The automotive ignition system has evolved from the old mechanical cap-and-rotor system into one controlled by a central computer linked to sensors monitoring crankshaft position, engine speed, exhaust and other variables. The computer processes the data and triggers the pulse for each cylinder’s coil.


To connect the coil to the plug, many digital electronic ignition systems rely on a spring-type device instead of the traditional mating terminals. While this system has improved fuel economy, boosted engine performance and reduced emissions, it has not been without problems.

Consumers have complained of severe engine skips and burned or deteriorated silicone boots. In some cases, the spring-type connector has actually welded itself to the plug. Although the latter issue certainly eliminates the problem of loose connections, it creates a new predicament. It impedes the use of a socket on the plug and destroys any possibility of reusing that connector.

Given the high cost per coil, ETCO’s research and development team embarked on an investigation of performance issues with industry-standard coil-to-plug connections. We compared an assortment of connecting devices. Our tests exposed each connecting device to the same conditions, providing an “apples to apples” basis for comparison. The data enabled us to gauge the performance of various connectors and develop ideas for improvement.

Samples and Measurements

After reviewing an assortment of OEM, aftermarket and prototype connectors, we selected 10 to evaluate. Prototypes were constructed using different combinations of terminals, terminal types, metals and wires.

We tested the following connectors:

  • Sample 1: OEM connector, stainless steel spring with resistor on one end.
  • Sample 2: OEM connector, stainless steel spring with ferrite rod.
  • Sample 3: Aftermarket connector, stainless steel terminals with ferrite bead on stainless steel cable.
  • Sample 4: Prototype, zinc-plated steel terminals on 9-millimeter silicone suppression wire.
  • Sample 5: Prototype, zinc-plated dimple terminals on 8.5-millimeter suppression wire.
  • Sample 6: Prototype, silver-plated terminal and disconnect on 8-millimeter suppression wire.
  • Sample 7: Prototype, zinc-plated steel terminal and disconnect on 8.5-millimeter wire.
  • Sample 8: Prototype, zinc-plated steel terminals on 7-millimeter carbon-core wire.
  • Sample 9: Prototype, gold-plated terminals on 8.5-millimeter wire.
  • Sample 10: Prototype, one-piece stainless steel terminal with disconnect, retention clip and ferrite rod core.

Our tests included:

  • Static resistance.
  • Voltage drop under load.
  • Coil output voltage at a 0.04-inch air gap.
  • Retention force.
  • Effects of vibration on the connection.
  • Visual evaluation of sparking events at a 0.4-inch gap.

Resistance and Voltage

A digital multimeter from Keithley Instruments Inc. was used to measure static resistance for each connector. We found that resistance varied widely among the 10 connectors. At 2,700 ohms, Sample 1 exhibited the most resistance of any connector. Sample 3 and Sample 10 had the least resistance, 0.37 ohm.

To measure voltage drop under load, we created a fixture that enabled us to obtain accurate measurements using a conventional oscilloscope. This was accomplished by attaching the conducting wire to the ground portion of the spark plug. Because it is a series circuit, the same current is required to jump the gap and travel through the plug and connecting device to ground. This allowed measurements over the connector in reference to ground, rather than 2,000 to 40,000 volts.

In addition to high-resolution 1X and 10X oscilloscope probes, a 1,000X high-voltage probe was used to monitor the coil output.

A regulated 13.8-VDC power supply provided the voltage to the control boxes, which regulated pulse rate, voltage output to the primary coil, and dwell of the signal to the coil. Radio-shielded cases, connectors and cable were used to reduce noise that could interfere with measurements.

We found that:

  • Sample 1 had the highest voltage drop at 365 volts. That’s almost 25 percent of the coil’s output at 1,500 volts.
  • Samples 2, 3, 5, 7, 9 and 10 all had voltage drops consuming less than 1 percent of the coil’s output voltage.
  • Samples 3 and 10 had voltage drops of less than 1 volt. In fact, Sample 10 averaged a voltage drop of 0.08 volt.

Retention, Vibration and Sparking

The Society of Automotive Engineers (SAE) has established minimum withdrawal force requirements for spark plug terminals. In accordance with SAE Standard J2032 3.2.1, anything measuring below 20 newtons is not an acceptable connection.

