1. Hello Guest. You have limited privileges and you can't "SEARCH" the forums. Please "Log In" or "Sign Up" for additional functionality. Click HERE to proceed.

LED Headlight with Atari

Discussion in 'XJ Technical Chat' started by SQLGuy, Aug 3, 2019.

  1. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    I asked about this a while back and was told the Atari would complain about the filament check. I thought I could probably build a circuit to compensate for that, and today I installed it. Seems to work fine, and this LED does fit in the stock Seca 750 headlight bucket. I was not able to install the rubber shroud, though, and instead applied some waterproof aluminum tape around where the lamp fits in to the lens.

    The circuit is a DPDT relay with 4.7 Ohm 3W resistors connected between the normally closed terminals and ground. The common terminals of the relay are connected to the high and low beam 12V lines, and these each also have the anode of a rectifier whose cathodes are connected to one end of the relay's coil. The other end of the relay coil is connected to ground.

    The way it works is that, when the headlight is unpowered, the relay's normally closed contacts are closed, so the Atari reads the 4.7 Ohm resistors as "filaments", and it is happy. When the headlight is powered on, the 12V to either the high or low beam causes the relay to close, disconnecting the resistors. Since the filament check only happens when the headlight is off, this works fine, and the current draw of the LED bulb is then sufficient to keep the Atari happy through the reed relay check once the headlight is on.

    The one down side is, if the relay should fail, on of those two resistors is going to get very hot. It's unlikely, but it would be nicer to have a safer failure mode. The new headlight only draws 12W, so a possible safety solution here may be to replace the headlight fuse with a smaller one... 2.5A or so.

    This is the LED I used: https://www.superbrightleds.com/mor...-with-internal-driver-2000-lumens/5323/12255/

    After taking this shot, I wrapped the relay/resistor thing in electrical tape.
     

    Attached Files:

    k-moe likes this.
  2. raskal

    raskal Active Member

    Messages:
    165
    Likes Received:
    72
    Trophy Points:
    28
    Location:
    Abbotsford B.C. Canada
    Have you ridden at night yet? I'm curious to hear about other drivers of your light. I've read that led lights generally require different reflectors in the housing than halogen, and when not used the led appears bright, but causes massive glare
     
  3. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    Not yet, though I replaced the halogens in my car with LEDs about a year ago and have been very happy with those.
     
  4. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    I'll give it a try tonight. My impression, though, is that H4 and other LED replacements for halogen lamps have come a long way in the last few years. They're no-longer just attaching a bunch of emitters to a cylinder. The new ones have reflectors and shields built into the lamp design.

    And, yeah, in the garage, it's plenty bright. The little fog lamp is an embarrassing dim yellow compared to the bright white of the LED headlamp.
     
  5. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    I am guessing you already tried this? It really shouldn't work because the filament check is done on the non operational beam. With the LED headlight the filament check has to be defeated, which presents just a bit of a safety issue. With the stock setup if I was driving along a dark country road running with the low beam, the CMS would set a warning if the high beam was out (filament open). Consequently the operator would not be suddenly surprised if the low beam were to burn out and a switch to high beam resulted in continuing down a dark road. That said, most of the non CMS bikes don't incorporate the warning, so I guess it could be viewed as not that big of a deal.
     
  6. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    OK, tried it out. Although the beam spread is wider than the halogen, I don't think it's unreasonable. It's still cut off at about the same distance with the low beam on.

    However, there is some problem with the filament check. No complaints for low beam, but when I switch to high beam I get a Head warning, which then goes away when I switch to low beam. That seems odd. I don't understand why I wouldn't have the Head warning regardless. It's acting like it only checks filament for low beam.

    So, yeah, I guess I'll need a pair of SPDT relays and no diodes, so that each terminal has its own filament check resistor... but I may first try just disconnecting from the high beam terminal, since it doesn't seem to be checking that one... or is it checking the high beam only at startup and the low beam anytime the high beam is on?
     
  7. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    The filament check is done on both the high and low beam.

