Ferrite bead on spark plug wire?

[Circlotron]

I'm interested in trying some ferrite beads, the split kind that clamp around a mains lead, on the spark plug wires on my car. Does anyone here have any experience doing this? I have never heard of people doing this before. Are there different grades of ferrite for different frequencies?

[ConKbot]

XY problem, what specifically are you trying to achieve that a resistive spark plug or spark plug wire doesn't achieve?
A single ferrite bead of reasonable size is probably at most going to be in the single digit k-ohm range of impedance, which is about the same range the spark plug resistance, or wire resistance should be. I'd have concerns over core saturation with high currents as the spark plug fires.

Edit:shot in the dark: if it's "my stereo has a whine that varies with RPM" Id look at ground bonding in the vehicle. Make sure your engine block to frame, and battery to frame ground straps are in good condition. And make sure the bolted connection for engine block to alternator is clean too.

[Xena E]

I'm interested in trying some ferrite beads, the split kind that clamp around a mains lead, on the spark plug wires on my car. Does anyone here have any experience doing this? I have never heard of people doing this before. Are there different grades of ferrite for different frequencies?

I've heard of it being useful for ignition emi suppression and less unreliable than resistive caps or leads, but it's anecdotal, I've no personal experience.

I'd have thought the ferrite in those things would be a mix generally suited to that kind of duty, though probably not particularly well controlled at manufacture.

It would be interesting to find out what the reactance would be in that situation.

Regards,
X

[David Hess]

A clamp ferrite core for EMI suppression of a spark plug wire will definitely work, but may not make a lot of difference if you already have resistive wires and spark plugs.  The ferrite type affects frequency range but the selection is more limited for clamp on cores because they are mostly for solving the same problems.

[T3sl4co1l]

Definitely won't work.  When a wire sparks, the sudden change in voltage propagates up the wire at its characteristic impedance, say 100-150Ω, with a rise time of perhaps a fraction of a ns.  (Which, if it's a spark of say 10kV, that's 100A peak easily.)  The impedance is high in relation to a ferrite bead (say 30 to 300Ω) so not much attenuation is had in the linear case, and the voltage is high therefore the flux rises very quickly, saturating the bead in a few ns -- more delaying rather than attenuating the edge.

Tim

[Xena E]

Definitely won't work.  When a wire sparks, the sudden change in voltage propagates up the wire at its characteristic impedance, say 100-150Ω, with a rise time of perhaps a fraction of a ns.  (Which, if it's a spark of say 10kV, that's 100A peak easily.)  The impedance is high in relation to a ferrite bead (say 30 to 300Ω) so not much attenuation is had in the linear case, and the voltage is high therefore the flux rises very quickly, saturating the bead in a few ns -- more delaying rather than attenuating the edge.

Tim

Ooh yeah...

Loaded question.

In the conjunction that wire ht leads are being used, (I'm looking at Circlotrons avatar of the T Ford).

What if they were wrapped around a ferrite core multiple times in an attempt to increase the reactance to the 5 - 10k needed to have a worthwhile effect?

Regards,
X

[David Hess]

Definitely won't work.  When a wire sparks, the sudden change in voltage propagates up the wire at its characteristic impedance, say 100-150Ω, with a rise time of perhaps a fraction of a ns.

Which brings up something I should have said.  The clamp needs to be applied close to the spark plug instead of at the coil output.  The fastest pulse edge originates at the spark plug and not the ignition.

[T3sl4co1l]

Ooh yeah...

Loaded question.

In the conjunction that wire ht leads are being used, (I'm looking at Circlotrons avatar of the T Ford).

What if they were wrapped around a ferrite core multiple times in an attempt to increase the reactance to the 5 - 10k needed to have a worthwhile effect?

Ha, because ferrite loaded? ...

Yes, many turns gets value, also more flux before saturation.  This has diminishing returns in two respects: 1. self-capacitance between turns reduces the impedance peak frequency, 2. the core itself is conductive, so is a factor in this as well: the winding is around the core, so always has some capacitance through it.  You need both enough turns and enough segments (separate cores?) to get really high impedances.  Or less conductive material (NiZn, but then the impedance is less, too).  Even then, capacitance to nearby material, or by itself to free space, ultimately limits impedance.  At relevant frequencies here (~100s MHz), I would guess achieving more than a kohm, over a broad frequency range (octaves), would be challenging if not outright impossible.

But more to the point, all we really need is filtering, isolation between plug end and cable; and an RC filter does an excellent job, both limiting peak spark current (controls wear, probably?) and offering more series impedance, and some distributed capacitance that keeps EMI down.

Any impedance divider, or ladder network, or distributed equivalent, will do; the resistor is just far more compact considering the high voltage and therefore high impedance.

Which brings up something I should have said.  The clamp needs to be applied close to the spark plug instead of at the coil output.  The fastest pulse edge originates at the spark plug and not the ignition.

Yup; and there's still some crud coming from the coil, but that's usually in the µs timescale as ignition driver transistors are designed for limited turn-off speed (for example, IGBTs with huge internal gate resistors)

Tim

[floobydust]

There's a bunch of youtube videos of people in third-world countries adding a clamp-on ferrite bead and the spark appears different "more stable" or +1HP gain!
If adding any inductance to the secondary-side was a good idea, I think it would have been done 70 years ago when radio hit the road, in aircraft etc when RFI was a big problem.

[coppercone2]

i mean whos the say the mechanical guy just threw some generic spark plug crap he knew in there and it was not engineered properly.


I dont think you should mess with this unless the engine was inspected by a engine technician though, if your still on gas. Like emissions test etc.

Seems like you would need to get quite into engines to figure out exactly what this does and how it effects your car.


my crystal ball shows a unregistered uninsured honda civic billowing red gas out the exhaust pipe before breaking down on the side of the road

[Xena E]

Ooh yeah...

Loaded question.

In the conjunction that wire ht leads are being used, (I'm looking at Circlotrons avatar of the T Ford).

What if they were wrapped around a ferrite core multiple times in an attempt to increase the reactance to the 5 - 10k needed to have a worthwhile effect?

Ha, because ferrite loaded? ...

Yes, many turns gets value, also more flux before saturation.  This has diminishing returns in two respects: 1. self-capacitance between turns reduces the impedance peak frequency, 2. the core itself is conductive, so is a factor in this as well: the winding is around the core, so always has some capacitance through it.  You need both enough turns and enough segments (separate cores?) to get really high impedances.  Or less conductive material (NiZn, but then the impedance is less, too).  Even then, capacitance to nearby material, or by itself to free space, ultimately limits impedance.  At relevant frequencies here (~100s MHz), I would guess achieving more than a kohm, over a broad frequency range (octaves), would be challenging if not outright impossible.

But more to the point, all we really need is filtering, isolation between plug end and cable; and an RC filter does an excellent job, both limiting peak spark current (controls wear, probably?) and offering more series impedance, and some distributed capacitance that keeps EMI down.

Any impedance divider, or ladder network, or distributed equivalent, will do; the resistor is just far more compact considering the high voltage and therefore high impedance.

Thanks for the clarification. I was going to ask the OP if it was actually for a T Ford (those having twin spark plugs/cylinder, and trembler coil ignition, initially I believe), and wondered about any special considerations they wanted to reveal...

I was interested in this for my own use, and did also ask an automotive electrical guy I know, who works in autosport about it after seeing this thread. Although he was light on the theory behind the ferrite bead suppression, he did concur that it would be of limited use.

What he did say was, that for a traditional Kettering ignition system, HT cables wound with resistance wire, over a glass fibre core, plus resistive caps was probably the best things would get for suppression, with the possibility of an outer shield helping a little further.

