kishy's 1985 Country Squire

[gadget73]

if the sweep on the sensor is flukey there isn't a lot of hope for that to work reliably. Hopefully a new one will sort it out for you.

honestly didn't realize the CFI cars used a different sensor than the SEFI cars but I can't say I've ever paid attention to it either.

[kishy]

Last night, in pursuit of a massive oil leak, I replaced the intake manifold gaskets. I used a technically "incorrect" kit for an 86+ engine because that's the only way to get the fancy silicone end pieces.

Process went fairly smoothly. I've done it before on this one, but it seems the cork rear gasket was leaking at one or both ends, so probably a poor RTV job on my part. Gaskets came off pretty cleanly with minimal scraping required, cleaned it all up, replaced a couple bolts that look ugly but fortunately haven't failed. I did use Ultra Copper gasket sealant spray on the gaskets between the heads and intake, which I don't think was necessary but hopefully hasn't hurt, and used Ultra Black RTV in the 4 spots where the 4 pieces meet each other. I allowed all sealants to set for a few minutes before assembly so it all tacked pretty nicely, and torqued all fasteners to a very low amount, let it sit for maybe an hour, then torqued evenly a bit more and let it sit overnight.

Today after work, I re-torqued everything and reassembled all the bits and pieces. While doing so, I selected one of the 1985-correct EGR valves, but swapped onto it the EVP from the valve which had been brought over from the 84. That valve is incorrect for this application entirely, but was working on the 84, so I wasn't questioning it. I tested the EVP with the analog meter and found it to be smooth, but didn't look at the actual values. It is worth noting that VREF on the 84 was always lower, as I recall, 4.something but not really 5, and on this car it's a solid 5.00. That could account for a sensor being acceptable on one car and not on another.

With it all back together, I briefly pulled the oil drain plugs to get the coolant out (but not changing the oil till I see there isn't a catastrophic leak), took it for a drive, and did tests. Still 44, 34, 32. The battery had been unhooked for over 24 hours, as well.

A NOS EVP as well as a NOS EGR valve (which may even include its own EVP as well) is waiting at my friend's house in Detroit, and two new-maybe-not-old-stock EVPs are being delivered to my work tomorrow, and then yet another NOS EVP is being delivered to my friend in Detroit on Monday. I kinda went nuts here. Oh well. Somehow, I will make the EGR work on this.

It did occur to me that the solenoids could be bad, and I will likely swap those from the 84 as a sanity check. Thanks to your screenshots above gadget, I was able to figure out the part number for the pair of solenoids as an assembly so I should be able to buy new if required.

I'm rethinking SEFI swapping this. If the EGR can be made to function without too much further insanity, AND it makes the fuel economy be "better", that will probably be enough for the CFI to stay until there's a major issue like a large-scale harness failure.

I also removed the HVAC inlet grille thing, cleaned up the area around it, and used RTV to seal it back down again. I am not sure that excess moisture entry into the car is being caused by this, but I'm sure it doesn't hurt to try to slow it down anyway.

[87gtVIC]

I admire your persistence and am eager to see the outcome...if the codes ever go away for you.

[kishy]

By all appearances the major oil leak seems to have been reduced or resolved. There is still a drip off the bottom of the bellhousing and I don't know where from, but not from inside, so it's a bit of a mystery. The valve covers seem to be sealed well.

Flushed power steering fluid. Changed oil and filter. One of those drain plugs is not long for this world. I threw a second sealing washer on it to space it further out from the pan and have it engage on some fresh threads. Since this has happened before I can only imagine the oil pan threads are trash. I have some oversize drain plugs if it comes down to it, but that's really more of an emergency problem solve should the plug fail to seal in the future.

The oil change interval was super early, but I had the required supplies and wanted to do it, so it's done.

Back to the EGR business.
Today, two Motorcraft sensors arrived at work. I figured, armed with these, I would certainly get to the bottom of it. Maybe not.

I investigated the vacuum end of things. The valve does not receive vacuum at throttle closed. With throttle partially open (engine running), vacuum to the valve pulses a little, but does not build (mind you, I had my gauge replacing the valve, not tee'd in with it, so maybe lack of valve movement makes the computer back off on the solenoid). It is clear the solenoid pair are doing something. It is not clear that the something is correct.

