Xenon HID headlight diagnosis

If you are not a PicoScope user and want advice with a diagnostic problem then post here. Forum members may be able to help.
User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Xenon HID headlight diagnosis

Post by FioranoCars »

Hi
I was just having a browse of some old posts, and noticed someone else has asked a similar question but with no answers given ... so excuse the repost but we've hit the same problem

When trying to diagnose the failure of a Xenon light, is there an empirical way to confirm the bulb verse the ignitor as being at fault? The previous post related to attenuation, but I guess this level of attenuation is pretty specific to this application and there maybe a totally different approach to take ???

Currently we take the bulb out and put it on our test ignitor if that fails then we assume the bulb (but can't say if the blub blew due to ignitor failure, or just died) and plug a bulb into the car, if it works, we cross our fingers ... well now we do, as recently had the car return a week later with another blown bulb (Ferrari charge £200 for a bulb), and we had no way to test the ignitor (need to be able to measure 27kv +) ... a new ignitor can be over £1000 !!.

It would be good to be able to make some proper tests, including volt drop to be able to say what's failing. If we'd been able to monitor voltage (stability over time as the light get up to temp) then we may have been able to spot the issue before the car left us, and saved the customer the extra visit and saved ourselves £200 in written off bulb.

thanks
Richard
Richard Lukins
FioranoCars

User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Re: Xenon HID headlight diagnosis

Post by FioranoCars »

Just found this Fluke 80K-40 High Voltage Probe

The 80K-40 is a high voltage probe designed to extend the voltage measuring capability of an AC/DC voltmeter to 40,000 volts peak AC or DC Overvoltage Category I. This means the probe can only be used to make measurements on energy limited circuits within equipment. Examples include high voltage within televisions or photo copy machines. DO NOT use this probe to measure high voltages on power distribution systems. The probe provides high accuracy when used with a voltmeter having 10 Megohm input impedance.

A high voltage probe that allows a multimeter to measure up to 40,000V respectively. Intended for low energy applications only
http://www.fluke.com/fluke/usen/accesso ... ?PID=55357

It says it is for low energy, but I have no idea of the peak HID current usage, or even it's stable operating use, though I suspect the stable operation consumption to be less than 20w so under 2 amps. Is this likely to be suitable?

Any feedback or help appreciated!
thanks
Richard
Richard Lukins
FioranoCars

Avdr
TwoWaves
TwoWaves
Posts: 157
Joined: Sun Jan 08, 2012 9:25 am

Re: Xenon HID headlight diagnosis

Post by Avdr »

I've got some saved waveforms on the other laptop of HID current requirements, although they do represent total current drawn by an aftermarket HID kit (including ballast), as opposed to a factory fitted kit and just current to sustain the light. If you think they would be helpful I will upload them tomorrow.

Andy

User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Re: Xenon HID headlight diagnosis

Post by FioranoCars »

any samples would be very handy, thanks.

On the basis of Ohms law could you calculate the volts of the Output from the Ballast in a WORKING circuit using 2 amp clamps, one on the known 12.x volt section to the ballast and one on the section from the ballast to the bulb?
Is the resistance going to vary as well or remain static ... Doh, I guess that is the whole point of the HID ballast to enable the bulb to work by overcoming the resistance to start the bulb!! The resistance then falls as it gets hot? I was thinking about using the consistent value of current (varied by volts) to trying and measure it, too clever for my own good!

Clearly I should think before I start typing!

Anyone got some pointers to web articles or others on the theory of Xenon/HID lights?

thanks, look forward to some samples
Richard
Richard Lukins
FioranoCars

Avdr
TwoWaves
TwoWaves
Posts: 157
Joined: Sun Jan 08, 2012 9:25 am

Re: Xenon HID headlight diagnosis

Post by Avdr »

Have read a few good articles about HID's, detailing types of metal salts used etc etc.Havn't got any links though sorry,
Problem with the method you describe is going to be ballast efficiency I would've thought.
I will post the waveforms tomorrow, and could take some more detailed ones for just the current to sustain bulb once its upto temp.

Further more, can you not see this data serially? I've got an old vectra B CDX with factory fitted HID lamps and can view a great deal of data serially, will have to see in the morning what I can and can't view (not had problems for a while) but will report back. On one side of the car the ballast is mounted on the innner wing, so is remote from the bulb, this should permit me to measure current in the wires to the bulb, and if I can match it to a Kv reading serially, then you've got a known good setup for a D2S bulb and ballast.

Andy

User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Re: Xenon HID headlight diagnosis

Post by FioranoCars »

thanks, appreciate the help and feedback, sadly not had time to play with a working unit, but I will after looking at your readings.

