Headlight PWM details?

theshadow27 · Sep 9, 2009

SwedeSpeed P1 (C30, S40, V50, C70) HID Retrofit Index

PLEASE do not revive this 2009 thread

This thread contains the original (c. 2009) investigation into the issues with P1 HID low-beam conversions. The Edited OP (c. 2011 - the real OOP was moved to post#2) was an older attempt to consolidate HID related info, and is included in the quote below only for posterity. If you are looking for the most accurate technical information on P1 HID retrofit issues, see the general HID discussion thread or https://www.skbowe.com/a/issues. Thanks!

<hr/>

theshadow27 said:
This post has information accumulated from a variety of sources including the statistics thread (here and on C30 World), and a variety of HID related threads (1, 2, 3, 4, etc...).

Section 1 - Symptoms and Problem with P1 (S40 2004.5+, V50, C30, and C70 2007+) HID conversions
Some of the common problems that arise with an HID conversion include:

Faliure to start - one or both bulbs

Flickering during operation

Dim output or excessive warmup time

Bulb Out Warning (BOW) message on the dash board LCD screen (DIM)

Windshield Wiper (WMM) failure

Any or all of these conditions can occur when attempting to install an HID kit in one of the applicable cars.

Section 2 - Cause of the Problems with HID retrofits
All of these problems come from one source. The computer module that controls the headlights, the Central Electronic Module (CEM) applies a pulse-width modulated (PWM) regulated voltage to the low beam halogen lamps to keep the voltage they see at 13.0V. It was designed to do this because halogen bulbs will last longer at 13V than the 13.5V or 14V at which the electrical system runs. Since the low beam bulbs are also the Daytime Running Lights (DRL) and are on all the time, Volvo wanted to maximize the life of the bulb. For halogen bulbs this is actually a good thing, and it does what it is supposed to.

The PWM works as follows: A halogen bulb is, in practice, a simple resistor. As such, the voltage across it is directly proportional to the current through it. By switching current on and off several times per second (about 80, actually), the CEM can control the average voltage across the bulb and thus keep it at 13.0V. The switching is not bad for the bulb, and the pulses are short enough (about 4/100th's of a second) that the human eye can't tell it's not a constant illumination.

We run into problems during an HID retrofit because of the HID ballast. What does the ballast do? An HID is an arc lamp which means there is no filament, but instead two contacts within a quartz envelope filled with xenon gas. Xenon is normally an insulator, meaning that it does not conduct electricity, but at a high enough voltage (between 800V and 2kV) the gas will ionize (just like a bolt of lightning), allowing electricity to flow between the contacts much more easily. We call this "lighting" or "igniting" the bulb. So the first job of the ballast is to ignite the arc.

If this high voltage was not removed, so much current would flow that the the contacts would melt and the capsule would shatter! So the second job of the ballast is to reduce the voltage to a more manageable level. HID bulbs are rated for a specific current (normally around 300mA) that they can operate at without damage. The ballast reduces the voltage to whatever level is required to keep this constant current flowing through the bulb.

When power is first applied to the ballast, it uses an inductor (in much the same way as the ignition coil sparks the spark plugs) to create the starting voltage across the bulb. Then, it gradually reduces the voltage until the constant current condition is met. To do this it uses a switch-mode power supply (SMPS) which is controlled by analog or digital feedback (this is where the "digital" in "digital ballast" comes from), that switches around 100KHz, or more. By measuring the current, it can switch on for longer or shorter to create the required average. This startup process takes about 3 seconds in most ballasts, although older ballasts can take much longer (30 seconds or more).

So there's a major problem when an HID ballast is hooked to the 80Hz PWM of the CEM. Every time the CEM switches the output on, the ballast begins it's startup procedure. When it shuts off, the ballast looses power.

Depending on the construction of the ballast and the duty cycle of the PWM (which depends on a number of factors, primarily battery voltage) the ballast might repeatedly try to start the bulb. Each time this happens, there is a large inductive flyback on the headlight ground - if you've ever touched the ground terminal on an old-fassion ignition coil while the car was running you know first hand - essentially a negative high voltage pulse. Normally this pulse would just go to ground, but it just so happens that the wiper motor (WMM) shares a ground wire with the headlights. Since there is an inherent resistance in the wire-chassies-battery connections, repetitive high voltage spikes can show up in the WWM and have caused many failures of a particularly sensitive component inside. This is the cause of symptom #4.

