Injector Info
Last updated 2007-09-18 11:11 PDT

Injector Testing

Always have at least two fire extinguishers on hand when testing fuel injectors, even with low volatility test fluids. Make sure one is well away from where any fire might break out (if your only extinguisher is in the heart of the fireball, well...).

First read this reprint of Neon John's article on building an injector flow bench (or get my local copy).

The rest of this section is directly from Dennis Novotny. (I might not know much about this EFI stuff, but I sure know who to ask, eh?)

An approved fluid for testing automotive fuel pumps (SAE spec J1537) is outlined in MIL-C-7024-B (type II). The stuff is basically Stoddard solvent (a pretty safe, high flash point cleaning fluid).

Alternate test fluids are:

SafetyKleen supplies a fluid that is pretty much the same stuff for use in their parts cleaning systems.

I recall during a project that I did with Visteon numerous years ago that "Stoddard solvent" was used for generic evaluation purposes.

But when it came to exact calibration, they used gasoline in carefully controlled and sealed benches.

In the mean time, there is also a fluid specified by SAE spec J1696 for use to test fuel filters. The fluid is called Viscor L-4264-V-93 and it is supplied by:

Rock Valley Oil and Chemical
1911 Windsor Rd.
Rockford, Illinois     815-654-2400

Note: J1537 clearly points out that there are difference between any test fuel and gasoline that may/can/will affect the veracity of the test results. J1537 indicates that it becomes the responsibility of the fuel pump manufacturer (injector tester in our case) to develop the correlation coefficient between the test fluid performance and that of gasoline.

One more thing or two:

If someone wants to "color" the fluid to make it easier to see, add some ATF to it. The amount you'd add to color the fluid "pinkish" shouldn't be enough to change the viscosity appreciably. This trick is done all the time to make it easier to CC heads.

Better plan is to inject in front of a black background with side lighting. By moving the light "just right" you'll get white spray pattern on a black background.

A thought in passing, isopropyl alcohol might also act as a decent test fluid. If the stuff is safe enough to rub on your skin, it should be safe enough to test injectors with.

Then again, one could also use Crown Royale or Seagram's VO and squirt the injector into a shot glass so as not to waste the fluid. I'd stay away from Ron Rico 151 rum as that stuff burns (151 proof = 75% pure alcohol).

Remember, if you do use 151, squirt the Ron Rico into the Coca Cola not the other way around.

Opening and Closing

I asked Ola Kristofersson to review a diagram of opening and closing characteristics of the injector mechanism that I had drawn and he graciously educated me on the subject.

Here are his raw comments, I need to fix my picture up to look as good as his.

Attached is a picture drawn after having again looked at scope traces, I have built a simulator/injector tester, so it is quite easy to refresh knowledge.

Yes, the flyback voltage gives a polarity reversal when the driving voltage is switched off. This is the famous Lentz Law that says that the driving EMF will always have the polarity that strives to maintain current direction and amplitude at constant level.

However, I am not sure that this makes much sense or is very important and I guess is best to just state it when explaining injector and injector drive behavior.

What is really important however is to state that the injector is a current controlled device and that the force on the pintle is a constant multiplied by current multiplied by the number of winding turns.

Also that the injector has quite a hysteresis, opening current for a hi-z injector is approximately 65% of steady state current and closing current is about 25% of steady state current.

Some comments on your text A to F
A is fine
B is fine ( maybe magnetic field strength)
C is fine and is probably around the time stated
D yes, Lentz Law and all that. BUT, the spike is polarity reversed.
E is fine
F is fine

(From later correspondence after Ola had the simulator done.)

My absolutely OT simulator with built in injector tester (electric test only) is now ready.

For your synopsis on the injectors I have a suggestion to demonstrate that, yes, the quirk in the current IS when the armature flies. Monitor injector current with a normally operating injector. Quirk comes after appr 0,7 to 1,3 ms depending on injector type and make.

Take a piece of hard plastic and push against the pintle where it protrudes. The harder you push the less of a quirk, and if you hold hard enough there is no quirk at all, ie a steady state magnetic circuit only influenced by the magnetic material and no change in reluctance since there is no armature ( pintle ) flight from closed to open.

Wish I had a scope camera, would send you the pics, but...

My picture below is all wrong, click on Ola's picture above to understand what is going on until I get them merged.

X axis is time in ms; Y axis is arbitrary.

Logic - The input signal from the ECU to the injector driver.

% Open - Physical position of the injector pintle, 0.0 is closed, 1.0 is fully open.

Voltage - Idealized voltage measured across the injector coil. 1.0 represents the driver supply voltage, nominally 12 volts. For purposes of illustration, I've chosen a saturated injector, so the injector output mirrors the input without any PWM or peak & hold logic.

A - Logic signal turns on voltage to the injector.

B - Field strength in the coil has reach a level that exceeds the spring force in the injector, and the pintle begins moving from closed to open.

C - The pintle reaches full-open in about 0.1 ms from time of first motion.

D - The logic signal is turned off, and the field in the coil collapses causing a flyback voltage spike.

E - The field strength has decayed sufficiently that the spring can now push the pintle closed.

F - (NOT ON PICTURE YET) The pintle valve has closed completely, time from E to F is again about 0.1 ms.

The sum total fuel injected from points B-C and E-F is the less than would be injected when the injector is full open for 0.1 ms.

Not2Fast hosted by TrackFiends