Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Advice on regulating a 6V motorcycle alternator.

Status
Not open for further replies.

Nick E

New Member
I am currently restoring a British 1964 Velocette LE motorcycle. It is fitted with electrical components made by the, now defunct, Miller Company (also British).

The vehicle has a 6V 3 coil alternator with permanent magnets mounted on the flywheel and is rated at something over 60W. It is a former police bike and has an uprated alternator which, unlike the standard 6V alternator, does not convert well to 12V.

Originally the bike just had a lead acid battery, a selenium rectifier and no regulator but the load was balanced by a very complicated 8 pole switch which switched in the appropriate number of coils depending on the electrical load required. This worked reasonably well with an open vented lead acid battery.

I wish to fit a Cyclon 6V 8Ah sealed battery and understand that this will mean I must fit a regulator. My problem is that nobody seems to make a 6V regulator for an alternator. Podtronics in Texas make a 12V unit which will also allow conversion of the Lucas 6V alternator to 12V but, as mentioned, conversion from 6V to 12V is not advisable with my alternator.

I have spent considerable time searching for a suitable device but have come to the conclusion I will have to build one. I have the necessary amateur skill to do so but really have no idea about the design i.e. circuit design and component values. I realise heat dissipation is important but the mechanics of that I can deal with.

Any advice would be very gratefully received having spent interminable hours searching for a solution!
 
Nick,

My guess is that you need some Schottky diodes (low forward voltage drop, for rectification) and a very-low-dropout regulator (because you do not have much headroom between the rectified output voltage and the actual battery voltage during charge).

Have you traced out the coils in the alternator? Are they wired wye, delta, or do you have access to all six wires?

Do you have any idea of what the per-coil open-circuit voltage is when the engine is running above a fast idle?
 
Last edited:
Thanks for your reply. The alternator has three output terminals. One terminal is common with one wire from all three coils. Two of the remaining wires are connected to the second terminal and the third wire is connected to the third terminal. These three terminals are then connected through the 8 pole switch which connects either one, two or all three coils depending on the switch position.

I am afraid I do not know the open circuit output voltage until I start the machine after the rebuild is almost complete by the end of this week.
 
Normally an alternator output is controlled by regulating the field current, but that can't be done, of course, with a permanent magnet field. I believe typically brute force regulation is used by a high power shunt regulator or zener diode to dissipate the excess energy. In your case that's a maximum of 60W which is a lot of power to get rid of.

A series regulator may be problematic since the alternator output voltage could get quite high when the output current is limited, causing the regulator to blow. But if you can measure the open circuit voltage and it is within the rating of a HV series regulator then that's the way to go.

A simple way to make an accurate shunt regulator for any voltage (unlike a zener which is only available in discrete voltage levels with a tolerance) is to use an IC shunt regulator with a high power transistor booster such as the TL431. https://www.electro-tech-online.com/custompdfs/2009/08/TL2FTL431.pdf whose voltage can be set by two resistors (see Figure 12). If the resistor values are set high enough (over 10k) then the bias current when the motor is off will not significantly discharge the battery.

You will need a darlington transistor or darlington connection so that it can carry the 10A maximum current when being driven by the 100mA max. output of the TL431.

The transistor will have to dissipate up to 60W so it must be mounted on a large heatsink to minimize its temperature rise (one with around 1C/W thermal resistance).

The transistor will need to be isolated from the heatsink (such as with a mica washer) or the heatsink isolated from the chassis to avoid shorting the transistor (which has the collector connected to the case).
 
...The alternator has three output terminals. One terminal is common with one wire from all three coils. Two of the remaining wires are connected to the second terminal and the third wire is connected to the third terminal. These three terminals are then connected through the 8 pole switch which connects either one, two or all three coils depending on the switch position.

I'm not seeing the picture. If you remove the switch, which will not likely be used with an electronic regulator, how many wires come out of the alternator coils? I would guess four, one from the common connection, and three more from the ends of the three coils. Is that correct?