He said that all the foo foo automotive ideas in the world are only claimed to work because those who spend their time and money on them want them to work.

Regards
X

[joeqsmith]

Shield
https://www.aircraftspruce.com/catalog/elpages/shieldbraid.php

Shield = capacitance?
https://thunderproducts.com/product/hotwire-spark-plug-wire/

Series gap?
https://www.steadygarage.com/products/flashing-spark-plug-cap-0-degree-straight-red-blue

Customers have reported up to 15HP (normally aspirated) & 35HP (s/c, turbo, & nitrous)
https://granatellimotorsports.com/i-30500949-granatellimotor-sports0-ohm-high-performance-spark-plug-wires.html

[David Hess]

There's a bunch of youtube videos of people in third-world countries adding a clamp-on ferrite bead and the spark appears different "more stable" or +1HP gain!
If adding any inductance to the secondary-side was a good idea, I think it would have been done 70 years ago when radio hit the road, in aircraft etc when RFI was a big problem.

Pulse shaping can be done with a lumped element transmission line to extend the spark duration and energy while reducing the peak power, but nobody bothers unless the extra complexity is worth it.

Resistive plugs and resistive wires are a simpler solution for controlling electromagnetic interference.

[Circlotron]

I was going to ask the OP if it was actually for a T Ford (those having twin spark plugs/cylinder, and trembler coil ignition, initially I believe), and wondered about any special considerations they wanted to reveal...

Regards
X

I used the term "my" car as a generic description. I actually make ignition systems for race cars, street machines, hotrods etc for a certain company and have done so for over 30 years. Usual installation uses spiral core plug leads and non resistor plugs. This is usually quite okay. Once in a while a customer will have EMI issues where an installation will misbehave. Often it is simply grounding issues. I was also looking for a way to get some extra margin against radiated noise. As T3sl4co1l pointed out, when the plug fires the voltage across it drops very suddenly, from maybe 10-20kV to ~1500V in nanoseconds. I was thinking if I had a bit of inductance at the plug end of the wire it would act with the stray capacitance of the wire to slow down the dv/dt upstream of the inductor at the instant of the initial arc. Typical voltage rise time across the plug is ~8uS but we are not talking about this, rather the sudden fall in voltage as the plug breaks over and arcs. And yes, I'm familiar with the difference between an arc and a glow discharge. I'm using the term "arc" loosely here.

[Xena E]

Circlotron.
Thanks for returning with the information.

I used the term "my" car as a generic description. I actually make ignition systems for race cars, street machines, hotrods etc for a certain company and have done so for over 30 years.

Cool job 

I did have one further thought. Although the multiple passes through the ferrite bead isn't practical, perhaps multiple ones along the lead may be helpful in some circumstances?

Of course the effect would be additive as far as reactance is concerned unlike the multiturn idea which would I assume follow the square of the turns.

I wonder, how deeply does someone such as yourself delve into this, as I had wondered about using something like a current probe to look at the effects of different ideas on the ignition pulse to the plug.

My interest is that I have a M/Cycle that I go on club rides with and we have short range intercoms, my machine is the only one in the group that affects the use of the intercoms.

I have no problem with the way the machine runs or performs.

The ignition system has been replaced pretty much, but the problem remains, (some) other people with the same machine I have contacted about it have reported similar putt, putt, putt, noise in their companions head sets on transmit.

I think the original leads may have been carbon, they were the reason behind the low price I paid for the machine and had to be replaced, the new replacement had failed intermittent within months, the ones in use now are coiled resistance wire with normal cap.

Changing the plug cap to a 5k resistive type made starting performance poor .

The bike is an '85 CBR500 and the intercom I'll have to check the make, but is a fairly generic 433MHz PMR type thing I think.

Today I think I may try finding a ferrite and try your idea, there's not much expectation of a positive result, but just to see if it does anything.

Thanks for an interesting thread.

Regards,
X.

