Headlight PWM details?

theshadow27 · Nov 10, 2009

Re: (ion)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote">Did you replace this with the Kyle BOW?</TD></TR></TABLE> Nope, my ghetto prototype is still working. I was thinking about building another one but (1) I'm too lazy and (2) if it fails (not expected) I'd rather not have to cut through silicone to figure out why. <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote"> 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. </TD></TR></TABLE> Both of the kits I bought didn't work without the caps.I didn't buy the error code eliminator as it was $40 at the time and I'd heard all sorts of bad things about them (like most of the DDM failures were the BOWE not the ballast) and I could build it cheaper. Looks like they are $20 now, but mine was $16 or something and Kyles was $14 in parts. Kinda nice to know the quality rather than guessing. If you are ok with 16v caps rather than 25v you can build it for $8.But they do work. A hand full of folks are running the DDM BOWE harness without issues, and a few others have had units replaced. My biggest problem with the DDM version is that it lives in the engine bay, which is not the best place for electrolytic capacitors. Also not sure if they have a diode, waiting on the one from Kyle to show up so I can rip it up. It should be here later this week, so if you want to wait until then I'll update this thread. The other thing DDM is lacking is a proper ground, which effectively introduces more resistance and drops the voltage available to the ballasts. So basically <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ForceFed Motorsports »</TD></TR><TR><TD CLASS="quote">bow harness is a band aide.</TD></TR></TABLE> Is correct<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote">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.</TD></TR></TABLE> I used 10k uF (pardon the SI annihilation, nobody measures in mF) caps in the original because I wasn't sure how much capacitance would be required. One worked intermittently so I decided to use all 10 of them... Not very scientific. After mine were installed I ran some SPICE simulations and it looks like 22,000uF is the sweet spot for 55w. Mouser is the best bet for high-value caps. The capacitors I used in Kyle's were $5 each (598-80LQ223M025K022) and I bought the last 3 <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/rolleyes.gif" BORDER="0">... After I ordered I was looking over the spec sheet and the 16v are transient rated to 30v so they could be used instead (598-SLPX223M016E3P3) and are only $2.36 each. The diodes were $1.30 from Digikey (497-2738-5-ND) but I've been eying the Mouser equivalent which is $1.20 (511-STPS8L30B-TR). The original capacitors I used were $1.65 from Digikey (493-1546-ND), and you'd only need two or three to get up to the required capacitance on each side. It would be about twice as expensive and take up more room.

theshadow27 · Nov 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"> After mine were installed I ran some SPICE simulations and it looks like 22,000uF is the sweet spot for 55w. </TD></TR></TABLE>BTW in case anyone is curious why the DDM's fail, the measurements I got from fone (thanks again m8 <IMG NAME="icon" SRC="http://********************/smile/emthup.gif" BORDER="0">) indicate that it's around 4.7mF. I was able to play with one out of 1SOUL's car in person and his was about the same (3.3mF). Here's what happens 82 times a second using the DDM BOWE:<IMG SRC="http://theeshadow.com/files/volvo/ddmproblem.png" BORDER="0">You can see the capacitor trying to hold the voltage up (blue), but it's just not enough. It dips below the 9v under-voltage cut off for 2ms each pulse. Also, the ripple current through the capacitor is a constant 6.5 amps! I've never seen a 4700uF cap rated for 6500mA ripple, maybe 2000mA, but that's way out of realistic spec. Add that to a hot engine bay and you can't expect an aluminum cap to last it's designed life.When we increase the value of the capacitor: <IMG SRC="http://theeshadow.com/files/volvo/ddmproblem2.png" BORDER="0"> the issue goes away. Also, ripple is reduced to 3000mA, the 22mF caps Kyle got are rated to 4300mA. Plus they aren't in the engine bay, so heat is significantly less. In all, they should last much longer...

ion · Nov 11, 2009

Re: (theshadow27)

Do you have the parameters for the input signal? Frequency, voltage? I'd like to mess around with some spice simulations of my own. The DDM ballasts just want a steady 13V, right?

theshadow27 · Nov 11, 2009

Re: (ion)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote">Do you have the parameters for the input signal? Frequency, voltage? I'd like to mess around with some spice simulations of my own.</TD></TR></TABLE> Input signal is PWM between 0v and the car voltage - 13.5 nominalFrequency I measured at 82hz, square waveDuty cycle varies between 70% and 90% depending on input voltage. Current for the load is ~4.2 amps, best simulated with a 3 ohm resistor. The real ballast has it's own switching supply which makes it non linear so it's not a perfect replacement.According to Volvo the output is suposed to be regulated to 13.0+0.2v. I measure it consistently less than that, around 12.5, but who knows my meter could be off. The ballasts will run from 9v to 16v without issues, below and above they have cutouts. That's why dropping below 9v is a bad thing. The goal is minimizing ripple, even if you loose a volt or two it's not a big problem. FYI I tried a bunch of configurations: multiple caps, ceramic and electrolytic, pi filter, delta filter, inductors, etc... Nothing worked as well as a single large aluminum electrolytic and a diode.

