HID kit comparison tests with measurement data

HID kit comparison with measurements

Optical Measurements

I recently got my paws on a Prolumen 50W HID kit from Devin and decided to make a few quick measurements of the kit for comparison purposes. The big burning question most people ask is how much light output does a given HID kit make. While the proper way to make the light output measurement is in a photometric sphere, I don’t have access to that toy at the office (and I’m not even sure ours would be able to measure an HID bulb without overloading). So, I did the next best thing. I grabbed a calibrated photocell (leftovers from a past life) and placed the bulbs under test about 40 mm away from the centre of the cell. The bulb to sensor spacing was checked before each separate test to ensure the same (give or take a mm) placement. The particular cell I used was adjusted for 10mA Isc (short circuit current) in full sunlight. The room I worked in had pretty dim lighting so the ambient light wouldn’t have contributed much error to the results. For reference, the room lighting generated only 50-60uA Isc.

Below is a photo of the setup. The big box at the top is a 60A power supply. The Fluke 83 meter was set to record min/max/average current from the cell. With each test, I fired up the bulbs and watched the arc (wearing a welding helmet). When the arcs were stable and the meter’s max reading no longer increased I reset the min/max/average values and made a measurement for about a min. Both 35W kits were tested with the same Prolumen 35W/4500k bulb. The 50w kit was tested with a bulb stamped 50W/5000k, which I was assured by the vendor, is actually close to 4500k.

Here are the light output test results. They results are listed as the photocell short circuit current (Isc). Isc is proportional to the illumination of the cell. Another thing to note is that the cell I used has a spectral response close to the human eye’s (ie: it see the effects of colour vs brightness much the way we would).

Relative Optical Outputs Using Prolumen Bulbs Expressed as Photocell Output

ApexCone 35W thick ballast Min 42mA Average 58.9mA Max 67.9mA

Prolumen 35W thick ballast Min 56.0mA Average 59.0mA Max 62mA

Prolumen 50W thin ballast Min 92.4mA Average 93.8mA Max 95.6mA

EDIT:

I had a chance to test two more kits for light output. These were tested at a lower temp (11c). Also note that the KDW kit had a higher colour temp bulb which usually reduces light output.

KWD 50W thin ballast with 6000k bulb Min 40mA Average 53 Max 68

ESTIMATED AVERAGE OUTPUT WITH 4300k bulbs 87 *** measurement incomplete

No-Name kit 35w thick ballast with 4300k bulb Min 44 Average 53 Max 74

during the tests the no name kit had some very large warm up output spikes. at one point it hit a reading of 96 for a few seconds. the KDW kit had a very smooth warm up. It sat around 10-20 for a few seconds then ramped up to mid 50s over a period of 25 seconds.

end of EDIT

If we use the 35W kit as a reference, the 50W kit is about 59% brighter.

I did also make a measurement of the ApexCone Ballast with their supplied 5000k bulb but it visibly made less light. To be fair, I used the same Prolumen bulb to test the two 35w ballasts.

Because the eye’s response to intensity is logarithmic, it would have been very difficult to visually observe the slow fluctuations is bulb intensity that occurred with the ApexCone ballast. Typically, HID bulbs do not last long if they are not operated at the manufacturer’s specified power. The few datasheets I’ve seen specify a +- 10% tolerance for the allowable power to the bulb. A bulb’s intensity usually varies a bit less than the power fed to it (mainly due to efficiency shifts at different power levels). We can therefore assume that the ballast power variation is equal to or greater than the percentage shift in light output. The percentage shifts from average to min and max are listed below.

Estimate of Ballast Power Variation from Nominal

ApexCone 35w thick ballast -28% to +15%

Prolumen 35w thick ballast -5% to +5%

Prolumen 50w thin ballast -1.5% to +2%

I couldn’t directly measure ballast output power because I didn’t have the necessary HV probes on hand and didn’t want to hack up the kit wiring. However, the light output variation is still a pretty good indicator of the power fed to the bulbs.

