Honda Spree and Elite 50 Forums

Here are a couple of flasher or winker relays for turn signals that IÔÇÖm studying. They are called thermal flashers (editorial comment, there is no thermal principle with these, solely magnetic, I may need to rewrite this entire entry) as compared to electronic. Several people have asked on the internet how they work, how to make them flash quicker, and usually the response is to just replace with electronic, donÔÇÖt complain, or just get a new one. There are U-tube videos showing how to adjust speed for electronics having the ÔÇ£555ÔÇØ timer chip.

That spurred my curiosity to draw a schematic and analyze how it worked. The rusty one is from a Honda Flush scooter and the other is from a Honda CL450 years back when it failed. A third is electronic and off a Chinese scooter I found. The latter two worked (I had renewed contacts for the CL450) with 3 to 4 amps of draw, but at 2 amps or less the signal lamp would stay constantly on. Also, if 4 amps was applied directly across the thermal element there was no blinking. How come?

IÔÇÖve concluded there are high surges of current created internally, exceeding 4 amps, that pass through the thermal element when it flashes. That comes through a quick discharge from the capacitor through a coil and thermal contacts in a loop, that is in addition to the 4 amps of lamp load. Sufficiently high amps, thermally it acts and contacts open. When the lamps go out after the contacts open, the capacitor quickly recharges. Contacts cool and the magnetic flux aiding opening dies off. Contacts reclose, lamps go on, and the cycle repeats.

The initial opening of contacts doesnÔÇÖt have the thermal effect as youÔÇÖll note when the flasher is first turned on, the turn signals donÔÇÖt light. The opening of the contact is totally magnetic. The capacitor is also brought up in voltage from zero. When charged the inrush current decays, less magnetic strength, and the held open contacts close. Now the light turns on for the first time and because of the kick by the capacitor makes enough heat to open the thermal contacts . The cycle repeats and the flashing rate increases slightly but then follows a repeatable pattern.

IÔÇÖve completely disassembled the rusty relay. Rust swelled the armature so it wouldnÔÇÖt move. I am about to unwind the coil. It is actually two coils of wire, one very thick on the outside and a fine, inner one. I found the fine winding broken. IÔÇÖll need to drill and tap the steel core and do some other work to even think of reassembly. I would like to see if there is a way to make it flash for a lesser amount of current. The scooterÔÇÖs electrical infrastructure canÔÇÖt handle lots of current (wire is small) so IÔÇÖll see what I can do.

There are slight differences in design of both relays but physically they work on the same principles, at least, so I think. Millions, if not billions of them made with no discussion that IÔÇÖve found. The engineers that designed them probably are probably no longer around.

I disassembled and reassembled the rusty relay.

But the relay works on magnetic principles and the changes in resistance of bulbs from off to glowing. The relay has two coils, one which changes in tensity as contacts open/close. The other reverses its polarity. When they add up the contacts open and turn signals go off. When they subtract as one coil reverses polarity, net magnetism is near zero, contacts close and lights go on.

This cycling event of on-off-on-off cannot be taken for granted. I'm still trying to get the rusty relay to work for 2 bulbs. I'm experimenting with various capacitors and 2 each 1000 microfarad capacitors seem to be best. I'm also adjusting spring pressure by bending the "strong back" to vary contact tension to close when closed.

I'm having difficulty when the contacts reclose. They don't do a smooth transition, contacts making a buzzing sound and they soon burn and lose electrical contact.

A glowing bulb has 6 times the resistance of one that is off. The circuit resistance changes when one of two bulbs is burnt out. I'm trying to work that out, too.

If this is too detailed or you want to know more, I'll take your advice.

I like this

Good work. Keep going.

Mousewheels used to do this kind of analysis. He's an EE.

Attached is a schematic I made of the rusty relay and the working relay with its blue capacitor. I still can't get the rusty one to work. I added more wire/turns to the thicker outer coil L2 and that caused the contacts not to open. I replaced the steel core of the relay with a nail so that the steel core could be closer to the armature, again no success. I can get the contacts to open for a split second, then they bounce and arc for which I can smell some ozone. Lamps have half brilliance as the contacts don't lock closed. They don't fully reclose. I've looked at the voltages on all components of both relays with an oscilloscope which is fast acting and appear to have the same response at the capacitor in charging and discharging.

Greetings:

Nice work David! The lack of resistance change may be why it's so hard to get LEDs to work as turn signals in circuits designed for incandescent bulbs.

Wheelman-111 wrote:The lack of resistance change may be why it's so hard to get LEDs to work as turn signals in circuits designed for incandescent bulbs.

It is. LED flashers work on a different principle entirely... more like a 555 chip blinker circuit that these thermal ones, techology which pre-dates silicone transistors, much less modern ICs.

I finally got the relay to work on my birthday last week. This is how it works.

L1 is an inner coil of very fine wire that has several hundred turns and draws milliamps. Multiply both and you have around 140 ampere-turns. It changes its flux direction in the schematic as shown in a blue trace during operation.

L2 is the outer coil of thicker wire, around 70 turns, higher amps (1.9 amps for this relay for it to work), the product of which is around 140 turns. Its flux direction in schematic is shown in red. It is either fully on or off.

When the relay is first turned on (Case 1) it quickly opens up the relay contacts and moves to Case 2. The inrush current into L1 keeps the contacts open, lamps off, while C is charging. When fully charged there is no more current to flow and the inherent relay spring closes the contacts. Now we're in Case 3. The lamps light and the flux of L1 counters, opposes, the flux of L2 for a very short while. Net flux is thus zero and the spring keeps the contacts closed as the lamps continue to light up. But C is discharging in its own little loop shown in blue. When near zero amps the constant flux of L2 opens the contacts. Jump back to Case 2. Now open, lamps go out, C charges, it is the only "flux in town" and wins over the spring action to keep contacts open. But as it charges up, its flux goes to zero and the spring wins to close the contacts. So you're going back and forth between Cases 2 and 3 as it blinks.

My problem was I was creating too much flux in L2, the outer coil, by running 3 amps or so or adding more coils to it, thinking I needed more ampere-turns. But that was wrong. I needed to cut down the ampere-turns by either reducing current or removing some windings. It was easier to remove load. I have a video of it I can send separately to another cell phone from my cell phone, or perhaps to your PC, if you want.

If one wants to go to LEDs and the current is half that of incandescent bulbs, the ampere-turns are thus half of what they should be. What to do is double the number of turns of L2 with wire about half the cross sectional area. Gets the ampere-turns back where they should be to match that of L1 plus they should fit in the relay without bulging out or hitting something.

Man I love this thread. I haven't been able to read the whole thing word for word yet which I can't wait to do when I have the time.

Photos show the two each 1000 microfarad capacitors attached with jumper wires outside of the relay and with them taped up. The nut on the new nail core of the relay is part of the L (Load) wiring and should also be insulated with tape or the like. The relay without tape has the original blue capacitor, value unknown, and was a helpful reference through this experience.

Included the marking on the relay cover as they are important. It shows a total load of 23.4 watts for which it was designed, to include the front/rear turn signal lights for either side plus any panel indicating lamp.