I accidentally test drove the car with the EGR line disconnected. While self testing, it gave 44, 34, 32. So these codes can be indicated even if no valve movement was detected at all.
I hooked up the vacuum line and test drove it with no change.
I swapped on a new sensor and took the car out for a test drive. 44, 32. No 34.
Thinking perhaps I was running into the "file the end of the sensor" problem, I swapped back on the sensor that had come from the 84TC and was apparently good-ish.
Test drove, 44, 32, no 34. Not sure where 34 went but he's gone now I guess.
I checked the voltage at the sensor with valve closed. 1.06V. A bit high for the chart indicating 0.83 as 0% EGR. The voltage being high does not seem consistent with code 32, but it is nonetheless the most obvious fault to try to address. I took the sensor out, used a cutoff wheel to shave perhaps 1/32" off the end, put it back in and checked my voltage. Came in a little low, 0.75ish, so I folded over a piece of electrical tape as a spacer on the end of the shaft. 0.85ish. I have no idea what the tolerance is here, but that seemed close enough (and maybe the 0.75 would have been too - I should try with the piece of tape missing).

Test drive found 44, 34, 32. So shaving the sensor did not fix 32 (logically it shouldn't have, but it was sort of what I was hoping would happen). It is unclear if shaving the sensor brought back 34, or if the condition that triggered 34 before has otherwise returned.

I'm not giving up, I'm just confused. And the precise definition of exactly what criteria can set a 32 and a 34 would be very useful, but this doesn't seem to be information that exists other than anecdotally.

32 means its too-closed. Most of the time thats a bum sensor. 34 is usually stuck partly open, or also could be a bad sensor. If its measurably above the voltage limit and you know its fully closed, see if the O ring under the sensor has gone totally flat. If thats good, take a file and remove a bit from the end of the sensor and keep doing that until its in the proper range if the sensor has a smooth sweep on the ohm meter.

http://www.grandmarq.net/oldfuelinjection/page35.html

Too closed = shaft extended too far = voltage too low
Stuck open = shaft pushed in too far = voltage too high
But are these codes only generated based on values at the time the valve is not being passed vacuum? Or can these codes be created any time when operating when the valve is "too open" or "too closed" for the commanded conditions?

As far as the O-ring goes, it has a track it lives inside of and is designed to be crushed, at least on this valve and sensor design. There is no viable way to have it be sealed and not have the parts flush together.

[gadget73]

thats weird that its not responding as expected. I don't have an emissions diag manual for a CFI car so maybe they changed the codes slightly along the way. I have one for a 91 but the position sensors are different by that point. Starting to get into funky territory here but maybe its worth an ohm check on the harness from the EVP to the ECM connector pins just to verify that what the sensor is outputting is indeed what the ECM sees. Also could be worth cobbling up a meter and a vac gauge to watch both while driving to see what its doing or trying to do, but my understanding is the 32 code is set with no vacuum applied at startup when the ECM checks the baseline output of the sensors. Basically its the "reset to zero" function so it knows where closed is. The TPS goes through the same process, it doesn't have to be at any very specific voltage, it just has to fall within whatever window the ECM accepts as normal and that becomes your zero position for that key cycle. If its outside the range it spits out a code and ignores the sensor.

Here's a wierd thought, have you verified that the ECM is in fact the right one for the car? I'm sort of curious if there may be some CFI ECM that wants the grey sensor.


also interesting that the intake gasket set with the rubber end walls is only listed 86+. The block didn't change physically in that area so far as I know so I would have figured they would just show the application as a much wider range. Unless you have one of the later 85 roller compatible blocks and there really is something different on the earlier ones?

[Tiggie]

I hate EGR codes. I went through a fiasco, not quite as detailed as yours, with my 88. Ended up swapping the valve and sensor off a junkyard donor and it got rid of all the codes.

[kishy]

At this point, ECM swap is looking like a valid troubleshooting step - and I hadn't even considered the black vs grey sensor issue but yes, if the computer is from some Fox car that used the grey one, that would probably explain some issues. We know the 84, while running on an 85 computer, has been happy with the black sensor at some point. I also have a box of 84-85 CFI Panther computers, none of which were verified as bad, all of which ran the 84, but I had random misfire problems and I never could fully rule out the computer. But as I recall the E5 computers were better behaved than the E4 ones which is why that car ended up keeping an E5 part.

The gasket set did not technically direct-fit correctly. The rear seal piece had an alignment nub that normally goes into a drilled/milled(?) hole on the rear uh...deck of the block. That hole did not exist on my block. It was the only apparent fitment problem. I just cut the little nub off and was mindful to keep the seal in the right place while aligning the intake. My head gasket tabs to align the intake gaskets are kinda destroyed on this so a lot of effort is needed using picks to keep things lined up while the intake drops down into the valley.

A NOS valve (possibly with a sensor attached to it) is at my friend's place to be picked up today. But I don't think the valve is my problem.

[gadget73]

ah, I know the thing you speak of. I didn't realize the older blocks didn't have that hole.


the way I do intakes on these is to use a set of guide pins. Its just a set of 4 too-long bolts that are the same thread as the intake bolts. I cut the heads off and cut a slot in the pin for a screwdriver. Thread them into the head and they will retain the head to intake gaskets, and prevent things from moving while setting the intake down. Stick a couple of the proper bolts in, then remove the pins. Normally they will just spin out but the screwdriver slots are a "just in case" deal.

[kishy]

That's a great approach. I'll have to keep that in mind. Could do selective dabs of RTV at bolt holes to keep things positioned exactly right, I suppose, as well.

Today while junkyarding in the US I also picked up a pile of NOS parts that have been accumulating for me at my friend's house. They are:

• NOS Motorcraft EVP
• NOS Motorcraft EGR valve
• NOS AC Delco EVP (the part is literally identical - I have a strong suspicion most of these under most brands all came from the exact same place)
• More NOS Ford transmission cooler line connectors
• NOS Motorcraft fan clutch

I had noticed in all of my EGR troubleshooting that the fan clutch was near-free-wheeling when hot (shut engine off, fan keeps going). That's not how that's supposed to work. Somehow in my part-number-searching activities on eBay I stumbled upon a Motorcraft one for small bucks, and bought it. I installed it today. Nice thing about having no smog pump present is the fan and clutch can sneak out without removing the fan shroud.

Results: fan is now moving much more air (can feel it and hear it), and stops very nearly at the same time as the engine when hot.

[kishy]

Continuing the EGR stuff:

I reviewed my inventory of computers. They are as follows:

• 84 CV 5.0 CFI, code SG (no other markings present)
• 84 TC 5.0 CFI, code SG, E4VF-12A650-G1A (original from my Town Car and suspected of having some issue I could never identify)
• 85 GM 5.0 CFI, code SG, E4VF-12A650-G2A (this is the Motorola version of the EEC-IV computer which is constructed differently inside)
• 85 CV 5.0 CFI, code D3MN, D3-E53F-12A650-N1

I disconnected the battery and removed my wagon's computer. It is as follows:

• 85 CV 5.0 CFI, code SG, E53F-12A650-G2A (and is also the Motorola version)

I distinctly recall that the SG coded computers behaved "less good" in the Town Car, so I started with the D3MN computer. I put it in, connected the battery and went for a drive. Starts, idles, and drives fine. No highway mileage yet. KOEO passes. KOER self-aborts very consistently. I remember running into this with the Town Car and I'm pretty sure it's because the idle is set too high. Some computers care about that and some do not. My idle speed is set high because the fast idle provisions of this throttle body are broken (screws broken off in the housing) so I needed it set somewhere it can idle without help.

I then came in to grab something to eat, post this, and develop further strategy.

From my perspective, a computer that can't be self-tested is a bit of a dangerous proposition, but I think it's worth taking it on the highway to see what the Tripminder says about economy. If it indicates a favourable number, that might be a sign that all the stuff I'm concerned about is doing what it's supposed to. I would also like to check what VREF looks like and also see what the EVP return voltage is with valve closed.

Edit:
Forgot to mention, unrelated, cruise has been acting up today. It's back to what it was doing before. Rapid accel/completely let off/rapid accel again. It switches between these with fairly consistent timing. I believe it's a problem with the module because a flaky sensor (such as the potentiometer in the servo, or the speed sensor) would not be so consistent in its failure.

As for the D3MN computer:
VREF is 5.05V with key on, engine off
EVP is 0.80V with key on, engine off
I believe this computer is running the EGR, but there's two edges to this sword.
Driveability is absolutely garbage. At highway speed, it is difficult to maintain speed because the pedal feels like it goes dead and stops responding. The amount of 'go' received for the amount of throttle is not as it was with the original computer, to the point that my TV setting now feels like it's off (more pedal but less go means the shifts are too firm). Issues maintaining speed were part of what I was chasing with the 84 when I first got it. This computer either has a problem, or its strategy is extremely picky about something that is hard to satisfy. On to another, I suppose.


