted to golf buggies, ride-on lawnmowers etc. It's what I use now with no problems, even on my 750GT. Chinese copies of these seem good and are very pocket friendly.
My 350 single has no battery but I have a voltmeter on it, which shows levels maybe a bit too high, but it doesn't blow bulbs and there's no battery to boil. The GT has no voltmeter to frighten me.

[mrkprsn]

As Jordan mentioned the early 60's Ducati singles with a DC ignition used just a rectifier(3 wire up to 200cc). That stator put out high single digit AC voltage and the rectifier would change it to DC and reduce the voltage a bit so the output would be around 7V. Then Ducati (250-350) went to a larger 2 wire stator that put out a higher AC voltage and used a R/R. That one had 2 diodes that clipped the negative voltage so only the DC waves persisted.

I ride twice a year in a Motogiro where we ride over 300 miles over 2 days. On 3 occasions my battery blew up due to over charging. I tried a number of different R/Rs and they behaved just like you mentioned in your post. I spent a lot of time studying this problem. Eventually I swapped out my battery for a capacitor and have never had a problem since. If you do that it really doesn't matter if your R/R puts out 12V or 16V.

Below is a link to a good mosfet R/R. They also sell a 6V R/R with a 5th wire used to sense the voltage. I really like their stuff.

https://sparckmoto.com/collections/char ... regulators