Aurorabucky

All,

So I did some searches on this subject, and found some good info, but I'm getting specific enough with the info I want that I'll start a thread.

First off, the question I'm asking:
Does anyone know or have a guess at what the oscillator frequency is inside a lot of LED bike lights (especially Niterider systems) so I can get ferrite beads optimal for the task? I asked Niterider. They never responded.


Backstory:
Basically, I have a Niterider Pro 700 wth 8-cell LIon, a wire, and a lighthead paired with a wireless Sigma ROX 9.1 (speed, cad,HR), mounted on a commuter. This is a combination from hell in terms of interference. The lighthead jams the computer within the range of a sphere like a basketball. The wire jams anything within about a 4 inch radius, and the battery jams everything within the sphere of a big beachball.

I have, however, gotten them all to work together. It's not pretty though.

I was lucky enough to find that hiding the computer on the top of my top tube immediately behind my headset and stem succeeded in shielding it from the lamp mounted at the handlebar. I moved the cadence sensor up to my seat-tube from my chainstay to minimize the distance. Then I found two ways to make the battery work -

A) Hide it in a saddlebag or on the rear rack and route the wire (extended) down my seat-tube, under my bottom-bracket, up my downtube, through the fork., and up to the lamp. UGLY. Effective.
But there are so many connections that the system is very vulnerable.

B) Built a Faraday cage inside said bikebag with battery and found a way to mount it on my front rack, which is great except that I use that rack...a lot..... for groceries, and that's valuable real-estate. The cage did a good job attenuating the signal from the battery to where the EM jam range is about 1ft in diameter (down from 3ft+), and the setup is simpler and more durable, however at low battery, the high-voltage oscillations are enough to jam the computer despite the cage.

So the solution I'd LIKE is to mount the battery behind my seat, and take the cord under my top-tube straight to the lamphead. In this position, the wire now comes into play. I've been trying to dampen some of the EM from the wire with a single ferrite bead at the battery, but those beads are meant for performance at specific frequencies, and I wonder if I can get it to work better by really optimizing what I use there. I could wrap the cord with foil, but I have to think there's a better way to do this....

Which returns us to our original question. Any idea what frequency I should aim for?



THANKS!

unterhausen

The light probably doesn't have any fixed frequency, but you just need to knock down the frequency that the radio in the computer uses. My guess is that it is one of the unregulated frequencies, probably the lowest one. 300 MHz or 900 MHz are both good guesses.

Aurorabucky

Mmmmmmm.. Good idea. I'm looking around for that spec, but having trouble finding it. I found one reference to a Sigma 1009 STS having a 112 khz frequency. On their (Sigma Sport) website, they claim that STS is a "low frequency" system. Far cry from 300MHz.

I'm pretty sure their "STS" sensors work with any of their STS computers so if the info isn't totally untrue, it might be what I'm looking for.

Looigi

It's not the fundamental frequency, which probably varies, but it's harmonics that cause the problems and which you need to knock down. Switching power supplies gain efficiency by very abruptly turning current on and off avoiding the linear regime where power is lost to heat. This creates rich harmonics.

There are more complex types of switchers that greatly reduce harmonics, but these aren't really appropriate/feasible for bike lighting, AFAIK. Existing types could likely benefit from some reasonable effort to reduce EMI, but I don't think this has been a design input/requirement for these lights.

pdlamb

Can't help it, I've got to ask.

If you're willing to put the computer on the top tube, and/or build a Faraday cage, and/or hack the innards of the system...
Why aren't you willing to deal with a wired computer?

Looigi

There are, of course, ANT+ and other digital protocol wireless computers that are essentially immune to RFI.

Road Fan

I would experiment with twisted-pair wiring for any point-point connections. Loop size has a lot to do with degree of noise coupling, and if you twist a 1-yard two-wire interconnect to say, 1 turn per inch, you've reduced the loop size by a factor of 36. That's very significant.

As far as frequencies go, I'd forget about optimized ferrite beads or other filters. Think of it as a broadband attenuator instead, perhaps 50 MHz to 500 MHz. If it's switching power supply or digital clock noise, you have very, very fast rise/fall times and hence a very wide spectrum of significant harmonics.

