6V DC motorcycle regulator

crutschow said:

For that reason I added a headlight modulator to mine, which causes a periodic change in brightness of the headlight (240 ±40 cycles per minute between approximately 20% and 100% of full intensity, and is legal in the USA).
Much harder for the vehicle driver to ignore a flashing light.
I haven't had anyone pull out in front of me since I installed it. (knock on wood).

Click to expand...

Wow that is something new. Haven't seen a modulated light like that in the UK. I do not think it would be legal here.

I get people, especially women pulling out in front of the car regularly. And our son has had two major accidents where women have pulled out in front of him at speed.

Mind you, I have done some stupid things on the road too.

spec

Spec,

pictures attached as requested, as I bought her, changed a little since then, though only details,

Still haven't had a chance to make anything for the regulator yet, looks like I will be going with Alec_t circuit design, any pointers on the exact Mosfet that would work?

Mike

One of these FETs would fit the bill.

Mike711 said:

Spec,

pictures attached as requested, as I bought her, changed a little since then, though only details,

Still haven't had a chance to make anything for the regulator yet, looks like I will be going with Alec_t circuit design, any pointers on the exact Mosfet that would work?

Mike

View attachment 100712 View attachment 100713

Click to expand...

N I C E Bike. Like the twin port exhaust.

spec

alec_t said:

One of these FETs would fit the bill.

Click to expand...

So this is how you include the original message, would the rest of the components need rating at 50W or more?

alec_t said:

Here's an ultra-simple shunt regulator :-
View attachment 100508
The MOSFET needs to be a logic-level type, generously rated to handle ~50W or more for extended periods and must be mounted on, but electrically insulated from, a big heatsink. The bike's chassis could be part of the heatsinking arrangement.

Click to expand...

Alec,

can I suggest that you fit a 330mili Ohm resistor in series with the drain of the NMOSFET: http://www.digikey.co.uk/product-de...y-amp-connectors/HSA50R33J/A102447-ND/2366305

And perhaps use this NMOSFET (much lower ThR junct/case= 0.4 DCW and higher VDS): http://www.digikey.co.uk/product-search/en?keywords=IRLB4030

This would make the regulator more robust, always a good thing on vehicles, especially motorbikes. http://uk.rs-online.com/web/p/heatsink-mounting-accessories/5040592/

Mike,

The resistor can be bolted to a heatsink without an insulator, but the NMOSFET will require an aluminum oxide insulating washer (the NMOSFET drain is connected to the metal part of the case), between the NMOSFET case and a separate small insulating stepped washer under the single retaining bolt: http://www.digikey.co.uk/product-detail/en/aavid-thermalloy/7721-7PPSG/HS418-ND/1625304 The mating surface must be perfectly flat and free from burs and sharp edges or the washer will be damaged. Note that other materials are used for insulating washers, but Aluminum oxide has a far superior thermal conductivity: **broken link removed**

It would also be wise to clamp the NMOSFET with a bar running over the plastic part of the case and held down to the heatsink by two bolts (just a strip of aluminum with a hole at each end). This will give more even pressure on the PMOSFET for improved heat dissipation and also make the PMOSFET less likely to vibrate lose, always a danger with the single hole fixing cases. http://uk.rs-online.com/web/p/heatsink-mounting-accessories/5040592/

A thin layer of heatsink compound will also be required on both faces of the insulating washer.


spec

spec said:

can I suggest that you fit a 330mili Ohm resistor in series with the drain of the NMOSFET

Click to expand...

It would do no harm, but hopefully wouldn't be necessary, considering the magnetos' coil impedance is probably higher than 0.33Ω at high rpms where the regulator will cut in.
That FET in your link does look better (albeit pricier) than the one I suggested, Spec.
Is the ref to a PMOSFET a typo? My suggested circuit uses just one NMOSFET.

Mike711 said:

would the rest of the components need rating at 50W or more?

Click to expand...

No. There is only 1 other component; the 10k pot. A 1/4W rated one will be fine.

Edit: Even a 0.2W rated pot will do. I suggest a small trimpot such as this. You could wire it directly to the FET legs.

alec_t said:

It would do no harm, but hopefully wouldn't be necessary, considering the magnetos' coil impedance is probably higher than 0.33Ω at high rpms where the regulator will cut in.
That FET in your link does look better (albeit pricier) than the one I suggested, Spec.
Is the ref to a PMOSFET a typo? My suggested circuit uses just one NMOSFET.

No. There is only 1 other component; the 10k pot. A 1/4W rated one will be fine.

Edit: Even a 0.2W rated pot will do. I suggest a small trimpot such as this. You could wire it directly to the FET legs.

Click to expand...

Hy Alec,

Yes, PMOSFET was a typo.

I was working on the dynamo producing 6.8V at 10A and thought that the 0.33R resistors would reduce the dissipation in the NMOSFET. As the regulator circuit regulates at 6.8V, by setting, and assuming that the dynamo puts out 10A, that means that the NMOSFET would be dissipating 68W, with a charged battery and lights off, which is a bit on the high side for comfort.

I was also thinking about fault conditions: battery disconnected while engine running, bad earth etc.

During my adventures with automobile, and especially motorcycle electronics, I found that a belt and braces approach is needed because the environment seems to be extraordinarily harsh: temperature, voltage, and vibration.

spec

Here is another regulator circuit- not quite as simple and elegant as your circuit Alec.

spec

Issue 7 of 2016_08_07



ERRATA

Shouldn't the opamp inputs be swapped in that circuit?

Edit: Ah, now edited.

alec_t said:

Shouldn't the opamp inputs be swapped in that circuit?

Click to expand...

ARRAGH- thanks Alec.

spec