Voltage Regulator for a Small Motorbike, 6V AC Current

[sign216]

sign216,

Secondly, when the brake light burns out, the end of the second coil that should be grounded, is not grounded, and that is what stops the engine. The brake light switch grounds that end of the coil when the brake is released, allowing the engine to run.

What's the purpose of putting the brake light as a part of that ground? Seems like inviting failure.

As I understand it: One end of the high tension (ignition) coil needs a ground, which is normally wired in. When the brake light switch is applied, that ground is broken, but the brake lamp is substituted as a path to ground. All is well, unless the lamp is burnt out and then there's no ground. Do I have it right?

 

[sign216]

Diver:

Thanks for jumping in. It's been lonely in here.

The shunt regulator around the brake light is an option, because when the bulb blows, the regulator would take the place of the 6V drop. You don't have to design it for 10 Amps, just a nominal value of 2.5A. It will only draw the full 2.5 A when the bulb blows. Otherwise it will be doing very little UNLESS you try to make a 6 VAC replacement LED bulb.

Kiss, my first goal is optimize the current 6v AC system. To get it to run well. After it's running right, then the modifications/improvements come.

I'm holding off on LEDs because it's 6v AC, which are not LED friendly. In theory in LEDs are optimum, but real world motorcycle tests show it's hard for LED to beat an incandescent, and that's on a 12v DC system. See here: https://www.webbikeworld.com/lights/led-brake-lights-for-motorcycles/index.htm

Switching to an LED headlight and taillight package would be great, but that must be a future project.

 

[sign216]

It can be. According to the 98ss diagram the main lighting switch receives the LT coil voltage either directly via the red (rosso) wire or indirectly via the choke IM and the green (verde) wire. The switch is shown as a 3-position one (centre off) and applies this voltage to the main/dip beam filaments (azzurro and marronne wires), the instrument light LPA (bianco wire) and the tail light LPP (giallo wire). All of these bulbs will be protected/dimmed by the choke presence if the indirect route is chosen. Perhaps the dimmed lighting is intended for daylight use?
You mean the brake light? Perhaps because the LT coil can't provide enough current for the extra 2.5A that the brake light draws, or for the reasons KISS gives, or even to drop the HT spark volts to reduce engine speed during braking.

I've been having a play with simulations of (a) a conventional car-type Kettering ignition system with the points in series with the ignition coil primary and (b) the system as per the 98ss diagram. I was surprised to find that (b) gives a higher HT voltage (I think because there's a greater current change when the points short the magneto coil). So the MC manufacturers are right and I was wrong on that score .
I've also come up with a possible way of regulating the brake-light voltage, using a couple of FETs. The 'regulator' (shown in the sim circuit below) clamps the voltage when it tries to go above ~6V (adjustable by resistor selection) and simply connects across the brake bulb/switch.
View attachment 88701

Alec, your three diagrams are 1) the system as it is now, 2) system w a regulator, 3) regulator design. Am I right?

 

[Diver300]

What's the purpose of putting the brake light as a part of that ground? Seems like inviting failure.

As I understand it: One end of the high tension (ignition) coil needs a ground, which is normally wired in. When the brake light switch is applied, that ground is broken, but the brake lamp is substituted as a path to ground. All is well, unless the lamp is burnt out and then there's no ground. Do I have it right?

The purpose of putting the brake light as part of the ground path is to light the brake light. It is inviting failure, because the brake light filament is in series when the brake is on.

Other designs would need more components. The 1983 Honda CG125 that I owned had a battery and a rectifier, but that was built 20 years after this bike. It also had no regulation on the battery. It tried to adjust the charging current by having different numbers of turns for the charging when the lights were on, and there was a ballast resistor to make sure that the coil had the same loading with the headlight on or off.

So it added a battery, a rectifier, a ballast resistor, two more connections to the coil, and a more complicated lighting switch, and it suffered badly from overcharging the battery on a long drive, and the battery going flat in town driving. It did have indicators as well. Actually, for 1983 it was a terrible design and I never understood why they didn't manage to make the electrics work as well as cars of that age worked, which had starter motors.

