Need Help, Running Rich on Right Bank.

Ok, trying to get my 91 TR ready for SMOG and discovered the right bank is running rich. I started out by checking the plugs and wires, replaced the plugs and the wires are doing their job perfectly. I tested the O2 sensor and it wasn't reading, so I replaced it, now it's reading, but it's reading rich. Anyone have an idea the next place to look?

Thanks

Eric

 

I've been helping Mr. Sleestak with his TR. Here's a bit of detail:
91 TR with US lambda KE Jetronic.

CO on the emissions report was over 2.5%
HC were over 2,000. The report didn't note left/right side readings.

In the shop, we used a timing light to confirm spark. Three dead cylinders were discovered. We swapped plugs and the misfire followed the plugs. Replaced the plugs and now have good spark on all 6 cylinders.

New plugs showed soot/right right away. So we unplugged and took readings at the O2 sensors (warm). Left side shows ~.6-.7v, right side showed 40mV.
Replaced right side O2 sensor. While plugged in, reading shows ~.9v now, steady. Haven't checked left bank while plugged in, only unplugged, and it still shows .6-.7v.

Left side plugs are tan/clean and look good. Right side continue to be black and sooty. Left (Driver side) exhaust (right engine bank) still smells strongly of fuel while right exhaust (left bank) smells normal.

Warning lights for the CAT temps do not come on. Car runs/drives seemingly OK, but has a flat spot around 3k rpm on the freeway when trying to accelerate from a cruise. Downshift and it runs away pretty well.

Both air flow sensors are at about the same height, move freely, are centered, and seem to have the same feedback/feeling when pressing down on them.

What we've not yet done:
Checked cylinder bank idle balance. (doesn't seem this would have an effect on cruise AFR/mix, more of an idle/part throttle issue).
Checked fuel pressure.
Checked return line.
Checked EHA current.
Checked that EHA isn't blocked.
Played with the mixture screw (would rather not go fumbling around with something sensitive that might not be 'wrong').

Car starts cold or warm very easy.

What we're trying to determine is the next best thing to check/repair while avoiding replacing parts for the sake of replacing parts, or chasing something that isn't a common problem on a TR.

 

not long ago I had an equal problem, but there the left bank was always not running well. the failure there have been wrong air over the brake vacuum hose.

but you may also check if right and left bank are synchronous, not only in idle, also when you accelerate

 

Ok, check 02 sensors and did you do a pre test yet ? and what are the readings PPM ?

also are they the original catalytic converters on it ?

 

Is it the original exhaust?

 

Don't see how you are going to avoid doing that (based on the "rich" reading of the "right" bank -- but better to use 1-6 bank and 7-12 bank to avoid communication confusion IMO). A key thing to confirm/deny is that you can manually swing the O2 sensor voltage (when unplugged) from lean to rich and back manually with the mixture screw -- if you can't do that manually (with the O2 sensor unplugged), there is no way that closed-loop operation (with the O2 sensor plugged in) can work.

Never hurts to measure the O2 sensor outputs when plugged in -- if the "rich" bank O2 sensor voltage stays high (and, for bonus points, the EHA voltage drops to 0V -- indicating that the system is trying to lean it out as much as possible) = another sign the mixture screw is just set way too rich. (Unfortunately, ICEs run beautifully when set rich -- if the cats don't set the coachwork on fire . You report no over-temp warning lights indicated -- but is the self-test always working perfectly, or are the warning light ECUs unplugged? Another possibility is that there nothing "catalytic" inside the pre-cats/cats -- however, that would only explain failing the test when the O2 sensors seem to report good values, not a wacky O2 sensor reading).

 

To answer more questions:
Stock exhaust, stock cats, given the relatively low miles on the ODO it all looks original. Most of the previous owner's hackery seems limited to the stereo system.
No pre-test was done, but it was checked for emissions. HC ppm was over 2,000 and CO was over 2.6%. There's a very strong smell of fuel/gasoline at the 1-6 exhaust tip. Enough to make your eyes burn. We've also both got perfectly round black spots on our jeans from standing behind the car as it's running (soot from the pipes).