Re Carl's suggestion of a shunt regulator, how about dumping the excess output of the alternator into the headlight/running lights (i.e. always having them on). The regulator could have daylight/night modes where in one case, the excess over what is needed to charge the battery goes into the daytime running lights, and in the other, the lights get priority over battery charging...
 
Last edited:
Re Carl's suggestion of a shunt regulator, how about dumping the excess output of the alternator into the headlight/running lights (i.e. always having them on). The regulator could have daylight/night modes where in one case, the excess over what is needed to charge the battery goes into the daytime running lights, and in the other, the lights get priority over battery charging...

Mike, that's a good suggestion but the problem is that it's likely the lights can't absorb all the charging current when the battery is fully charged and the excess will be forced into the battery, which is not desirable.

A solution would be to add a power resistor in parallel with the lights with a value that would insure all the generator power could be absorbed. For daytime running, the lights and resistor would be connected to the collector of the shunt regulator transistor (the other ends go to ground of course). The night switch would remove the lights connection from the transistor (but leave the resistor) and connect them directly to the battery.

The transistor dissipation should thus be no more than about half the maximum generator power or 30W, which is more manageable.
 
Mike, that's a good suggestion but the problem is that it's likely the lights can't absorb all the charging current when the battery is fully charged and the excess will be forced into the battery, which is not desirable.
Modern bikes do use the "forced shunt" regulator technique typically using SCRs. The regulator feedback drives the gates of the two SCR's that go from the field windings to ground. The winding current is "OR" connected and can either go to the bridge rectifiers (and thus the bike's electrical system) or to the SCR shunts to ground. The regulator constantly adjusts the shunt current to make the system voltage hold at 14V.
 
Last edited:
Modern bikes do use the "forced shunt" regulator technique typically using SCRs. The regulator feedback drives the gates of the two SCR's that go from the field windings to ground...

The OP's alternator has no field winding; it uses permanent magnets, so there are only two possibilities: shunting the excess current to ground, or using series-regulation. Which method is better depends on what the open-circuit voltage across the coils turns out to be. The OP is going to test it.
 
Even with a 6v perm magnet alternator the open circuit AC voltage is still about 50v at highway speeds, and regardless of what people say they will convert fine to 12v (just by using a modern 12v regulator) provided you don't need to produce a lot of current at 12v.

It should run a standard "classic" Lucas 12v headlight at 3.5 amps and a taillight and brakelight etc. On the other hand if you plan to run a lot of lights etc (like the Police issue running lights) you mat have to stick to 6v or rewind the 3 alternator coils.
 
If I remember rightly, the're more constant current devices than constant voltage. The regulator is just a big zenner diode which acts as a shunt regulator to stop the battery from overcharging.
 
The OP's alternator has no field winding; it uses permanent magnets, so there are only two possibilities: shunting the excess current to ground, or using series-regulation. Which method is better depends on what the open-circuit voltage across the coils turns out to be. The OP is going to test it.
Well, if it is like a basic motorcycle system, it has what I called a field winding, it could also be called a "stator" winding because it is stationary. In a generator where the armature is permanent magnet (there is no armature winding), I believe the stator is the field winding but terminology aside: bike's do use SCR's to shunt the excess current in the windings to ground and the regulator is used to enable the gates of the SCR's so the system voltage is regulated to 14V.

I am not sure what winding configuration his is, it sounded like what we used to call a "star" winding:

The alternator has three output terminals. One terminal is common with one wire from all three coils.
 
If I remember rightly, the're more constant current devices than constant voltage. The regulator is just a big zenner diode which acts as a shunt regulator to stop the battery from overcharging.
Correct, but the shunting typically occurs on the AC side (output of the stator windings) before the bridge rectifiers. That way you can use SCR's because the AC means the SCR's will turn off after any cycle when the gate voltage is taken away. That's the most prevalent design these days. There may be some old designs using a giant Zener on the DC side.
 