Your understanding is correct, except that the ignition generator coil isn't a high tension coil. That coil generates spikes of a few hundred volts only when the contacts open, and that few hundred volts is transformed into 10 - 30 kV by the ignition transformer (usually called a coil). That is the high tensions coil, which is not part of the generator, and is close to the spark plug so that the fat spark plug wire, which takes the 10 - 30 kV to the spark plug, and needs to be fat to have lots of insulation, can be reasonably short.

 

[KeepItSimpleStupid]

Your understanding is correct, except that the ignition generator coil isn't a high tension coil. That coil generates spikes of a few hundred volts only when the contacts open, and that few hundred volts is transformed into 10 - 30 kV by the ignition transformer (usually called a coil). That is the high tensions coil, which is not part of the generator, and is close to the spark plug so that the fat spark plug wire, which takes the 10 - 30 kV to the spark plug, and needs to be fat to have lots of insulation, can be reasonably short.

One of the schematics used H.T. (High Tension) and L.T. (Low Tension) to distinguish the windings, so I'm guilty of using that terminology. I don't know what to call them. Main (Ignition) and Aux (Lighting). They are likely different voltages because the coils are drawn with different lengths.

We will agree that "ignition coil" is connected to the spark plug and develops the High voltage for the spark, even though coil is a mis-nomer because it is a transfomer. Everyone calls it an ignition coil.

Main?: reguired to run bike - Labeled H.T. in some drawings.
Aux?: Mostly lighting and Horn. - Labeled L.T. in some drawings

 

[KeepItSimpleStupid]

What's the purpose of putting the brake light as a part of that ground? Seems like inviting failure.

Reverse engineering suggests:
1) Limit the energy stored in the cap at lower RPM's
2) Lower the temperature of the H.T. (Main winding) when stopped (i.e. Low RPM)
3) Add dynamic braking. Your adding (1/2 the load of the AUX (lighting) winding (15 W)) to the IGNITION winding when you brake. The brake light is 15 W - It cannot be supported on the normal lighting winding.
4) Lower costs: e.g. One 45 W winding and one say 3 W winding will cost more than a 30W+18W windings. Plus other stuff.
5) Stupid - Costs over Safety (What's new?)

Yes, it is inviting failure. A shunt regulator there would fix that.

Running right
You need to start with a 6V 15W bulb for the brake light. With a bulb failure, you probably have some points damage.

 

[KeepItSimpleStupid]

Kiss, my first goal is optimize the current 6v AC system. To get it to run well. After it's running right, then the modifications/improvements come.

I'm holding off on LEDs because it's 6v AC, which are not LED friendly. In theory in LEDs are optimum, but real world motorcycle tests show it's hard for LED to beat an incandescent, and that's on a 12v DC system. See here: http://www.webbikeworld.com/lights/led-brake-lights-for-motorcycles/index.htm

Switching to an LED headlight and taillight package would be great, but that must be a future project.

That was an interesting article. So, far my experience with LED replacements are:
1) Dome lights in automobiles (Nice white light - not yellow, seems brighter)
2) Map lights in automobiles (Nice white light - not yellow, seems brighter) Runs cooler.
3) Feet lights in automobiles.
4) And the Switch 3-way LED bulb from Earth LED. It's just a fantastic bulb. Rugged. The CFL's that I had in the application burned out in maybe 3-6 months time.
5) I almost tried an appliance bulb, but I got the wrong envelope. The bulb was so fragile, it did not survive shipping.

I think I would agree, you have a parabolic reflector and light coming from 180 and 90 degrees. Not exactly efficient and you end up focusing at nearly a point source. So, "rolling your own" is an option where many bulbs are used on a flat plane. 6 VAC LEDS seems to made from unobtainium, but that doesn't mean your doomed. You can modify or "roll your own",

I also agree, that it necessary to have a properly functioning bike before undertaking modifications. Eliminating the "dead bike syndrome" should be priority two.

With Alec's regulator, you should not have to rob Peter to pay Paul although you may have to design your own PCB.

==

There are some proprietary PCB houses that are more expensive, but provide relatively easy to use layout programs with instant pricing. e.g. http://www.pad2pad.com/?gclid=CLTVz8G6p8ECFYMF7Aodu3oALg The "standard" uses Gerber files and a "RX271 or whatever drill file". Gerbers were at one time created with "apertures" and photo-plotted before there was the ability to plot directly to film. The proprietary houses can enforce the "design rules" of the house because your using their proprietary software. some proprietary houses may provide Gerbers for a fee.

