Thanks for the great write up on the CAN BUS, I'm sure that people will definitely find it helpful.

 

For a concrete example of what happens, let's go through the flow of what happens when you turn the headlights on. Let's assume this is a manual manipulation of the switch, and not the Auto mode.

1. You turn the headlamps on using the multi-function switch stalk on the left-side of the steering wheel. This multi-function switch is hard-wired to an electrical module called the Steering Control Module, or SCM.
2. The SCM senses the switch change, and sends a message to an electrical module called the ElectroMechanical Instrument Cluster, or EMIC. This module is also known as the Cab Compoartment Node, or CCN - either way, this is your instrument cluster. The SCM tells the EMIC about the headlamp switch over something called a LIN data bus, which is a very simple 1-wire serial link bus that runs at a speed of 9.6 Kbps. It is typically used for switches to communicate to an ECU that is monitoring them.
3. Now that the EMIC knows about the headlamp switch being activated, it sends a headlamp and headlamp beam request message to the Total Integrated Power Module, or TIPM, over the CAN interior bus.
4. (This text provided directly from the service manual.)The TIPM responds to these messages by providing a pulse-width modulated voltage output to the headlamps through high side drivers on the right and left low and high beam feed circuits to illuminate the selected headlamp filaments.
5. (This text provided directly from the service manual.)The TIPM also sends the appropriate messages to the EMIC to control the illumination of the high beam indicator.

So as you can see, the digital networking busses in the car (the LIN and CAN busses) are only used to communicate to the ECU about what action to take, and what happened, but the actual voltage control of the headlamps is a simple PWM driver to the headlamps themselves from the TIPM.

 

I believe the misconception occurs because new cars use this network "BUS" technology where as before in older cars you flipped a switch that went through the fuse hit a relay and provided voltage direct from the battery, no modules to provide this integrated power, so the word CAN-BUS should be swapped with TIPM to make everything sound legit. PWM could not occur in a car without a BUS type system because nothing would regulate the voltage to pulse. Sometimes computers make our lives worse than better. So maybe I'm to blame for the misconceptions of this forum sort of, but I still understood it was the PWM causing the headlights to be terrible.

A wiring harness gets rid of the PWM because its pulling power direct from the battery and the harnesses that LED lighting people sell just use a capacitor to help store the voltage to correct hold voltage that tricks the PWM.

 

Yeah you are mostly right - you're not likely to see PWM without a micro-controller somewhere in the mix, but it's not a mandatory thing. The reality is, that's what's been going on in automotive since the 1980s so this isn't a new thing, really. Memories of how bad cars really were back then when all of this was new...

The engineering decision to go with PWM for the headlights seems weird to me, so I wish we had some real data from the engineers to share and speculate about. I mean, the TJ didn't do that, correct?

The harnesses that use direct battery feed and use the main headlight driver just as a signal switch is totally the right way to go as a work-around. It's clean and simple. The CAN bus controller project I'm working on is another way, but it's more complex/powerful so we'll see if that ever comes to fruition or not.

 

This is awesome, thank you. Definitely sticky material.

 

^x2 Mods can you sticky this please?

Excellent overview!

.

 

For the lights in particular, the TIPM uses a pulse-width-modulation signal. If you look at this signal with a oscilloscope, you will see a square wave. The end result is that the lights get approximately 8 volts of power sent to them, in pulses, rather than a constant 12 volts. This increases bulb life and reduces fuel consumption (probably on the order of 1-2%). For incandescent bulbs, this is a great trick as the bulb won't start to dim before the next pulse of power is received. For LEDs or other lights, it doesn't work the same, so that's why you get a flicker.

Question: How does this reduce fuel consumption?

.

 

The theory would be less power draw, the alternator does not need to work as hard. The reality because of the computer systems we now have a standard 160 amp alternator inside of the old 35/60 amp alternator.

 

Yeah light bulbs running can actually be a bigger draw on fuel than you'd expect. I remember reading some GM data years ago that DRL's were 2-3% economy hit. This is old, old data (pre-2000), so who knows what the numbers are today.

