Voltage Regulator for a Small Motorbike, 6V AC Current

[alec_t]

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?

It doesn't, unless while stopped you have the revs > ~3000 rpm .

 

[sign216]

Points pitting: Bad condenser or too close.

The pitting is not a result of the broken brake light?

Thank you tremendously for the debugging/construction info.


Fyi, I notice that the 3w parking light is broken, as well as the 0.6w hi-beam indicator. Both of these are on the headlight circuit, and I don't think they're an issue for figuring out the system.

 

[sign216]

Alec, thanks for giving me a user-friendly FET data.

Do they need to be mounted on a heat sink, or thermally insulated from the sink?
And electrically insulated from the sink too, or not?

 

[KeepItSimpleStupid]

Yes, because they are dissipating about 15 Watts (6 V * 2.5 Amps). The case can become a heatsink as I suggested. The size of the heatsink can be calculated using thermal resistances, Max ambient temperature and max junction temperature.

You NEVER thermally insulate a semiconductor device unless it's used as heater. It destroys the term Heat SINK or a place for heat to go.

Alec covered the electrical insulating and he suggested it, The case bottom is one of the tabs and it's not at ground potential. Electrical connections to some devices are made using the mounting screws.

Here http://gardentractorpullingtips.com/ignition.htm , you can see some ideas that were used at the bottom 2/3 of the page for TO-3 transistors. They typically could not easily touch ground. I'm suggesting mounting the FETS to the underside of the inside of an aluminum case after making it thicker for the final version.

In terms of a PCB, you probably don;t need one. Solder directly to the leads. Use heat shrink and mounting becomes the rigid base.

The pitting is not a result of the broken brake light?

Don't know. "the OP" is short for "the Original Poster"

 

[sign216]

Kiss, thanks for that. I'll follow your suggestion, and forget the PCB. Try mounting it on the metal base or cover of the enclosure.

 

[Diver300]

The pitting is not a result of the broken brake light?

I wouldn't think so. The points operate every revolution of the engine, so up to 150 times a second. The brake light switch operates every few seconds a most, and as soon as you realise that the brake light bulb has failed, you'll either stop or use just the front brake. When the brake is not being operated, the broken bulb has no effect. When the brake is operated with a blown bulb, no current flows in the coil, so there can be no pitting of the contacts. The only possible time that the brake light bulb could have any effect is when the brake is first operated or released, so just a few times a minute. It just doesn't happen enough to cause pitting.

I think that pitting is due to the condenser being faulty, as KeepItSimpleStupid says.

 

[sign216]

Damn, I was hoping to kill two birds with one stone by fixing the brake light circuit.

Actually, I'm really disappointed to see the beginning pitting already. There's probably less than 3 total hours of running time on those points.

I recall reading that points erosion shouldn't be an issue on magneto ignition engines, because the points carry less current than battery+coil engines. Is this a false assumption?

I bought one or two of the Velleman electronic ignition kits, but when I read that point pitting should be a non-issue for magneto engines, I put the kits away. All my "project engines" are magneto, so I thought the kits would be irrelevant.

 

[KeepItSimpleStupid]

This https://www.aviationpros.com/article/10383985/magnetos-installation-tips attributes severe point potting to a bad capacitor and supplies reasons.

Not too many years ago, I converted a lawn mower to electronic ignition. That system eliminated the points and just fired based on the magneto coil. It's the Mega-fire system in the link in post #104. It's also known as the Atom. Stens sells it too. That's where I got mine from. There are a few versions of the ignition module. It definitely improves starting.

The Velleman ignition isn't applicable for 6 VAC. We MIGHT be able to get it to work though - who knows.

==

When I briefly looked for a case, I looked for an IP rating of 65 or IP 67. See: https://en.wikipedia.org/wiki/IP_Code

I hope you generally understand what I'm getting at with respect to construction. An IP 65 or IP 67 case should be sealed, but I don;t expect the case to be very thick. You could end up with an extruded case where only the end caps are sealed.

So, if you put sealed first, you can build using that. The sealing doesn't have to survive 100' in the ocean, but it does have to survive water and salt spray. A thick plate with good thermal contact COULD be a good start. Rivets and thermal grease would provide that thermal contact.

But I'm starting backwards. In the forward direction:
1) Cut a 1/4 inch thick piece of aluminum to fit the inside cover. (assuming two pieces)
2) Make 4 mounting holes for rivets.
3) Drill and tap holes for the TO-220 mounting hardware.
4) Get or cut screws, so they would fasten the FETS, but will not protrude.
5) You could use these https://uk.farnell.com/concord-electronics/1125-44-0519/terminal-solder-ptfe-4-40/dp/2210324 for some of the resistors. Use countersunk holes on the bottom of the plate.
6) Use clear heat shrink tubing for the connections.

So, you have essentially built a circuit on a flat plate with turret terminals. There are so few components.

 

[KeepItSimpleStupid]

debe did the "dead bug" approach for an ignition module here: https://www.electro-tech-online.com/threads/atom-diy-module.142530/

 

[alec_t]

I was thinking perhaps a diecast alu box like this to house the components, plus bolt on two of these finned heatsinks? This gives about the minimum size/configuration I reckon would be needed for 15W dissipation with prolonged braking at high revs. The box/heatsinks would be exposed to the airflow. Overall size will be about 2" x 2" x 4".
Since the drain of one of the FETs will be grounded, that FET drain-tab won't need electrical isolation from the box/heatsink, thus improving heat conduction and enabling non-isolated chassis mounting of the box: but the other FET drain-tab will need electrical isolation.

