Re:HID bulb/module testing

I brought this up to the group chat awhile ago, but wanted to revisit it with the community. Around here, when a vehicle comes in with an inop HID headlight, it can become a mess. :silly: Try a known good bulb, then sell a new headlight assembly or HID module when that doesn't fix it. It goes from cheap to $$$$ quickly.

What's the better way? I started playing around with a used Subaru HID headlight assembly for giggles. Headlights themselves are becoming more and more inaccessible, so I figure it'd be awesome if we could use a current probe at a headlight fuse to do our testing. I hooked up our test headlight and got a current waveform during normal, working operation:

 

And here's a zoomed in view of the bulb 'lighting off' for lack of a better term. I don't know HID bulb operation like I should. :blush:

 

So we know how much current a working bulb should consume. Great. Now I cut the outside of the electrode to simulate an open bulb. I added a voltage trace on channel #2 to show when voltage was applied.

 

 

It would seem to me that the HID ballast/module is trying to light the bulb off, fails, then gives up. :silly: This behavior was repeatable. I don't know about the internal construction, but there would seem to be some kind of logic at work.

Then for giggles, I damaged the module itself, and retook the capture:

 

Can we infer that, if we see an attempt by the module to ignite the bulb, that the module is good? And that power/ground to the module is good? I'd welcome your thoughts.

It does appear that way, but I wonder what real world damaged modules look like if they may act all together? In what way was this module damaged ? Shorted output? Over voltage? Secondary ignition spike? Did you try testing voltage on the output before cooking it?

Sent from my SM-G781U using Tapatalk

wrote:

It does appear that way, but I wonder what real world damaged modules look like if they may act all together?

Good question! I took this capture off an '05 Acrua TL with inoperative headlights:

 

Based on previous experience, I took a shot that the bulb was blown and that the rest was OK. Customer took the vehicle. Never got to fix it.

In what way was the module damaged? I don't remember. :blush: :lol: I actually took these captures in January, but never posted them immediately. I've slept since then. :silly:

 A couple of months ago I watched this interesting studies by Hans Jorgensen. I have not tried it myself though..

Figures that TBT beat me to it. :silly: But I do feel better that I'm on the right track.

I'll have to find anther sacrificial headlight to play with.

wrote:

Did you try testing voltage on the output before cooking it?
 

I wouldn't recommend taking voltage measurements on the ballast output. IIRC it takes several thousand volts to start these things.

I think it was Rick who mentioned that's a great way to kill a scope...

Nevertheless you could just use a capacitive secondary pick up with a 1:5000 (snap on) or 1:10000 (pico, hantek, Autel) attenuation.

Another example off a 2011 Acadia. The left front is working correctly:



The right front is not:



Zoomed in, it appears the ballast keeps trying to light the bulb and can't do it:



I'll follow up when this gets a new bulb next week.

New bulb went in. This is the right low beam after:



Note that this bulb only comes as a bulb/ignitor assembly.

Tyler wrote:

New bulb went in.
Note that this bulb only comes as a bulb/ignitor assembly.

This takes away the pain to distinguish between faulty igniter and bulb. Glad the manufacturer didn't decide (yet?) to sell a complete replacement car when the headlights go south :whistle:

The fault capture looks like the igniter tried it's thing but the bulb was bad. If so, did the igniter stop its fruitless intents after a while or went it on forever?

Thank you for the update.

juergen.scholl wrote:

The fault capture looks like the igniter tried it's thing but the bulb was bad. If so, did the igniter stop its fruitless intents after a while or went it on forever?

It wasn't clear in the capture, but yes, the ignitor (or the control module?) eventually quit trying to light the bulb on its own. I can go back and look at the capture, but it wasn't long. Two or three seconds. It would not attempt to light the bulb again until I cycled the headlight switch.

Some kind of logic in the ignitor or control module knew there was something wrong.

Following up to this thread as I encounter more broken vehicles.

2011 Grand Cherokee 3.6L comes in with a complaint about the headlights. Left no workie, right is fine. B162C backs up what we already know.

Zoomed out shows a strong in-rush spike, followed by...very little. Some noise, really. Customer has already put their own bulb in it. I suspect the ballast is smoked. As in, it's not capable of igniting the bulb.

If this customer goes for the repair, I'll follow up with an after waveform.

Soooooo the Jeep from my last post left without a repair. Oh well!

I came across another opportunity in the form of a used headlight assembly from a late 2010's Hyundai... something. :silly: Sorry I don't recall the exact YMM. Anyway, I extracted the HID bits from the headlight itself and got them out on the bench:



Pretty simple, really. Switched power and steady ground.



The ballast/module is made by Osram. I'm kinda mad now that I didn't take note of the OE manufacturers on all the other systems I've tested. The way different OE's react to failures might be worth knowing.

Also worth knowing that the stated output voltage is 42V. I thought these things ran several thousand volts? Or, does the ballast supply 42V to the ignitor, and the ignitor is the high voltage component?

Anyway, onto the testing. This is a known good shot of the bulb in normal operation:



Some good current flow going on, and a couple solid spikes when power is first applied.

Now I remove the bulb from the ignitor:





While this is happening, there's some pretty good arcing going on in the ignitor socket. :silly: More than enough to light your grill. The really fun part is that this three-arc pattern is repeatable. Apply power/ground, three arcs, then nothing, even if you leave the power connected. More evidence that the module knows something is wrong and moves to protect itself.

Then I disconnect the ignitor from the ballast/module:





Again, the three-arc event is repeatable. Just a lot less noise. The module still knows something is wrong.

If anyone can think of a good way to mutilate this rig to tell us something we don't already know, I'm listening.