HID Startup current: Stock 55 watt, DDM 55 watt, HID50 50 watt

September 2009

Introduction

The purpose of this experiment was to determine any detrimental effects that higher-powered HID ballasts may have on the Volvo S40 electrical system, which is designed to support two 55 watt H11 base halogen bulbs for low beam headlights. The CEM in this vehicle regulates the voltage going to the low beam to 13.0+0.2v using PWM modulation of a JFET, and monitors current flow presumably through a low value shunt resistor. JFETs are sensitive to reverse-voltage and EMF transients, and the CEM is an expensive component. For the JFET and CEM, the worst part of an HID ballast's operation is the startup, so I needed to figure out exactly what was going on.

I originally bought both a HID50 "50 watt" H11 kit, but then purchased the DDM "55 watt" H11 kit out of sheer curiosity for why one could cost $280 shipped and the other $85 shipped... Both came in relativity good time - HID50 kit took about 6 days to get here (VA) from across the pond, the DDM kit was delivered in 2 days from Cali. What shocked me the most was how similar the two kits were, they used essentially the same aluminum case. DDM used a embossed-logo casting for the top cover, and painted their ballast; while HID50 labeled theirs with a sticker and left the ballast with a glass-bead blasted finish. The weight of the two units is identical, 305 grams.

Experimental trials

For the electrical tests, there were four test subjects:

A OE, brand new Phillips 55 watt halogen H11 bulb - $11 local, so $22 for 2.
An old, beat up "Xentec" 35 watt HID unit - as a control. 3rd generation phillips knock-off. This puppy has lived on my bench as a worklamp for about three years after being ripped out of my last car which was totaled. Probably Aprox $160 at the time, for 2.
The HID50 50 watt HID unit. $280 shipped, for 2.
The DDM 55 watt HID unit. $85 shipped, for 2.

Procedure

Tests were prepared using a 50 amp shunt (1 mV = 0.27 amps) and a 20 mhz oscilloscope. The power supply is military surplus, 60 amps @ 12/24 volts, and is rock solid (more so than any bench supply). It was calibrated to 13.0 volts for the tests (what the S40's CEM is suppose to regulate voltage at). All tests used the same HID bulb, a broken-in (runtime > 100 hours) H7 rebase. The bulb was allowed to cool for 30 minutes before each cold test, sufficient for it to reach room temperature. Hot restart tests were cycled off-5 seconds-on. Here's a picture of the test setup:

Data

Each subject was tested from a cold start and from a warm start. The observation duration was 3.5 seconds. The collected data follows:

Trial	Cold	Warm
Halogen	Cold ~4.52a average (15a peek < 20ms)	Warm ~4.52a average (14.25a peek < 13ms)
Osram H11 55w
Control	Cold ~6.63a average (20a peek < 15ms)	Warm ~3.38a average (14.73a peek, < 30ms)
Xentec 35w
HID50	Cold ~5.1a average (10.91a peek < 25 ms)	Warm ~4.09a average (14.73a peek < 20ms)
50W Slim
DDM	Cold ~4.36a average (13.64a peek, < 20ms)	Warm ~3.95a average (14.93a peek < 20ms)
55W slim

Analysis:

The fuzzy bits* during the HID50/DDM cold starts are the ballast essentially "feathering the clutch" to bring the arc envelope up to temperature, without dumping too much current in. This is a feature of digital ballasts designed to extend bulb life. This means that both DDM and HID50 ballasts take significantly longer to warm up than the cheap Phillips knock off. The 'scope doesn't go wide enough to capture it, but we're talking 30-40 seconds VS 10-15 seconds.
The HID50 kit and the DDM kit are *very* similar in construction and operation. However it appears that their guts are not the same (see warmup frequency note).
From a warm start, the current draw characteristics for the OE bulb, the DDM and HID50 kits are very similar. The HID50 kit appears to generate less back EMF, but not quantifiably less.
From a cold start, both 50/55 watt HID kits draw significant current, over 10 amps for almost 250ms. This is not long enough to blow a 7.5 amp fuse.

* Frequency note The warmup modulation frequency of the HID 50 and DDM ballasts is different.

The HID50 ballast did this at 275 khz (25ms/div):	While the DDM ballast worked at 535 khz (50ms/div):

This indicates that some internal components (flyback inductor) and software are different.

Conclusion

This experiment did not test the quality, longevity, or durability of the kits, simply startup/warmup current characteristics. The data collected shows that both the HID50 and DDM HID kits may be safely operated using the stock headlight wires in a 2005 Volvo S40, on the original 7.5 amp fuse (each side). Due to the large amount of high-frequency noise, a filtering circuit is highly recommended: