Edmunds dealer partner, Bayway Leasing, is now offering transparent lease deals via these forums. Click here to see the latest vehicles!
Diagnostic procedure for "no starts" (domestic cars)?
It is my understanding on domestic EFI gasoline that if the vehicle cranks good but won,t start, then you check for spark at the plugs, no spark, then you check for a signal at the injector/injectors, if you have neither, then you look for an rpm signal. It doesn,t matter Ford, Chrysler, GM all the same, no rpm signal, then you have no spark, no injector pulse, and other than the initial prime, no action from the fuel pump relay, correct? With the symtoms I listed, the logical process starts with the rpm signal, correct? I realize the rpm signal or the processing units ability to process the signal. I also realize if I take B+ away from the ignition module on my vehicle, there is no spark or injector pulse. Any thoughts? thanks
Tagged:
0
This discussion has been closed.
Comments
http://tools.batauto.com/?crn=104&rn=214&action=show_detail
it's a $400 item, though, so would seem to make more sense for fleets and shops to me.
as far as your analysis of the engine manglement systems, sounds about right to me. if you can get some probe leads into the crank position sensor connector, and can look at it for pulses on a scope, that should determine whether you are getting position pulses off that sensor.
similarly, you could put a voltmeter or a scope on the low voltage feed to the coils and see if they are being pulsed by the computer.. if the coils are being pulsed and there is no spark on the other side, the coil(s) are bad. little tough to yank an injector and see if it sprays, but you can at least see if the LV connectors are getting fire pulses from the computer.
but if there are no crank position sensor pulses, there is no rotation fed to the computer, and no reason to sequence any of the other events, so that's a great place to start, I agree. RPM is worked out by the period between the pulses in the computer.
it would appear to me considering how I'd program a microprocessor for this service that if you have crank pulses in, and are missing any series of output pulses, or they are mis-shapen compared to a similar engine on the same car make, the computer becomes a suspect. but I'd lift the battery cable and pull and poke the connectors to the computer (and the missing inputs) first before swapping in another computer from, say, a junkyard pull.
Basically you are checking for
1. Spark
2. Fuel
3. Compression
I'd say in most cases, if you are getting spark and fuel, you are probably getting them at the right time.
Then the questions are
Am I getting enough spark or fuel?
Do I have enough compression.
I went through this recently on my 87 LeSabre. Just drove it 2100 miles round trip to Baltimore, MD and back. Parked the car Saturday and Monday morning it quit on the way to work, two blocks from the driveway.
I spent two days checking that I had spark and fuel. They all looked good, so my conclusion was probably a slipped timing chain or insufficient fuel volume.
Took it to my mechanic and $5xx later, I had new timing chain, the gears and other etc. The oil pan was dropped too, to retrieve the 25% of the cam gear teeth that were in there.
But, the bottom line is even in the age of the computer controlled engine, you still need three things, spark, fuel and compression.
FWIW,
TB
No starts are always the first basic checks.
As was said, fuel, spark and compression.
Here is an article that may shed some light on this. It is appropriately named, Where Do I Start??
Another article of interest would be What Do I Tell My Mechanic?
oilyspill,
I think fuel pressure readings can be miss-leading because of the initial prime.
Incorrect.
If you look in engine specs for fuel pressure, you will see a key on spec and and engine running or cranking spec.
Fuel pressures are NEVER misleading.
If they are not with in the manufacturer's specs, there is a problem and engines today are extremely finicky about fuel pressure. If it is not correct, too high or too low, the engine will not want to run.
Most fuel pumps are "2 stage", prime and run.
key on, the system charges. Cranking or run, it is normal pressure. The oil pressure doesn't affect until rpms come into affect.
I know, I wrote it.
Also, [quote]With a noid light, determine if the injector is getting a signal from the engine.[endquote]
A noid light would have shown the CKP sensor loss.
Proper diagnosis is the way to figure out where to start.
I had a crank sensor that was faulty and I had fuel pressure, but it was just the initial prime, missleading.
Not if it had been checked properly. As stated, there are 2 pressure specs. Key on and cranking.
Of course, I get paid to install, maintain and repair large computer systems, so I have to know or learn a great deal about both hardware and software.
We took a class about a year ago about diagnostic techniques. Some may have heard of Kepner Trego (I think that is the name of the outfit.) Anyway, the first example, before they taught you anything was to fix the "Square Donut Machine"
Seems the scenario was to fix two problems with the machine that made this trademark square donuts.
Without knowing anything about the machine or donuts other than how to eat them and that raw materials go into one end of the line and donuts come out the other, I was the only one who fixed my donut machine.
However, I didn't fix it the least expensive fashion because I wasn't an expert in the system.
However, I did get it fixed.
It comes down to asking questions about what it is doing, or isn't doing, or what changes have been recently made, etc.
On the donut machine, I simply divided the assembly line in half and looked at the donuts at that point to determine of the problem was upstream or downstream. By dividing the system and looking at parts intelligently, I quickly isolated the area at fault.
Fixing the no-start problem is basically the same game.
What is the car doing?
What is the car not doing?
For example, mentioned above was cranking or not cranking.
And if cranking, does it crank fast or slow?
I'm not going to check fuel, spark or compression if the engine doesn't crank.
Also, I'm not going to check the crank position sensor or cam position sensors if I get spark and injector pulses. (Unless I know these can occur without these rotational position sensors.)
But to fix a problem, you have to isolate the problem.
If you throw enough parts at it, eventually you will fix it. However, it costs more to do it this way.
TB
sounds like fun, eh? cars aren't quite that bad, although if you have an interference engine, you have chances to make yourself poor real fast, no matter where you start from...
Many years ago, I had a tv problem, turned it off, took back off, and stuck my hand in there. I removed my hand real fast. I don't do that any more.
I know a broken timing belt can cause mechanical damage on an interference engine. Can a faulty cam or crank sensor cause mechanical damage on an interference engine? As far the piston hitting the valves. Right off hand, I don't know.
Some give a vehicle referance signal in voltage, but some..... give a frequency signal, so you need a wave pattern to determine if they are faulty.
I would expect just for the halibut that the structure off a hall-effect sensor comes off the end of the wire as a damped square wave with a varying duty cycle, depending on how fast the engine goes.
much more useful is to have a screen shot from a scope in the service documentation, and when you see a wrong wave on the scope from a test car, you know it's bad, rather than saying what the pulse duration and duty cycle are, and the period of delay in the leading and trailing edges of the pulse. nobody wants to learn that crap, it puts engineers to sleep in the shop. they want to see the picture.
complete service lit should have that picture to compare with... and the vendors of scopes to garages put typical screen shots on the quick-reference cards, so you don't pay for the mechanic putting himself to sleep determining with drafting tools what a decadent square pulse leading edge of 35% looks like.
as Joe Shade-Tree, you can sorta fake your way around it... see what a pulse off a running car with a similar system looks like, then hook up the victim and see what that pulse looks like on the same scope. good enough in peacetime.