[Circlotron]

Most automotive ignition coils used in a distributor setup have the unseen end of the secondary winding connected to the primary (+) terminal. Not sure if your bike is like this, but if it is, make sure it has a film capacitor of at least 0.47uF between coil + and cylinder head. Make this connection as sort as reasonably possible. This will give a path for the HF component of the spark current to get to ground without going via the wiring loom. This will help prevent upsetting on board electronics and also reduce the chances of the entire wiring loom acting as a transmitting antenna.

This capacitor is very different from the one that goes from coil (-) to ground, i.e. across the breaker points or equivalent.

The above is assuming the bike has an inductive discharge ignition, not a CDI.

[T3sl4co1l]

I used the term "my" car as a generic description. I actually make ignition systems for race cars, street machines, hotrods etc for a certain company and have done so for over 30 years. Usual installation uses spiral core plug leads and non resistor plugs. This is usually quite okay. Once in a while a customer will have EMI issues where an installation will misbehave. Often it is simply grounding issues. I was also looking for a way to get some extra margin against radiated noise. As T3sl4co1l pointed out, when the plug fires the voltage across it drops very suddenly, from maybe 10-20kV to ~1500V in nanoseconds. I was thinking if I had a bit of inductance at the plug end of the wire it would act with the stray capacitance of the wire to slow down the dv/dt upstream of the inductor at the instant of the initial arc. Typical voltage rise time across the plug is ~8uS but we are not talking about this, rather the sudden fall in voltage as the plug breaks over and arcs. And yes, I'm familiar with the difference between an arc and a glow discharge. I'm using the term "arc" loosely here.

Inductance has a first line effect, but mind that no inductor is ideal, and without a shunt capacitor to set the impedance lower (above a cutoff frequency), you need a real big value to have much effect (and again, cored probably won't do, so we're talking air core).  Note the cable capacitance is tiny in relation, again because it's just Zo * length, but also the wave only sees a bit of the cable at a time, it's somewhere between a lumped element and transmission line case here.  Put another way, if you're relying on the cable to act as a capacitance, that means its filtering value is derived from the electric field around it; it necessarily will have HF content in the air around it.  Less than without the coil, perhaps, but not as much with a capacitor proper.

Needless to say, a bypass cap at ignition voltages isn't exactly practical either...

You also need some damping, else you just get a resonant peak at the (cable + coil) capacitance + sparkplug inductor frequency; you might have less spectral noise overall but it's been squashed into one or few taller peaks, worse for anything in those bands.

"Spark" and "arc" are adequate terms here; "spark" more colloquially as a brief discharge, and "arc" a correct description of the conditions (V/I, distance, pressure).

Tim

[Xena E]

Most automotive ignition coils used in a distributor setup have the unseen end of the secondary winding connected to the primary (+) terminal. Not sure if your bike is like this, but if it is, make sure it has a film capacitor of at least 0.47uF between coil + and cylinder head. Make this connection as sort as reasonably possible. This will give a path for the HF component of the spark current to get to ground without going via the wiring loom. This will help prevent upsetting on board electronics and also reduce the chances of the entire wiring loom acting as a transmitting antenna.

This capacitor is very different from the one that goes from coil (-) to ground, i.e. across the breaker points or equivalent.

The above is assuming the bike has an inductive discharge ignition, not a CDI.

Thanks.

I have to correct myself (again) caught out by typo, the Bike is a XBR500 not CBR. There's not many about in the UK.

It's a single pot engine, it's obviously not distributor, but I don't think any Hondas are, I think the multi cylinder ones just use individual coils or wasted spark systems?

It is a CD ignition system so perhaps it was a trait of the system

May be worth trying to look at the waveform, the system is durable but I'm now wondering if the coil switching is vicious? I'll have to look at something well behaved to compare it with as I've zero experience with this kind of thing 

What's the best way to check, pick up coil round the HT lead or toroidal core probe?