Other ballasts have a "warm start" condition in which they can tolerate short breaks in input power and continue operating so long as the gas remains ionized (about a second or so). These ballasts may work intermittently, again depending on the duty cycle of the PWM. This may seem OK, except that there is a filter circuit inside of the ballast required for proper operation of the SMPS. This circuit consists of an inductor and a capacitor, set up to "cover" the high-current spikes produced by switching, and allowing the ballast to draw a relatively steady current once running (see graphs). The filter is designed to deal with the operating frequency of the ballast, which is 10,000 times faster (or more) than the the PWM frequency of the CEM. In effect, subjecting it to low frequency modulation causes a large AC current to flow through the capacitor in the filter, several times more than it was designed for. This causes excessive heating and premature failure. The various stages of this degradation will cause the other symptoms listed above.

In all cases, when the ballast is no longer able to maintain the arc within the HID capsule, the capsule acts as an insulator, preventing current from flowing. This reduces the current that the ballast draws, which trips the BOW warning in the car.

Section 3 - How to fix the problem

There are three ways to get around these problems:

DRL disable does NOT disable PWM (just the DRL 1/2 brightness)<s>Disable the PWM modulation. This can be done by having a dealership flash the car to disable DRLs. This is the best option if DRLs are not a legal requirement where you live, and you don't mind the $80 for the dealer to do the flash. </s>

Bypass the CEM all together by using a battery harness with a relay. This gives the HID ballasts a constant DC voltage, since the relay stays switched on regardless of PWM modulation. The downside to this is that the relay does not draw as much current as the halogen bulb did, and the "bulb out warning" message will continue to be displayed.

Smooth the PWM modulation to approximate a DC voltage at the ballast. This is the cheapest solution that doesn't cause an error message, but because of the low frequency (80Hz) and large current (3A+), an extremely large (22,000uF or more) capacitor must be used. See here.

The third option is what developed from this thread (and the others). Essentially we designed a "bulb out warning eliminator" (BOWE) of which the first was built for SwedeSpeed member Kyle (ForceFed Motorsports), which for lack of better nomenclature became known a "KBOWE". The KBOWE consists of a large bypass capacitor and a diode, situated between the CEM and the HID ballast.

The capacitor serves to smooth the PWM voltage by shunting the AC components to ground. It's specification isn't supercritical, but it must have a ripple rating of at least 5A, which generally precludes sizes smaller than 22,000uF, and a voltage rating of at least 16V, but 25V or even 35V provide a better safety factor. Several smaller capacitors can be used in parallel if a larger unit is not available, provided they are all rated at least 16V.

The diode is situated to prevent current from flowing back to the CEM from the capacitor when the PWM switch is "off". While the exact mechanism is not known, empirical data suggests much better performance with the diode than without, and in some cases omitting the diode prevented the circuit from working at all. This is probably due to the current measurement mechanism within the CEM. The diode specification, again, is not critical, but it must be rated to at least 5A (10A or 15A provides a better MOS) and 16V or higher reverse voltage. A low forward voltage is also desirable, so most KBOWEs have employed Shottkey rectifying diodes.

The ground terminal of the headlight, CEM, and capacitor are all tied together and then directly to the chassis at the location of the KBOWE. Both the capacitor (which can absorb negative transients) and the grounding combine to eliminate the possibility of damage to the WMM.

One of these diode-capacitor assemblies is required per ballast. The most straightforward approach is to combine two assemblies in one box and tap it into the wires where they come out of the connector. This allows both capacitors to share a ground connection. Such an assembly is described in detail in this tutorial. Another approach with a slightly higher cost would be to create two identical units, hooking up between the vehicle wiring and the ballast at the headlight assembly, and grounding each independently.

The KBOWE has been extensively tested with over ten vehicle-years of operation since 2009 (as of 2011). The KBOWE is designed specifically for the challenges inherent to the P1 cars and, at the time this article was written, the only "adapter" type design that provides both HID functionality and WMM protection.

Section 4 - How not to fix the problem

A number of "solutions" have been proposed to allow retrofits to function normally. The majority users end up purchasing what have been called "BOW eliminators", "error cancelers", "harnesses", ect... They typically consist of a small box with two connectors, one male which plugs into the HID ballast, and one female which plugs into the power supply wires. For the sake of notation, we will call this class of solution the BOW-3, which is the model number of the product produced by DDM tuning.

The BOW-3 is essentially a beefed-up version of the filter inside of the HID. It contains another inductor and capacitor, to reduce the small 100KHz ripple imposed on a DC supply voltage. It works well with cars that (1) supply a DC voltage, and (2) are confused by this ripple.

The capacitor in the BOW-3 filter is about four times bigger than the capacitor in the ballast, and it has a correspondingly greater ripple (AC) current rating. However it's nowhere near big enough to deal with 80Hz PWM - and even if using a BOW-3 initially allowed an HID retrofit to work acceptably, most fail within a year or two because of this ripple current. About 75% of BOW-3 units work the first time, of those 65% fail within one year (sample size = 18 installs) and 20% take the WMM out when they fail. If you have a BOW-3 and it is still working, consider yourself on borrowed time.

