Slow reacting indicators?

[jono035]

LOL--I guess you're seeing where Im going with this then dude--the other thing i'd be interested to see is if you connect up a car battery to the bike battery if things change

Yep Haven't noticed the headlight dipping when the indicators light and the battery seems to hold up strong under the headlights/cranking. Cranked happily for probably a minute or 2 without slowing when I first had the carbs to bits and was trying to start it with the hand-across-the-air-filter choke method...

[jono035]

Turns out it was replaced, but with a crap bi-metal one, not an electronic one. Even shorted it takes a couple of seconds to actuate from cold, but then stays blinking at a normal rate. Uses an external heating element in parallel with the contacts rather than one in series...

Got a new electronic one for $25 from an auto-electricians while I was grabbing some other stuff for my car. Will see what happens...

Figured out a simple bootstrap circuit for using a microcontroller and MOSFET in place of a 2-terminal flasher.... Tempted to make it now, just out of spite...

[CookMySock]

Turns out it was replaced, but with a crap bi-metal one, not an electronic one. Even shorted it takes a couple of seconds to actuate from cold, but then stays blinking at a normal rate.

The bi-metal ones do a useful turn mostly. Only thing is they are load and voltage sensitive, they have odd failure modes and they fail quite regularly.

Figured out a simple bootstrap circuit for using a microcontroller and MOSFET in place of a 2-terminal flasher.... Tempted to make it now, just out of spite...

circuit diagram?

Steve

[jono035]

The bi-metal ones do a useful turn mostly. Only thing is they are load and voltage sensitive, they have odd failure modes and they fail quite regularly.

circuit diagram?

Steve

Will throw one together tomorrow in Protel if you're interested. If you make the 2 terminals of the flasher your +v and ground, then your circuit is a diode forward conducting through to a 5v regulator which has a 100uF cap to ground. Running an AVR (one of the 8 pin tiny series would do it) off that supply at 1MHz off the internal oscillator will consume less than 1mA at 5V and will operate down to 1.8V so the micro will keep running for a good 4-5s off the charged capacitor depending on your code.

The micro can directly drive a logic-level N channel FET (found a couple of 5A capable ones for <$1 from digikey) to short the 2 terminals for the 'lamps on' portion of the cycle, then turn the FET off again for 'lamps off'... As long as you turn the lamps off for however long it takes to charge the cap (probably a ms or 2) every 5 seconds (depending on cap size) then you can keep the micro running and in control of the FET... If the voltage on the cap folds too far, then the voltage on the FET gate will go too low, it will turn off, allowing the bootstrap capacitor to recharge and the micro to run again.

With a bridge rectifier instead of the initial diode and a more careful selection of FET (JFET possibly) or a relay (latching possibly?) as the driving element then you could make the circuit completely non-polarised. There are tons of options for decreasing the current draw of those microcontrollers too.

Nothing new, it's just a bootstrap power supply from a high-side MOSFET driver really... I've pulled the can off the old one and might make up a circuit to fit inside just for a laugh... Could do a timer-based auto-turn off, I've got an accelerometer spare somewhere that I could fit too, turn the indicators off after a slow-down followed by a speed-up...

The other way to go is to do a minor re-wire and be able to drive the indicators independently, would give you the ability to flash them as hazard indicators...

[jono035]

Just threw this together, should give the idea of roughly what I was thinking.

Could also add a current sense resistor and amplifier to measure the load (for failed-bulb indication)

[CookMySock]

Good fun. Yeah it would be interesting to wire the indicators individually, especially if they were high-power chip LEDs. They would have to "flash" at the required rate for WOF, but it should be possible to pulse-modulate them much faster so they still looked "on" but had a real high-energy output. If you could "talk" to the controller with the indicator switch and select some options like a high-viz, or fog mode, that would be interesting. Also maybe it could automatically detect a day or night and adjust brightness accordingly.

Steve

[jono035]

The device that I'm working on at the moment is a switch-mode, dimmable, high-brightness LED controller for the HB leds in the 1W to >5W range... You can wire the switch directly up to the controller easily, ambient lighting sensing is easy enough to add as well... A button for a High-Viz or fog mode could be a good choice and would be a simple addition.

In terms of pulse-modulating them to have higher output, the human eye naturally integrates whatever light it sees anyway, so a light source that puts out 2 lumens 50% of the time will appear exactly as bright as a light source that puts out 1 lumen 100% of the time (assuming the flashing is fast enough that you don't notice it). Most of the LEDs can be over-driven safely providing you have adequate heatsinking, this could allow for a temporary 'boost' effect if required for situations where you want to be seen...

Wiring the indicators individually would be a major bonus, especially with LEDs as for high-bright LEDs you can't run them in parallel.

[CookMySock]

In terms of pulse-modulating them to have higher output, the human eye naturally integrates whatever light it sees anyway, so a light source that puts out 2 lumens 50% of the time will appear exactly as bright as a light source that puts out 1 lumen 100% of the time (assuming the flashing is fast enough that you don't notice it). Most of the LEDs can be over-driven safely providing you have adequate heatsinking, this could allow for a temporary 'boost' effect if required for situations where you want to be seen...

I read somewhere you could pulse them hard for a short duration and end up with more light.

You could also pulse them VERY brightly in a really irritating fashion, so it was annoyingly fucking clear to cars that you were there. Not a nice thing to do, but hey everyone will see you.

Steve

[jono035]

Most high-bright LEDs can be run higher than their rated current. The reason for pulsing them would be so that they didn't overheat, so would be on the order of LED on really bright for a few seconds then off for a few seconds. They do a similar thing for using LEDs as camera flashes, run them at absolute max power for a very short time with no heatsinking, not giving them a chance to heat up.

With enough heatsinking capacity you can still run the LED at this higher-power indefinitely though. Would probably be perfectly fine running a HB LED at absolute max power for indicator use though...

The efficiency suffers a bit though, an extra 50% input power for an extra 20% light output for instance...

[bimotabob]

A good bike shop (and probably some not so good ones) will stock the common aftermarket relay/capacitor type for around $16.
This type is small and will do the more common 21-23 watt incandescent bulbs.
I can't remember the brand but they are worth it, often people go and buy genuine for plenty more when they don't have to.
Used to see them all the time.
They even had a wiring diagram on the box for those electrically challenged.

[jono035]

A good bike shop (and probably some not so good ones) will stock the common aftermarket relay/capacitor type for around $16.
This type is small and will do the more common 21-23 watt incandescent bulbs.
I can't remember the brand but they are worth it, often people go and buy genuine for plenty more when they don't have to.
Used to see them all the time.
They even had a wiring diagram on the box for those electrically challenged.

Yeah, thats what I ended up getting while at the local auto-electricians getting some other stuff looked at. Was a Bosch one, works well.