Failed Again! 1977 Seville

[MarkV]

Replaced the thermostat with the 195 and coolant sensor on the 77 seville. And it failed smog again for being too rich, so I am at a loss, I took it into another mechanics shop, he seems to be pretty confident he will find out what's up with it. 

[old car fan]

Yes,anyone can throw parts at it.find a new shop.

[pfloro]

As "old car fan" said:  Stop throwing parts at it.  This is a random and expensive way to troubleshoot any problem. 

 

<begin soap box>

 

I'm sounding like a broken record (actually it's called a "locked groove") but once again:  Check the fuel pressure at the fuel rail.

I did some research and too high a fuel pressure (everything else being equal) will squirt more fuel at the intake port during the 

period of time the injector is energized.  The fuel pressure regulator might be malfunctioning.

 

You need a mechanic who can think outside of the box.  Remember, this is ancient electronic fuel injection technology (although it was sophisticated for it's day).

There is no diagnostic port (OBD II) to which a scan tool can be connected to read out stored trouble codes and real time data.  You can't even "flash out" the trouble codes (OBD I)

via the Check Engine light because there isn't one. There is no CPU and there is no memory because it's an analog control system.

That being said, there are ways to troubleshoot this system with a digital Volt / Ohm meter.  I used to have more information about this Bendix/Bosch/Delco system and will search for it.  

 

</end soap box>

 

I need another beta blocker now...!

 

Paul

Edited June 15, 2018 by pfloro (see edit history)

[MarkV]

7 minutes ago, pfloro said:

As "old car fan" said:  Stop throwing parts at it.  This is a random and expensive way to troubleshoot any problem. 

 

<begin soap box>

 

I'm sounding like a broken record (actually it's called a "locked groove") but once again:  Check the fuel pressure at the fuel rail.

I did some research and too high a fuel pressure (everything else being equal) will squirt more fuel at the intake port during the 

period of time the injector is energized.  The fuel pressure regulator might be malfunctioning.

 

You need a mechanic who can think outside of the box.  Remember, this is ancient electronic fuel injection technology (although it was sophisticated for it's day).

There is no diagnostic port (OBD II) to which a scan tool can be connected to read out stored trouble codes and real time data.  You can't even "flash out" the trouble codes (OBD I)

via the Check Engine light because there isn't one. There is no CPU and there is no memory because it's an analog control system.

That being said, there are ways to troubleshoot this system with a digital Volt / Ohm meter.  I used to have more information about this Bendix/Bosch/Delco system and will search for it.  

 

</end soap box>

 

I need another beta blocker now...!

 

Paul

The guy I took it to got his start on these he is going to check the reg. First

[padgett]

People are too used to modern computer cars. Can toss higher flow injectors on and the O2 sensor will detect the "too rich" and correct the PW. AFAIR you do not have an O2 sensor so the "computer" is guessing. I suspect you will need an adjust able fuel regulator so you can tune the flow to meet emissions.

https://www.ebay.com/itm/Black-Adjustable-Fuel-Pressure-Regulator-with-6AN-Fitting-End-100psi-Gauge-WT-/272675970012?_trksid=p2349526.m4383.l4275.c10#viTabs_0

 

May also need a sniffer.

 

 

[pfloro]

Quote

The guy I took it to got his start on these he is going to check the reg. First 

This is good news...  As has been said, half of the problem is finding someone who has real field experience with the system in question.

 

This system had it's idiosyncrasies but it worked reasonably well...

 

Keep us posted.

[padgett]

OK there is about 30 pages on the FI system starting on page 6-39 of the '76 Seville manual. I expect the 77 is the same.

 

There is an Air Injector Reactor. If it fails the exhaust will be too rich. There is an EGR. If it fails the exhaust will be too rich.

 

I have a Sun Sleuth II with exhaust gas analyzer. Haven't needed in years since Florida does not test emissions and would be easier to ship the Cad here than it there. Maybe you need a street racer shop, they check the HC and CO all of the time.

[Bloo]

If Cadillac made this thing run too rich when it was new, I'll eat my hat.