 

If any truth to the above would LED headlights improve Fuel Economy since their draw is so low?

 

This is helpful info & I agree it should be a sticky.
Good write up..

 

Thank you for the great write up!!!

 

Thanks for the information! Just a quick question, is that true of all lights in the vehicle? Suppose I put an led 3rd brake light, same thing?

 

I think this is where it gets confusing for most people. Some light sockets don't have any issue with a flicker from LED bulbs, and some don't have an issue with the LED bulbs itself, but the lack of a higher voltage causes it to believe there is a bulb out and cause other issues.

For example, the dome lights in a 4 door JK from front to back are -
map lights on the mirror
b-pillar lights over the front seats
rear cargo lights over the rear seats

The mirror lights are already LED, the front seats lights are two pods with a 194 sized bulb in each facing sideways, and the rear cargo is a festoon bulb.

You can change out the front bulbs with any 194 LED and it works fine. No flicker, no "CANbus" issues.

If you change out the rear cargo light with a basic (not CANbus rated) LED bulb you'll find everything works but the map lights will stay on after the doors are closed. In doing my LED project, I found the map lights will stay on constant when the Jeep is running, and will time out after about 15 minutes when turned off and parked. Very similar to what happens if the door isn't closed tight and the lights stay on for about 15 minutes before timing out.

To fix this issue with the cargo dome light I ran a tiny 12v standard bulb inline on the hot wire to provide the feedback for the brain to figure out there's a bulb there and it's ok. Now whether that brain is the TIPM or the CANbus, that's where most just call it a CANbus issue and leave it at that. That includes myself, as I understand what's happening but I could really care less as the I know the fix too.

The biggest problem I see is that back when cars began having CANbus systems and bulb issues came up - the label for compatible bulbs was "CANbus compatible" or "CANbus error free", so this pushed the reason for bulb issues to being a CANbus issue.

Canbus Error Free LED bulbs Frequently asked questions

 

Wow! Somebody's making it rain knowledge up in here. Sometimes I miss the good ole days of my '68 Chevelle. Thanks for explaining this to us dummies.

 

Somebody call ? :whistling:

 

To the OP - You should also link your work back to this thread, you already have a lot of followers.

http://www.wranglerforum.com/f202/hacking-the-can-interior-bus-466730.html

 

Thanks!

 

Id think using a PWM on the vehicle's bulbs would have other benefits aside from less parasitic loss and longer bulb life. Some ideas that come to mind, not sure if they're true:

- Better regulation of headlight performance. Regardless of load on the electrical system the PWM logic can maintain the same brightness by adjusting the pulse timing / amplitude

- Programable behavior, factory can set things like DRLs to any light desired and any brightness desired

- Potentially simpler wiring, why not have a tail light be two wire? Simply reduce the voltage for running lights and increase for brake / turn lights

Just imagining the other potential advantages to a system like this, not sure if any of these things are actually being used.

 

My thought of the night - and I'm tired, so maybe it's way off but..

If PWM is good for bulbs, and can extend the life of the bulb, why does turning a old fashioned incandescent bulb off and on repeatedly, shorten its life? It's been said it's better to leave a bulb on because the heating and cooling of the filaments weakens them over time, causing the bulb life to shorten. This would not really apply to the LED bulbs, but the blinkers, dome, tail light/brakes, and headlights are incandescent bulbs as stock. The main difference is, does the research found in household bulbs, apply to vehicle bulbs - or is this a moot point.

Like I said, I'm tired, and this was a random thought tonight.

 

I believe it has more to do with heat up/cool down cycles, then actual on/off operation.

Household bulbs are operating on ac power at 60 hz so in effect they are turning on and off 60 times per second. But they don't cool off between "pulses" enough to change the apparent light or cause flickering. Cars are dc power, so it's a constant 12v.

 

This was kind of my point. The AC old style lightbulb would run on a 60 Hz setting, the pulse would be quick enough to not normally dim the bulb but how does this translate to vehicle PWM heating/cooling? It's been said you can sometimes see the stock headlights flicker, so it that causing enough heating/cooling of the bulb filament to weaken it? The "constant" in this would be against a vehicle with no CANbus/PWM.