Section 5 - Random Comments

Some users (6 of 44 recorded) have reported kits to function "perfectly" without any adapter, relay harness, or software flash. The majority of these users have "Kaixen" kits obtained from a group buy. Since I have not found an active group buy and the company does not seem to have a functional website, I can't obtain a kit for more rigorous evaluation. The Kaixen kits from various group buys (not just Volvo forums) have different styles of ballast, and each ballast style may have different components within them.

While there is no evidence to the contrary, I believe that (1) these kits are not designed to work with PWM supply voltage, and (2) that operating outside of design requirements will cause undesirable effects. Until I get my hands on a Kaixen kit I can't provide more information, but I do not recommend running any HID kit without one of the corrective measures listed above.

Jacob

theshadow27 · Sep 9, 2009

I have a set of hid50 50 watt HIDs on the way for my '05 T5, but before I think about hooking it up I'd like to know some more about the PWM circuitry in the CEM. I have been looking through the HID threads on here for 72 hours, and there's a lot about which different brands work with and without harnesses, dummy loads, frying the CEM, etc... but nothing about why they work or fry. According to the wiring diagram, pins E:25 and E:37 on the CEM (4/56) go directly into the logic, and are denoted by a JFET symbol which I assume means that they are highside switched. According to the <A HREF="http://theeshadow.com/files/S40MY2005/S40MY2005-Electrical.pdf" TARGET="_blank">electrical spec</A> (page 20), the headlights are regulated to 13.0 volts + 0.2 volts via PWM: <IMG SRC="http://theeshadow.com/files/volvo/hid_pwm.png" BORDER="0"> And the CEM monitors lamp faults.To me (without seeing the guts of the CEM) this means that inside there is: 1) A voltage monitoring feedback loop comparing output voltage with a fixed voltage source hooked a PWM modulator to maintain 13.0 volts. 2) A current shunt, probably a low value (<0.1 ohm) 5 or 10 watt resistor, with another comparator checking for current flow. Further, based on the nature of halogen lights: 3) There is no ripple reduction (inductors, capacitors, etc...) in the outputs because filaments don't care. 4) There is no reverse protection because there are no (significant) inductive components in a halogen bulb. On to the questions. Assuming the above is accurate (if it's not, please chime in!) then the things that could affect functionality are: A) HF/RF artifacts from low-quality ballasts may produce reverse current spikes in the line, and the JFET is not protected against such. B) The large initial current from the ignition pulse is not what the current shunt pickup is looking for (bulbs ramp up linearly, their resistance decreases as they warm up, and current increases as resistance decreases, until they reach equilibrium). This could, if the monitoring circuit was fast enough (< 200ms sample time, w/ poor ripple isolation maybe longer) be perceived as a short, and cause a shutdown/bulb malfunction indication. C) The stock halogen wiring is undersized for the current inrush (looks to be about 16ga, no?) and the 10-15 amp current in addition to the the 0.5-1.0volt drop between the CEM voltage meter and the HID will cause line voltage at the ballast to fall below it's low voltage cutoff, and prevent said ballast from sustaining an arc. D) The ballasts may not like unfiltered PWM, and lower quality units without additional filtering/isolation could blow up pretty quick. The reason I'd like to know this stuff is so I can design a circuit to counter any ill effects before I have to replace my CEM. There's three strategies that I can think of: I) Add a schottky diode in series with the halogen bulb output on the CEM, and a second schottky reverse-biased between the anode of the first and ground. On the anode side, add a few mil-spec high-temp 10,000 uF caps as filters. Then hook this filtered output to the HID. This assumes the capacitors could store enough charge before the ballast's under-voltage cutoff was exceeded to sustain the ignition current. In any case, it wouldn't hurt anything, the question is if it would help. Total cost: ~$4 / side II) Add a tuned inductor in series with the halogen bulb output on the CEM, picking a value based on the PWM frequency. This should take out most of the ripple, but the circuit described in (I) would most likely also be required after the inductor. Total cost: ~$6 / side III) If that is insufficient, the madness starts. Assuming that the CEM is totally and utterly incapable of providing the required current safely, we need to devise a system that guarantees the CEM sees 4.23 amps at all times. This would require intercepting the output from the CEM and isolating it with a diode. A P-type MOSFET connected to a +30 terminal after the diode would provide additional current if draw CEM draw exceeded 4.23 amps at any point. A N-type MOSFET, with the source connected at the P-type drain, and the drain connected to a 15-30 watt 1 ohm shunt hooked to ground, would be used if the current fell under 4.23 amps. All of this would need a +0.1% 0.1 ohm shunt resistor in series with the whole ordeal. Said shunt would drive a pair of comparators, or even a MCU with dual ADC's, to determine what FET was active at any given point. This circuit would also need an inductor before, and capacitors after. Like I said, worse case, but guaranteed to not damage anything ever. Total cost: ~$35 / side. Any thoughts from folks that have tried HID conversions? I'm not looking for "Oh, kit blah-blah-blah works fine", because that doesn't help answer the question. Much more useful would be what didn't work at first, and how you fixed it. Or if something broke, how and why. For the Volvo geniuses out there, any flaws in my logic? Electronic nuts - am I missing something?Thanks much, I love this car, and would rather run halogens than risk messing up the CEM. If I can't figure this out, I'm seriously considering getting the other engine harness and buying the OEM bi-xenons ($2K+ labor <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/frown.gif" BORDER="0"> )... Cheers, JSD 
Modified by theshadow27 at 6:48 PM 11-11-2009

