New LED Head Light - casuing issues

I have a 1982 XJ750J Maxim (the one with the LCD dashboard / LCD fuel gauge).
I changed out the standard Halogen head lamp bulb with an LED bulb. I like the new light, nice and bright - however, there must not be enough draw for the Headlight Warning light on the dashboard to turn off - the system thinks my headlight is burned out. I read through a lot of the electrical, switch, relay, diode information on this website. I see other models have a headlamp relay (with a yellow tab) and the instructions to bypass (jumper wires) that relay are in the readings... however, my model does not have that relay - I am wondering if there is still a way to bypass the head light warning light on the dash.
Once this is solved, I will move onto the next electrical issue - provided this issue is not the cause of both.

 

Remove the light bulb (in the dash) under that particular display.............

 

It is not a light bulb for the warning indicator - it's a section of the LCD screen

 

Install a load resistor inline of the power leads of the headlight.

 

This. Unless someone else knows a method for bypassing the circuit.

 

thanks everyone... I am willing to try a load resistor if you have more details, like what size.

 

this might be you, from the xjcd

What happens in the computer? The current goes onto the main circuit
board where each wire (high and low beam) goes to a coil which is
wrapped around a reed switch. When current is running through one of
the coils, a magnetic field is induced which triggers the reed
switch. The reed switch is connected to the computer, and allows the
computer to know when a headlight is burnt out. If a headlight is
out, no current will be in the coils and the reed switch won't
trigger--the computer sets a warning "HEAD" and a flashing light.

now how to fix it, get a magnet, find those reed switches and close them with your magnet.

 

I am not aware of any way to bypass the headlight warning circuit. It's all about current, and when you switch to a lower wattage device that draws less current the CMS will sense this and set a fail. It doesn't matter if you have a headlight relay or a set of contacts in the right control to apply power to the headlight.

I would have also thought that the headlight circuit worked similar to the tail and brake sensing circuits, which utilizes low value current sense resistors (not reed switches as described in the XJ CD) to convert the current to an expected voltage. SQLGuy did a fairly detailed explanation of this for the tail and brake light in this post:

http://www.xjbikes.com/forums/threads/how-to-convert-to-leds-and-still-have-everything-work.14068/

If the headlight circuit does incorporate the same design, then the same modification could be made using the diodes in place of the headlight sensing resistors in the CMS. Since the headlight is very similar in wattage (55/60) to the brake light (54), it is highly likely the modification could be done but I would consider it quite ambitious.

The size of the load resistor will depend on the wattage of the new headlight. You will want to size the resistor to return the system to the power of the original headlight - so new headlight wattage plus load resistor equals original wattage. Unfortunately, this somewhat ugly modification X2 for low and high beam will also increase the line drop and reduce the voltage to your new headlight causing it to dim slightly.

Edit: Power requirements for load resistors and headlight reed relay - see next post

 

Polock for the win...........

 

Not sure if the OP is still looking, but what bulb did you use?

OK, so this sparked my curiosity a bit and since I had an older Atari assembly I decided to take a look. It's a fairly tedious process to disassemble the Atari to this level.

I also tested the reed relays to see what the current requirement would be to detect a headlight fail condition. The reed relays appear to be the same device for high and low beam, but at this level it is impossible to see any markings to be sure. They contain a set of normally open contacts.

The low beam reed relay closed the contacts at 1.1 amps, the high beam reed relay closed the contacts at .9 amps. The fact that the high beam current was lower than the low beam current suggests they are the same part and part tolerance is a factor. This is well below the headlight current for the 55/60 watt setup, so it appears Yamaha wasn't concerned with a dim headlight excepting they did use the reed relay vs. a resistor to limit line loss.

So, if load resistors were to be used, keeping in mind this much lower current requirement to set a safe condition, then the load resistor could be reduced significantly compared to the load of the light.