 

34 minutes ago, pfloro said:

I'm sounding like a broken record (actually it's called a "locked groove") but once again:  Check the fuel pressure at the fuel rail.

 

Yep, that DEFINITELY  needs to be checked at this point. I cannot stress enough that it should be done the way the factory shop manual calls for, whatever that is.

 

Also, if it has not been done, the vacuum hose needs to be pulled off of it and checked for fuel in the vacuum hose. If there is no fuel coming out of there, and it is safe to do so, the vacuum hose should be pulled off with the engine running. The reading on the fuel pressure gauge should jump up about 5 pounds.

 

While I agree with pfloro that the pressure should be checked, my fear is that pressure will be out of spec by two pounds (just because cars often are), or five pounds off because someone did not check the fuel pressure the same way the factory said, and we will all assume we have found the problem, and it will fail again.

 

If it passes the hose pull test (no fuel in the vacuum hose, pressure jumps up roughly 5 pounds when you pull the hose off), there is a 99 percent chance there is nothing wrong with the fuel regulator. Yes, I know 99 is not 100.

 

If the spring was weak from age, the pressure would go down, not up. I have seen the fuel pressure go way up without leakage out the vacuum fitting, once. I suspect the spring broke and stacked up in some way that limited motion. IMHO you have a better chance of being struck by lightning than encountering that.

 

I may have mentioned this before, but check the timing, and be sure you are using the factory method. Back in those days, methods for checking varied, sometimes with a vacuum hose on or off, sometimes there was an electrical connector that had to be disconnected, etc. etc. If the timing is wrong, the vacuum could be low. If the vacuum is low, the MAP (or VAC if thats what it is) would richen the mixture at less throttle than normal, causing a failure.

 

 

 

[Bloo]

28 minutes ago, padgett said:

There is an Air Injector Reactor. If it fails the exhaust will be too rich.

 

If the smog equipment were completely missing on this car, it would STILL be too rich. Trying to cover up a richer than normal condition with AIR often just melts the catalytic converter.

 

28 minutes ago, padgett said:

There is an EGR. If it fails the exhaust will be too rich.

 

I hate to be so contrary, but I would expect the opposite, and a very small change, unless the failure was a leak (it often is) causing a misfire at idle, AND that reduced the vacuum enough that the MAP compensated with more fuel. Possible, but I doubt it in this case. EGR problems usually change the mixture WAY less than you would expect. Exhaust is mostly inert, and exhaust from EGR takes up space in the cylinder that would otherwise be occupied by air. The air is much more voluminous than the fuel. If you take the EGR away, that space will be occupied by air, leaning the mixture slightly, all else being equal. The part that will amaze you is how little difference.

 

28 minutes ago, padgett said:

I have a Sun Sleuth II with exhaust gas analyzer. Haven't needed in years since Florida does not test emissions and would be easier to ship the Cad here than it there. Maybe you need a street racer shop, they check the HC and CO all of the time.

 

YES THIS^^^  A gas analyzer, preferably four or 5 gas is what is needed to troubleshoot stuff like this.

 

Edited June 15, 2018 by Bloo (see edit history)

[padgett]

OK way back in the other thread I recall mention that you changed out the injectors with ones from a 'vette. The Cad injection has no feedback. It assumes stock injectors (17# ?) and if you replace with larger (say 21#) then you need to reduce the stock fuel pressure (39 psi) to about 32 psi - where I would start then tune.

 

Back in the day I once turned a 455 Buick with cat so tight it squeaked - when they put the sniffer in for the '75 FEA, it read zilch. Checked twice against other cars. Zero. Still have the plaque.

[mike6024]

Does this system have a cold-start enrichment device which is not turning itself off? Reason I ask is my '81 FI does, and you need it to come on to start when cold and go off after it warms up, and they often malfunction.

[mike6024]

Looks rather complex if this is it.