Samples 1 and 2 had no measurable retention force. A Chatillon force gauge was used to compress the spring-type devices 0.5 inch from their extended state to about the length they would be under the boot. The compression force required to do this was less than 3 pounds for both Samples 1 and 2.

The remaining eight connectors were comparable. Each required 63 to 70 pounds of force to separate them from their spark plugs.

A secure connection should maintain uninterrupted function despite vibration. Breaks in a connection can cause arcing or totally stop the flow of current.

We developed a fixture to subject the connector-plug assemblies to vibration. The fixture enabled us to adjust both the frequency and amplitude of the vibrations.

Although our test design was rigorous, it’s important to note that spark plug connectors might see much more traumatic vibrations during actual operating conditions. Vehicles are exposed to country roads and potholes daily. High-performance vehicles at high speeds might easily see much more shaking.

Samples 1 and 2—the OEM connectors—were the only connectors to exhibit arcing during our vibration tests. Sample 1 showed very little mating contact with the spark plug terminal. The inside diameter of Sample 2 easily cleared the outside diameter of the spark plug terminal. A loose fit between mates, low spring pressure, and lack of support inside the insulator boot contributed to arcing problems with Sample 2. Arcing occurred at the top and bottom of the hourglass-shaped terminal.

We also assembled a fixture to visually compare sparking events for each connector. The fixture enabled us to view any two connecting devices as part of a functioning circuit.

Simultaneous triggering of the coils with adjustments to speeds and gaps showed no visually distinguishable differences between any of the devices. The coil’s ability to deliver enough voltage appeared uninhibited by any combination of speed, gap, device or coil.

Our apparatus did not expose the device to the load required to jump a gap inside a combustion chamber. However, it does offer a comparison of each device to an extended air gap. With every connector, all sparking events created in our experiments were bright and blue. In every case, the coil’s output appeared to overcome any inefficiencies of the device.

Reviewing Connectors

Our tests revealed that certain connectors used with coil-on-plug assemblies separated at forces lower than those required to separate mated terminals. During experiments, these separations resulted in arcs to bridge the gaps. In the confines of the lab, these gaps were sporadic and measured several thousandths of an inch. The frequency and gap size would be expected to be larger with exposure to real-world driving conditions.

The use of an individual coil per spark plug could help overcome the frailties of a connection for a period of time by sending more voltage. However, the long-term effects of high-voltage arcing will degrade these intermittent contact points in the circuitry and lead to deterioration of the insulator boot or coil failure.

Our research shows that terminal-style devices demonstrated the most efficient connections. No arcing was observed with any of the terminal devices. Sample 1, a spring-type connector, had a voltage drop measurement of 365 volts. It lost 24 percent of the coil’s output voltage to complete the circuit.

Sample 3, an aftermarket terminal-style connector, had a barely perceptible resistance of 0.37 ohm and a voltage drop of 0.5 volt. This device delivered 99.04 percent of its coil output voltage.

Sample 10, a prototype terminal-style connector, has features designed specifically for mating to a blade and a spark plug. One end of Sample 10 has the disconnect form for mating to a blade and the other end has the spark plug terminal with a clip. The coils used in this comparison have a blade terminal that mates with a connecting device. Although there is an interference fit between this blade and every device tested, the surface-to-surface contact between the circular devices and the blade terminal is not as encompassing as the disconnect-type terminal.

Sample 10 was created using a combination of terminal features that are ideal mates. This terminal has a 0.25-inch ferrite rod in its center for noise reduction and had a static resistance measurement of 0.37 ohm. It recorded a voltage drop of only 0.08 volt and delivered 99.92 percent of its coil output voltage. It was the most efficient of all devices in this comparison.

A good connection to a circuit has always been essential, perhaps even more so with the complicated and delicate circuitry of today’s vehicles. Connectors of the coil-on-plug system are shielded from the accidental disconnections that long ignition wires were once susceptible to, but they remain exposed to all the bouncing and vibration from road conditions and high speeds.

Under these conditions, a connecting device should be retained by a physical coupling that offers resistance to movement at the point of connection. The mated terminal proved to be the best at this when exposed to the road-like test conditions. Connectors using terminals attached to quality wire or cable exhibited superior power-conducting efficiencies.