Electrical Measurements

Next, I decided to measure the power consumption and startup current transients for the 3 ballasts. To do this, I set the power supply to 14.5V and introduced a 0.1ohm sense resistor in series with the supply leads. A Fluke 123 was connected across the sense resistor and set to record at one horizontal division every 5 seconds. The vertical scale was set to 200 mV per division, which scales to 2A per division. The tolerance on the sense resistor was 5%. Regardless, the same sense resistor was used for all three ballasts. Some readers will note that the additional voltage drop in the sense resistor will alter the results somewhat and a smaller one should have been used. I need to point out here that the car’s headlight wiring likely amounts to a couple tenths of an ohm so we are actually simulating real life conditions better than if I had used a smaller sense resistor.

The start-up current plots are below. Sorry about the cell phone camera shots. I don’t have a data capture kit for the scope I was using.

All three kits generated a large but short current spike when first turned on. This is usually just some input filter capacitor charging. The two 35w kits generated a spike a bit over 14A and the 50w kit was just under 14A. The initial current spike is what tends to burn headlight relay contacts. However, none of the kits had a spike big enough to worry about. A regular H7 bulb has a cold filament resistance of ~0.7 ohm. At 14 volts, that would generate a 20 amp spike.

By eyeballing the average current each ballast draws after warm up we can get an estimate of the current consumption. The different ballasts had varying warm up times and behaviour so the full plots were useful for doing this. The final settled power consumption (read off the flat part of the plots) for each ballast is as follows:

Power Consumption From Electrical System

ApexCone 35W 5A 72.5W @ 14.5 volts

Prolumen 35W 3A 43.5W @ 14.5 volts

Prolumen 50W 3.75A 54.4W @ 14.5 volts

Assuming the ballasts all deliver their rated powers to the bulbs, the heat load and efficiencies are as follows:

Wasted Power as Heat & Efficiency

ApexCone 35W 37.5W 48%

Prolumen 35W 8.5W 80%

Prolumen 50W 4.4W 92%

The higher the heat load, the greater the heatsink requirement to maintain the same internal temperatures. Generally speaking, hotter running electronic parts fail sooner.

The claimed efficiency for the Prolumen 35W kit was 80%. I couldn’t find figures for the other two.

Assorted Notes

The 35W kits had an earlier bulb design where the return wire would cast a shadow in the headlight beam. The 50W kit had this corrected by placing the return wire on the same side of the bulb as the locating tab (on the bulb base).

The 35W kits had longer wires making stealth installation a bit easier.

The 50W kit was thinner, increasing the placement options.

The 50W kit is much noisier than the two 35W kits. With the engine running though, you would never hear it.

The arc stability and confinement was much better on the 50W kit. I couldn’t get a photo of this (there was no way I’d take my welding helmet off with the bulb burning) but anyone with a welding helmet can confirm that the arc wanders less and is narrower with the 50W bulbs. A narrower arc is easier to focus as it more closely mimics the filament light source. Another thing to note is that the 35W bulbs are glowing red-hot right after shutdown while the bigger 50W bulbs are not.

A buddy of mine just picked up a super cheap no-name kit. I might have a chance to measure it at some point.

Another friend has some OEM ballasts and D2S bulbs. If his car is ever down for a couple days, I’ll probably grab his setup and make a similar measurement.

Sooo.. In short?

This is a bit too technical for me

Buy the 50w prolumen it looks like if price is not an issue.

Much more efficient than the the apexicone in terms of power consumption 54W @ the 50W light output vesus 72W at 35W light output.

Sooo.. In short?

 

This is a bit too technical for me

Coles notes version:

both apexcone and prolumen 35w kits make about the same light with the same bulb. with the supplied bulb the apexcone makes less light. I didn't bother with the full measurement of the apexcone bulb when I had to turn my welding mask down to shade 9 just to see the arc (was up around 11 and 13 for the other tests).

the apexcone kit generates much more waste heat so make sure it has tons of cooling if you chose it. it was actually drawing more power than a stock H7 bulb would

expect the apexcone kit to smoke bulbs faster due to the variation in output power.

the 50W kit generates more lumens per watt than the 35W kits and will likely have good bulb life.

What about for D2S factory OEM ballast bulbs Denso's Phillips Orsrams and such?

I'll test whatever happens my way

The sets I tested were what I had on hand or could get my hands on easily. My buddy has a D2S setup with OE ballasts but they are on his car.

There are a couple of problems with this comparison: one is you're using older bulbs and presumably ballasts (most Apexcone/DDM use the thin ballast now), you're testing 2 brands and 2 wattages and comparing them, and it's basically impossible to draw any conclusion from this other then not to get the particular Apexcone kit you used.