 

http://www.cadillacseville.org/elctroubleshoot78.html

 

[mike6024]

Rich operation

Back to the top

1. Short circuits in the following wires can cause this condition:

a. Coolant Sensor
b. Air Sensor
c. ECU and Jumper Harness
 

2. A leak in the MAP hose will cause this condition as it will alter the signal received by the ECU. To check the MAP hose, disconnect the hose at the throttle body and connect a hand vacuum pump such as 1-23738 to the hose. Develop a vacuum of 15” in the hose and check for leak down. A damaged hose can be replaced without replacing the entire line by splicing in a section of hose which conforms to GM spec. 6107 M.

Also check the vacuum hose at the fuel pressure regulator as a leak in this line will alter the MAP signal as well as increase the fuel pressure.

3. Start engine and run for at least 1 minute. High fuel pressure will cause rich operation as a greater quantity of fuel will be injected with each injector opening. Normal fuel pressure is shown in Fig. I.

4. If the car has high fuel pressure, this fact proves the capabilities of the chassis pump and isolates the problem to either the regulator or return hose. To check these components, remove the fuel return hose from the pressure regulator and run a line between the regulator and a suitable container. Run the engine and observe the fuel pressure. If the pressure is normal, the restriction (return hose) has been eliminated. However, if the pressure is still high, a noncontrolling regulator is indicated.

5. The coolant sensor provides low resistance when it is cold. If the sensor is shorted, this low resistance will continue to be supplied to the ECU which will continue to provide a long (rich) pulse width. The resistance of the sensor can be determined with analyzer 1-25400 in step 10, or can be determined with a VOM (volt-ohm meter) at the connector. A defective sensor is indicated if the resistance value does not compare with the values shown in Fig. 3.

If the sensor checks OK, it is still necessary to insure that the resistance signal is being delivered to the ECU. This can be accomplished with analyzer 1-25400 or by probing terminals “D” and “G” of the red 9 way connector with the VOM. This value should be the same as that measured at the sensor.

6. Vacuum leaks at the throttle body will alter the MAP signal the same as a leaking hose. Inspect the throttle body for this condition.

7. An injector which is sticking open will cause a rich condition even though all other components are operating normally. This condition can be analyzed as described on Pages 44 through 49

If no defects have been found at this point a new ECU should be installed on a trial basis.

 

http://www.cadillacseville.org/technical.html#11

[mike6024]

testing the coolant sensor. if it reads too low will run rich. http://forums.cadillaclasalleclub.org/index.php?topic=107883.0

[Bloo]

1 hour ago, mike6024 said:

Looks rather complex if this is it.

 

THANK YOU. Finally, some actual documentation of the system.

[Bloo]

1 hour ago, padgett said:

OK way back in the other thread I recall mention that you changed out the injectors with ones from a 'vette.

 

WHAT?? I missed that. This issue seems to have a lot of threads.

 

If they are higher flow, there's the problem.

Edited June 16, 2018 by Bloo (see edit history)

[padgett]

Actually looks pretty simple. No PID feedback.

[mike6024]

Yeah, the most likely guess would be a bad coolant temperature sensor. That would make it run rick, because it's thinking like the choke needs to be pulled for a cold start situation, and the choke never gets turned off. The way to test it would be to completely remove it, and disconnect both wires from it, and put a "voltmeter" on it but measure the resistance. DVM - digital voltmeter I suppose. Need to locate the coolant temperature sensor first though of course.

[bobg1951chevy]

Regardless of how  out of sync the car may register, the bigger problem is living in a state where a 41 year old car has to be smog checked !

 

Sad.

Edited June 16, 2018 by bobg1951chevy (see edit history)

[joe_padavano]

People need to READ the factory service manual before working on this car. As I pointed out waaaay back in the OP's first thread about this car, the port fuel injection used on this 350 Olds motor is ancient.  It DOES NOT use an O2 sensor. There is NO feedback on the A/F ratio.  It runs open loop all the time, by design. The ECU is an ANALOG computer, not digital. I doubt there are more than a handful of mechanics left in the country who understand how this EFI works, much less are able to troubleshoot it. Also, the injectors used in this system are unique to this system.  You cannot simply install "Vette" injectors - they do not fit in the holes in the manifold.

