So where exactly is the high voltage return in this automotive ignition circuit?

[Ben321]

I noticed that in this diagram https://www.adriansmodelaeroengines.com/catalog/pages/images/Spark-Ignition/ASI-19.jpg all possible return paths for the high voltage going to the spark plug are obstructed by other components. Even though the spark plug's ring connector is grounded, that ground doesn't appear to provide a clear path for high voltage return. Ground isn't some magical location where electricity is dumped and it disappears. The current will try to complete the circuit back to its source (the source for the high voltage being the ignition coil).

One possible return path goes through the battery (not something you want to subject to a repeated sequence of high voltage surges).
Another possible path is the ignition capacitor but that's only designed for a few hundred volts, not 10s of thousands of volts.
The last possible return path (at least that would seem possible in this diagram) is the ignition points. But those open up to generate the high voltage surge, so they are already open when the high voltage surge attempts to return to the coil so that would force it to arc across this switch, damaging the switch contacts over time (something that the capacitor is actually designed to prevent on the primary side of the circuit, as without that capacitor to absorb the primary's inductive kickback, the hundreds of volts on the primary would arc across the switch contacts, damaging them).

So I don't see any possible return path for the high voltage in that diagram, that doesn't involve the current flowing somewhere that would damage the components. So is there some other possible return path just not taken into account in this diagram? If so, I'm curious what it might be.

[Slh]

Isn't it a flyback circuit?

 Energy gets stored in the coil when the points are closed. When the points open, the coil secondary flashes over in the spark plug thanks to the high turns ratio.

It's a high frequency pulse so passes through the capacitor easily. The spark plug gap takes all the volts so the capacitor doesn't have to.

[Circlotron]

Simple.
The diagram is wrong.
Any three-terminal ignition coil for negative ground usage has one end of the secondary connected to primary positive. And preferably a capacitor of 1uF or so connected from primary positive to engine metal. This is to stop the high frequency component of the spark current from traveling through the vehicle wiring loom as it finds its way from ground back to primary positive.

Edit -> Proof that the diagram is wrong is when the points open the primary negative terminal spikes positive several hundred volts but the HV terminal goes negative many kV. If the other end of the secondary winding was connected to primary negative as per the diagram then the primary voltage spike would *subtract* from the secondary HV spike.

[langwadt]

return path is the battery, which is fine

[ejeffrey]

The mindset that "return path for a high voltage pulse is special or difficult" is wrong.  The coil and the spark plug see the high voltage but what matters for return path is the current. The current is relatively small, a small fraction of the primary current.

[Circlotron]

The mindset that "return path for a high voltage pulse is special or difficult" is wrong.  The coil and the spark plug see the high voltage but what matters for return path is the current. The current is relatively small, a small fraction of the primary current.

Ignition coils most commonly have a 1:100 turns ratio, so if you interrupt 5 amps in the primary you will produce 50mA in the secondary circuit at whatever voltage is needed to make that current happen. The secondary side is a current source, not a voltage source.

[Ben321]

Isn't it a flyback circuit?

 Energy gets stored in the coil when the points are closed. When the points open, the coil secondary flashes over in the spark plug thanks to the high turns ratio.

It's a high frequency pulse so passes through the capacitor easily. The spark plug gap takes all the volts so the capacitor doesn't have to.

Yep. Current flows through the coil when the contacts are closed. When the contacts then open the primary current is broken, which produces a high voltage primary-winding pulse of a few hundred volts, and an even higher voltage secondary-winding pulse of a few tens of thousands of volts. The output of the secondary drives a current through the spark plug and back to the coil through SOME ROUTE. It needs a return path to the coil.

[Ben321]

Simple.
The diagram is wrong.
Any three-terminal ignition coil for negative ground usage has one end of the secondary connected to primary positive. And preferably a capacitor of 1uF or so connected from primary positive to engine metal. This is to stop the high frequency component of the spark current from traveling through the vehicle wiring loom as it finds its way from ground back to primary positive.

Edit -> Proof that the diagram is wrong is when the points open the primary negative terminal spikes positive several hundred volts but the HV terminal goes negative many kV. If the other end of the secondary winding was connected to primary negative as per the diagram then the primary voltage spike would *subtract* from the secondary HV spike.

Still that puts the spark current from the secondary-winding high voltage traveling back through the ignition capacitor. The ignition capacitor is intended only to delay the high voltage kickback on the primary so that it doesn't spark across the point contacts when they open. This allows the contacts to open without a spark, and thus avoiding wasting energy in that spark (as well as avoiding cumulative damage to the point contacts from all of the sparking that would otherwise occur). The ignition capacitor isn't designed to take HV return current from the spark plug. That return current is high enough, that even though it only lasts about a millisecond per pulse, that it will charge the capacitor momentarily to a very high voltage, that could potentially exceed the breakdown voltage of the capacitor's dielectric layer.