No clue on the self-test. Will have to confirm that the warning lights function properly. The cats never seem that hot, under hood temps (judging by standing next to it) don't seem terrible either. Stock US exhaust, but no idea if the cats have been gutted either or that they're not fouled from so much soot and raw fuel having run through the pipes by now.

Sounds like the next logical step is to check voltage and/or current at the EHA for the 1-6 bank to confirm that the O2 sensor reading is causing feedback. Will also need to confirm that the O2 reading on the 7-12 bank oscillates while running, or that the EHA voltage does while running to ensure that the ECU is operable. So far, the only O2 voltage check we've done on 7-12 has been when unplugged.

After looking over the manual, it appears the base mixture will alter the enrichment across the board, right? It's not just an idle mixture adjustment? If so, then it would seem reasonable to check there next to see if an effect can be observed.

I'm assuming that if altering the mixture adjustment doesn't net a result then that would count as evidence of a fuel pressure issue. Either a plugged return line, or EHA, or regulator, right?

 

These symptoms sound familiar. I had the soot spots on the garage slab after cold starting and the gasoline fumes, but it was my left bank. The right bank was fine.

Check the fuel distributors. After 25 plus years the internals may need a rebuild along with the injectors. The air plates get sticky. Makes for sloppy throttle response, especially when cold.

 

Luckily, throttle response is actually good. Cold starts and warm starts are very easy, driving it seems good except for 'down low' when the throttle gets stomped. The air plates move easily, all cylinders on 7-12 read the same at the plugs, and all of 1-6 read the same as well black/soot).

This morning we gained access to the mixture screw on the 1-6 bank. We tweaked it and it definitely responds at the O2 sensor. The 7-12 bank is oscillating as I'd expect from my EFI experience when plugged in, and reads ~.91v when unplugged at idle (~1080 rpm).

The 1-6 bank was reading .93-.98v when plugged in. Turned the mixture screw out by over a half turn in the end, and have it reading .89-.95v when unplugged. Any 'leaner' and it drops to 100mV or less. So at least each side matches the other. The exhaust no longer makes my eyes burn either.

Sleestak took it for a drive and says its much better across the board. Sounds like now we just need to check bank-to-bank balance and have the CO set. From there, I'd imagine it will pass emissions.

 

Which mixture screws are you turning? You can also connect a manometer and that will help to balance the banks tremendously. I have made one that works great for a couple of dollars: $1.55 Carb Sync Tool by Marty Ignazito

 

The screw under the tamper plug that's between the air sensor and the fuel distribution connections. The procedure is described starting on D59 of the WSM. The one we adjusted corresponds to the 1-6 bank. The 7-12 bank was left alone since we don't have a CO meter and the plugs on that side are not showing the same symptoms of running very rich.

 

Good.

 

OK, since every dealer in town wants to replace everything before they'll just tune the engine, we picked up a wideband oxygen sensor.

Sleestak took it back to have the emissions checked, and the HC are back to normal. However, the CO at cruise was beyond the 2% state limit. We leaned it out and got it to pass CO but then it failed idle HC (lean misfire).

So we installed the wideband, and we can tune it to 14.2 AFR (was reading 22.8 when first installed, after the HC fail) but when we pick the revs up to 2,000, it drops to 12.5. If we tune cruise for ~14, we get too lean to idle again (misfire at idle).

So we checked the 7-12 bank. Idle is a perfect 14.2 AFR, when picked up to 2,000 rpm it's reading 14.9 to 15.2. Perfect.

We then put a couple of vacuum gauges on the intakes. They read near-balanced, but the gauges are not 'mercury stick' style and are instead dial gauges.

If I push down on the air sensor plate, the idle picks up and the AFR richens (of course), but this is basically what the idle mix screw does, right? The height of the two sensor plates appears equal when judging by their placement within the throat of the round opening, and both plates are centered.

