Permanent Magnet Alternators - PMA

I have recently been working on a Virago with a PMA and stock regulator (thyristor type shunt). I noticed a lot of heat generated at the regulator so have spent considerable time searching for explanations...

I have learnt the following which may come as a surprise to some:

1. A PMA essentially puts out a constant current above say 2000rpm. The impact is that as the revs go up so does the voltage output. Hence the need for a voltage regulator.
   
2. The recommended way to regulate voltage on a PMA is to use a shunt type regulator. These generally come using one of two technologies - Thyristor (SCR) or MOSFET.
3. When the voltage output goes too high the shunt regulator shorts out the stator windings so dumps a lot of current.
4. The voltage output of a shorted PMA stator is very close to 0 volts. This means the power dumped through the stator is also very low since Power = Volts x Current.

So where does all the regulator heat come from then??

The way in which a thyristor and MOSFET switch on and off is quite different. MOSFET switching is very fast and precise so can provide very good voltage regulation. Thyristor switching is slower and produces a lot of heat. Some estimates I have seen put the heat generation by thyristor shunt regulators at between 40 to 60 Watts. Heat generation by MOSFET switching is quite low but I have no figure to hand.

My Virago's stock thyristor regulator when idling with the headlight off gets hot to the point of being uncomfortable to touch - approximately high 40s Celsius. I have now replaced this regulator with a modern MOSFET and under these same conditions I can barely feel any heating at all. So is it really necessary to run a high load with additional lights to soak up power that might otherwise overheat the stator??

There are many reports of shunt regulation on PMA systems damaging the stator windings. However, this is most unlikely due to the PMA shorted voltage being close to 0 Volts. So what is causing the problem?

It is thought that the heating from thyristor shunt switching stresses the rectifier diodes. This constant heating leads to slow decay of the diodes resulting in increased reverse current leakage and potentially a shorted diode. When a short occurs this can dump significant current back through the stator and damage it. Often the main fuse between the regulator and battery is high so it is often suggested to drop this value close to the max. current output of the PMA. For the Virago 750 - 1100 it is suggested to go to 15 Amp so if a short occurs the fuse will blow before the stator windings get too hot from leaking battery current.

For those in the UK I bought my MOSFET regulator for £72 from Electrex World Limited:

https://www.electrexworld.co.uk

The regulator is nicely finished and has "Made in England " engraved on it. They have a range of MOSFET regulators available but mine is quite basic with 3 phase wires, red and a black earth wire.

On their Technical and Troubleshooting page they provide a detailed procedure for testing rectifier diodes for reverse current leakage. This information may be helpful for those with PMAs when they experience unexplained voltage fluctuations or persistent issues with batteries not holding their charge.

Paul Sutton said:

So where does all the regulator heat come from then??

Click to expand...

As I understand it... .... The thyristor's are basically a solid state switch. When we reach about 14.5V they close, shorting (shunting) all that stator current to ground. Think of it this way... the alternator is still producing max current and that current is passing through the thyristor's on it's way to ground. Large current flow through a solid state device is gonna generate a lot of heat.

Thanks Jim. I have yet to find the technical explanation but the thyristor seems to have a unique property where it switches on slightly then relies on the current flowing through it to complete the switching process. There must be a significant resistance during this process to lead to heat generation. The MOSFET switches are more like a traditional switch - bang on bang off with no intermediate stages.

Wikipedia does a fair job on explaining 'em.

Found this on mosfet regulators that explains the heating difference.
https://mosfet-regulator.com/79_how-does-a-mosfet-regulator-work.html

I was under the impression that SCR regulators were shunt types and MOSFET were series types. Looks like I was wrong (again).

Don't know where i found this......... Even a dummy like me can understand what hes saying............

https://www.superhawkforum.com/forums/knowledge-base-40/mosfet-regulator-rectifiers-why-how-25117/

The "series" type regulators are an improvement over the mosfet. The SH775 by Shindengen or the newer SH847 shut down phases instead of shunting to ground.
They run cool, mine doesn't get more than 10* or so above ambient. Mine will hit 14.3v and stay there, I data log for the EFI tuning and voltage doesn't fluctuate once the battery is topped off.