Another edit:
I swapped in the other Motorola computer. The car behaves exactly the same with this as with the original one (drives great but EGR isn't working). 44, 34, 32.
I decided at this point to install the new valve and one of the new sensors. No filing, no modifications, just put it together and bolted it on.
Then noticed that since my last troubleshooting last night, the line was unhooked at the solenoid end.

Anyway, hooked the line back up, so now we have the non-original Motorola computer and a new valve and new sensor. Test drive showed no changes. Test still gives 44, 34, 32.

I think my next tests will be:

• Verify the functionality of the EGRC and ERGV solenoids as per the Emissions Diagnosis manual.
  If not working, swap from 84 and re-test.
  If working, thoroughly inspect vacuum lines and maybe re-run the whole lot of them as related to this system.
• As a curiosity, maybe put back the D3MN computer and see how it behaves now that the EGR vacuum line is connected (if will now self test and drives correctly, leave it)

I honestly feel like I'm pretty close to an answer, and I suspect it might be the solenoids. One thing I've definitely noticed is the 84 makes a loud clicking noise at part throttle which I believe is the EGRC/EGRV alternatively opening and closing to pulse the vacuum as required. This car has never done it once.

[gadget73]

I believe the egr is disabled if it gets a code so not hearing anything might not be definitive.

the original cruise control unit in the Continental behaved the same way. There was no return signal from the pot inside. Whether that was from the pot actually being bad or the damaged harness I never bothered to figure out, but a junkyard unit with a good pot and not-damaged harness fixed the problem. Easy enough to check, just see if the 3 pins that make up the pot give reasonable ohm readings. Panther cruise is less annoying, the servo is where you can get it and you can do all this from a standing position, Fox cars hide it down inside the fender which means removing the inner liner, and the harness is part of the servo so it has about 2 feet of wire that fish from under the dash with a large grommet attached in the middle of the cable to plug the big hole that the connector has to pass through. On mine someone cut the grommet to get the wires out of it, then ran them pinched between the grommet and the body so the wires all got sliced.

other thing I've had do this is a bad vacuum dump valve, not bad enough to trip the cruise out but leaky enough that it didn't hold properly. Also easy to check with a vac pump.

[kishy]

I did swap the module (because I found one before I found a spare servo). Cruise is maybe working properly again. But it's hard to be sure if this is random chance servo behaving again (pots can be like that), or if it's because I disturbed the connections at the module, or if it's because the module was the problem.

I believe I know what's happening with the EGR now. I do not know why, yet.

The solenoids check out OK. I found a break in the vacuum line which feeds EGRC and the one that maybe has something to do with evap purge (eating my past words here, it seems there is an electronic solenoid for that). That vacuum line was pretty much completely open. I'm not sure how I missed it before, but I'm certain it has been broken the whole time I've had the car.

With that connected, I drove around a bit and the test gave me 44 and 32 only, 34 is gone again.

I backprobed the sensor to see what it's telling the ECM and I saw 0.86-0.87V (pushing on the sensor flexes things slightly and it'll go between these values). This is above the minimum of 0.83V, plus a reasonable margin. I decided to use the breakout box and see what the computer sees. It sees the same.

I then piggybacked my analog meter on my digital one and noticed that it (I do expect they apply a small load) pulled the voltage down to 0.84V. I then suspected harness resistance might be messing me up here, but I checked it and it's very healthy. I forget the value but it was low enough I immediately wrote it off as an issue.

I then swapped all of my new EVPs and verified they all start at about the same voltage. Within 0.02 of each other. Went back to the one I started with today (the NOS one from yesterday).

I then started shimming the end of the shaft on the EVP progressively with aluminum tape bits, until I was at 0.971V. Did a couple tests incrementally while building up the tape bits and consistently kept the code 32 (and did not gain a 34). But I did notice that my digital multimeter, still connected to the BOB, dipped to 0.966V after the computer cycled the valve during the test.

I think that's what's happening here. The sensor is staying in its overall range, but it drops below the value at startup. How or why this happens is beyond me. The valve has no ability to close further than it was at first startup. If something is thermally expanding I'd expect it to go "less closed" (e.g. higher voltage) rather than more closed/lower voltage.


