Combustion Studies

Since the 1920’s, the knock phenomenon in the internal combustion engine has been the subject of intense study. Many theories with different nuances have shown that knock results from turbulence and the combustion chamber geometry. High temperature and auto-ignition can intensify the knock depending on the chemical mixture (air/fuel). Studying the combustion process is complex because there are so many variables, causing scientists to use the most advance research tools available.

In the past, researchers found that recording combustion chamber pressures showed a transient (knock) correlated with a sharp metallic noise that is known to cause damage to pistons and values. Shown to the right is a cut-away of a modified combustion chamber with a transparent piston and an optical head. A mirror is on a fixture mounted at 45 degrees inside the moving piston which has an internal relief area that allows the piston to travel without hitting the mirror. The transparent piston provides a full view of the combustion chamber during ignition. The optical slice at the top of the cylinder allows for illumination with a YAG laser, double pulsed for making Particle Image Velocimetry (PIV) measurements of the flame front velocity. The head has a pressure sensor that produces a signal proportional to the internal pressure within the chamber. This signal is typically sampled every one or half a degree of crankshaft angle. This sampled data is then plotted on a diagram of pressure vs. crank shaft angle. This plot is known as a pressure diagram. Since the flame front propagates at supersonic speeds, very short exposure times (microseconds down to nanoseconds) will reduce the motion blur and clearly image the wall of the flame. Scientists often have to trade off between shorter exposure times vs. sensitivity vs. sampling enough of the burn sequence (full combustion cycle) to produce enough data to reach a conclusion.


Recent introduction of Photron’s FASTCAM SA-X camera which has high-speed, high-resolution and high-sensitivity performance in the required spectral range is opening new opportunities for non-intensified combustion studies. Visualizing the ignition process with the Photron FASTCAM SA-X camera has never been easier or more accurate due to the camera’s sensitivity and advanced synchronization methods. Scientists using the FASCTCAM SA-X can study the stability of the flame with various air-fuel mixtures by observing small color flame variations as mixtures are adjusted while accurately synchronized to a specific crank angle. Past research efforts with other high speed cameras have fallen short of expectations due to equipment (camera) limitations.

One noted problem found in several research studies [1] was the inaccuracy of synchronizing the high speed camera to the engine with precise and repeatable results. What was noted was the image sample rate frames per second (fps) was not as precise as the study required which forced data extracted (flame velocity) from multiple combustion cycles to be averaged to compensate for the synchronization variability with previous high speed cameras.

The FASTCAM SA-X has advanced synchronization electronics which precisely lock to a sync input signal from the engine’s shaft encoder. The variability found from cycle-to-cycle with the engine is not a problem because the camera images are truly synchronous, capturing precisely at the shaft encoder sample rate. The synchronization error will be less than +/- 18.5 nanoseconds! The SA-X can synchronize easily with the highest resolution shaft encoders used in combustion studies, up to 324,000 frames per second in the synchronized recording mode. Typical encoders are at least 1/10 degree or more in resolution [2].

A 10,000 line encoder can drive the FASTCAM SA-X synchronous record rate for a full cycle at an image resolution of 1024 x 1024 pixels, 512 x 512 pixels at 30,000 lines, 256 x 256 pixels at 72,000 lines and at 324,000 lines as a resolution of 128 x 16 pixels is achieved. In addition to the advanced synchronization recording, the FASTCAM SA-X has the capability to fine tune the point at which an image is captured. The ability to delay either the sync signal & trigger signal in increments of 100 nanoseconds is unprecedented in high speed cameras. This fine adjustment allows a researcher to ever so slightly adjust the camera capture timing. If you have ever adjusted a strobe light and observed how you can stop motion when in sync, this is the same principal except you can remain in sync and adjust the observation point through a process called ‘Slip Sync‘. This is extremely powerful in research where past high speed cameras failed to capture a transient due to the exposure time not being synchronized with the event. These controls also provide a far easier method of synchronizing with a YAG laser used for PIV combustion studies. The start of exposure for the FASTCAM SA-X can be moved within the frame to be centered on the maximum light output point from the laser as shown below.

Image intensifiers have been used for flame propagation studies because in the past the fastest high speed camera’s electronic shutter speeds were around one microsecond and the hypersonic flame front was too fast to capture clearly. Image intensifiers were also used to increase the sensitivity since a fast shutter time reduces the amount of time the sensor has to collect available light. Image intensifiers have been useful but their useful life is limited, and they can be easily damaged optically by illumination on photocathode. The FASTCAM SA-X sensitivity is sufficiently high so as not to require intensification for flame propagation studies. The ISO rating of the camera is an incredible ISO 25,000 for monochrome and ISO 12,500 for color when measured with the recognized ISO 12232 Ssat method. The SA-X Advanced CMOS Image Sensor (ACIS) has a huge 20 µm pixel yielding 12-bit images. The SA-X can shutter electronically to 293.2 nanoseconds or 1/3,410,526 seconds independent of the frame rate selected. Although this is not as fast as some intensifiers [3] used in some flame propagation studies, it is better than other high speed CMOS cameras available today. The FASTCAM SA-X has either a color or monochrome ACIS. Color can be very important in studying the flame front. Below is a SA-X combustion image sequence shot at 5000 fps with a shutter of 1/5070 or 197 µs and a resolution of 512 x 512 pixels. The sequence starts at the upper left corner. Every 4th frame thereafter is shown below for illustration purposes only with text numbers added to the image. This sequence was also reduced 4:1 in resolution for this paper. You can see the spark on frame 4 and clearly the flame front propagating. Frame 106 & 110 show an after burn as a faint blue flame, possibly residuals of NO (Nitric Oxide).

The FASTCAM SA-X can produce high quality flame propagation images using an optically ported combustion chamber. When the piston has traveled as close as possible to the combustion chamber’s the crankshaft angle at this point is called Top Dead Center (TDC) or 0°. The actual firing of the spark gap normally occurs Before Top Dead Center (BTDC) by several degrees (-10° TDC) in a normal combustion cycle. When a knock transient occurs, the pressure within the chamber rises sharply and pressure oscillations are measure and viewed in the image. Researchers have generally accepted that knock may be caused by too much energy being release at an accelerated rate resulting in auto-ignition of the “end gas” ahead of the normal spark ignition or with a dilution of the bulk charge in HCCI (Homogeneous Charge Compression Ignition) engines. To avoid effects from a previous combustion cycle, researchers will run a skip-fire cycle for [7] non-combustion cycles. This clears out any residual. If a researcher is able to provide a signal that indicates only the cycles where there will be an ignition, this signal can be used to enable recording by connecting to the General In READY POS input.

Image analysis of events during engine combustion has been performed by researchers in four ways:

  1. summation and averaging to obtain cycle-to-cycle averaging of flame images at the same crank angle
  2. contour forming of the flame front based on the brightness levels
  3. contour superposition of the flame front at various crank angle positions and
  4. integration of flame area for statistical study of the behavior of flame propagations.

The FASTCAM SA-X can be easily interfaced with Matlab or LabView for such analysis. Photron provides the documentation & code for interfacing Photron cameras to these programs. Image analysis of combustion images are well supported in the industry on these two software platforms.

The FASTCAM SA-X can be used to study gas or diesel injectors using shadowgraph (see above) where very high frame rates are required to capture the spray characteristics. Illumination and fouling (auto ignition) correction can be accomplished by calibrating the SA-X black level during the non spray period. In other high speed systems, this has to be a post image process but the SA-X stores the image as calibration and subtracts from all captured images.

Synchronizing the injection time and the SA-X is very precise due to the advanced synchronization capabilities previously discussed. The uncertainties are greatly reduced over all crank angles, even at slower RPMs. Frame rates from 10,000 to 60,000 will provide detail of the initial stages of the spray. The SA-X can also be calibrated in terms of geometries for analyzing different nozzles by using the Grid found in Photron’s PFV (Photron Fastcam Viewer) software provided with the all Photron high speed cameras. The grid is calibrated by entering the distance between two know points in the image. This is a 2D calibration and assumes the spray is in-plane with the camera. Measurements can be made as a quick look before archiving the images for later analysis with more sophisticated programs such as Matlab or LabView. As shown above, the spray penetration depth is measured from the injector tip to the end of the pattern. The spray angle is measured at the halfway point upstream and tangent to the maximum spray cone. Besides doing a quick look using the PFV software, these measurements can be automated through Matlab or LabView programs.