[Bloo]

1 hour ago, bobg1951chevy said:

the bigger problem is living in a state where a 41 year old car has to be smog checked !

 

It doesn't have to smog anymore in WA. In WA you can just keep driving it broken until all that extra gas washes the rings out of your engine. It is pretty hard to screw up an Olds 350, but that might do it.

 

I just don't remember issues getting one of these Bendix systems through smog. It had better fuel distribution than a carburetor. Like most cars, once I got it running right, I didn't need to worry too much about a tailpipe test.

[John348]

Was there a difference the levels from the emission test before?

[MarkV]

Coolant sensor replaced be before

 Smog, it tested higher cos than before. They are putting it on a gas analyser on Mon. They will be checking the vacuum fuel pressure, etc. It frankly got worse between the smog checks therefore the problem is getting worse whatever it is. Leads me to vacuum, failing regulator or injector.

[MarkV]

I have the fi manual and did some of the trouble shooting myself before giving up. 

[misterc9]

In New York we don't have the tailpipe sniffer. We do have to have the check engine light not on to pass inspection.  A vehicle with a check engine that won't stay off can get a waiver to get an inspection certificate. The vehicle owner after spending a certain amount of money on repairs can furnish receipts for emission/check engine light work to an inspection station. These receipts would need to be for work relevant to the problem and from a registered repair shop. The inspection station can then apply for and then get the waiver to issue an inspection certificate.  California must have something similar to this.

I don't know what gas is available in CA. I would think a non feedback car would run leaner on gasohol than no alcohol gasoline. I might even try adding a small amount of E85 to dilute the fuel. Nobody wants to put gasohol in their old car but. I don't know how far over the limit you are rich. Maybe the fuel difference wouldn't even help it.

[mike6024]

California has different standards for different vehicles and years. When I had my 1976 with carburetor and no catalytic converter it could pass if hydrocarbons (HC) were below 300 ppm, but it could be adjusted to give about 100 ppm or even below that so I would get it to pass easily even if at 120 ppm. My 1987 Japanese import with fuel injection and catalytic converter was I think 100 ppm or less to pass, and i failed and needed a new Cat which brought it below 100 ppm easily. And now my 2002 Toyota pickup has 58 ppm HC as allowable max and it passed last time with 19 ppm.

 

This Cadillac has max HC 129 ppm I see.

 

Edited June 17, 2018 by mike6024 (see edit history)

[padgett]

Just noticed it failed on excessive carbon monoxide, not HC. It is called "too rich" but is really incomplete combustion. See here. If fuel pressure checks out maybe time for something else ?

[Bloo]

High CO results from a mixture that is too rich. High HC is generally caused by a misfire of some kind, such as a lean misfire or an ignition problem (among other things).

Edited June 21, 2018 by Bloo (see edit history)

[joe_padavano]

Have the injectors been cleaned? It doesn't take much buildup on the pintle to hold them open enough to cause excess fuel to be dumped into the intake.

[60FlatTop]

Being an old boiler operator, I have seen elevated CO levels when the products of combustion "lick" a relatively cool spot. I wasn't kidding about blocking the radiator with a piece of cardboard on the way to the test. Get everything as hot as you can. You are close, just a nudge will do it.

 

When taking any kind of test, and knowing what the answer should be, doesn't sound like a big problem.

Bernie

[padgett]

Try this or similar.

[John348]

On 6/15/2018 at 5:17 PM, MarkV said:

Replaced the thermostat with the 195 and coolant sensor on the 77 seville. And it failed smog again for being too rich, so I am at a loss, I took it into another mechanics shop, he seems to be pretty confident he will find out what's up with it. 

 

I think now it is pointing to the MAP sensor circuit. The fact that it is located in the ECU, and there is a vacuum connection inside of it is suspect. Vacuum hose sitting inside an ECU in a car subject to 41 years of California heat makes me think.... I would suggest that you locate the vacuum hose that supplies the MAP information to the ECU and explore further. You are almost there,,,,,,,  

Edited June 18, 2018 by John348 (see edit history)

[John348]

1 hour ago, padgett said:

Try this or similar.