    The diagram below will defeat the filament check and allow the CMS to report a headlight failure if the selected beam is not illuminating. The LED headlight bulb selected must draw at least 1 amp to close the reed switches. The original wiring of the high beam indicator also causes an issue, and must be changed as indicated below as well as staying with the stock incandescent indicator bulb. It's a fairly easy mod on the Maxim and Turbo that already incorporate bullet connectors on the H4 connector. It also requires two extra connectors to re-wire the high beam indicator as illustrated below. All of the mods are done external to the CMS.

    upload_2019-8-4_10-27-8.png
     
    Last edited: Aug 4, 2019
  8. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    I only see IMG instead of an actual photo for what you posted.

    If both filaments are always being checked when not powered, then I don't understand why my bike is only displaying Head when the high beam is on.
     
  9. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    Edited above post and reloaded image, hopefully it is viewable.

    I can only speculate on that since I can't quite make out what you have wired, so I won't go there. Maybe it will make more sense once you view the diagram.
     
  10. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    OK. So, your mods are a 1K 1/2W in parallel with the low beam, and the high beam lamp in parallel with the high beam, and that was sufficient for the filament test?
     
  11. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    Yes, that should do (1/4 watt is fine) but you have to use the standard incandescent bulb in the high beam indicator so that it effectively supplies a shunt to ground to remove base current from Q1.
     
  12. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    I should add also if you really want to keep the LED high beam indicator bulb you can. It will still need to be rewired as above and another 1K resistor can be used on the high beam side to ground.
     
  13. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    I just experimented with mine. It only seems to be checking the low beam filament when the high beam is on. It does not care about high beam filament when low beam is on. I don't know why, but that's what it's doing. 1K across low beam is not enough to make it happy when high beam is on. 5 Ohms is.

    IOW, I start it up with nothing connected but the LED lamp itself. The CMS is happy. I switch to high beam and Head comes on. I add 1K Ohm from LB to ground and it makes no difference. I add 5 Ohm LB to ground and Head goes away.

    Maybe different revisions have different lamp detect circuits? Mine's an '82.

    So, I'm going to revise my relay circuit to switch for only LB load when LB is unpowered. I think I will do that by powering its coil from the HB... so I have a safer failure mode as well.
     
  14. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    Verified. With the 5 Ohm load resistor across LB/Gnd applied by the NO contacts of the relay when the HB is powered on, I no-longer get a Head warning under any conditions.
     
  15. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    And is that with the high beam indicator light wired as above? What happens if you don't re-wire the high beam indicator it takes a much lower value resistor to defeat the filament check. Also if you don't rewire the high beam indicator 5 ohms will have the high beam indicator on all the time.

    If I remember correctly you can verify the internal CMS voltage that is used for the filament check by measuring on the H4 connector with the headlight disconnected. Both the high beam and low beam should be about 4 volts with just the key on and the headlight not illuminated. When a standard headlight is installed, the filaments effectively ground those two points taking the voltage to near zero, this way the CMS works on Euro bikes that can switch power completely off to the headlight.

    I also have an 82 Seca so I doubt there is a revision change. I have been running with the above for several years now.
     
    Last edited: Aug 4, 2019
  16. SQLGuy

    SQLGuy Well-Known Member

    Messages:
    2,140
    Likes Received:
    175
    Trophy Points:
    63
    Location:
    Colorado Springs, Colorado USA
    That doesn't make sense. The only load you added to the low beam side of things is the 1K resistor. The rewiring of the high beam indicator lamp has no effect on the low beam side.

    Are you sure that diagram is an accurate representation of the changes you made?

    And, no, my high beam indicator is not on all the time. The 5 Ohms is in parallel with the low beam "filament" only when the high beam is on, which makes the LED high beam look like an H4 high beam from a resistance perspective.

    I don't really feel like opening it up again to measure socket voltage... but maybe I will. The only other things I can think of that would explain our observations would be if:

    1. My LED presents a lower forward voltage drop for the high beam side, which is enough to satisfy the filament check and

    2. Your LED presents more load at low voltage for the low beam, so that 1K is enough extra load to pull the comparator below its threshold.