Nice thing is polaris used them OEM and you can pick up a used genuine one for around $50-75.

Signal, nice little summary on that RMStator link. Sums it up well.

The shunt reg doesn't send the excess current to ground - it fires it back through the windings.
The stator has three sources of heat to deal with
1. Engine heat
2. Heat generated by the electrons being ripped through it and out
3. More heat being generated by the unused current being shunted back down through the windings.
You can see why stator windings don't last long in a lot of bikes where the headlamp is not in use. That headlamp load reduces the current being shunted back down the windings.
There is NO electrical connection between the stator windings and the frame (ground, earth, call it what you like) - if there was, it's a sign of a damaged winding and the full voltage will not be produced.
Series regulators aren't a new idea - it was obvious 40 years ago that they were the perfect solution, but the power control electronics of the day weren't up to it, and the cost of developing them were far outwith the likes of Suzuki, etc.
So... Suzuki, and others, developed the shunt type - a quick and dirty (and in some ways elegant) solution to the problem. Did I mention it was cheap?
It got the bike electrical generation system working well for the first customer to ride around happily on it, the second owner might have got away with it too, but the third and subsequent owners all paid the price, with mysterious boiled batteries, fried electrics, dead stators and dead reg-recs.

Post #70.

http://www.xs650.com/threads/solid-state-power-box-for-xs650.33037/page-4#post-418251

Grimly, you are quite right that the current is not shunted to ground. But the ground wire does play a part, it acts as a common point to short out the stator windings at the R/R: See TM's link above to a circuit diagram. I guess it is just a convenient point at the R/R. The heating from the shorting of the stator is not major due the voltage being zero for a shorted PMA.

Using Watts = Current squared x Resistance suggests that even at 0.5 Ohm Stator and 15Amps the stator can get quite hot from shunting. But this is a DC argument and we need to really know the AC reactance. I read on the Virago forum that the voltage of a shunted PMA is zero and that this is unique to a PMA. I do not understand the reasoning behind it but that is the reason that shunting is the preferred method to control voltage output on a PMA system.

I have also seen comments stating the use of the the term series regulator is a marketing term and the mode of voltage regulation is still shunting similar to the SCR system, just better suited electronics with lower heat output.

There is another source of heat due to eddy currents. These are the result of the magnets moving quickly past the metal components in a similar manner to induction heating. But this is frequency depended so at 3000 rpm it will be similar to mains frequency, so very little heating will occur. The largest heating contributors are the thyristors/SCR and removing these from the system makes a substantial increase to reliability.

I have no data to support the view that damage to the battery/stator originates from heat damage to the rectifier diodes causing battery current to flow back through the R/R to the stator. It is a view I have seen expressed on other forums and coupled to this is the idea to reduce the fuse size. The observation being the fuse blows before battery current can damage the stator when the R/R fails due to heat damage.

I notice that many circuit diagrams show SCRs, but they would only shunt half the AC wave cycle. I forget the real device used but is effectively two SCRs connected in opposite directions so that they shunt both the overvoltage from the -ve and +ve phases of the PMA output. Its the one used in universal motor controllers. Must try to find the link to the detailed explanation on a forum I visited.

But the proof is in the eating, I hope my MOSFET lives up to the manufacturer's claims. It does run very cool.

Hey, Paul.
Now that yer brain juices are warmed-up,

http://www.xs650.com/threads/stator...wires-at-the-phase-connections-melting.30742/

Here is one reference which makes interesting reading:

http://www.zefox.net/~bob/mc/vfr/alternator.html

TM, I am off to mow the lawn but will return to the link you have given. I will try to find the PMA zero voltage reference I saved somewhere?

Edit: Also...

https://www.superhawkforum.com/forums/knowledge-base-40/mosfet-regulator-rectifiers-why-how-25117/

I stand by what I said.
1. Generating current causes heat
2. Shunting the stator current back into the windings causes heat - it's still there, because the only way that current ceases to exist is if the PM field suddenly stops, and it can't.
Things are different with the series RR, which is why stators live much longer.
These two factors mean the stator is running at or near full load all the time and its days are numbered.

Grimly, I agree with most of what you wrote. The only part I am still researching wtihout much luck is why the voltage of a shunted PMA is zero. The part that I am at odds with is your point two. That does not mean you are wrong, just that it conflicts with certain information I have found. If we assume that the stator resistance is 0.5 ohms and the current is 15Amps then YES a lot of heat is generated. But I have seen it stated that the voltage of a shunted stator in a PMA is zero so therefore no heat is generated. Perhaps this zero voltage only relates to an ideal stator with zero resistance?? But then the PMA is a constant current generator so a stator working at full output sees the same current flow through it as when shunted.

I hope you see the conflicts in where I am coming from. More advanced research is where this leads me.

Bye for now.

Paul Sutton said:

The only part I am still researching wtihout much luck is why the voltage of a shunted PMA is zero.

Click to expand...

Here's a circuit and a good description:
http://gpzweb.s3-website-us-east-1.amazonaws.com/RegRec/GPZvRegMagnetField.html

In the pic below, when the voltage determined by Z1 is reached (on the capacitor), Q1 turns on, which turns on the SCRs and all three phases of the Alternator are connected to ground/0v. So it's shunted and you could say its voltage is zero...
http://gpzweb.s3-website-us-east-1.amazonaws.com/RegRec/VoltRegMagnetField.gif

Lay a box end wrench across your car's battery posts, and measure the voltage.

And watch the wrench glow and begin to melt.

If power is V * V / R, then we're close to a zero divided by zero.

Enter the mystical universe, near the extremes of measurement.

Every conductor has some degree of resistance, even superconductors.
Whenever there's current, there's always voltage.
Just gets tricky to comprehend when values approach zero...

TM, certainly is a mystical Universe.

Xjwmx, that first article you gave a link to is very good with good descriptions. Several months back when I was planning on making my own MOSFET R/R I was going to steal the voltage control section from this design. What puzzles me with this circuit diagram, and many other similar ones, is the use of an SCR. An SCR is a diode so allows current to flow in only one direction. In these regulators the SCR must deal with the overvoltage in both the +ve and -ve phase of the AC wave. Are 3 SCRs really used, or are 6 SCRs really used in pairs to allow for conduction in both directions, or is a Triac used, or some special bi-directional SCR???

Warning - Back Track: I previously commented on the current being shunted to ground and referred to the ground wire being a convenient place to tie the 3 phase windings together. In light of my recognition of the SCR being uni-directional this means the current must flow through the earth wire via the rectifier back to the stator winding - extra loading on the rectifier. If a bi-direction "SCR type" were used then my argument would have been correct and the rectifier would see less loading and less heat:



I think the mysteries of shunting PMA is centered around the concept of zero voltage. I have seen PMA properties described as being "counter-intuitive". Is the shunted zero voltage aspect counter-intuitive or just the result of how the 3 phases cancel eachother out? Answer this question and then I will sleep again...

Please correct me where I go off the rails and talk BS. No offense will be taken....

Paul Sutton said:

Are 3 SCRs really used, or are 6 SCRs really used in pairs to allow for conduction in both directions, or is a Triac used, or some special bi-directional SCR???

Click to expand...

3 SCRs are used. In the link is a circuit diagram and a parts list for something we have to assume works. It only needs to conduct in the direction toward ground.
Frankly, I'm surprised that the SCRs have to be so rugged. I'd think along the lines of both ends being so close to the same potential that it's not functioning as a generator anymore. Then when it melted down, I'd put in bigger SCRs

Paul Sutton said:

Is the shunted zero voltage aspect counter-intuitive or just the result of how the 3 phases cancel eachother out?
Answer this question and then I will sleep again...

Click to expand...

Neither one. Think of the SCRs as gating the generator away from the battery. Apparently there is some problem with simply opening the legs, which could be easily done, so instead they're grounded. The rectifier keeps it from grounding the battery.