Cadsoft EAGLE (free version) is a popular package, This http://imall.iteadstudio.com/open-pcb/pcb-prototyping.html is probably the cheapest proto service.

 

[alec_t]

Alec, your three diagrams are 1) the system as it is now, 2) system w a regulator, 3) regulator design. Am I right?

Not quite. The left part (1) is a Kettering system with the points in series with the primary of the HT ignition coil (transformer). The middle part (2) is the system you presently have in the 88ss diagram, with the points shorting out (i.e. in parallel with) the magneto coil. The right-most part (3) is indeed the regulator, and although it's shown separately it is an active part of the simulation by virtue of the connection marked 'reg'. The waveform shows the regulation effect at high revs.

Edit: If you wanted to try the suggested regulator a PCB as KISS mentioned would be ideal; but for proof-of-concept I'd build a prototype on Veroboard instead.

 

[KeepItSimpleStupid]

Alec:

What's you design equations for the regulator? Suggested diodes?

==

I do agree with the "proof of concept" approach. It just has to show it works.

 

[alec_t]

What's you design equations for the regulator?

Diodes: any general purpose silicon, e.g. 1N4148, 1N400x, ...
Resistors: 1/4 Watt
FETs: as shown, or other automotive type, low Rds(on), each able to handle at least 15W for as long as the brake is 'on', package easily mountable to heatsink, must have body diode which can handle >5A.
Note: Without electrical isolation the FETs would need separate heatsinks and the heatsink(s) could not be bolted directly to the bike chassis.

 

[KeepItSimpleStupid]

alec:

Vout = f(r1,R2,R3) ; What's f. Theory of operation?

Do, you really want a 1n4148? The diode doesn;t carry the full current?

I thought it was something like in one direction, the diode conducts and flows through the voltage-divider like thing and then through the non-conducting MOSFET body diode? But I can;t quite figure out the design equations.

 

[alec_t]

Theory of operation?

In the sim, f is the magneto frequency. For the particular instance shown I chose f=125Hz (corresponding to 7500rpm, modelling one cycle of magneto waveform per flywheel rev). Open circuit output voltage vm of Mag2 is modelled as proportional to f. Wave period tp is 1/f. Points-open phase delay td and open time to are fixed fractions of tp.
Consider a positive half-cycle where V(reg) is >0. Current flows via D3, R3, R4 and the body diode of M2 but is only about 4mA peak (essentially just to charge the gate capacitance of M1). M1 conducts if the gate voltage (set by divider R3,R4) exceeds its turn-on threshold (about 4V in this example) and so bypasses the D3,R3,R4 string. So the vast majority of the current is through M1 drain-source and the M2 body diode.
For a negative half-cycle, current flow is instead via D4,R3,R4,M2 and the body diode of M1.

Well, that's the theory .
In practice we would hope the two FETs have fairly well-matched Vgs(threshold), so that they share the power dissipation evenly. If they are unmatched then each FET could have its own 'R3' to set the respective gate voltage appropriately. A bit of resistor tweaking will be needed so that the desired clamping voltage causes the divider to supply the thresold voltage to the gate.
Edit: Although a trimpot could be used to set up the divider when bench-testing a prototype, fixed resistors should be used if the regulator is bike-mounted, because a trimmer would be unreliable when subjected to vibration.

 

[sign216]

Aargh, I ran the bike last night just to check for blown bulbs, and sure enough, the brake is out again. So...I need to make Alec's regulator or else the bike isn't workable.

Kiss, thanks for asking Alec to flesh out some of the design details. Looking at the latest conversation, I can see I'm at the deep end of the pool here.

I've never designed my own PCB boards. I'll make a working model with a generic circuit board, and explore a custom PCB later on, to shrink the size of the unit. That will be interesting.

What amp rating should the diodes have?

Any other issues or design parameters I should be aware? Remember I'm a novice.

 

[KeepItSimpleStupid]

Aargh, I ran the bike last night just to check for blown bulbs, and sure enough, the brake is out again. So...I need to make Alec's regulator or else the bike isn't workable.

Have you checked the points?

<snip>

What amp rating should the diodes have?

I'd use a 1N4003 which is 200 PRV (Peak Reverse Voltage). I'm pretending it's a automotive regulated alternator, which it isn't.