Here's (rough) schematics for circuits I, II, and III <A HREF="http://theeshadow.com/files/volvo/hidcircuits.pdf" TARGET="_blank">http://theeshadow.com/files/volvo/hidcircuits.pdf</A>

theshadow27 · Sep 9, 2009

Re: (theshadow27)

Measured 82 hz, 12.02 volts at the bulb. I'm very suspicious of that measurement... PWM (for LEDs anyway) is normally in the 100-300khz range. If that's the real measurement, there isn't a big enough coil available to do solution II practically. That leaves I and III.

jdsr917 · Sep 9, 2009

Re: (theshadow27)

What LOL i do not understand not one thing on this page <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/confused.gif" BORDER="0"> <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/rolleyes.gif" BORDER="0">

theshadow27 · Sep 9, 2009

Re: (jdsr917)

Sorry, that's probably a bit hard-core for most non-Electrical-engineering folk <IMG NAME="icon" SRC="http://www.vwvortex.com/zeroforum_graphics/screwy.gif" BORDER="0"> Here's the cheat sheet:HID = High Intensity Discharge, a kind of lamp that uses an arc to create light instead of a filament. See <A HREF="http://en.wikipedia.org/wiki/High-intensity_discharge_lamp#High-intensity_discharge_lamps" TARGET="_blank">wiki</A>.HID Ballast = the box which supplies the necessary high voltage(s) and control current within the HID lamp (see <A HREF="http://en.wikipedia.org/wiki/Electrical_ballast" TARGET="_blank">wiki</A>.PWM = Pulse Width Modulation, a form of digital control used commonly by microprocessors to regulate voltage. See <A HREF="http://en.wikipedia.org/wiki/Pulse-width_modulation" TARGET="_blank">wiki</A>.CEM = Central Electronic Module, the brains in the S40. Basically a duel core computer. Dummy load = another name for a resistor that simulates a 55 watt headlamp bulb to trick the CEM into thinking one is connected. Converts 55 watts to heat. Pins E:25, E37 = the CEM pins that go to the left/right headlights in the S40. The diagram is taken from the wiring diagram.Lamp Fault = "bulb burnt out" indicator message. Triggered when the computer sees something wrong with the headlight outputs.JFET = A type of transistor that can be used for PWM. See <A HREF="http://en.wikipedia.org/wiki/JFET" TARGET="_blank">wiki</A>. Sensitive to reverse current fluctuations and ESD.Fixed Voltage Source = A special electronic component consisting of a fixed current source (see <A HREF="http://en.wikipedia.org/wiki/Current_source" TARGET="_blank">wiki</A> and a current-to-voltage converter (see <A HREF="http://en.wikipedia.org/wiki/Current-to-voltage_converter" TARGET="_blank">wiki</A> that can be used to regulate voltage by comparing it's generated voltage to the measured voltage.Current shunt = A low value resistor inserted in a circuit to measure current flow. DC current cannot be directly measured, but rather we measure the voltage drop across the shunt and extrapolate current via ohms law (see <A HREF="http://en.wikipedia.org/wiki/Ohms_law" TARGET="_blank">wiki</A>). Also see <A HREF="http://en.wikipedia.org/wiki/Current_sense_amplifier" TARGET="_blank">wiki</A>.Halogen light = The standard type of headlight that uses a bulb with a filament. See <A HREF="http://en.wikipedia.org/wiki/Halogen_light" TARGET="_blank">wiki</A>.Current ripple = A byproduct of PWM control (see the PWM wiki) where voltage drops between "on" modulations. See <A HREF="http://en.wikipedia.org/wiki/Ripple_(electrical)" TARGET="_blank">wiki</A>. Can cause problems on electronics which expect clean DC voltage.Reverse bias = A state where the voltage on the negative (ground) rail has greater potential than the voltage on the positive (12v+) rail. This can occur during an inductive flyback cycle (see <A HREF="http://en.wikipedia.org/wiki/Flyback_transformer" TARGET="_blank">wiki</A>) that the HID ballast uses to step 12 volts up to 20,000v or higher.HF/RF = High Frequency and Radio Frequency noise, AC currents that ride over the DC power wires running between the headlights and the CEM, caused by the ballast flyback cycle.Monitoring time = The CEM (or any computer for that matter) can't monitor values all the time. Instead, it uses an A/D (analog-digital) converter (see <A HREF="http://en.wikipedia.org/wiki/A/D" TARGET="_blank">wiki</A>) to measure values every so often. The longer the monitoring time, the less chance it has a picking up the current spike created when the HID ignites the arc.Voltage drop = Although they are conductors, wires have resistance, and longer wires have more resistance. High current flow through small wires will drop the voltage measured at each end, sometimes significantly. This is why battery cables are thicker than audio cables, for example. See <A HREF="http://en.wikipedia.org/wiki/Voltage_drop" TARGET="_blank">wiki</A>Blow up = In point D. I mention blowing up. This is caused by resonant feedback (like audio feedback, <A HREF="http://en.wikipedia.org/wiki/Audio_feedback" TARGET="_blank">wiki</A> between the oscillator driving the CEM PWM circuit and the oscillator driving the ballast transformer circuit. This will create alternating high and low current spikes which easily exceed component ratings and cause said components to explode. Filtering = Various electronic circuits designed to reduce current ripple. Schottky = A type of diode (see <A HREF="http://en.wikipedia.org/wiki/Diode" TARGET="_blank">wiki</A>) that has a low forward voltage, and a high current capacity. See <A HREF="http://en.wikipedia.org/wiki/Schottky_diode" TARGET="_blank">wiki</A>. A diode lets current flow only one way.Series = A particular configuration of electronic components within a circuit. See <A HREF="http://en.wikipedia.org/wiki/Series_circuit" TARGET="_blank">wiki</A>.Capacitor = An electronic component which stores electrical current, and can release it rapidly. See <A HREF="http://en.wikipedia.org/wiki/Capacitor" TARGET="_blank">wiki</A>. Capacitors are commonly used to reduce ripple.Tuned inductor = An inductor is an electronic component commonly known as a coil. It also stores electrical current, but in a different way as a capacitor. See <A HREF="http://en.wikipedia.org/wiki/Inductor" TARGET="_blank">wiki</A>. They can be used to filter a specific frequency of ripple very effectively, selecting the correct inductance value for a particular frequency is called tuning. The part III is basically saying "Use the CEM to power the light, BUT if the ballast wants to take more current than it's expecting, instead connect straight to battery, OR if the ballast wants to draw less current than it's expecting, hook up a resistor as a shunt". This will trick the CEM into thinking there's a bulb connected, but not waste 55 watts as heat. <IMG NAME="icon" SRC="http://********************/smile/emthup.gif" BORDER="0">