Regards,
X

[Circlotron]

What's the best way to check, pick up coil round the HT lead or toroidal core probe?

I use a 100 ohm 5 watt wire wound resistor and a 33nF capacitor in parallel. One end goes to ground and aim a 5-10mm spark at the other end. Don't connect to the plug wire directly. Put your scope across the parallel R&C to find current and duration. Haven't played with a bike CDI but a rough guess I'd say about 200mA current and 100uS duration. Guess tolerance is +/- 50%. Of course, you can't run the motor like this, but just cranking should be okay for a measurement.

A 5 watt wire wound resistor is way more powerful than necessary, but is is insurance against going open circuit and allowing your scope to get a surprise.

[David Hess]

What's the best way to check, pick up coil round the HT lead or toroidal core probe?

An oscilloscope current probe which is a toroidal core probe will work to measure the current.

Measuring the voltage could be done with a *real* oscilloscope high voltage probe, like a Tektronix P6015A.  A high voltage capacitive divider might work but the mechanical construction would be difficult unless you have a machine shop.

[Xena E]

OK.

Thanks to all for their sage advice,
That set up that joeqsmith has is seriously impressive 

The discussion gave me the incentive to get on and do this thing and investigate rather than put up with it.

I was considering the probe route to have a look at the waveform on the HT lead, and was thinking how to liberate the relevant equipment from work, but then considered what I wanted to achieve if the investigation was successful.

I don't think my popularity was going to improve if I wrecked any of the work kit, and the CRO I have at home is an old friend, and I'd be pissed at myself if I managed to zap it.

My thoughts about this being a job for some parallel capacitance as suggested, and relatively easy to implement for a first (cut and) try.

I'd found some close weave braiding in the waste that was stripped from some coax and used to sleeve the bikes previously removed HT lead to the plug with. I didn't so far measure the core to braid capacitance which I should as it may give clues to what's happening.

I switched on my air band radio, (can't find the PMR one at the moment!), and started the motor. The unscreened cable made the radio swamp with ignition noise.

With the "screened" cable however and with the braid clipped to the head, the level of noise was reduced to the point that it was possible to hear the alternator whirr on the radio as well: the difference was not markedly great in volume, perhaps 6db less, (a quarter), but that didn't take into account the effect the AGC would be having in the set as well.

@ Circlotron...
Both times I tried with, then without a randomly chosen (that fitted) clip on core from my junk box. Really I have to say no difference.

Nowhere near perfect, but a Crusty the Clown approved good enough for me.

Thing is, this was just purely empirical research... I now want to see those HT waveforms more than ever...

[Xena E]

My own test equipment is inadequate for this.

I really can't use any of the capable stuff at work for understandable reasons, it would set a precedence, cause possible damage, insurance, heath and safety 



If a department boss is doing it, why can't everybody?

I can see the use of the pressure vessel, though if I rigged a probe to see the breakdown point then would a reasonably fast real time scope not show the difference, A-B with testing on a running motor ?

With respect to the comments about plug gap: before messing with the plug, a new one is ordered, I took the old one out to identify it, (the jobs a pain even with the right socket as its recessed). The one fitted is a copper cored part that was what was sent when ordered for the bike at its last service, not even a considered choice. In checking the manufacturers recommended part I find that it is specified as a resistive core (NGK), and the fitted plug is definitely not resistive, (tested)... that could have a bearing on the problem.

[ Attachment Invalid Or Does Not Exist ]

The bikes previous keeper did very little mileage with it, and sold it on to me for quite cheap, sold as seen, because of bad starting and rough idle. After new ignition parts that I'd had put on and servicing it had been OK apart from the interference.

I've ordered materials for a service now, and I'll post up the results when I try the new plug... I'll also regap the old plug to try that, as well as measuring the braid to core capacitance of the HT cable too.

Regards,
X

[mr ed]

Piston aircraft use fully shielded plugs and  wires as the main navigation and comm radios are AM only. This works well, it has to.