Comparison Test: LED drop-in headlight bulbs vs. halogen bulbs [pics inside]

Yeah, to have proper projectors, you have to have something engineered to work in your vehicle properly. That doesn't mean a product such as this will not work in a reflector. There are benefits to this type of tech and there are drawbacks too. Also, you'd be surprised at what is possible with today's tech. I'm sure that if you had the know how, you could make one yourself with a raspberry pi or the like.

This is now on hold. I had dust incursion so I disassembled the headlights to clean them. I didn't break anything, but with just the water, there are portions of the chromed finish that decided to lift off. While dabbing dry other portions of the fixture, more of the finish came off.

Aside form either needing to refinish the reflectors or outright buying new headlight assemblies, I need to put the focus of this off until I have replacements. I did happen to toss in the lamps to see how they performed and they did surprisingly well, btw.

Digital Corpus said:

.........I did happen to toss in the lamps to see how they performed and they did surprisingly well, btw.

Click to expand...

that's kind of not fair to at least show a pic of the results. I'm hurt.

I would've had to have taken one

Edit: I only put one in too...

This may be a meaningless side note, but do you have stock headlights or one of the newer aftermarket sets with "clear" lenses? I guess what I'm getting at is do your lenses suffer from aging as I see you have a 1997.

Aftermarket glass lenses on E-code type headlights.

Teaser post to prove that this is not forgot. Also, sorry, I have too many things in motion to do a complete post on this. I don't know how common of an audience I have with my posts but some may be aware that I have a few projects with my TDI here on the forums... Anyhow, enough rambling.

LED Driver consumption: 1.5 A @ 13.25 V = ~19.88 W
LED Array consumption: 1.21 A @ 5.65 V * 2 = ~13.67 W
(I measured the current of one branch of 4 LEDs so this stands to reason that it needs to be doubled as you see)

Efficiency = ~68% peak (This improves to this as things warm up)


LED Headlamp Schematic:


The power supply is a simple buck regulator with a constant 5.6-5.7 V output. LEDs "should" be driven at constant current, but that complicates this type of product and will result in cascading failure if 1 of the 8 LEDs fails. However, if one LED fails with this setup, it will take out it's partner it;s in series with and the failure stops at there. The datasheet specifies ~700 mA as nominal drive current though it was clear that the measurements I made places that @ ~600 mA, much to my surprise actually. This puts the luminair at ~1750 lumens whereas a standard H7 bulb is ~1500 lumens for the 6500K LED.

Teaser pics taken with one of my reflectors w/ no glass, some dust, and a few flakes of reflector missing and taken with my cell phone. I have new reflectors arriving next month.

6500K, 70 CRI LEDs


And with 4000K, 90 CRI LEDs


The square pattern you see is actually the front of the individual LEDs. Distracting? Quite? Practical problem in real life use? I doubt it.

the 6500K looks like more usable light to me. need to see it on the road to really tell. was the room completely dark during the pics?

deejaaa said:

the 6500K looks like more usable light to me. need to see it on the road to really tell. was the room completely dark during the pics?

Click to expand...

Looks, especially from a cell phone where exposure and color accuracy are unknown, are deceiving. Especially since I shined the beam on a non-neutral colored surface.

The room had some partial curtain drawn midway through the day.

I've not finished my write-ups yet, but the TL;DR is that 6500K light will bleach the scotopic portion of your mesopic vision making it harder for you to see at night. The 2500K light from a standard halogen bulb won't do that but is so far removed from scotopic vision that you'll get better efficacy in the midrange, 4000K to 4500K, the former of which is what I'll be testing against.

There there is CRI which is also a measure of useable light, which is much more finite...

I'll be tuned in because this is interesting.

I'm watching this too.

So in the last 72 hours I've been able to track down the final piece of information regarding the human eye I've needed to know to account for when rods bleach and actually determine S/P ratios. Real life is a bit of a mess with our house torn apart, but I've been able to run the numbers with photopic & scotopic efficacy, spectral power distribution of ideal black bodies from 2500K to 6500K, the area under the curves when joining those two, how mireds make CCTs mathematically simple to digest, and cone & rod sensitivities dependant upon the luminance of a scene (the units of lux that is). Being able to break all of this down and generate regressions with R^2>=0.999998 or better and then mesh all of the data together is incredibly satisfying for the mathematics nerd in me.

I have been rather surprised that every study I've seen only deal with S/P ratios and does not try to account for whether or not rods get bleached. I do need to verify that the material I've reviewed is being used accurately by myself, but preliminary findings are quite encouraging. If they are accurate, I do need to involve one last piece of information regarding white balance and spectral power distribution, and then I'll pull the data off a calculator and into a spreadsheet to share and generate higher resolution graphs.

All in all, 4000K is currently the highest CCT most LED manufacturers generate 90+ CRI products in. Though 4500K has a tad better result than 4000K, we're already at diminishing returns, thus putting 4000K CCT sources with 90 CRI values or higher as quite literally the best cost vs visual performance lighting solution for the human eye under mesopic vision.

Does the 4000k value hold for other, non-LED types of lights?

In the end, are there any current products (LED or non-LED) you'd recommend?

Great work, thanks for the information!

I'm not quite done. I'll be wrapping the mathematics end by New Years, hopefully. Until I have all of that properly sorted out, I'm going to hold off on answering the question because it's not quite a simple answer. You're welcome from what I've listed so far.

Do you have a link for the 4000k 90 cri?
I would be interested in trying those in my e-codes. Or if you took them off let me try them

Here ya go!
http://www.mouser.com/Search/Produc...-3virtualkey58270000virtualkey997-LXZ2-4090-3

If you want complete assemblies, that's a whole other story

Digital Corpus said:

Here ya go!
http://www.mouser.com/Search/Produc...-3virtualkey58270000virtualkey997-LXZ2-4090-3
If you want complete assemblies, that's a whole other story

Click to expand...

Excellent research Digital Corpus!
It’s been almost 2 years now...
Is there any LED headlight bulb assembly product on the market right now, with excellent heat sinks, made with these ideal 4000K, 90 CRI, high power neutral white, LXZ2-4090-3 Lumileds ?!
Asking for many TDI owners we want to upgrade all of our car’s lighting to current high quality low power 2018 best quality LED bulbs possible.

I had some data loss about 6 months ago, outside of this being one of many projects. I have slated time (flexible, by year's end) for addressing everything present and rewriting or creating a new thread approaching the topic from a scientific perspective regarding the human vision system.

That said, no manufacturer produces end-user products with 4000K LEDs in headlight applications. You'd have to buy a set and modify them to match your needs.

not sure if that is really the issue:
it's glare to the oncoming drivers on the road.

That matters if you buy a crap bulb. There are plenty that aren’t, to which will be demonstrated soon enough.