Aurorabucky

LOL Well, to be honest, I was using a decade-old wired system that thanklessly worked day in and day out until one day the sensor simply didn't work anymore. At the same time, I had a full set of sensors and a HR monitor that were orphaned when a Sigma wireless computer on my road bike met an unfortunate end (under a bus no less....not the bike....just the computer). So instead of buying two new systems, I opted to replace two computers with one system for use with BOTH bikes and make use of my spare sensors. (the ROX computer-head will assign a group of sensors to a Bike #1 or #2, and tell which bike it's on at any given point).

Other than the light issue, having two bikes of mileage records in one place and downloadable to a computer has been really handy. The top-tube mount is just fine with me for my commuter because it keeps it more protected and I shouldn't be fiddling with it in-flight anyhow. Drawback is that it's not ANT+ because Sigma just HAD to be different.

Why keep track of my records? Because it's fun.

Aurorabucky

So thanks everyone for weighing in.

So that loop comment made a good point. I have two extension cables for the Niterider. I could take the outer rubber off of one of them and risk trashing it to add some twist into the wires and then shrinktube it sealed again.

Apart from that, it's sounding like general off the shelf ferrites will work about as well as anything then?

If I can cut the cable RF, I can mount the battery in that saddlebag behind my seat but keep that cord safer than it is now. I'd be happy with that.

Road Fan

Hi, Bucky sorry I've been silent so long! I didn't know you were asking me for something.

As far as trashing cables, I'd have to take a good look at the cable itself. It's always a risk to start homebrewing on a low-cost cable that has nevertheless been given good engineering tests when the system was designed. For example, if the OEM contacts are held in a rubber or plastic header, soldering will only distort the header and then you'll have nothing useable. I think I'd try to score an alternate set of cables or at least connectors, and build the twisted-pair-with-jacket cable from scratch.

Considering these are low-cost connectors, you need to twist the wires in a stress-free way. The time-honored approach of chucking a hook in a drill, looping the wire over it, and tying the other end to a doorknob will add stress to the wires, and that stress may try to twist or distort the contacts after you build up the cable/interconnect. Partly for this reason, some industries have a standard of not using wire finer than AWG 24 in some cases, AWG 22 or even 20 in other cases. But the wire has to suit the contact.

If I'm not making much sense, sorry! Post some pictures of what you're actually dealing with and I can be more focused. A sketch of your overall planned electrical layout on the bike, or the options for the installation, would also be useful.

I see you're on an island in Washington State. Salty air? Any history of cable or connector corrosion due to that? Gross-looking green or blue crud?

Aurorabucky

Speaking of waiting long, I just took the cake.

So I found a solution, and ferrite beads were my salvation.

So part of the delay was my finding a messy but stable way of putting the battery on my front rack for a time...and getting sick of fiddling with it....but the annoyance finally got me to try moving the battery back to the seatpost with the power cable running along the top-tube again. In the end, I was too chicken to try cutting open one of my cables, but between my own scavenging and a buddy of mine, I ended up with four ferrite beads.

Installing one bead at each end of the cable near the head and battery came up negative. The same occurred with doubling up the ferrite with two on each end. So for the heck of it, I placed three ferrites on the wire where it made its closest approach to the Sigma....and.....Speed, cadence, etc lights right up as if the lamp was turned off, even at full power. If that hadn't worked, my next step would have been to get larger ferrites, and do some wraps - I read something about that cubing the effectiveness per wrap.

So being that I'm an engineer, but not an engineer in THIS field of study, I have two theories as to what I just did.

A) Either the ferrite signal attenuation effect in the power cable itself wears off with distance, and ferrite proximity to the stretch of wire actually causing the jamming matters (how could this be?).

or

B) The ferrites are the wrong strength to really clean up the signal in the cables to the satisfaction of my computer, and the ferrite surrounding the wire in that area acts like a Faraday cage, one that knocks out all those harmonics, but does not block the low frequency signal from the cadence transmitter. I guess I could test this by wrapping the wire there in foil and see if I get the same effect. I find this the most likely.