 

That tool is not designed for this type of application and heat generated from the exhaust might cause a problem for the tool. These are designed to be used in confined spaces such as manholes for personal life safety. It would have to be modified for the type of use you are suggesting

[60FlatTop]

35 minutes ago, John348 said:

These are designed to be used in confined spaces such as manholes for personal life safety.

 

Like one of these.

[capngrog]

29 minutes ago, John348 said:

 

That tool is not designed for this type of application and heat generated from the exhaust might cause a problem for the tool. These are designed to be used in confined spaces such as manholes for personal life safety. It would have to be modified for the type of use you are suggesting

 

You are correct about that meter being designed for testing air quality in confined spaces.  They are also designed to operate in an environment of little or no gas flow (velocity), so introducing the meter's sensor into the exhaust  pipe of a running engine might confuse it.  That's just my feel for the situation, based on little other information.  A call to the manufacturer of the meter should answer these questions.

 

The exhaust pipe temperature of a running (1400 rpm) 1977 Seville should be fairly low, but, again, I'm not sure what those values would be.  I'm guessing that the exhaust pipe temperature at 1400 rpm would be fairly low ... on the order of 200°F +/-.

 

Cheers,

Grog

[Harold]

Have you checked the plastic air intake  'accordian' for cracks?  That could be affecting the air/fuel mixture.

[Bloo]

The air temp sensor should be verified just in case. It would richen things up if bad.

 

John348 is correct that this points to the map. I am almost sure at this point that the solution is going to revolve around the MAP somehow.

 

I did not originally understand how California's test works, and could not understand why there were two lines with about the same engine speed. First of all, apparently California does not test at idle in a case like this... at all.....  (they must have solved all those traffic jams. I should look into moving there). No, this is an "ASM" test (only), run on a dyno.

 

I should take a moment to explain that power enrichment on cars is not usually gradual. It is either on or off. There are exceptions, like 60s Chryslers with 3 stage fuel metering, but even those aren't really gradual in the transition to power mixture. On a carburetor with a typical style power valve, like a Holley for instance, the enrichment is either on or off. GM CCC shuts off the fuel control solenoid altogether at wide open throttle (full rich). Some cars have a "three dimensional fuel map". They really do, but it is just little changes to improve things under different conditions. For all practical purposes, the enrichment is either on or off. This applies to almost every car ever made that is new enough to have power enrichment. You could do it with a switch. In fact it has been done. If anyone doesn't believe this, hook a vacuum gauge to your car, tape it to the windshield, and go for a drive. the transition from cruise to acceleration is several inches of vacuum. You'll see.

 

Now back to the California test. MarkV's post shows that it is an ASM test. Whats that? Well according to this (page 12):

 

https://www.bar.ca.gov/pdf/Smog_Check_Manual_ENG_2013.pdf

 

           

"The ASM emission inspection includes two loaded mode sequences known as the 50/15 test and the 25/25 test"

 

So then, what the hell is a 50/15 test and 25/25 test? California didn't bother to say, but according to this (page 2):

 

https://www.saddleback.edu/uploads/atas/autotech/courses/bar-97-transition-class-student-workbook.pdf

 

"ASM testing equipment tests the vehicle at two constant speed and load levels

          (A). 50% load at 15mph (known as a 50/15 test)

          (B). 25% load at 25mph (known as a 25/25 test)

 

Now we are getting somewhere. I see what they are up to. When you take off from a stoplight, you probably give the accelerator a pretty good stab. When the car gets moving you probably let up and continue accelerating, but with less pedal until you get to your desired speed, then you back off on the pedal some more and cruise. The are trying to see if you are getting into the power enrichment mode during acceleration in normal driving (not hard acceleration).

 

Remember that mode is almost like a switch.

 

A perfect cruise mixture (IMHO) is stoichometric, or 14.7 to 1 air to fuel by weight. This corresponds to 0.3% CO (theoretical) or 0.7% CO (practical). Leaner cruise mixtures give greater fuel economy, until you get to the point of lean misfire. GM was apparently trying to do that in the late 70s on carbureted vehicles (it did not work out well). They very well could have calibrated this car leaner than 0.7. They wouldn't have calibrated it richer. These are numbers all for the engine itself. The tailpipe CO emissions (after the catalytic converter) would be lower.