I Saw 50 · Sep 10, 2009

Oh jeez that is some in depth stuff! Way beyond my curse at electricity when it shocks me knowledge. But judging from that, and not to side track you or anything, but would that be able to be done backwards?For vehicles already with the HID lights except for the ones with the DRL's removed they are constantly on. Would some sort of relay be able to shut the lights off and on based on the photocell on the dash and the car's day/night scheme?Just thinking, don't really know the inner workings of the modules.

theshadow27 · Sep 10, 2009

Re: (I Saw 50)

Unfortunately no. The factory Bi-Xenon system is significantly more complicated (hard to believe, but it is) than the halogen system, so much so that it requires a completely different engine harness. In the stock Bi-Xenon system, the CEM is connected to the headlights by a single wire over the (EDIT:GDLs run off LIN direct from the CEM) LIN bus (low speed in-car network). This wire sends the "GDL Master" module (passenger side headlight computer) the signal to turn off/on and low/high beam. The "GDL Master" module relays this information to the "GDL Slave" (driver side headlight computer) over another wire using the LIN protocol. So basically the signals for the factory Bi-Xenons are all computer codes over the car's network. The day/night sensor a photodiode connected directly to the CEM, so intercepting it would be difficult. Even with that sensor though, the most you could do is cut all power to the GDL modules. This would, with overwhelming probability, cause a warning code to be stored and a check engine light. The CEM isn't very fond about loosing modules.I guess it'd be easier to take it into the dealership. No easy outs for less than the $70 or whatever they charge. Won't start this here, but I'd personally never take out the DRLs, especially in Xenon, they look so f-ing cool. Modified by theshadow27 at 7:09 PM 9-9-2009 
Modified by theshadow27 at 7:09 PM 9-9-2009