The advantages of electronic imaging in combustion studies are the timeliness of viewing combustion chamber events in real time and the immediate analysis of images as digital data.

  • viewing real time combustion chamber events at a designated point in the cycle
  • viewing the combustion chamber events in relationship from image to image
  • interacting with the test variables while view real time combustion events
  • increasing productivity by faster turnaround of data
  • having the image data in a format that enhances the image processing techniques

The following table of electronic imaging parameters is given as a guideline for these types of studies based on the industry’s experience with electronic imaging and its applications.

Parameter FASTCAM SA-X Comments
Frame Rate 3000 – 32000 fps diesel
1500 – 100000 fps gas
6000 – 24000 fps SG
1000 – 324000 fps Framing rates depend largely on the test engine type, the revolutions per minute (RPM), and the type of event which will be studied. A good rule-of- thumb is the minimum framing rate should be 5-10x that of the engine’s RPM or the expected flame velocity.
Record Time 7 mS to 30 mS 5.46 sec – 43.68 sec Record time depends on the event of interest and the test engines RPM
Shutter Time 1us – 50 us
20 ns-1us ns intensified
Standard 1.0 us
Optional 0.293 us
Flame velocities of 50 to 100 Meters/Second are common. The hot side of a flame front can rise to sonic and supersonic levels. The shutter times depend on the velocity and the amount of light to image. 12 microseconds for a flame front with a velocity of 100 M/s will produce about 60mm front and 2% motion blur. For a flame front moving at 50 M/s, the shutter time should be 24 microseconds.
Shutter Interval Frame period ~ 500 us Frame period > 3 us
intraframe 2.8 us
Synchronizing the shutter to some interval of the crank shaft angle will significantly help to correlate the image to the combustion cycle. Shutter intervals of 1/2 to 1 degree are common. As an example, an engine running at 1500 RPM has a combustion cycle of approximately 40 ms. A one (1) degree interval would be 110 microseconds. This is a very fast interval. However, to expend most of the unburned fuel, three out of every four or seven cycles are used for exhaust. This allows for a homogeneous fuel mixture and cuts the firing interval by 4 to 1. The corresponding sampling rate (frame rate) for each degree can be slowed to approximately 2000 fps.
Sensitivity ISO 400 – 1000 ISO 25,000 mono
ISO 12,500 color
The blue-end light peaks at approximately 460 nm and has a spectrum between 390 to 510 nm. Researchers have determined that a luminescence of approximately 1000 cd/m2 is needed for direct imaging
Resolution 1024 x 1024 1024 x 1024 Minimal requirements for resolution largely depend on what needs to be resolved. Research has shown that resolving less than 1mm required approximately 10 line pairs/mm.


The FASTCAM SA-X is ideal for observing the flame structure through the visible light spectrum. One customer has found that the Photron high-speed camera’s high resolution and frame rate allow for studying the sensitivity to the second spatial derivative of the index of refraction of the edge-flame propagation using a shadowgraph.

With the great sensitivity of the FASTCAM SA-X, it is possible to compare the intensity, or brightness, of the images to qualitatively compare the light intensity between samples in time. This qualitative comparison can provide insight into the number of combustion reactions as well as the amount of heat released by the flame. This insight gives researchers valuable information concerning the spatial distribution of heat released in a cycle. In order to compare one test configuration to another, it is critical that the camera is stable and the test results repeatable. The camera settings need to be the same between tests and the lens settings including focus, f-stop and Field-Of-View (FOV) must remain the same for valid comparisons. The analysis of these images both from a temporal and spatial viewpoint can be very valuable for researchers.