It's interesting information, it just doesn't tell us enough, basically. You'd have to run the same tests on a whole battery of brands/wattages to reach any useful conclusions.

So whenever you want to get started on that, we'd appreciate it.

Edit: I re-read this and it sounded annoyingly critical, which isn't how I meant it. I think it's cool that you have the means and knowledge to do these tests. I'm jealous, too, and now want to test my lights to see if I made a good or bad buying decision...lol.

hmm, I have the apexcone 55W kit but, at least from my driver point of view, it doesn't seem brighter than oem lights out there. I know this could just be the crappy design of my projectors

What color are you using?

I'm not super familiar with the STI lamps, but you should notice a difference with the 55w lights in there. Have you looked on NASIOC for help? To really make the most of mine, I had to take some time and carefully re-aim all of the lights, including the fogs, and that's when I really noticed the difference.

Yeah, I took time to aim the headlights, really wide pattern, but not as intense as I would have thought compared to a 35w kit. Sometimes people with the 6000k headlights seem brighter, but that's probably cause the colour is just more noticable.

a lot of what you are noticing is the eye's non-linear response to light. it takes a pretty big increase in brightness to percieve a difference.

i have a techone 35watt kit if you want to test those.

I'll bring some gear the next time we meet up.

I was actually worried that I'd be testing the apexcone ballast unfairly so I kept the best result it scored (which happened to be with a prolumen bulb). I do agree the tests are not as complete as I would like. Hopefully I'll get a few more tests done over the holiday. Al and Patriki have offered their techone kits for measurement. we might be able to see the lot to lot variation there. I might also get a chance to measure a 3 year old OEM d2s setup AND a 1 month old setup. That should yield some info on bulb fading.

I can probably get my hands on a prolumen slim 35w kit for test but I don't know anyone around here with the apexcone slim kit.

There are a couple of problems with this comparison: one is you're using older bulbs and presumably ballasts (most Apexcone/DDM use the thin ballast now), you're testing 2 brands and 2 wattages and comparing them, and it's basically impossible to draw any conclusion from this other then not to get the particular Apexcone kit you used.

 

It's interesting information, it just doesn't tell us enough, basically. You'd have to run the same tests on a whole battery of brands/wattages to reach any useful conclusions.

 

So whenever you want to get started on that, we'd appreciate it.

 

Edit: I re-read this and it sounded annoyingly critical, which isn't how I meant it. I think it's cool that you have the means and knowledge to do these tests. I'm jealous, too, and now want to test my lights to see if I made a good or bad buying decision...lol.

I may also get you a set of our KWD slim ballasts and bulbs for testing if you are up to it

Hey Devin,

I did a quick optical test on the ones Sin picked up. The numbers look pretty good given the higher colour temperature bulbs he selected. I'll redo the tests after he runs the bulbs in a bit + when it is nicer out. The low ambient temps probably dropped the light output figures a couple percent.

His projectors are fantastic. I think the only better ones I've seen were on a TSX or s2k.

I have some photos. Will post them when I can.

the numbers are in the first post.

I just got and install HID conversion kit from XENTAC with 6000K color temp H7 bulbs and they are sweet. Easy plug and play install in 30 min. Only small mod was adding a hole in headlight bulb access cover for weather tight wire plug and disabling DRL and that was most of my work. Function and look great. Pics will follow. Price was not bad $60 plus shipping from EBAY.

PS I would NOT go with any New Slimmer HIDs Conversion Units that are on market since their bulb igniter are relocated on line, not in one unit, making them smaller. WOW !!! and because my friend have a bad experience with them.

one thing to note that doesn't jump out without a closer examination of the numbers is the amount of light that is lost going to higher colour temperature bulbs.

during the first round of tests, the apexcone kit did quite badly with its supplied 6k bulb. at the time I had assumed the bulb was crappy and proceeded to test the kit with the prolumen ~4.5k bulb. with the second set of tests on the 50w KDW ballasts + 6k bulbs showed that the light output is about the same as a 35w ballast with a 4.5k bulb.

the bulb colour temp seems to have quite a big impact on light output. possibly more than the advertised data would suggest. if you are on a budget and want light output, go with the lower colour temp bulbs. if you must have the bling bling look, then try and spring for the higher wattage ballasts.