Either way, I've gotten a really noisy multi-day endurance lamp to run with a fancy-pants wireless computer in a way that looks pretty clean, and that's success. The battery is cushioned and protected in a Timbuktu saddle bag, and the wire is about as intrusive as any of my brake or shifter cables - a lot nicer than running the dumb cable down the downtube, under the bottom bracket, and up the seat-tube using TWO extensions.

Really interested to know what some of the experts think I did though.....

So I'd like to thank everybody for the good tips and help with this. I looked at a lot of forum posts in trying to get this to work, and hopefully this thread becomes yet one more to help the next rider get their gear to get along.

For illustration, I'll post some pics.

Road Fan

Ok, I'll go, since I'm an EE with some experience in interference coupling and control.

As a material, ferrites have properties somewhat like iron, in that it tends to concentrate magnetic flux when in its presence, like wrapped around a wire carrying current. The magnetic field in the ferrite has an oscillation, since the cause of your interference was a noise current. This oscillation transfers noise energy into the ferrite and reduces the amount of it broadcast into the rest of the universe.

Many ferrites also are able to convert energy at certain frequencies into heat, like a resistor. This is in part where the low-pass filtering you noted (dc gets throught, high-freq noise does not) originates. This action is the "strength" of the ferrite as you called it, which as an EE I'd call energy dissipation. When the ferrite converts that energy into heat, the noise energy is no longer electrical and is not able to radiate and cause trouble to your ROX.

It's not a Faraday cage. Your wire was carrying a noise current from point A to point B, and was radiating noise. I think you said the wire was not going directly to your ROX, yet the ROX was experiencing interference. Your ferrite stopped the noise component of the current hence stopping the noise from interfering with the Sigma any more.

Well done, though it seems like you just dinked with it until it was all better! Professionals in noise control often depend on "experimental skills," too.

What you did was a technique of filtering out the noise, not shielding the sensitive device (the ROX) or sshielding to contain the noise source. No problem, what works, works. And a low cost filter like a freebie ferrite donut or three is a lot easier to set up than a shield is!

Will one ferrite donut work just as well? who knows? If you want to test with other ferrites to see if you can find one piece that works as well as three of them did, great, but don't lose the "good" ferrites. Mark them or store them in their own little plastic bag.

tedboone

I wish I'd seen this thread a few weeks ago! Just picked up a Sigma Rox 5.0, and I like it. Trouble is (of course) it doesn't play nicely with my new light: NiteRider MiNewt.500 Cordless LED Headlight. Reading through the thread, it looks like the original poster solved the problem by reducing the noise emanating from the cabling connecting the lamp to the battery supply. Given my light is cordless, that's not an option for me. Anyone have any suggestions on how to shield the NiteRider from the Sigma? Or, conversely, any suggestions on alternative lights that play nicely with Sigma wireless computers?

Thanks in advance.

Aurorabucky

Well, I use my ROX on my road bike too, and I typically run a Light and Motion Urban 550 out at the end of the aerobars. If I bring it in near the ROX, it jams it, but I keep it upside-down and about 6 inches away, and it does fine. So for the cordless light, distance was all I tried. Granted I have no idea if the Niterider is dirtier or cleaner than my L&M Urban.

If nothing else and appearance means nothing, the pie tin wrapped around 3/4 of my big LIon battery hade an effective shield. You could make a similar cage around the Minewt. I know....unsatisfying.....

While the thread is alive again, I'll give a follow-on to what worked with the big light and say that ALL ferrites are needed, and I have to keep them such that they stay in a zone directly between my ROX and the cadence sensor. My patience "dinking" with it ran out soon after I found adequate success! lol Since I get use out of HR and cadence functions on my roadie, I almost think I might punt the cadence for the commuter and just have it read the speedometer. The ferrites are working, but the connections of the wire extension shake loose easy. That's happened a few too many times after potholes.

tedboone

Thanks for the info. I tried moving the Niterider around, and it seemed to have a "dirty" radius of at least a foot, so there was no manageable solution.

Just bought a 4 Mode 1200 Lumen CREE XML LED Headlamp, and it seems to work just fine with the ROX. A cheaper solution as well ($24 on Amazon, vs $80 for the Niterider). It has an external battery, which I was trying to avoid, but that's a small price to pay.