So what we know for sure is:

• The ECM can and does command the EGRC and EGRV solenoids until the code 32 event happens.
• The solenoids, when commanded by the ECM, operate correctly (edit: the correctness of their operation is now suspect, see next reply).
• There is reliable vacuum feeding the solenoids and the solenoids are passing it to the EGR valve, as well as venting it, as required.
• The EGR valve is opening and EGR flow is occurring (passages are clear, and the engine idle changes when the valve opens).
• The EVP sensor is reporting values to the computer and its value at initial KOEO is within the specs in the chart.
• The EVP sensor sweeps through its range properly, and fits on the EGR valve properly.
• The EVP sensor, when the valve has cycled, sometimes returns a lower voltage than what was given at KOEO, by about 0.005V, which the computer may be recognizing as 0.01V.

That last bullet point is now the target, for sure, I think.

I have largely discounted the revised version of the O-ring because it spaces the sensor further away from the valve. I had assumed this would cause a 32 because the sensor will always be below the 0.83V number...but maybe I was approaching this wrong. If the revised O-ring design (more wide and flat) stabilizes the body of the sensor against the valve, it might reduce the flex which may be at fault for dipping below startup voltage.

[kishy]

Another observation, in addition to the previous reply:
I swapped in another ECM and found the same behaviour with the sensor voltage dropping a little below.
I then pulled the EVP and changed over to the newer, wider style of seal.
When I re-attached it to the valve, I used wide flat washers under the screws to spread the clamping force better (trying to eliminate flex in the sensor body maybe making up my voltage discrepancy).

That did not solve the problem. I swapped to a different sensor and verified it is also the same. So this is not some oddity of a specific sensor.

I then checked VREF. VREF (at the breakout box) becomes lower when the engine is running than when it is not. Not much, say it was 5.00V, it'll be 4.95.
But I have to wonder if 0.05V reduction of VREF would be enough to drop the output of the sensor by 0.005-0.01V and make the computer sad about it.
Less volts into the sensor means less volts out of it.

Since the 5V is regulated within the ECM I am not really sure what I can do to stabilize it, or if it's even possible. Two ECMs though, so if it's something like bad caps, it affects both of them about the same.

I have no idea if I'm even barking up the right tree here.

Edit: I might not be. A code 32 on a car that uses the EVP sensor (and specifically does not use the 1-piece EVR such as SEFI does) might actually indicate "EGR not responding properly during test", not necessarily the voltage being below closed.
This opens up an entirely different can of potential problem worms. My take on some possibilities:

• solenoid problem e.g. they open and close but are too laggy to follow the computer's instructions
• my choice of vacuum line used in the repair is too thin (and it is thinner than stock) preventing the valve moving fast enough

Speaking in terms of simple probability, the things I haven't touched (and verified against other copies of themselves) are limited to the solenoids and the engine harness. Looks like the solenoids from the Town Car are going on an adventure to another car.

[gadget73]

have you tried to tee in a vac gauge right at the egr valve and watch it while watching the voltmeter on the EVP and probably the analog meter on the EVR line? Its pulse width modulated so it doesn't *quite* read normal but you can crudely watch pulse width increase with an analog meter because the average voltage will climb. Just curious if you can see any sort of delay between when the pulse width increases to command a valve open event, the vacuum starts to increase to actually pull the valve, and when the EVP registers valve movement.

I wouldn't think that 0.05 volts of drift on the 5v reference line is a problem. I don't know this for sure, but my assumption is that the voltage goes to an internal input on the ECM as well and rather than monitoring the various sensors for an absolute value, its monitoring them for a % of the reference voltage to come back. At least thats how I'd have thought to design it rather than just assume the reference was always 5.00 volts under all conditions. Component tolerances make that a bad bet. Not sure how that is even generated but in the 80s I'd guess either an LM7805 type regulator, or maybe with discreet bits using a pass transistor and a zener or something for the voltage reference. Likely nothing super exotic or incredibly high precision.

[kishy]

I do not have appropriate fittings to assemble the gauge alongside the valve. I bought a kit of pieces like that, but couldn't tell you where it is.

I swapped the solenoid assembly over. Self test reveals the same solenoid clicking noises the Lincoln has always made. The only code received is 44. Highway fuel economy over a fairly short distance of very flat level ground is 21-22MPG at 55, and 17MPG at 70.

This would appear to be unchanged from before, but only longer distance will tell me anything truly meaningful I suppose.

Anyone dealing with an EEC-IV CFI car (model years 84-85 in applicable vehicles) would be well advised to consider that code 32 does not mean the same thing when your car has the 2-in-1 EGRC/EGRV solenoids, than it means for cars with the 1-piece EVR. 32 means below closed value for EVR, but it means generic "valve didn't operate as expected" with EGRC/EGRV.