 

[alec_t]

BTW, when buying components I'd reckon to buy a couple of spare FETs .....just in case .

 

[sign216]

Thank you for the instructions. I''ve built kits and things before, but I'm still a beginner.

Kiss, I like the Atom module Debe did. It's more compact than the Velleman kit.

I heard that points on a magneto ignition shouldn't wear out because of the low current passing through them. Info from a BMW antique motorcycle page, that said "The current going through the points is so low that the contacts should (almost) never wear out. ... The points on a /2 BMW motorcycle are more reliable than the crankshaft." See the bottom of: **broken link removed**

I'll replace the condenser to see if that fixes the point pitting. The points are behind the flywheel, making access a chore, and gives me a reason to change to electronic ignition as my next project, once the brake light issue is resolved.

 

[sign216]

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.

The trimpot should replace R4?

Once I figure the optimum value I'll replace in the final model with a fixed resistor.

 

[KeepItSimpleStupid]

Troubleshooting

You can use a 12 V DC supply to troubleshoot. Variable even better. Variable current limit better yet.

Suggestion:

Get a series resistor 12/0.1 ohms or 100 mA. Check power dissipation. Place in series with 12 V supply.

You should check for ~ 6V with the polarity reversed.

Then you can actually use two 12 V #1157 lamps (Brake side). One in place of the resistor above and one where the brake light goes.
You can solder wires to the bulbs. You can even test with the tail light side first.

They should both glow at equal brightness (6 V across each) if the circuit is working. Reverse the DC polarity and it should do the same.

If the outer one glows at full brightness, something is wrong. If they don't glow, something is wrong.

==

Fuse in the final circuit. 3 A mini automotive might work.

==

Alex uses external heat sinks - which could work for a test, but I don't like that construction for the final version.

==

If you decide to try to POT in silicone, you can't use silicone with acetic acid.

==

I forget if I covered this stuff.

 

[sign216]

Kiss, thanks for much needed details.

I need to get a new condenser, and since you guys know much more about electrics than on the motorcycle forums, I'll ask here:
Do I need a 6v condenser, or can I just got to the local autoparts store and get a 12v VW condenser (which is the only one they usually stock).

 

[debe]

With the condencer across the points the value may be any where from .15 to .30 uf & at least 400V. The condencer value is selected by the manufacturer to match the coil. If the value is not right then the points will burn.

 

[KeepItSimpleStupid]

debe is right again, although the value could go as high as 1 uf, I grabbed a very old automobile capacitor (probably) and it measured 0.3 uF

Although I'm not sure of the type and rating, I would actually consider a 400 VAC rating. A rare breed. Lawn mower caps might exist in your locale.

I don't know the sizing criteria. Capacitors add in parallel, In series they follow the 1/Ct=1/C1+1/C2+...1/Cn

 

[sign216]

So.....can I use a 12v condenser as a substitute for a 6v system?

The bike's manufacturer, Gilera, went out of business years ago, and getting the factory specs is a quest like the holy grail.

 

[Diver300]

So.....can I use a 12v condenser as a substitute for a 6v system?

The bike's manufacturer, Gilera, went out of business years ago, and getting the factory specs is a quest like the holy grail.

As the others have been saying, you need a condenser that can stand around 400 volts, so it is irrelevant whether it is for a 6 V system or a 12 V system.

When I had my 1983 Honda which had a similar ignition system, the ignition coil (the one with the HT lead to the spark plug) was labelled "6 V," but there was no point where you could measure anywhere near 6 V, as it was just being used as a pulse transformer. There was no connection, except ground, to anything that ran at 6 V. The battery was 6 Vdc and the lights were 6 Vac ( as long as the correct power bulb was used) so the coil was labelled "6V" just to mean that it could be fitted to a bike that had 6 V systems.

Honda labelled the coil "6 V" because customers, and probably dealers, think that electrical devices have to have a voltage and you must stick to that voltage and the rest of the bike was 6 V. For some devices that is true. If you have a 6 V bulb, it will blow if connected to a 12 V battery, and a 12 V bulb on a 6 V battery will be so dim that it is useless. However for many items, like magneto systems, or current-driven devices like LEDs, then the voltage is not so important.

 

[KeepItSimpleStupid]

Nobody on the web is giving values. Way back 45 years ago, I think the condensers were marked

So, I would start with a lawn mower condenser. It would be neat if you could measure it.

Nigel just bought one of these: https://www.jackssmallengines.com/Products/BRIGGS-STRATTON/Ignition-Part/CONDENSER

For the cost, they are effectively free.

Are you using the original condenser?

Maybe alec can figure out a way of simulating to get a good value for the condenser, if you can get some info: Inductance. Not sure if wire gauge would help, but resistance of the winding would help more. Low resistance measurement will take some effort.

With a 100 mA current source ( a few parts) and a battery and a DVM that has a mV range, you can get a decent number

Alec:
I took the cap I had and did the following measurements at
1kHz: 318 nf D=0.48 deg
R = 24 @ -87 deg
Z = 501 @ -87 deg
70 mH Q= 20.3
ESR 25 @ -87 deg
DCR >120 m ohms

This contrasts to a newis lawn mower cap @ 1 kHz
C= 196 nf D=0.003
R = 3 @ -87.7 deg
Z= 808 @ -87 deg
ESR = 3.9
DCR = OL

ESR or Effective series resistance is the better way of determining a defective capacitor, The 50 year old cap is about 6x higher. The top one has a measurable inductance which isn't good, so I would bin the first one.