    Edit: Just checked. Mine doesn't work like yours. Key on, engine off, lamp disconnected, I get 4.2V on the HB connector, and 0.9V on the LB connector. Flip the high beam switch, and they reverse. So, mine is only checking filament on the non-illuminated side. This still doesn't exactly explain why it's passing the filament check on HB with no load, though...

    On the other other hand... I do have the relay coil as a load. It's 360 Ohms. Maybe that's enough. Except that, even when I had nothing connected to the HB side, it still passed the filament test. So, I'm back to the theory that the HB terminal has a lower forward voltage drop than the LB terminal.
     
    Last edited: Aug 4, 2019
  17. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    Well, it has a huge effect when trying to disable the filament check for the low beam. With the original wiring the high beam indicator bulb is always hot on one side, the other side is connected to the low beam filament. To defeat the low beam filament check the internal CMS BIT voltage must be pulled low, somewhere around 1 volt - I have the precise number somewhere or could re-measure. The high beam incandescent indicator bulb has an effective resistance of around 47 ohms, so in order to pull the low beam filament low the resistor value is in the order of 5 ohms as you suggested, although a bit high as with 5 ohms Vout would equal 1.23 volts. The same thing occurs when trying to use an LED bulb in the high beam indicator light, although not as drastic. I had tried the LED approach and found notes suggesting in the order of 60 ohms. With a 60 ohm resistor the wattage goes up to about 3 watts when the low beam is powered so I gave up on that idea as I wasn't interested in using high watt resistors. Couple that with the variables of LED bulbs and it could make it problematic.

    Yes, I believe it is accurate. The resistor values noted for the CMS are measured values. As you are aware there are no schematics available so those values are bound to be slightly off.

    Yea I had that backwards, so I agree

    Yes, mine works as you stated, and checking the filament on the non illuminated side is part of the original design. The reason for the voltage reversal with no power to the headlight is that the diode in the headlight relay provides a current path to ground through the headlight relay coil - so about 100 ohms to ground. Pull the headlight relay and you will get 4.2 volts on both the high and low beam.

    And I did go ahead and re-verify the failed conditions I am seeing. With the H4 connector removed, headlight power off, the HEAD warning is set. Ground only the low beam terminal on the H4 connector and HEAD warning remains, ground only the high beam terminal on the H4 connector and HEAD warning is present. Ground both the high and low beam terminals on the H4 connector and HEAD warning disappears.

    Headlight connected, disconnected 1K resistor and high beam indicator light – HEAD fail

    Connected 1K resistor HEAD fail when low beam selected

    Disconnected 1K resistor and connected high beam indicator HEAD fail when high beam selected.

    Measured voltages with head light off and 1K resistor and high beam indicator connected

    Low beam = .85o mV, high beam = 80 mV

    No doubt that differences in LED bulbs exist. To test that theory you could measure the voltage on the non operating beam keeping in mind the approximately 1 volt threshold to disable the filament check. Of coarse the high beam indicator plays a role here and would at the very least need to be disconnected to get meaningful measurements specific to the LED bulb assembly.

    This is what I measured with the LED bulb I am using - note resistor and high beam indicator disconnected.

    Low beam selected - high beam terminal = 2.9 volts
    High beam selected - low beam terminal = 4.2 volts

    So both are above the approximate 1 volt threshold, however the high beam filament is supplying a bit of a load. I'll go measure that:

    LB - capacitive starting at 87K up to open
    HB - a solid 3.25K, which calculates to a Vout of 2.9 volts as measured
     
  18. Rooster53

    Rooster53 Well-Known Member Premium Member

    Messages:
    1,852
    Likes Received:
    723
    Trophy Points:
    113
    Location:
    North, FL
    I forgot to add all testing was done with the CMS manually enabled. It can make a difference when selecting resistors as the added voltage from the AC generator will make a bit of a difference.
     
  19. Rodney Brown

    Rodney Brown Member

    Messages:
    73
    Likes Received:
    5
    Trophy Points:
    8
    Location:
    Bronx, n.y.
    Yea I've ridden @ night. Light looks ok 2 me. Cant account for oncoming drivers.
     

Share This Page