I don't have much knowledge on LED bulbs, but the quick look I did at superbrightleds.com the recommended LED bulb would consume enough power that a load resistor would not be needed. However, in their specification they don't identify the terminals, and the recommendation was for a single brightness. Since the CMS checks the filaments looking for a high and low, perhaps selecting a different LED bulb would eliminate the need for load resistors.

 

Thanks Rooseter53 -
Here are the details of the bulb I am using:

30W 2000LM H4 BA20D H6M
LED COB Motorcycle/Bike Hi/Lo
Xenon White Headlight Lamp (Color: White)
Material: Aviation aluminum
Built-in radiator
Voltage: DC 12-85V
2000LM,and 360 lighting, more bright,30% improved.
Color Temperature: 6000K, White light

However - that data is from the website [https://www.wish.com/c/5629b1a2f16f26204bbda8c3]
I will try to put an amp meter on the bulb in the next few days.

Just a thought. Do you think it's possible (after digging deep into the assy) to by-pass the reed relays with a jumper?

What's funny is; the draw must be just barley under the limit because the LCD "HEAD" indicator stays on, but there is no red warning light that I need to clear.
So it is really just a minor inconvenience at this point - but I still would like to resolve.
thanks for all your help so far - I'll post what I discover about the amp draw on the bulb

 

Great pics! Very interesting.

 

do you have a photo or link to the bulb

 

I wonder if it is something to do with the basic function of a diode. Since it is just a light emitting version after all.
hmmm hmmm...

 

That is possible, but hopefully it can me made to retain the functional warning as headlights are important, even in the daytime. If a much smaller load resistor can be used than I originally speculated on, then that may be a good way to keep the functionality of the CMS and not ugly things up with massive heat producing load resistors.

This link did not work. Can you provide a link so we can look at the exact bulb that you have?

I will have to try and test the internal logic of the headlight out and warning light (unless someone already knows this). Per the FSM, any failure resulting in a fault should activate the flashing warning light. However, it could require that neither filament be working to set the flashing light, whereas only one filament out that is checked at power up would just give the LCD warning. This will take a bit of creativity so possibly a couple of days or more.

Edit: no special check at power up. The bulb current and filament are monitored continuously and will set the head warning and flashing warning light if either condition occurs.

That would be some good information and along with the LED specifications could add some insight to what might be occurring.

 

Try this - at wish.com - type in the search for "30W 2000LM H4 BA20D H6M LED COB Motorcycle/Bike Hi/Lo Xenon White Headlight Lamp " - that is the exact bulb I have. Again, I'll through a meter on the bulb tonight and come back with some data on amps.

 

I think I had found it earlier on Ebay - this looks like the same one. With wish.com you have to create an account.

http://www.ebay.com/itm/30W-H4-BA20...ash=item43e760072a:g:fMYAAOSw~bFWJf-m&vxp=mtr

The wiring instructions seem to contradict how to wire it up when comparing the annotated lines to the label on the white wire. So do you have a functioning high and low beam?




Edit: Not the OP's bulb

 

Could be, particularly if there is an initial power up test the LED may be too slow compared to the incandescent in the initial current draw. I need to verify or have someone confirm this sequence and logic.

Edit: The LED bulbs contain a constant current source. The voltage used for filament check in the CMS is sourced from and internal 5V regulator. Consequently, the filament check has no hopes in working with the LED bulb.

 

Not broken link:

https://www.wish.com/m/c/5629b1a2f16f26204bbda8c3

It's reading the trailing ] as part of the link in the post above.