theshadow27 · Nov 11, 2009

Re: (theshadow27)

Okay, got the broken <A HREF="http://www.ddmtuning.com/products/Error_Code_Eliminator_3_PAIR-130-94.html" TARGET="_blank">DDM Error Code Eliminator 3</A> from Kyle today (thanks!): <IMG SRC="http://theeshadow.com/files/bow2/ddm_error_code_eliminator1.jpg" BORDER="0">Let's go surgical eh? <IMG SRC="http://theeshadow.com/files/bow2/ddm_error_code_eliminator2.jpg" BORDER="0">I was pretty surprised to find an inductor in there! It's a T-80 iron powder toroid with 22 windings of 16 gauge wire. If it were one wire, according to some extremely rusty math that comes out to about 1uH. However it looked to be split in two windings, and I'm not sure exactly how it was wired. The capacitor (which was across the positive and negative, only way to hook it up) was EXACTLY what fone measured, 4700uF. Nice. It's also a decent brand and rated to 105c (another surprise!). <IMG SRC="http://theeshadow.com/files/bow2/ddm_error_code_eliminator4.jpg" BORDER="0">Couple of hunches on how the circuit is rigged up, my first guess was this: <IMG SRC="http://theeshadow.com/files/bow2/ddm_error_code_eliminator3.jpg" BORDER="0"> but the more I think about the dual winding the less I'm sure about it. No easy way to tell now <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/biggrin.gif" BORDER="0"> EDIT: Some more looking into the bits that fell off it. Thinking now more along the lines of a inductive <A HREF="http://en.wikipedia.org/wiki/Ladder_network#Ladder_topologies" TARGET="_blank">t filter</A>: <IMG SRC="http://theeshadow.com/files/bow2/ddm_error_code_eliminator5.jpg" BORDER="0"> 
Modified by theshadow27 at 7:17 PM 11-11-2009

-ForceFed- · Nov 12, 2009

Re: (theshadow27)

Interesting, so im assuming the only reason they would actually fail would be due to the heat of the engine bay... being that they actually use decent materials?

theshadow27 · Nov 12, 2009

Re: (ForceFed Motorsports)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ForceFed Motorsports »</TD></TR><TR><TD CLASS="quote">Interesting, so im assuming the only reason they would actually fail would be due to the heat of the engine bay... being that they actually use decent materials?</TD></TR></TABLE>Oh sorry no - I didn't mean that at all. The circuit is constructed correctly, but is not appropriate for our car.With a 4700uF capacitor and a diode there would be 6.5A of AC current ripple running through the cap at all times. If you look at <A HREF="http://www.farnell.com/datasheets/79228.pdf" TARGET="_blank">the datasheet</A> the 25v 4700uF cap isn't even ripple current rated! The 80v version is rated to 3A, so the 25v is probably around 2.5a, so when the car is at 13.5v it's getting 250% of it's rated ripple current. Looking at the tan delta coefs it'd be running at close to 150c without the engine bay temps affecting it at all. Being coated in epoxy doesn't help heat get out.The 22000uF <A HREF="http://www.cde.com/catalogs/SLPX.pdf" TARGET="_blank">caps you got</A> are rated to 5.3A ripple continuous. The higher value capacitance reduces ripple so there is only 3A of actual ripple current. It's running at 60% of it's rated value, so heat dissipation is minimal.Besides the caps, there's a couple of other problems with their circuit. First, it grounds through the headlight wires, which I've measured the resistance of (including connectors) to be around 0.3 ohms. This doesn't sound like much, but it constitutes a 1.2 volt drop at operating currents. Effectively, their cap is working with 11V which makes everything worse (capacitors store charge exponentially), where as ours are working with 13 since the wires to the CEM are huge. Second, there's no diode in the DDM box. It's pretty hard to explain why this is necessary, but try to think of a capacitor as a tank of water with a fill and a drain at the bottom. The drain goes to the HID, and the fill goes to the CEM. The CEM feels the need to maintain the level of water in the tank by hooking a garden hose up to the fill. The level goes up, and after a certain point the CEM removes the hose. Well water still goes out the drain, but it also goes out the fill because there's no hose connected. The diode is a one way valve that stops water from back washing out the fill when the CEM disconnects the hose. (jdsr917 - I tried, I really did)Third the inductor is a great idea, but is way too small for our PWM frequency of 82hz. As I went over in the original posts, you'd need a huge inductor (several pounds) to dampen such a low frequency significantly. 1uH is nowhere close, it'd take more like 300mH (someone check my math, it's up there though).To summarize, if we had 10KHz PWM the DDM box would work perfectly. Even 1KHz it would probably be okay with. But 82Hz? We're lucky it works at all! 
Modified by theshadow27 at 6:42 PM 11-11-2009

theshadow27 · Nov 12, 2009

Re: (ion)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote">How do you know it was the error code eliminator that failed? My guess would be that the failures are a result of poor assembly (e.g., bad solder joints).</TD></TR></TABLE> Because when one BOWE fails you can move the other to the failed side and it will work. Also, both my car and Kyles now work with my circuit and without the DDM BOWE's, but neither will run without a BOWE. The reasons for failure are outlined above, which is I'm guessin why you deleted the post. Sorry buddy I got the email from SS <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/tongue.gif" BORDER="0"> Don't subscribe to many threads, but I was pretty sure nobody would ever reply after the OP so I had to keep an eye on it lol

ion · Nov 12, 2009

Re: (theshadow27)

lol, and here I thought I got away with it.Your capacitor current rating explanation makes sense, but I think the circuit might be acceptable if the ballasts have circuitry to handle high frequency PWM input. One possible arrangement of the inductor and capacitor can simply increase the frequency of its pulse input. It'd be interesting to see what exactly the DDM can run off of: maybe by hooking it up to a supply and signal generator.Anyways, I ended up buying the DDM eliminators just because I think their solution looks cleaner. If they fail, maybe I'll just get better capacitors for them (if the capacitors are what caused them to fail).

theshadow27 · Nov 12, 2009

Re: (ion)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote"> One possible arrangement of the inductor and capacitor can simply increase the frequency of its pulse input.</TD></TR></TABLE>Passive components can change amplitude and phase, but they can not change frequency. At the very least it takes a few diodes to double frequency once, and even then it's not rated for anything more than a few mA. The headlights run at 4.23 amps each. Not sure where you're going with that, but by the time you build a circuit to increase the frequency of the PWM to what the internal filter capacitors (arranged like mine, just smaller) in the ballast can handle you've far exceeded the cost of my circuit and the DDM...<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote"> Anyways, I ended up buying the DDM eliminators just because I think their solution looks cleaner. If they fail, maybe I'll just get better capacitors for them (if the capacitors are what caused them to fail).</TD></TR></TABLE> A lot of people are running the DDMs without issue, but there's also a good number that fail. I agree that it is a "cleaner looking" solution.I wish there was a way to correctly do it the way DDM did with an in-line box, but there's not. I'm working with a company called Retro Solutions to come up with a commercial variant that looks less DIY, but it will at the very least require a real ground tap. Not sure how you would replace the capacitors in the DDM units, the whole thing is potted in epoxy. If it came apart easily I wouldn't have attacked it with a saw to take pictures. If you somehow manage to get it apart, you might as well start from the beginning and do it correctly with bigger caps and a diode...

ion · Nov 12, 2009

Re: (theshadow27)

LC circuits are commonly used as oscillators. I mean yeah... the output is no longer a square wave (an oscillation at each edge that lasts the duration of the pulse), but that doesn't really matter, it's a higher frequency signal that the ballast might (apparently can) process.As for replacing the capacitors, you said "semi-rigid". Is it kinda rubbery or completely solid? I was thinking maybe it could be heated up and scooped out or something. Might not be pretty, or worth it, I didn't really think it through.

theshadow27 · Nov 12, 2009

Re: (ion)

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote">LC circuits are commonly used as oscillators. I mean yeah... the output is no longer a square wave (an oscillation at each edge that lasts the duration of the pulse), but that doesn't really matter, it's a higher frequency signal that the ballast might (apparently can) process.</TD></TR></TABLE> I see where you're going, but a filter and an oscillator are two entirely different ball games. A pulse can induce an LC circuit into oscillation, but it's at very small voltages and currents. To get a decent 4ms long ring-down that could supply 4.23A would take a... I'm not even going to try to do the math... Godzilla capacitor and a Stay Puft Marshmallow man inductor. Plus it'd have to take voltage up to 24 volts (since it's a DC biased AC signal). The only practical way to use oscillation to increase the frequency between modulations would be to use a LC behind a MOSFET. If you're going that far, you might as well use an RC behind the MOSFET so that it gave a solid current rather than a rapidity pulsating one. At that point you have transistors and resistors and capacitors and circuit boards and a +12v source and it's just all not worth it. <TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by ion »</TD></TR><TR><TD CLASS="quote"> As for replacing the capacitors, you said "semi-rigid". Is it kinda rubbery or completely solid? I was thinking maybe it could be heated up and scooped out or something. Might not be pretty, or worth it, I didn't really think it through.</TD></TR></TABLE> Yeah semi rigid epoxy. That's the stuff that if you put in a vice and squeeze it won't crack (tried that) but it's completely solid. There's no good way to remove it - epoxy is a polymer reaction so heat won't matter, and it's probably impervious to most solvents other than 40% HCL which would destroy everything inside anyhow.

-ForceFed- · Nov 12, 2009

Re: (theshadow27)

Being that my BOWE harness is sitting above the CEM id say thats pretty clean <IMG NAME="icon" SRC="http://www.vwvortex.com/vwbb/wink.gif" BORDER="0">

theshadow27 · Nov 12, 2009

Re: (ForceFed Motorsports)

I think the DDMs are "cleaner looking" outside the car since they have non-destructive connectors. Installed, the DDMs have an unappealing aftermarket look to them, where as mine if done correctly look to be part of the stock harness until you remove the connector. No boxes dangling in the engine bay. I'll try to get a pic of where I mounted my ballasts, it's ridiculous.

-ForceFed- · Nov 12, 2009

Re: (theshadow27)

How our fenders roll inward water just gets dumped on the top of our head lights, so I cant mount my ballasts like factory on the back of the head lights I dont think. Probably make an extension harness and mount them behind the bumper or further down into the fenders.

theshadow27 · Nov 12, 2009

Re: (ForceFed Motorsports)

The factory GDL modules went in the funny shaped cut outs behind each headlight (bent rectangle) I think, and the ballast was integrated in the headlight assembly. I'd be a bad idea to try to put it inside the case itself because there's not enough cooling (I thought about it long and hard though) and I don't think they fit. The other option is to screw the ballast to the plastic cover, and drill two holes for the power and HV wires, then seal the whole thing with massive amounts of silicone. I was also about to do it this way, but the wires would have to be bent very tightly to clear the wire cover retainers and I wasn't happy with that. What I ended up doing was drilling three holes in the body next to the GDL mount points, and mounting the ballast on the inside of the wheel well (facing the liner) so the cap head bolts are facing the front of the car, run through the body, then into the ballast. The ballast was threaded so I didn't need nuts or anything, just some loctite. The wires come in through the GDL cutout. Not too worried about water since the HID50 is rated to run completely submerged. I used split-loom to join the ballast wires to the headlight harness. Only thing I wish I had done differently is I drilled the hole for the ballast in the cover, rather than the assembly. That's the only part that looks aftermarket. Maybe if I get around to a retrofit I'll buy new covers and move the hole above the factory connector...

theshadow27 · Jan 24, 2011

Thanks to guru's help we've got another piece of the puzzle - the inside of the generic 35W slim ballast:

Note the circled 1000uF (1mF) 25v bypass capacitor - which explains why these don't work without the BOW-3 (which adds another 4000uF or so). Considering the minimum calculated value is 20,000uF, I maintain that it's unsafe to run these without further bypass capacitance, plus it's effectiveness is reduced without the diode.

Thanks guru!

Bgrob58271 · Jan 7, 2012

Hi Shadow, Im not sure if you're still subscribed to this thread, but I've read the whole thing several times just to get the big picture. I dont know if everything sunk in yet but I think i've got the jist of it.

I posted my question in a new thread, but I wanted to ask someone with real electrical knowledge of the CEM and the wiring of our cars.

Ill just copy what I posted in the other thread here:
Since the low beam circuit is the one with the PWM issue, and the fog lights circuit is not PWM monitored. would it be possible to swap the wires leading into the CEM - so that the fogs are then the circuit that is PWM and the low beams are not?

in theory this would make the fog lights DRL, which i would be happy to keep as halogens, and the low beams would be on the fog light switch which i could turn on and off at will.

Now what if i disabled drl at the dealership, and then got the euro fog update so that the low beams would stay on when i switched to highbeams.

Would this be a possible work around?

If not, do you have a spare set of harnesses that you made similar to your own, or kyles? Or would i have to build it myself. Being a 1st year engineering student myself, i have very limited electrical knowledge and experience
Thanks.

Bgrob58271 · Jan 7, 2012

i guess for 80 bucks or so, ( whatever they charge to disable drl's and update the software ) it wouldnt be worth it, when you can just put your harness in for much cheaper, and have everything inside the car work the way it should.

theshadow27 · Jan 9, 2012

You are right about the fogs not being monitored, and yes, swapping the low beam with the fogs will work, if you run 55W bulbs in the fogs (maybe 35W will work, maybe not).

Edit 01/2017: DRL Disable alone is NOT a safe alternative to the KBOWE. . <s>However with DRLs disabled, there is no need to swap headlights for fog lights, use any harness, or do any euro fog SW. You can simply install the kit as designed. </s>

I don't have any more KBOWEs. I have posted instructions to fabricate them, however the DRL disable is much easier if you have not done similar projects before.

Cheers,

Jacob

Headlight PWM details?

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered

Registered