Do you think the Fluke 80K-40 with a 40kv but "low energy" (what ever that means?!) is suitable to use post ballast?

I guess my fear testing a faulty unit is, will the voltage output of the ballast remain in the kv ranges if the bulb is duff, ie maintain trying to 'start' it, or will it drop to a lower level after a set amount of time? and as I don't have a suitable probe I'm reluctant to poke it and find out by being knocked flat!! (could replicate this by just taking away the bulb once I have a suitable probe).

But the sole goal is to differentiate between a faulty bulb or ballast, without any risk to equipment, car or staff.

I guess a good ballast with duff (or no) bulb could be measured pre-ballast for amps over time using the "turning on" as a trigger, and measure for 10minutes, verse the same for a fully working unit, and then see if the results show a ballast consumption without the bulb that might be missing in a duff ballast, but it really would be better to test post ballast.

thanks
rIchard
Richard Lukins
FioranoCars

Avdr
TwoWaves
TwoWaves
Posts: 157
Joined: Sun Jan 08, 2012 9:25 am

Re: Xenon HID headlight diagnosis

Post by Avdr »

Sorry, its been a busy day, haven't yet had a play with serial data on the vectra. Will aim to do that on friday now,

In the meantime here is a HID start-up current from cold, this is total current to ballast, for a cheap aftermarket kit, with a supposed light power of 35W.
HID PICTURE.gif
Interestingly there is no change in current required from a cold start to a fully warm bulb.

RE: The fluke clamp, I think the devil must be in the detail, exactly what is meant by low energy? Perhaps a call to fluke's tech department could clear that up.

I would expect the ignitor to keep supplying a voltage in the order of kV until the salts had vapourised. Will have to look into it before I comment further haha.

EDIT: could be interesting to use a plug lead clamp on the ignitor wire?

User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Re: Xenon HID headlight diagnosis

Post by FioranoCars »

Thanks
yes, the 35W seems to be borne out by the current being used, I too expected the current to reduce after getting up to temp, or maybe a big peak at start up ...

I guess we'll take a look at a 5 minute trace and see if things change with an amp clamp and maybe try the Ignition lead clamp at the same time, as that should give us an approximately safe range?!

Busy new year testing ideas!
If you find that document of the HID theory, feel free to post a link or PM me with it.

thanks
Richard
Richard Lukins
FioranoCars

Avdr
TwoWaves
TwoWaves
Posts: 157
Joined: Sun Jan 08, 2012 9:25 am

Re: Xenon HID headlight diagnosis

Post by Avdr »

Have just got a capture from the aftermarket HID kit, The factory fitted kit does have a data block for lamp voltage, but appears to ignore the initial ignitor phase. I did a short data log from a cold headlight, through to a warm up and then re-striking a hot burner. You can see the voltage applied to the bulb to control its intensity.

Pico file
Aftermarket HID kit CURRENT kV.psdata
(4.85 MiB) Downloaded 874 times
Was suprised to see the current wasn't continual, also did another run with the other channels turned off and just the kV channel in use, interesting result...
AFTERMARKET HID KIT kV.psdata
(13.8 KiB) Downloaded 828 times
Thats 1 micro second of time on the screen...

Here is the data log from the factory fitted system

User avatar
FioranoCars
TwoWaves
TwoWaves
Posts: 386
Joined: Thu Dec 13, 2012 11:26 am
Location: London, UK

Re: Xenon HID headlight diagnosis

Post by FioranoCars »

Very interesting
1. It looks like the HT clamp for voltage is going to be fine, that saves an expensive probe (although it's on my Christmas list anyway ... for 2016 ...!)
2. The switching voltage/current, will add this to my tests when we do them
3. in the first file:
a.the amp clamp was on the 12v pre ballast section?
b. Engine not running? (from voltage and the dip upon turning on lights?)
Questions:
c. what was causing the erratic HT clamp readings pre lights being turned on? (assumed it might be alternator ripple, but engine appears to be off, or is that wrong?)
d. Once past the "on" moment, the voltage stays under 500v, so could be measured using an 20:1 attenuator with some accuracy? (on the later scopes you have 2000v to work with, I have an earlier 50v native, but still 1000v with 20:1 fitted) so a more direct measurement could be made?

Well I think I can now do some further tests myself without any further kit and be able to form a simple but effective method of diagnosing a bulb verse a ballast (on the assumption the ballast will send out a High Volt spike to try and ignite the bulb). Clearly need to test the theory, but after a year of thinking and asking experts, the solution came from here, so many many thanks!

If anyone has bought the Fluke 80K-40, I'd love to hear any results, issues etc before I eventually do get one!

thanks
Richard
Richard Lukins
FioranoCars

Post Reply