[IanB]

I agree with Circlotron. The diagram is very poor. It quite unclear how the far end of each coil is connected inside the can.

In the actual ignition circuit, the low voltage end of the HV coil is connected in series with the negative end of the primary coil, so the high voltage return path is through the primary coil, then through battery to ground (the battery having effectively zero impedance).

[IanB]

One possible return path goes through the battery (not something you want to subject to a repeated sequence of high voltage surges).

This is, in fact, the normal return path. There is nothing to worry about, since it is impossible to subject a lead acid battery to a high voltage surge. The internal resistance of the battery is so close to zero that high voltages across it cannot happen. It is effectively a short circuit. There is no need to worry about high currents either, since the battery is designed to supply hundreds of amps to crank the starter motor. There is no way a piddling little ignition coil is going to exceed the starter motor amps.

[bdunham7]

I couldn't readily find an online reference, so I drew you a picture.  There are other possible arrangements, C1 is optional and many 12V systems will have additional ballast resistors, but this the basic, common setup.  And yes, by positive current convention I suppose I_spark goes the other way.

[Ben321]

If I were to design an ignition circuit like this, I'd make sure that one end of the ignition coil was always grounded, like in this diagram, which is a revised version I made of the diagram (the original version being the one in the opening post of this thread). Why isn't the circuit done like this in actual cars?

Note that in the diagram the capacitor is going across the switch, but it would also work being put directly across the primary of the coil.

[bdunham7]

Why isn't the circuit done like this in actual cars?

Probably because insulating the breaker plate would greatly complicate the design of the distributor and drive up the price of a set of points without gaining any advantage.  If you wanted to redesign it to comply with what you want to happen here, it would be easier to just isolate the two windings of the coil and ground your HV coil directly instead of having it go through the battery (and low voltage section depending on the exact configuration).  So same setup, different coil design, no-fuss drop in replacement to conventional system.

[andy3055]

An ignition coil is basically like a transformer with a very high turns ratio. The primary has a few turns and it is fed with the battery 12 volts. The secondary has thousands of very fine wire and results in providing an extremely high voltage that makes the spark at the spark plug. I know this because I have dissected one when I was may be 15 years old.
If the primary is left connected to the battery, it will burn up after some time and that is why in older cars, people do not leave the ignition on, if the engine is not running. The transformer action happens only if the primary is interrupted. This is done by inserting the contact breaker points in the primary circuit. To prevent the contact damage due to arcing there, a capacitor is put across the points. It does not serve any other purpose. So, basically, the primary circuit goes from the battery (+ve), starter switch, the contact breaker points and ends up connected to the ignition coil primary. The other end of the primary is grounded. The battery negative is grounded, so completing the circuit. When the engine turns, the contact breaker points are opened for a split second and closed again. If you think of the time, the “open time” is smaller than the “closed time.” The magnetic field generated in this “pulsing” scenario builds up and collapses accordingly. When the circuit “makes,” there is a back emf generated in the primary which generates an opposing current there and this makes the filed weaker compared to when the circuit “breaks” when no back emf generated.
On the secondary side, the HV goes through the distributor to each spark plug in sequence of the firing order. The spark plug body is grounded as we know. Other end of the secondary coil is grounded along with the primary coil grounded end, which we call the common end (connected inside the body of the coil housing). The three terminals on the ignition coil is the common end, the primary coil input and the HV output.
The timing of the primary “break” can be made to coincide with distributor rotation such that the distributor makes the connection to any given spark plug at the time the “break” occurs, by adjusting the distributor position.

[IanB]

If I were to design an ignition circuit like this, I'd make sure that one end of the ignition coil was always grounded, like in this diagram, which is a revised version I made of the diagram (the original version being the one in the opening post of this thread). Why isn't the circuit done like this in actual cars?

Perhaps because then you would only have one coil contributing to the high voltage? And the current induced in the other coil would be wasted. With the standard design, both primary and secondary coils are in series, and the collapsing magnetic field induces a current in both, thus ensuring maximum efficiency.

[bdunham7]

Other end of the secondary coil is grounded along with the primary coil grounded end, which we call the common end (connected inside the body of the coil housing). The three terminals on the ignition coil is the common end, the primary coil input and the HV output.

On a 'conventional' system neither end of the primary is grounded.  The secondary could be, of course, and still work OK. 