I Saw 50 · Sep 10, 2009

Aaand my eyes almost went cross eyed. But that's some real good info. Shame there's no real workaround except for the dealer. I've already had two bulbs go out on mine.

theshadow27 · Sep 10, 2009

Re: (I Saw 50)

For what it's worth, here are the schematics for the two systems: Halogen: <A HREF="http://theeshadow.com/files/volvo/3978202_58.pdf" TARGET="_blank">http://theeshadow.com/files/volvo/3978202_58.pdf</A> Bi-Xenon: <A HREF="http://theeshadow.com/files/volvo/3978202_59.pdf" TARGET="_blank">http://theeshadow.com/files/volvo/3978202_59.pdf</A>

-ForceFed- · Sep 10, 2009

Re: (theshadow27)

Kudos for breaking this down but i feel at this point you might as well just buy a car with factory Bi-xenons before taking the time to rewire the car for them. Ive never looked that in depth as to why the failures are occuring but your facts are pretty solid. Has there actually been a recorded CEM failure due to HID's, i had only heard of the WMM?

theshadow27 · Sep 10, 2009

Re: (ForceFed Motorsports)

The WMM? The Wiper Motor Module?I have a 2005 T5 AWD with the M66 and sport package. This is my second (2005 T5 AWD M66), the first was totaled in an accident this spring. The problem is, when I went to find another T5 AWD with M66, there were 2 on the east coast. Volvo apparently only made 6 speed AWDs from 2005-2007, before they figured that anyone with enough cash to buy an AWD didn't want a MT. Now they don't sell any (T5, T5 AWD, or 2.4i) 6 speeds in the US <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/mad.gif" BORDER="0"> Trust me, if I could go down to the dealer and pick up a '09 or '10 T5 AWD in 6 speed manual and Bi-Xenons, I'd do it in a heartbeat. But I can't, so I'm stuck with this one.

fone · Sep 10, 2009

Re: (theshadow27)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by theshadow27 »</TD></TR><TR><TD CLASS="quote">Measured 82 hz, 12.02 volts at the bulb. I'm very suspicious of that measurement... PWM (for LEDs anyway) is normally in the 100-300khz range. If that's the real measurement, there isn't a big enough coil available to do solution II practically. That leaves I and III. </TD></TR></TABLE>Wiki link says 100 to 120 Hz? Not sure if the PWM requency differs from car to car, could be the reason why some cars are able to fix the HIDs directly, while others needed the DRLs disabled.Heck, I'm an Electrical Engineer and I only understand half of your post. Not even sure if my low end (free) multimeter even measures frequency.<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by theshadow27 »</TD></TR><TR><TD CLASS="quote">C) The stock halogen wiring is undersized for the current inrush (looks to be about 16ga, no?) and the 10-15 amp current in addition to the the 0.5-1.0volt drop between the CEM voltage meter and the HID will cause line voltage at the ballast to fall below it's low voltage cutoff, and prevent said ballast from sustaining an arc. </TD></TR></TABLE>From my limited EE knowledge, a 55W HID bulb should only take as much, if not less, current as a standard 55W bulb? The BOW harness simply makes up for the differences in the current (ie additional large resistor in parallel) between a 55W HID and the stock halogen so that the CEM thinks the bulb is not faulty. I have not checked the voltage / current that much to know the variations, but a 55W HID on a line rated for 55W at the same voltage, should be able to handle the currents involved.I would vote for solution I, but won't be able to use it on my setup as I'm using a battery harness, aka relay switch.<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote »</TD></TR><TR><TD CLASS="quote">II) Add a tuned inductor in series with the halogen bulb output on the CEM, picking a value based on the PWM frequency. This should take out most of the ripple, but the circuit described in (I) would most likely also be required after the inductor. Total cost: ~$6 / side</TD></TR></TABLE>Shouldn't a capacitor do the trick? I'm not too familiar with inductor. ie smooth out the ripple effect of the PWM? 
Modified by fone at 8:50 PM 9-9-2009