Last edited:
Correct, but the shunting typically occurs on the AC side (output of the stator windings) before the bridge rectifiers. That way you can use SCR's because the AC means the SCR's will turn off after any cycle when the gate voltage is taken away. That's the most prevalent design these days. There may be some old designs using a giant Zener on the DC side.
So the power is dissipated in the SCR's? I woud assume the short circuit current of the generator output would be quite high.
 
Last edited:
If the stator winding act as a "current source", then the dissipation in an SCR would be the SCR's forward voltage drop times the coil current, which is managable.

The stators in a wound-rotor field type (automotive) do act as a current source. The open-circuit voltage (with max-field current applied) from an automotive alternator is like 120VAC.

Wonder if you could just use some big NFets? They seem to be cheaper and more readily available than SCRs nowadays.
 
I'm not seeing the picture. If you remove the switch, which will not likely be used with an electronic regulator, how many wires come out of the alternator coils? I would guess four, one from the common connection, and three more from the ends of the three coils. Is that correct?

No not quite. There are three stationary coils and six permanent magnets on the flywheel alternating in polarity i.e. N -S - N - S - N - S The coils are mounted at 120° to each other in a triangle arrangement. The magnets are spaced such that each coil is excited at the same time by a pair of magnets as the flywheel rotates which I guess makes it a single phase alternator (Lucas used to make a three phase alternator).

The connections are one terminal post with a common connection to one leg of each of the three coils and two other terminal posts. Of those two terminals one has connection from the remaining leg of one coil and the other has a connection from the remaining leg of the remaining two coils. The 8 pole switch, at the same time as selecting ignition, ignition and pilot lights or ignition and headlight also switches the alternator coils bringing into circuit one, two or three coils as required. To put it another way one coil is connected alone, two are connected in parallel and all three have one common leg. I hope that makes sense!

Regarding converting the bike to 12V. The reason I am reluctant to do so is that the owner's club had a very knowledgeable adviser on all matteers related to old fashioned vehicle electrical systems. He wrote many articles for the club magazine and his background was in this area. I was in correspondence with him when he recently, sadly, died quite suddenly. However, in his last letter to me he enclosed graphs of the output of the various models of alternator, fitted to the Velocette LE, and mine, being ex-police, has a higher output at 6V than the standard alternator and performs very poorly when converted to 12V. I was going to ask him to explain further but that, of course, is no longer possible.

I have had the flywheel magnets remagnetised and this evening started the bike after its rebuild. The voltage from the coils at idle showed around 10.5V. I was not able to measure the output at higher revs due to lack of time, hands and enthusiasm.

I have this evening been told where I can find an off the shelf 6V regulator. I will know for sure tomorrow. In the meantime I thank you all for yoru interest, advice and comments and will keep you all posted (if I haven't already bored you to death).
 
Wonder if you could just use some big NFets? They seem to be cheaper and more readily available than SCRs nowadays.
Don't see why you couldn't.
 
So the power is dissipated in the SCR's? I woud assume the short circuit current of the generator output would be quite high.
The SCR's do dissipate a lot of power but they are part of the regulator assembly on a large heatsink. They dissipate only the shunt current, not the total current output of the alternator. The engine's ignition and headlight/tail light/running lights are always on with the key on modern bikes, so there is always a fairly large draw anytime the bike runs.
 
Wonder if you could just use some big NFets? They seem to be cheaper and more readily available than SCRs nowadays.
Dangerous because the feedback would try to operate them in the linear range which would have huge power dissipation in the FET. The SCR is either ON or OFF, and when ON the drop is only about 1.2V. A FET in the linear range might have 10V across it and still have a very large current, which would be instant smoke...
 
Today I collected a purpose made module from a supplier of parts for old motorbikes. In conversation he told me that he sells around 6 per year because most people convert to 12V. The unit appears well made but is of sealed construction. The whole thing is within a large finned heat sink which measures about 3" x 4" and I am assured will fit the bill.

Thanks again for all the contributions to this thread.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top