 

Guys - I really appreciate the help so far. Here is what I discovered this evening:

Yes, the high and low beams both work on the new LED as designed. (except for the LCD warning indicator)
Here is a photo of the light

Now some more details.
The bulb is marked High Beam 18W, Low beam 12W
At 12 volts that means I am drawing
low beam = 1amp
high beam = 1.5amp

I placed a load resistor in-parallel with the low beam wire.
It's a 50W 25 ohm load resistor.
However, this did not rectify the LCD warning indicator "HEAD" on the dash.
I can try another load resistor tomorrow - next up is a 50W 3 ohm resistor.
I might be chasing the wrong things here, but at least I am discovering what won't fix it

 

in parallel with the low beam wire? that divides resistance
try putting it in series with the ground wire or low beam wire this will add resistance

also did you mean 2.5 ohm not 25 ohm?



http://www.calculator.net/resistor-calculator.html


any resistor you put inline diminishes the savings of going to led .
just like using a load resistor in led turn signals

stock bulb specs

this link caculates watts of the resistor
http://www.sengpielaudio.com/calculator-ohmslaw.htm

3 ways to calculate to see how many amps(watts) are used by resistor

2.5 ohms uses 4.8 amps at 12 volts =57 watts

3 ohms uses 4 amps at 12 volts= 48 watts
w/v=amps

I think you need about 5 ohms to equal oem bulb 2.4 amps would add 28.8 watts

5.5 ohms would be 2.18 amps would add 26.1 watts

remember you are running 14 volts and there is voltage drop across the resistor

 

Should be pretty easy to just jumper the circuit and disable the headlight indicator.

A suitable load resistor in series with the light should work, but will completely negate any power savings leaving "it looks cool" as a reason to run LED, which is certainly still valid.

I'd argue the best route is NOT using a load resistor, to be honest. That wattage could be used to maintain your battery better, run heated grips, or any number of things.

But, I respect wanting to upgrade, and keep the CMS functionality.

An everybody wins alternative would be jumping the existing circuit, but adding your own. Add a relay after the headlight, use that signal to the CMS? Headlight burns out, no signal from the relay.

 

his best bet would be Polocks suggestion of the magnets

 

That's pushing the limits of the wiring when combined with the bulb wattage, but probably OK

The load resistors need to be installed in parallel in order to return the current back to the original spec. If you install the resistor in series you will exacerbate the situation you are trying to correct.

 

for some reason the numbers here don't work, first we have

then there's this

so the high beam should have worked, the low may or may not have worked. is it possible to hook up a regular headlight, just to prove the CMS ?

 

at 18 watts and 12 watts with 14 volts when running

you get .7 amps and .8 amps

 

Both the high beam and low beam are continuously monitored so there will need to be a load resistor on both the high and low - still unsure of the value on that. With the regular incandescent bulb the circuit for the non illuminated side relies on the low resistance of the filament to effectively ground that side. That way when you are running along with the low beam on and if the high beam filament should break, then a "HEAD" warning would be set even though the high beam is not in use.

Still a work in progress but some thoughts:

The illuminated side of the LED bulb draws sufficient current so the CMS should not flag the "HEAD" warning - the 1 amp low side is marginal though. However, as mentioned above both sides are continuously monitored. So on the non illuminated side the LED bulb presents a problem as insufficient voltage is present (4V) provided by the CMS.

Shorting the reed switch or closing them with a magnet would not solve the problem as the logic appears to be an OR function:

Reed switch closed on illuminated side OR filament present on non illuminated side.

Edit: See below shorting reed switches appears to work but disables CMS monitoring of the light - explanation to follow as to why.

 

Electrical tape. Cover that part of the LCD. Done.

 

Electrical tape. Cover that part of the LCD. Done.

That was my first suggestion to him. Tape
I have an idea we are going to try in the next few days to fool the current sensing relays in the LCD panel. I'll let everyone know if it works.

 

But then you lose one of the nice safety features of the bike letting you know if the lamp is out. I don't think anyone would argue that illuminating the headlight in the day time will increase the riders visibility and reduce the risk of someone pulling out in front of you.

Sure, it's an easy check to stick your hand in front of the headlight, just as it is easy to say I will always remember that the side stand safety feature I disabled is not a problem as I will always raise the side stand before taking off. Murphy's law - why risk it?

 

+1
Get it to work. Don't disable it.

 

So a bit further testing on this. You can actually go quite high on the load resistors to bypass the non illuminated filament check. Values as high as 1K will work, which is less than a quarter watt of power consumption. A load resistor is required on both the high and low side connected to return. However, this does negate part of the safety feature in that the non illuminated side will not be under constant monitoring. Given the electrical characteristics of the LED bulb I presently don't see a way around this.

Note: I have three of these assemblies, and one is unique in that the low beam floating voltage is 12V as compared to 4V on the other two. This assembly, which I speculate has an issue requires a much lower value load resistor to bypass the non illuminated filament check. The floating voltage can be checked by disconnecting the headlight plug and measuring the high and low terminals to return - measure low with headlight set to high beam and high with headlight set to low beam.

As for the high current side (reed switches) as long as the LED bulb chosen is of adequate wattage (approx. 1 amp minimum) then the CMS monitoring of the lit element should still function.

I could never duplicate this. All situations that produced a "HEAD" warning resulted in a red flashing warning light.

 

you could unplug high or low on led light to simulate broken filiment

 

how about some running lights to draw a little more power

 

+1
that would give you more light for your wattage use than just burning it with a resistor.

 

I would put down money that the warning lamp bulb is burned out; that's why it's not flashing.

 

Okay - here is what I settled on.
I think MLew has higher expectations of my Maxim than I do. I was willing to live with the "HEAD" warning on the dashboard, but he was not. (but because of his encouragement, I have such a kick-ass bike - so thank you MLew)

Looking at all the options, and playing around with load resistors, we settled on soldering in two jumper wires on the reed switch leads that protrude from the bottom of PCBA. Bypassing the reed switches. See photos below.



He lightly removed the epoxy coating on the ends, and then soldered two wires, one for the low beam (tan colored wire), and one for the high beam (red wire).
He then used a little hot glue to secure the wires to the board.



It was a tight fit putting the board back into the case, but we squeezed it in.
What I like about the solution we selected is we made the jumpers about 10 inches long so I can route them into the headlight bucket. If I ever want to return to a standard bulb, I can always easily access the jumpers in the bucket, cut them to break the jumper, and re-instate the warning indicator as designed.

Thanks everyone for your help on this - I suspect this fix will last for a long time, but I will update again at the end of the 2017 riding season.

 

You are now required to do a full write up of the process and put it in the DYI section.

 

Epon as we always called it. Yamaha did not want anyone repairing these as the component side is also covered in the stuff. It is also perhaps a vibration deterrent for the components, but would be overkill for that purpose. Heat is the only way to remove it - like with a hot soldering iron to isolate to the point of interest.

I was about to post about this again and provide an explanation as to how the CMS is monitoring the lights and what could be done to keep the functionality - at least the working beam.

Some more observations:

The CCA from your CMS appears to be different. Perhaps just a different revision, but I have looked at two and they were the same. I have the Seca, but thought it would be the same as the Maxim? The Seca P/N for the CCA is 5G2-8358E-00. You can look at this pic and see the difference in the etch on the backside of the CCA and see that it is different:



If you happen to have a pic of the top to the CCA I would like to see it.

The reed relay coil is high current, but the actual switches that were bypassed with the jumpers are handling low current TTL signals. Therefore, if disabling the CMS in this manner, then a smaller gauge wire could be used to make the fit easier as well as using only three wires as one side of the reed relay switches share a common between the two on the CCA.

Glad you got it working and are happy with it. Did you consider doing the mod for the brake and tail LED? I have only seen one other person on the forum do this other than SQLGuy, and although it is advertised to keep original functionality, my current speculation is it would not function as originally intended.

 

So here is a circuit cutout of how the Headlight warning is monitored.

Q1 must be turned off to go high and indicate a safe condition. The pull-up for Q1 is provided by the Control Chip on the main CCA.

Ground through high and low filament would apply to UK bikes that have the headlight switch, but active monitoring would still occur when the switch is placed in the off position. Ground through one filament and one reed would apply if the headlight switch is set to ON.