theshadow27 · Sep 10, 2009

Re: (fone)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> Wiki link says 100 to 120 Hz? Not sure if the PWM requency differs from car to car, could be the reason why some cars are able to fix the HIDs directly, while others needed the DRLs disabled. </TD></TR></TABLE>Different PWM frequencies are used for different applications. For example, DC-DC converters work at ~100khz (100,000 cycles/second) and fuel injectors PWM has to be at least 20khz if not higher, to keep up with intake valve operation. Lower frequencies are suitable for incandescent lights, since the human eye can only pick up frequencies lower than ~30hz. In any case however, the frequency is fixed by the MCU's clock, which is driven by a crystal oscillator. Unless they change the speed of the CEM processor (not likely during a production run) the PWM frequencies will be the same.I doubt the PWM frequency has anything to do with using the stock harness for HIDs vs disabling DRLs. Even with DRLs disabled, the low beam is still PWM modulated (when on). I suspect it has more to do with the HID kit, and said kit's ability to handle transient voltages. <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> From my limited EE knowledge, a 55W HID bulb should only take as much, if not less, current as a standard 55W bulb? The BOW harness simply makes up for the differences in the current (ie additional large resistor in parallel) between a 55W HID and the stock halogen so that the CEM thinks the bulb is not faulty. I have not checked the voltage / current that much to know the variations, but a 55W HID on a line rated for 55W at the same voltage, should be able to handle the currents involved. </TD></TR></TABLE>Theoretically you're right. A 50 watt HID *should* draw the about the same current in continuous operation as a 55 watt halogen. The difference is not continuous operation, it's the initial startup. When a bulb starts up, it draws very little current at first (~2 amps), then ramps up slowly as the filament warms up (~200 ms for this to happen), and eventually gets to 4.2 amps. A 50 watt HID runs steady-state arc at around 700 volts, 3-5ma (milliamps). The problem is to get there, the ballast has to ignite the arc within the HID capsule. This requires ~7,000-15,000 volts depending on the quality of the ballast (higher ignition voltage=better ballast=more reliable starts). The ignition pulse is created using a flyback transformer (like an ignition coil), and if the ballast is poorly designed this will send a transient spike down the power leads. It will also draw 20-30 amps for the first few milliseconds while it strikes the arc. After all that madness is over, you are right, the wiring will be fine. It's the initial spike I'm worried about. I dislike the idea of the parallel resistor ("BOW harness") since that's essentially running a 110 watt heater in the engine compartment. This would result in a 100% increase in total current consumption, and HID's are suppose to be more efficient, not less. <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/wink.gif" BORDER="0"><TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> I'm using a battery harness, aka relay switch. </TD></TR></TABLE>Out of curiosity, did you get a bulb-out indicator? <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> Wiki link says 100 to 120 Hz? Not sure if the PWM requency differs from car to car, could be the reason why some cars are able to fix the HIDs directly, while others needed the DRLs disabled. </TD></TR></TABLE>Different PWM frequencies are used for different applications. For example, DC-DC converters work at ~100khz (100,000 cycles/second) and fuel injectors PWM has to be at least 20khz if not higher, to keep up with intake valve operation. Lower frequencies are suitable for incandescent lights, since the human eye can only pick up frequencies lower than ~30hz. In any case however, the frequency is fixed by the MCU's clock, which is driven by a crystal oscillator. Unless they change the speed of the CEM processor (not likely during a production run) the PWM frequencies will be the same.I doubt the PWM frequency has anything to do with using the stock harness for HIDs vs disabling DRLs. Even with DRLs disabled, the low beam is still PWM modulated (when on). I suspect it has more to do with the HID kit, and said kit's ability to handle transient voltages. <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> From my limited EE knowledge, a 55W HID bulb should only take as much, if not less, current as a standard 55W bulb? The BOW harness simply makes up for the differences in the current (ie additional large resistor in parallel) between a 55W HID and the stock halogen so that the CEM thinks the bulb is not faulty. I have not checked the voltage / current that much to know the variations, but a 55W HID on a line rated for 55W at the same voltage, should be able to handle the currents involved. </TD></TR></TABLE>Theoretically you're right. A 50 watt HID *should* draw the about the same current in continuous operation as a 55 watt halogen. The difference is not continuous operation, it's the initial startup. When a bulb starts up, it draws very little current at first (~2 amps), then ramps up slowly as the filament warms up (~200 ms for this to happen), and eventually gets to 4.2 amps. A 50 watt HID runs steady-state arc at around 700 volts, 3-5ma (milliamps). The problem is to get there, the ballast has to ignite the arc within the HID capsule. This requires ~7,000-15,000 volts depending on the quality of the ballast (higher ignition voltage=better ballast=more reliable starts). The ignition pulse is created using a flyback transformer (like an ignition coil), and if the ballast is poorly designed this will send a transient spike down the power leads. It will also draw 20-30 amps for the first few milliseconds while it strikes the arc. After all that madness is over, you are right, the wiring will be fine. It's the initial spike I'm worried about. I dislike the idea of the parallel resistor ("BOW harness") since that's essentially running a 110 watt heater in the engine compartment. This would result in a 100% increase in total current consumption, and HID's are suppose to be more efficient, not less. <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/wink.gif" BORDER="0"><TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by fone »</TD></TR><TR><TD CLASS="quote"> Shouldn't a capacitor do the trick? I'm not too familiar with inductor. ie smooth out the ripple effect of the PWM? </TD></TR></TABLE>If you look at the PDF above ( <A HREF="http://theeshadow.com/files/volvo/hidcircuits.pdf" TARGET="_blank">http://theeshadow.com/files/volvo/hidcircuits.pdf</A> ) you'll see all three circuits have capacitors. Inductors are used commonly for filtering (look inside a UPS, or a computer power supply, or a TV, etc...) in the form of an RLC circuit (see <A HREF="http://en.wikipedia.org/wiki/RLC_circuit" TARGET="_blank">wiki</A> is pretty universal and can be tuned to a specific frequency. I ordered the caps and diodes for method I. If I can get a decent scope trace after that circuit, I'll try the HIDs.