If we go back to the OP's first post, this was stated -

"For the lights in particular, the TIPM uses a pulse-width-modulation signal. If you look at this signal with a oscilloscope, you will see a square wave. The end result is that the lights get approximately 8 volts of power sent to them, in pulses, rather than a constant 12 volts. This increases bulb life and reduces fuel consumption (probably on the order of 1-2%). For incandescent bulbs, this is a great trick as the bulb won't start to dim before the next pulse of power is received. For LEDs or other lights, it doesn't work the same, so that's why you get a flicker."

Based on this, the lights are pulsing in square wave of 8 v, so is this square wave something that would lessen a stock bulb's lifespan and is using PWM worth it because of ?? I can't see this saving gas, or extending the life of a bulb, rather shortening the life of a bulb by causing premature wear of the filament due to the pulsing voltage heating/cooling the filament. Maybe I'm just looking at it differently, but that's my first thought.

Now if someone can explain that the wire would actually be staying cooler and holding a fairly consistent temperature due to the PWM lessening the voltage and therefore lessening the heat pull into the filament, then that makes sense too. Lower heat on the filament could cause less fluctuation in the filament and prolong its life.

The question then becomes, what exactly is occurring? (or which theory of mine is closer to being right)

 

I wired in a tiny, rice sized bulb, inline with the cargo wiring and it works perfect. Here are a few photos I took during the project. The overall way I modded this one was a bit ghetto, but it looks good and function properly and unless you took it apart, you'd never know.

Jeep Cargo LED Mod Photos by Lusus_Naturae | Photobucket

Yes, the map lights are COB (chip on board) construction. I changed it to red using glass beads and gluing them into the cone tunnel for the light output. Works perfect, again, a bit ghetto in how I did it. I have heard of a few trying stained glass circles and siliconing them into the cone or tint. Here's my info on this mod.

As a side note, test the light before completely reassembling it again. The photo showing a few beads in the bottom of the cone was the first test. I did find a bit of white light coming through and went back and added one more bead afterwards.

http://s1318.photobucket.com/user/Lusus_Naturae/library/red map light LED mod?sort=4&page=1

 

I suppose it depends on the voltage and frequency of the wave. If the bulbs are flickering(incandescent) then it sounds like either the voltage isn't high enough, or the "on time" isn't long enough. In reality, it should be a fairly easy code change for Chrysler.
Led's on the other hand will need a cap inline to smooth out the voltage supply into more of a flat line vs a wave.

 

When I get back to the Jeep this weekend I'll post some scope measurements of the headlights, some other lights, and a CAN bus connection so you can see the general difference (those are on the forum elsewhere but hard to dig up).

As far as bulb life, the pulse is happening fast enough it's helping the bulb not hurting it.

 

I would love to see what you find. Fascinating stuff. Thanks for the work!

 

I am going to install driving lights on the front bumper, the relay for them is supused to be triggered by the high beams . Can a rellay be operated with the pmw to the headlight?

 

They can, but generally you'll need a capacitor to even out the voltage flow so that's enough to drive the relay trigger without it going nuts. There's some HOWTOs on how to do that out there. It's pretty easy.

 

I have a question that may or may not be related to CAN bus. I purchased front and rear Ace bumpers on a Black Friday sale last November. These bumpers have integrated LED light bars, a 20" two row in the front and two 6" single rows in the back. I wired the front into the fog light harness and the rears into the reverse light harness. The LEDs seem to work just fine. But ... now my HVAC goes nuts when I start my car. Nuts, meaning that for several seconds (30 or so) it cycles through various temperature, vent and fan configurations while the rear window defrost light flashes. Related? Maybe not. But it seems like a CAN bus thing.

Thoughts?

 

Easiest thing to do is to temporarily disconnect the LEDs just to see if you have a difference in behavior.

Chances are it isn't that, but something else related to the bumpers that is confusing it. Did you perhaps damage the front ambient temperature sensor (mounted on the grill frame)?

 

Scribed...got the same ACE Bumpers coming!!