Any other issues or design parameters I should be aware? Remember I'm a novice.

You could use 20 mA as a "design current" and size the power of the resistors to be > (I^2)*R or (0.02)*(0.02)*R; common sizes are 1/8 W, 1/4W and 1/2 W.

 

[alec_t]

1n4003 would be fine (200V, 1A). Even the ubiquitous 1N4148 (75V, 200mA). Worst case scenario, with brake bulb blown, is that D3 and D4 would see ~7V and 3mA (rms) when braking, and the dissipation in R3 and R4 would be <25mW each. The FETs would each have to dissipate ~ 10W-12W.
I'm thinking braking time is unlikely to be more than 10 secs or so at a time? Or perhaps we should design for descents of Mont Blanc, with brakes on for ages? . Braking time will determine the heatsink size. Intuition (unreliable) tells me we could get away with just a few square inches of heatsink.

 

[KeepItSimpleStupid]

Another comment: heatsinks/mounting

The FET's have package names. e.g. TO-220, so make sure you get the package you expect.

The TO-220 package has a mounting kit available for it. e.g. https://uk.farnell.com/aavid-thermalloy/4880sg/mounting-kit/dp/1313880

The purpose is two-fold: To thermally conduct and electrically insulate the semiconductor from the heat sink A thermal insulator goes against the flat portion and nylon collar insulates the screw hole The washer used is a Belleville washer that applies constant pressure.

The thermal insulators were traditionally mica (clear, brittle), but other materials are used now like Kapton. hese require a THIN layer of heat sink grease. It's usually white and can get all over everywhere. Silpads (a thermal insulator type), look like grey rubber with a pattern font require grease.
The lower the thermal resistance, the better.

Some heatsinks you have to drill. They can be vertical or horizontal etc. Some can hold two or more parts.

 

[sign216]

Kiss, I looked at the points, as you surmise, they are beginning to show pitting. This is disappointing, since the engine has less than 4 hours of use since I freshened the points a while ago.
Thanks for the advice on mounting the FETs. This is good information that I was unaware of.

Alec, braking time could easily be 30-60 seconds for many applications. It's common when stopped to keep your foot on the rear brake, so that the brake light is illuminated for approaching traffic. How does this change the specifications?

I wasn't planning on descending Mont Blanc, but is the Isle of Man Tourist Trophy out of the question?

 

[alec_t]

Re packages and heatsinking, I've just realised the FETs used in the sim are surface-mount types and have D-Pak (TO-252) packages, which would be hard to mount and heatsink. A better FET would be the AUIRF3205, which is also Automotive Grade and has a TO-220 package, making it easier to heatsink properly.

Some can hold two or more parts.

.... but would require electrical insulation kits for the two FETs in the proposed regulator.

 

[KeepItSimpleStupid]

Initial debugging

You can use a 12 V DC source and a series resistor to debug. e.g. Set the max current to be 100 mA or R < 12/0.1 and select the wattage appropriately.
You would then check for 6V dropped across the regulator. Reverse the leads and do it again.

Then you could graduate to a 12 V lamp in series and a 12 v bulb where the 6V bulb goes. If the device were shorted, only the series connected bulb would be at full brightness and reverse. Otherwise, the lamps should be about the same brightness. If you had to, you can solder to n 1157 bulb.

You might include a fuse in the final design. Say a 3 A mini-automotive fuse.

==

Another aspect might be to use a connector on the bike. The GM weatherpak might be nice, but they are probably too big. You could try to select a case. With the proper metal case, you may not have to use a separate heatsink. I know sockets are available for the TO-220 package too. I'm not advocating them, but suppose the PCB plugged into an

Probably too big, but ideas:

https://www.aliexpress.com/item/201...ges-and-4-screws-180-140-55mm/2004037959.html

**broken link removed**

So, some ideas might be to put a fairly large 1/4" thick Aluminum plate on the LID/bottom or whatever using heat sink grease and 4 screws.
Tap the plate for the FET mounting and you effectively have a blind 1/4" threaded hole to use to mount the FETs. The holes to mount the Aluminum plate can easily be sealed. Rivets are another possibility.

Just ideas.

 

[KeepItSimpleStupid]

Points pitting: Bad condenser or too close.