Ground through one filament and one reed would apply to versions where the headlight is always on.

The filament check should not take much of a load resistor to defeat, values as high a 1K should be adequate.

So, it is important to note it always takes two to create a safe condition. The LED bulb electrical characteristics will defeat the filament check and require a load resistor (1K or so) for the non-illuminated side. Subsequently, the LED bulb must also draw enough current on the illuminated side to activate the reed, which to my testing to date is just above 1 amp.

Ground though two reeds would also work, but not a normal condition and jumping the reeds will disable the CMS headlight monitoring for all conditions.



Edit: Added more details to drawing

 

speaking of headlights, could these be any good at this price?
https://youngstown.craigslist.org/mpo/6018169928.html
their close enough, i might have to check them out

 

seems interesting............

 

Here are a few things about this project;

The circuit board components are epoxied in place. Any attempt to remove a component will destroy it. I originally wanted to make a new coil for the reed switch to get it to close at a lower current. They are single wound coils and replacing with a 3 layer coil should have reduced the trip current. It is possible but the reed switches are glass and will break if removed. He did not want to go through the time and expense. I found the reed switched for just a couple bucks but the time to do it was not there.

Adding load to the headlight with resistors will work but negates the current savings of the LED.
Adding load to the headlight with running lights was also discussed but the current draw of the LED headlight and LED running lights was not enough to trip the sensors.
We decided to add the jumpers as a quick and easy solution. The wires were brought out of the LCD panel and routed to the headlight bucket so if needed in the future they can be cut and normal function of the LCD panel restored. The wires were labeled and noted as such during reassembly.

 

But those won't fit into the square hole.

It can be done, but with great care using a temp controlled heat gun or a soldering iron to soften the epoxy for removal. Not sure it is worth the effort though and all attempts will not be successful.

Did you try to do any measurements to see at what current value the reed tripped? I know the two I have tested will trip at just above 1 amp. It the CCA you had was the same, then the high beam wattage of the OP's bulb (18/12 = 1.5 amps) should have worked. It could not be done as a standalone test though since both nodes must be taken low - bike switched to high beam and a load resistor (value debatable but quite high) on the low beam side. The point of the load resistor in this case on the low beam side is not to trip the reed relay but to simply sink enough current to fool the filament check.

Using an LED running light would add very little current (approx. .1 amps) so it would do little to obtaining enough current for setting the reed switch.

 

The magnetic reed switches on the board are made from thin glass . Any attempt to remove epoxy from them will break them. I have tried in the past on other projects.

No we did not actually measure the current needed to trip . I don't have a low resistance high wattage pot to do that. We did estimate the needed current and it was just above 2A but under 2.5 for both high and low. The LED bulb was 1.4A adding a load of 25ohm(.56A) did not trip, but adding a 12.5ohm( 1.2A) did.

 

Post expired

 

So by adding a magnet on top of Reed swith might work ? So I am assuming I'll have 2 disassemble the control head and figure out which ones do the job hey anybody got any pictures

 

You do have to dissemble the unit and remove the circuit board and yes a magnet will work but keeping it in place is not easy. I found that the magnet has to be almost touching the switch to activate it. The circuit board has to come out of the enclosure for the jumper or magnet so it's easier just to solder a jumper wire on the switch.
The headlight reed switches are the ones on the far left on the circuit board. There are 2 more in the center , they are for the taillight. Look at the pic posted by Rooster53 on the first page, it shows the switches.

 

Seca CMS, it employs two reed switches for the headlight and current sense resistors for the brake and tail light.



Maxim CMS, it uses four reed switches, two for the headlights and two for the brake and tail light.



I should be able to post more later today, going next door to help a neighbor with electrical issues on his 2003 BIG DOG. Just like a Harley, everything on it vibrates loose, even to the point of breaking the smaller wired connections on the battery.

And, I still believe you can keep the functionality of the CMS excepting the filament check, and no need to take it apart.