 

Now lets have a look at MarkV's numbers again.

 

HC         47ppm    43ppm

CO         1.60%      2.00%

O2          0.00%     0.00%

CO2       14.1%     14.0%

NO(x?)  108ppm  97ppm

 

Now, what would I expect? On a healthy engine, BEFORE the catalytic converter, something like this:

 

HC         188ppm

CO          0.7%

O2          2.2%

CO2       13.6%

NO(x?)   400ppm?

 

So what do the numbers show us? They show us that the catalytic converter is working. HC is below 100ppm, and you generally don't get below 100 without a catalyst. Low 02 proves you really do have a sample of the exhaust. CO is 14, and you don't usually get that high without a catalyst either. You might get 13.8 if the mixture wasn't too rich, but it is too rich. 1.6% CO proves that it is too rich. NO(x) wasn't looked at much in my day, but MarkV's numbers are lower than what a healthy engine emits, so that also points to a working catalytic converter.

 

Since the catalytic converter is working, we can infer that it is reducing the CO also. From MarkV's numbers, I would guess the actual CO is 6% or higher. 6% is in the neighborhood of a perfect full throttle rich fuel mixture for maximum power. Some engines may like a little more, and many are calibrated more like 10-12%, sometimes causing a slight reduction in power.

 

Yes, that is all wild speculation, but I used to do this for a living and I'll bet it isn't far off.

 

Since this looks like a full throttle power mixture, there is a good chance that this car IS into it's power enrichment. How does the system decide when to do that? With the MAP sensor. How does the MAP sensor determine that? By measuring the Manifold pressure (vacuum). When vacuum gets low enough, the power enrichment comes in.

 

Why would the vacuum be low?

 

Cam timing could do it (probably not on this car, it has a new cam and cam chain)

 

Ignition timing could do it. (It needs to be checked, or re-checked, and set in the right way, whatever that is. Many cars of the late 70s and 80s have some "gotcha" in the timing procedure).

 

Vacuum leaks could sure do it. Either a leak in the map hose causing the MAP to measure a lower vacuum, or just general vacuum leaks everywhere, causing the vacuum to be low. Bad diaphragm in the purge valve? Bad diaphragm in the diverter valve? Hole in the brake booster diaphragm? Just leaky old 1977 vacuum hoses that don't fit on their nipples tight enough?

 

The MAP itself could be bad, but I would suspect dirty connections or vacuum leaks first

 

Remember that this car was backfiring through the throttle body due to the bad camshaft. That could have blown out a diaphragm somewhere.

 

Edited June 21, 2018 by Bloo (see edit history)

[padgett]

Well sorta. GM computer cars of the 90s had a "lean cruise mode" of 15.7:1 and "Power Enhancement" (PE) may drop to 10-12:1 (why it is better to downshift for a hill than let it go into PE in top gear).

 

However since the Seville has no feedback (O2 sensor), it is just guessing and evidently needs tuning. Fuel line pressure is the easiest thing to change without impacting other things.

[John348]

2 hours ago, padgett said:

Well sorta. GM computer cars of the 90s had a "lean cruise mode" of 15.7:1 and "Power Enhancement" (PE) may drop to 10-12:1 (why it is better to downshift for a hill than let it go into PE in top gear).

 

However since the Seville has no feedback (O2 sensor), it is just guessing and evidently needs tuning. Fuel line pressure is the easiest thing to change without impacting other things.

 

If that is done then all he would be doing is compensating for a problem without finding the problem? Secondly it really is not a "GM" computer car as we know it, it is a Bendix unit and is similar to the K-Tronic that was used in European cars in the 70's.

[padgett]

Ok so find someone who repairs/tunes Volkswagen Squarebacks.

 

ps was commenting on A/F mixtures in GM P4 systems not Bendix/Bosch K-Jetroninc.