fone · Sep 10, 2009

Yes, I have a bulb out indicator, but that would be the least of my concerns as I'm not going to put a resistor there that draws almost 55W by itself, wasting too much energy as another 55W is drawn directly from the battery already.

theshadow27 · Sep 10, 2009

Re: (fone)

That would drive me nuts, but I'm <IMG NAME="icon" SRC="http://www.vwvortex.com/zeroforum_graphics/screwy.gif" BORDER="0"> about stuff like that (obviously). I ordered 10 of each component (should only need 4) so if it works I'll send you the other set.

fone · Sep 10, 2009

I'm actually fine with my current set up. Also, you might want some of the standard HID harness to make the set up plug and play. I might have some spare ones from the damaged ballasts, if you would pay for shipping.

theshadow27 · Sep 10, 2009

Re: (fone)

Hmm... I was planning on mounting the electronics inside of the CEM's enclosure, and mounting the HID ballast to the plastic cover plate behind the headlamp. Then it's just two holes in a $5 piece of plastic, and no wire tapping or cutting. I'll let you know if something comes up. Until they get here though, I guess there's nothing else to do but wait <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/frown.gif" BORDER="0">

theshadow27 · Sep 17, 2009

Re: (theshadow27)

I bought both a <A HREF="http://www.hid50.com/" TARGET="_blank">HID50</A> "50 watt" H11 kit and the <A HREF="http://www.ddmtuning.com/index.php?p=product&id=101&parent=60" TARGET="_blank">DDM "55 watt" H11 kit</A> out of sheer curiosity for why one could cost $280 shipped and the other $85 shipped... I'll leave the results for another thread, but for the record I installed the DDM kit. All factory wiring from the headlight assemblies to the CEM. As I suspected, the PWM modulation from the CEM's voltage regulation prevented the HIDs from working at all. With just the DDM components installed, neither bulb turned on and there was a "low beam failure" message. Also, as I suspected, a simple capacitor ripple filter was sufficient to get things cooking again. Here's the circuit I used: <IMG SRC="http://theeshadow.com/files/volvo/capfilter.jpg" BORDER="0"> Each ballast got 5 capacitors and 1 diode. Total cost of parts was $16.30. You can do it for cheaper, but I got all industrial rated components (105c operating temperature). It's possible that 50,000 microfarads is a bit excessive, but it did the trick without breaking a sweat. The capacitors being assembled: <IMG SRC="http://theeshadow.com/files/volvo/capfilter1.jpg" BORDER="0"> With diode board installed: <IMG SRC="http://theeshadow.com/files/volvo/capfilter2.jpg" BORDER="0"> Not the smallest thing in the world, but not that bad either: <IMG SRC="http://theeshadow.com/files/volvo/capfilter3.jpg" BORDER="0"> Here's how I tapped into the headlight wires. It's the connector labeled "Front" on the CEM. The wires are green-violet (passenger) and green-white (driver). Be careful - there's two green-white wires. The one next to the green-violet wire is the correct one. Or just flip it over, the pin number are 25 and 37. <IMG SRC="http://theeshadow.com/files/volvo/capfilter4.jpg" BORDER="0"> Encapsulated it in silicone, wrapped in electrical tape, tucked it in behind the passenger carpet, and voilà: <IMG SRC="http://theeshadow.com/files/volvo/hids1.jpg" BORDER="0"> Starts and stops with no issues. Flashes pretty fast too <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/embeer.gif" BORDER="0">

ion · Nov 10, 2009

Re: (theshadow27)

Did you replace this with the Kyle BOW?I was considering buying the 55w kit you linked to, but I'm wondering if DDM tuning's "error code eliminator" works. Did you also buy/try their "error code eliminator" or did you just work on making your own from the beginning.Also, where'd you get your capacitors from? Digikey is always expensive and I'm hoping you know a better/cheaper site considering you probably buy stuff like this all the time.

-ForceFed- · Nov 10, 2009

Re: (ion)

Id recommend shadows PWM fix rather than the DDM BOW harness if it works with your setup. Its a more proper system than the bow harness which is kinda a band aide.

Headlight PWM details?

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered