Headlight modulator causing voltage spikes

Status
Not open for further replies.

Thread Starter

jfairman

Joined Jul 2, 2010
22
I have a headlight modulator on my Gold Wing motorcycle ( 12V ). When it is actively switching ( 240 cycles per min ) mfr confirms that it causes "power surges". I suspect that the resulting voltage fluctuation is the cause of an associated clicking noise in the audio system.

I believe I need to install a capacitor in close proximity to the modulator to effect a "transient load decoupling" from the rest of the bike's electronics. The headlights are two 45 W bulbs on a 10A circuit.

If I'm off track, I'd appreciate general guidance.

If I'm on track, how do I decide what size ( farad ) or type of capacitor to use?
 

Ante

Joined Dec 25, 2009
10
I might be wrong but I believe what you also need is an inductor in series with this particular 10A circuit to even out the current peaks. Just a capacitor might not be enough to eliminate the interference.
 

Thread Starter

jfairman

Joined Jul 2, 2010
22
yes, thanks. thinking of putting a ferrite loop inductor and capacitor beside the headlight modulator.

i need some guidance on choosing the right capacitor and inductor.

how should these be sized?
 

SgtWookie

Joined Jul 17, 2007
22,230
Ahh, is the PWM at a 240 cycles per minute rate? Or is that the rate for the high-frequency PWM that dims the headlamp?

You need a pi filter on the input side of the modulator, not on the output to the headlamp. Simply using a large inductor on the output of the modulator risks frying the modulator and burning out the headlamp simultaneously.

We usually don't talk about automotive lighting projects on the forums, as there are lots of safety implications.

There are certainly safety benefits to a motorcyclist, if a modulator that is properly designed and built gets a proper installation.

It sounds to me like the modulator you purchased was either designed poorly, or was designed for lower power headlamps. I would not be surprised in the least if your modulator's guts consisted of a 556 dual timer and a few resistors and caps - just enough for a gated PWM circuit.

Simply adding a large capacitor across the modulator's input power/ground terminals may be enough to get rid of most of the noise.
 

Ante

Joined Dec 25, 2009
10
Just to clarify a few points; I did not suggest any filter on the headlamp side but on the 10A feed circuit, only on this side it can protect the rest of the electronics on the bike. I did suggest an inductor in combination with the capacitor to even out the current spikes.
 

SgtWookie

Joined Jul 17, 2007
22,230
Just to clarify a few points; I did not suggest any filter on the headlamp side but on the 10A feed circuit, only on this side it can protect the rest of the electronics on the bike. I did suggest an inductor in combination with the capacitor to even out the current spikes.
OK, but you were really rather vague on your initial post, which reads:
I might be wrong but I believe what you also need is an inductor in series with this particular 10A circuit to even out the current peaks. Just a capacitor might not be enough to eliminate the interference.
It's really impossible to tell from your wording on which side of the modulator you meant for the inductor to be inserted, or whether a cap was necessary.
 

Ante

Joined Dec 25, 2009
10
This doesn’t point to the headlamp side of the circuit as I see it; “in series with this particular 10A circuit” more like the circuit breaker (fuse) side. Should not be confusing but I’m sorry if that’s the case English isn’t my first languish. :)

 

SgtWookie

Joined Jul 17, 2007
22,230
Fair enough - but remember that "current" begins with the current source, and ends at the current source.

We have to be rather specific here. I suppose I'm as guilty as anyone about not explaining things thoroughly enough. Schematic diagrams help a great deal to eliminate confusion.
 

Thread Starter

jfairman

Joined Jul 2, 2010
22
i have been researching this for months with Gold Wing riders, Honda parts folks, electric do-dad purveyors ... just about anyone that knows a little about eletronics. it turns out Gold Wing owners have this problem with modulators affecting the factory audio units regardless of what brand modulator they use. most owners i've talked with have simply figured it would be too hard to figure out how to solve the problem ( compared to just ignoring it ). i happen to love a challenge!

i appreciate your suggestions, your thoughts, any refinements on your suggestions! I'm not jumping into this as I'm not in a hurry and I don't want to fry anything.

the 240 frequency is simply the specifications from the mfr of the modulator on how many times per min the light changes low to high. this is setup to modulate the high beam circuit. the mfr suggested putting a "large" ( 20 farad ) capacitor near the audio components that are effected, but i think most guidance i've received suggests that isolating the irregular votage/current at the modulator by putting a capacitor on the power source side of the modulator makes more sense. the additional guidance i've had is to put an inductor on the + power source side just before the capacitor. my understanding is that this should steady the current draw feeding the modulator/capacitor.

i'm tempted to try the capacitor first, then add the inductor if the problem remains.

but still ... don't know anything about how to choose the right size for these things!

sgt.. you mentioned a pi. what is that?

not that it helps, but you may be curious to see what this modulator looks like. photo attached. There are four plug connectors. They connect between factory wiring F plug and M headlight connector for each of the two headlight bulbs. the additional long black wire has a light sensor at the end which de-activates the modulator in low light.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
I mentioned a pi filter; so named because it resembles the Greek symbol pi: Π

A pi filter consists of a capacitor connected from supply to ground, then an inductor in series, and then another capacitor to ground. It's reasonably effective in eliminating noise.

I don't think you need a lot of inductance; just enough to slow down the transition times so that your receiver can reject the broadband noise created by the switching circuit in the module.

Trouble is, you have a pre-built module, and I have no idea what kind of ability you have in regards to electricity/electronics. Failure of lights on motorcycles is a critical safety issue, as I'm pretty certain that you are aware of.

I think I can "fix" this problem, but IMHO, the best way to do so is to start off fresh with a known set of parameters, and a long series of careful tests where I can observe the results.

If for some reason this thread is closed, you have my permission to contact me via E-mail. This is a very unusual circumstance which merits attention.
 

Thread Starter

jfairman

Joined Jul 2, 2010
22
i recognize and respect your careful approach to offering guidance to persons of unknown skills on critical saftey equipment.

the way this modulator is built, if it fails, it fails "on". i had this happen with the first one i installed that had an early life failure. the modulation stopped in one light and just stayed on. maybe obvious to you the headlight circuit is three wires, high beam power, low beam power and ground. i expect that I will be only making changes / connections to the hi-beam ( the modulated power ) and the ground. this should mean that the low beam circuit will be undisturbed and always available as a backup in case the hi-beam circuit fails.

i have done lots of simple wiring projects including installing on the bike a relay switched accessory circuit & battery charging circuit of my own design, installed a modulator ( which also creates noise ) on a pair of brake lights at the back of the bike, installed CB and radar detectors into auto factory wiring on a series of cars i've owned, along with installing a manual override fuel pump switch.

this project will be my first opportunity to work with capacitors / inverters.
 

SgtWookie

Joined Jul 17, 2007
22,230
It appears that our Moderators have formalized and posted the guidelines for automotive-related discussions.

That posting can be reviewed here: http://forum.allaboutcircuits.com/showthread.php?t=40361

In keeping within the spirit of the guidelines, as far as the headlamp modulators - the only thing we can recommend is maintaining the lamps and wiring in the same configuration as it was when delivered by the dealer, as it had to meet all applicable standards/regulations existing at that time. Modification of such items may lead to problems of safety or non-compliance with either the standards/regulations of the time, or subsequent standards/regulations.

If left in the original configuration, it would be "grandfathered" in for subsequent regulations. However, if you make modifications, then it must comply with all standards/regulations, however often they are changed/updated.

I don't believe that "filtering out electrical noise" falls under the category of modification of lighting systems. However, modification of the wiring harness to the modulator supply (which powers the headlamp) DOES fall under that category - and modification of the harness would be necessary to insert a filter.

So, it appears that this is now a dead topic.

If you remove your after-market modulator and restore the harness to as-delivered configuration, your noise problem will disappear.

This is the only course of action that I can suggest to you.
 

beenthere

Joined Apr 20, 2004
15,819
Yes, post a guideline, and immediately it seems to cause a problem.

The headlamp modulator is a safety issue, in that it can increase visibility, and aslo mess up your electrical system if it fails badly. The module appears to use proper connectors, so it would not be ruled out for that reason. Our intent is to discourage someone from just soldering in a connection and wrapping a bit of black tape over it for insulation.

This could be a worthwhile problem to deal with. The OP's modulator is not the only one that will have these problems. Given that it appears to attach with approved connectors, that is not an issue. Solving the problem is a safety enhancement.

If the OP has not given up on us, this gets a green light.
 

SgtWookie

Joined Jul 17, 2007
22,230
Thanks kindly, Beenthere.

The guidelines are not a problem; however this particular situation needed clarification as to whether to proceed (or not).

One of the difficulties in designing a filter of this type is the input and output impedances. I'm making an assumption here of 5 Ohms for the headlamp, which is in the ballpark for the power consumption vs the typical system voltage; and I could be off considerably. The dynamic nature of the load (switching on/off etc.) makes the design rather "iffy", particularly since I do not have all of the information about the supply and the load, nor a similarly equipped motorcycle to test it on.

However, here is a start:


The power source is on the left; the lamp is the 5 Ohm resistor on the right.

Attached is a plot of the transmission and reflection of the filter. There's a deep notch right at 240Hz, better than 70dB rejection. Since it's highly likely that the modulator is using square waves for the PWM, the better-than 40dB attenuation should keep thing pretty quiet. Even if the components have somewhat sloppy tolerances, it should at least do some good.

The inductor - 19.5mH is a large value of inductance. However, you could get some of these ferrite toroids from Electronic Goldmine:
http://www.goldmine-elec-products.com/prodinfo.asp?number=G6683
These toroids have an Al value of 9900. 44 turns of AWG 20 magnet wire (about 1.5 yards, 4.5 feet or 54", make sure to use some extra), or will nearly fill the toroid, and have an approximate inductance of 20uH.

Alternatively, you can use four strands of AWG 26 magnet wire, which will have the same circular mil area, will be easier to wind, and will be more resistant to breakage due to vibration. Make certain that the wires are not twisted when winding them on the toroid. Make certain that you don't "tie a knot" on the toroid either, or you will tend to kill the Q of the inductor.


Radio Shack carries an assortment of magnet wire:
http://www.radioshack.com/product/index.jsp?productId=2036277
There is a 75' spool of AWG 26 in the assortment. You would need 4 x 5'=20' of wire for one inductor.

459uF is not a standard value of capacitance; however 470uF is. The capacitor on the input side must have a rating of 60V or better, or else you risk destruction of the capacitor when load dumps occur - such as your headlamp turning on and off. Also, 22uF is a standard value.

The capacitors should be rated for 105°C operation or higher. Low-ESR are preferred. You need to buy the capacitors from some place like Digikey, Mouser, Newark, Avnet Express, Allied, etc. who are large volume manufacturer authorized distributors for a huge variety of components. Do not attempt to use caps from an auction site or a local mom & pop or Radio Shack store, as they will be stale from sitting on the shelves.

Please do not modify any portion of your motorcycles' factory wiring harness that involves your safety lighting. Instead, place this filter on the harness of the modulator. That way if you experience any problems, you can quickly remove the modulator and filter as a unit, and be back in business in moments.

For wiring connections, use Sn63/Pb37 solder, rosin flux. Isopropyl alcohol for cleaning. Use heat shrink tubing for insulation; two layers. The type that has glue inside is preferred; as it won't leak. Don't use "electrical tape" as it will get gummy and fall off. Silicone tape that sticks to itself is suitable, but you should secure the ends using zip-ties or lacing tape to ensure that it does not try to unravel before the silicone has cured.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
Rifaa,

To be more specific, it is a low-pass LC (inductor/capacitor) third-order Cauer filter. The stopband is actually -40dB at 208Hz, however since the fundamental frequency is approximately 240Hz, I positioned a zero at that frequency.

Ideal square waves are composed of ALL of the odd harmonics of the fundamental frequency; the even harmonics are absent. The first odd harmonic will be 240Hz x 3, or 720Hz. It's difficult to discern exactly 720Hz on the plot, but suffice it to say that it will be at an amplitude of less than -40dB. Since every 3dB is a half power point, the noise should be negligible - if I have the proper impedance matching going on.

On the plot, the blue trace is the transmission, or the level of signal attenuation at the output in dB vs frequency, with respect to the signal at the input of the filter. You can see the transmission characteristics using just a spectrum analyzer with a frequency generator that is capable of sweeping a range of frequencies.

The red trace is the return loss vs frequency. Basically, this is reflected energy, and is a good measure of how well-tuned a filter is for a given frequency. Return loss of -15dB to -20dB in the passband (frequencies that are desired to pass through the filter) indicates a well-tuned filter. In this plot, return loss exceeds -30dB at all points in the desired low frequency range.

In order to see the return loss, you really need a spectrum analyzer that has an S-parameter test set. With such an instrument, transmission is reported as S21 (what is measured at the output port 2 as referenced from the input port 1) and return loss reported as S11 (what is measured at the input port 1 as referenced from the input port 1.) There is also reverse transmission (S12) and reverse return loss (S22). In some high-end commercial and military filters, all four parameters need to comply with the specifications.

The land of LC filters is not an easy one. It's pretty easy at low frequencies, as long as you know what your input and output impedances are. Things start getting pretty finicky when you're up in the 100MHz - 500MHz range. In the GHz range, it practically seems like magic. Inductors are tiny, smaller than this -> "_n_" - just a little kink in a small-gauge wire, and capacitors are tiny bits of ceramic with metal plating on both sides.

But, I'm now on a tangent to our OP's problem.

You would not likely experience such problems with a small motorbike [eta] that is in stock configuration.

A Honda Gold Wing is a large touring motorcycle, with all kinds of accessories. Such a large motorcycle would probably not be practical in the Maldives. However, here in the States, you can start off on the east coast, and ride west for perhaps 3,000 miles before you come upon the west coast. It's not something you would want to do on a small motorcycle, unless you have masochist tendencies.

A number of years ago, in order to make auto/truck drivers more aware of the presence of a motorcyclist, someone came up with the idea of what is called "headlamp modulators". These headlamp modulators alternately run the headlamp at full brightness, and then reduce the brightness to perhaps 1/3, and then goes back to full brightness. This happens at a rate of 3 to 5 times per second. Turning the headlamp completely off would not only be illegal, but would lead to early burn-out of the filament.

The flashing (or "modulated") headlamp attracts the attention of other motorists, which is very important to motorcyclists. Most people driving cars and trucks are not expecting to see a motorcyclist at all, and motorcyclists present a much smaller profile than an auto or truck. The modulated headlamp makes them much more visible, and helps to ensure that oncoming drivers won't turn in front of them.
 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Just got done with a bit more testing of the toroids I mentioned a couple of posts back. I stacked two together, and secured them using polyester adhesive tape (think masking tape, only much stronger, permanent, and far better insulating qualities.

Click here for a PDF on 3M 1350 tape: http://multimedia.3m.com/mws/mediawebserver?mwsId=SSSSSu7zK1fslxtU48tx4x_Uev7qe17zHvTSevTSeSSSSSS--

You could also simply use a quality epoxy to keep the toroids firmly stacked together.

Stacking the toroids (like stacking donuts) doubles the Al value. This reduces the number of turns required from 44 to 31. That may not sound significant, but by reducing the number of turns, the wire gauge can be increased.

You can instead use 3 strands of AWG-22 magnet wire, included in that Radio Shack assortment. Each strand will need to be about 65" long; that includes a couple inches on either end of the windings. It won't hurt to add a few more inches; better to waste a few than try to cheap out and have to start over.
 
Last edited:

Thread Starter

jfairman

Joined Jul 2, 2010
22
Sgt Wookie!

Finally getting back to you. The cold weather is upon me and the riding season has ended. I've pulled the headlamp(s) off the Gold Wing and purchased the toroids you suggested. I will purchase capacitors next, but want to offer you as much additional insight into this project as I can. I've attached photos of the headlight assembly, from the front, from the rear, and a close up of the wiring that will need to be fit with the pi filter.

You should be able to observe that the modulator plugs between the headlights and the factory headlight connectors. It has an additional wire connected to a daylight sensor that disables the modulation when ambient light is low. The modulator is configured to "blink" the hi-beams when they are on. The modulator has no effect on the low-beam circuit.

The close-up photo shows how the factory wiring splits the connection for hi-beam ( blue wire ) just after the white factory connector, to service both bulbs in the headlight. The factory wiring makes the same split for the ground ( green wire ), low beam ( white wire ), and marker lights ( brown wire ) just after the white connector. I propose to add the pi filter to the hi-beam ( blue ) wiring between the factory connector and the "split". I am not sure if my PI ground should tap into the ground wire just after the white connector.

I have reviewed the diagram and your description of the PI circuit. It makes sense, but I have a couple of questions about the details. I understand the concept of stacking two toroids. I understand the concept of wrapping the magnet wire and not overlapping previous wraps. I am not sure I understand the 3-strand wrapping implementation. Are all three connected ( soldered together ) at their ends? How does my installation of three strands need to be different from a single strand?

With your suggestion of minimizing changes to the factory wiring, my plan is to insert the PI filter into the circuit using single wire plugs for input, output and ground so that I can remove it and reconnect ( plug ) the factory wiring together.

I am keen to get your additional suggestions or questions now that I've provided a little more info about the project.

Lastly, you mentioned that the resistance of the light bulbs might influence your circuit recommendations. My ohm meter was not able to detect a resistance on either bulb between the ground pin and either power pin. Could it be that the resistance on these bulbs is too low for my device to measure? The ohm meter is working fine as I tested it on a standard 60W bulb and measured 5 ohm.
 

Attachments

Status
Not open for further replies.
Top