[IanB]

Here's an approximate schematic of how the system would be arranged:

[Circlotron]

Here's an approximate schematic of how the system would be arranged:

That setup would work, but the normal thing is for the HV pulse to be negative going because the hot centre electrode of the spark plug emits electrons more easily than the colder ground strap. On your diagram the junction of the two windings will spike positive as the points open, and that would only be beneficial if the HV end of the secondary was positive going. Seeing you want it to be negative, this will subtract from the HV output as previously mentioned.

Edit -> I see your battery is positive ground so it would work okay. The primary pulse now spikes negative and this *adds* to the secondary voltage. That is how it was done in olde positive-ground tymes.

[langwadt]

Here's an approximate schematic of how the system would be arranged:

That setup would work, but the normal thing is for the HV pulse to be negative going because the hot centre electrode of the spark plug emits electrons more easily than the colder ground strap. On your diagram the junction of the two windings will spike positive as the points open, and that would only be beneficial if the HV end of the secondary was positive going. Seeing you want it to be negative, this will subtract from the HV output as previously mentioned.

Edit -> I see your battery is positive ground so it would work okay. The primary pulse now spikes negative and this *adds* to the secondary voltage. That is how it was done in olde positive-ground tymes.

polarity also changes which electrode that gets eroded, but many cars are/were wasted spark so every other plug have the opposite polarity and both electrodes are made from precious metals to account for that

[joeqsmith]

Here's an approximate schematic of how the system would be arranged:

"The center electrode is hotter than the outside electrode because of the thermal resistance of the ceramic sleeve that supports it.  If the center electrode is made negative, the effect of the thermionic emission from this electrode can reduce the required ignition voltage by 20 to 40 per cent."

[HighVoltage]

 Why isn't the circuit done like this in actual cars?

Many new cars these days have the secondary ground of the high tension side either connected directly to ground or through a circuit to ground to measure ionization current.

Older systems had the return path either on coil positive or coil negative. For practical purposes it does not make any difference in the car.

[Circlotron]

This fairly energetic coil I pulled out of my junk box has one end of the secondary available for grounding via the black wire.

[amyk]

This page might be useful: http://gpzweb.s3-website-us-east-1.amazonaws.com/Ignition/IgnitionTypesAndCoilWiring/IgnitionTypesAndCoilWiring.html

[langwadt]

This fairly energetic coil I pulled out of my junk box has one end of the secondary available for grounding via the black wire.

can't imaging there is many cars left that use coils like that, isn't almost everything now wasted spark, coil on plug, or a combo?

[Circlotron]

This fairly energetic coil I pulled out of my junk box has one end of the secondary available for grounding via the black wire.

can't imaging there is many cars left that use coils like that, isn't almost everything now wasted spark, coil on plug, or a combo?

Yeah, that's racecar stuff, not the sort of thing you would use for a daily driver. Plug life would probably be unacceptably short.

[Ben321]

 Why isn't the circuit done like this in actual cars?

Many new cars these days have the secondary ground of the high tension side either connected directly to ground or through a circuit to ground to measure ionization current.

Older systems had the return path either on coil positive or coil negative. For practical purposes it does not make any difference in the car.

When did ignition coils ever get a separate HV return connector? Can you show a picture of what such a coil looks like?

[HighVoltage]

When did ignition coils ever get a separate HV return connector? Can you show a picture of what such a coil looks like?

BMW started already very early on with the 12 cyl. engines.
Almost all car manufactures followed.
Since the last 20 years at least, almost all manufacturers have a coil with a return path directly to engine ground.

Here are some sample pictures.
All of these modern coils have one pin in the wiring harness connector that goes directly to engine ground.

Some modern coils even have 3 separate ground connections:
- High Voltage engine ground (Spark Return)
- Battery ground (For high current)
- Logic ground, just for the ECU input signal

Some modern coils have very high primary winding currents of almost 30A peak,
resulting in about 300 mA peak current in the spark. These coils also have some
clever processing inside the coil with lots of software. Knowing how, you can even
update the firmware inside a coil.

[Circlotron]

Internal diagram of a GM LS1 coil.

[Ben321]

When did ignition coils ever get a separate HV return connector? Can you show a picture of what such a coil looks like?

BMW started already very early on with the 12 cyl. engines.
Almost all car manufactures followed.
Since the last 20 years at least, almost all manufacturers have a coil with a return path directly to engine ground.

Here are some sample pictures.
All of these modern coils have one pin in the wiring harness connector that goes directly to engine ground.

Some modern coils even have 3 separate ground connections:
- High Voltage engine ground (Spark Return)
- Battery ground (For high current)
- Logic ground, just for the ECU input signal

Some modern coils have very high primary winding currents of almost 30A peak,
resulting in about 300 mA peak current in the spark. These coils also have some
clever processing inside the coil with lots of software. Knowing how, you can even
update the firmware inside a coil.

You mean the coil itself has an embedded microcontroller? Or do you mean the car's computer simply has connections to sensors in the coil to monitor its performance?

[james_s]

This fairly energetic coil I pulled out of my junk box has one end of the secondary available for grounding via the black wire.

can't imaging there is many cars left that use coils like that, isn't almost everything now wasted spark, coil on plug, or a combo?

New cars from the dealer? No, probably not any at all, they're all coil-on-plug now but there are LOTS of cars out there still using setups like that. Pretty much any American V8 made up into the early-mid 2000's. That particular coil is an aftermarket unit people put on hotrods and vehicles they wish were hotrods.

[xavier60]

When I bench tested some Falcon ignition coils, I noticed that they could randomly give weak spark. I soon realized that the primary is polarity sensitive. I'm assuming that the core is magnetically biased.
https://www.supercheapauto.com.au/p/goss-goss-ignition-coil-c198/631721.html?gclid=CjwKCAjw8JKbBhBYEiwAs3sxN3xBojCWDzflm-hFUgF4A2PUKFCCGwI_I9AqYciuCuAddPVHB8DFnhoCoskQAvD_BwE&gclsrc=aw.ds

[SeanB]

Plenty of vehicles now integrate the coil driver into the coil. Makes design of the ECU easier, as there are no high current high energy circuits to route through it, and you only need a relatively low level signal to provide the drive. Also they integrated this into a coil per pair of plugs as well, and also made the whole assembly a single unit, so that for MB you have a large unit that sits above the engine, containing not only the 2 plug per coil units, but also a DC Dc converter to provide a 400VDC bus to drive the coils, so they can use standard off the shelf IGBT transistors to drive them, but use low inductance coils, few turns of heavy wire, and still get both multiple firing pulse per cylinder, up to around 6 pulses per firing period at max RPM, enough to start even the weakest of mixtures reliably in stratified burn engines. So now you have both 2 coil packs, a pair of DC DC converters, around 400uF at 450V of capacitor, and a CAN bus microcontroller, all sitting exposed to 120C on top of the engine.  And these units will typically fail shortly out of warranty, they are that well designed even though the components are running on the edge of their temperature envelope in normal operation.

[HighVoltage]

You mean the coil itself has an embedded microcontroller? Or do you mean the car's computer simply has connections to sensors in the coil to monitor its performance?

Yes, some modern coils have a microcontroller inside the coil and a totally clever feedback system.
The voltage pulse width of the signal tells the coil what it should do: (Normal spark Multispark, Ion feedback, Slight retard ...)
And the current of the input signal gives feedback to the engine computer if this mode was successful by adjusting the current in 5mA steps from 5mA to 35mA for instance.

[HighVoltage]

Plenty of vehicles now integrate the coil driver into the coil. Makes design of the ECU easier, as there are no high current high energy circuits to route through it, and you only need a relatively low level signal to provide the drive. Also they integrated this into a coil per pair of plugs as well, and also made the whole assembly a single unit, so that for MB you have a large unit that sits above the engine, containing not only the 2 plug per coil units, but also a DC Dc converter to provide a 400VDC bus to drive the coils, so they can use standard off the shelf IGBT transistors to drive them, but use low inductance coils, few turns of heavy wire, and still get both multiple firing pulse per cylinder, up to around 6 pulses per firing period at max RPM, enough to start even the weakest of mixtures reliably in stratified burn engines. So now you have both 2 coil packs, a pair of DC DC converters, around 400uF at 450V of capacitor, and a CAN bus microcontroller, all sitting exposed to 120C on top of the engine.  And these units will typically fail shortly out of warranty, they are that well designed even though the components are running on the edge of their temperature envelope in normal operation.

The DC/DC converter system has been abandoned for a while already, even on expensive cars. It was far too costly and had too many problems.

These days the most reliable coil system has the IGBT and driver circuit and microcontroller inside the coil, to reach 30A primary current, for the goal to get the maximum spark current for best combustion results.

[Circlotron]

so that for MB you have a large unit that sits above the engine, containing not only the 2 plug per coil units, but also a DC Dc converter to provide a 400VDC bus to drive the coils, so they can use standard off the shelf IGBT transistors to drive them, but use low inductance coils, few turns of heavy wire, and still get both multiple firing pulse per cylinder, up to around 6 pulses per firing period at max RPM,

What car had that setup?

[HighVoltage]

so that for MB you have a large unit that sits above the engine, containing not only the 2 plug per coil units, but also a DC Dc converter to provide a 400VDC bus to drive the coils, so they can use standard off the shelf IGBT transistors to drive them, but use low inductance coils, few turns of heavy wire, and still get both multiple firing pulse per cylinder, up to around 6 pulses per firing period at max RPM,

What car had that setup?

Maybach comes to mind and SAAB