My gut feeling is that there's a vacuum leak somewhere. There are only a few vacuum sources (as far as connections to the intake). Are there other places to check that cause issues on the 1-6 bank?

The WSM also talks about setting the static height of the air sensor plate by punching a pin. I'm hoping this isn't something we'll have to tinker with? Seems like a very one-way setting.

 

More information and perhaps some more input on what to do next?

We've checked primary fuel pressure and read 81 psi on both FD.
For 1-6 we checked the lower chamber pressure and get ~76 psi, so we've got the differential pressure in the right ballpark.

Next thing to test seems to be the EHA. I'd like to swap side-to-side, but the 7-12 EHA seems like a real PITA to reach/remove. So instead It seems like we'll have to rely on testing the voltage/current.

What is the typical failure mode of the EHA? Would it fail in such a manner as to cause a rich condition? Or would it typically cause a lean condition? Reading the WSM indicates that if there is no power to the EHA then it would default lean, so my initial feeling is that it's unlikely to fail into a rich mode.

Can anyone confirm this? Suggest other checks we should perform?

 

this might be of help.... or may be wildly off base. Use caution.

Thanks for sharing your ongoing efforts!

Buona Fortuna,
Vincenzo

from:
EHA Valve original/factory adjustment(setting) - Mercedes-Benz Forum

Original setting (on new EHA):
depth from EHA body to adjusting screw is 6.6 mm

Current at actuator (only ECE/RÜF):
ignition "ON" = + 70 mA
idle, coolant temp. +20 C = 2.6 mA
idle, coolant temp. +80 C = 3-4 mA
post-start enrichment, +20 C coolant temp. = 7-13 mA
acceleration enrichment at +20 C and blipping throttle = >20 mA
full load enrichment at approx 5000 rpm = 4-6 mA
part load mixture adaptation = from -9 to 0 mA
decel fuel cutoff = approx -60 mA

 

Unfortunately, those MBenz EHA current values are for (I believe) KE3-Jet -- which uses an EHA current that normally varies from -10 mA to +10 mA. KE-Jet on TR uses an EHA current range of 0 mA to 20 mA (i.e., negative current values are not possible, and KE-Jet TR don't have the "decel fuel cutoff" feature).

 

So I've read. Seems the only way to replace an EHA these days is to contact CIS FlowTech and he's got a hack/re-calibrate method to use a Benz unit in it's place.

My biggest curiosity at this point is whether the EHA commonly fails in such a way as to cause a rich condition? Or is it more common to fail and cause a lean one? Or is it 50/50?

I'm afraid that even if the EHA measures out correctly for voltages, it may not be functioning internally the way it should. Is there some method to tweak it or test it's actual function?

 

Measure the lower chamber pressure (or the flow rate of fuel exiting the injectors for a given size slit opening) vs EHA voltage/current (e.g., close the WOT switch -- EHA voltage/current should go up = EHA blocks the inlet to lower chambers more = lower chamber pressure should go down = upper chamber pressure goes down = increases pressure drop across slits = more fuel comes out injector for the same size slit opening) although this takes a little bit of plumbing to connect from the lower chamber test port to the Bosch CIS pressure gauge. Alternatively, you could just swap EHAs and see if the trouble follows it, or not.

 

easiest way to check is to change left side to right side and otherwise before sending out for expensive rebuilding

 

Is it possible to swap without removing the intakes? It looks tight in there.

My concern is that if we 'test' the EHA, all could be well, but if it's obstructed, corroded, or just not working correctly, it may still be a fault even if it's behavior is correct. Swapping it side-to-side seems the best way to isolate, just not sure how feasible that is.

 

IME, no

 

you have to remove the collecting tubes, but not the complete manifolds. so only 8 srews each side. it is a little work I did already. problem is the aluminum water collector Y. but the compelte meters and distributors are only fasten with 4 bolts in rubber. so about 2 or 3 hours work to check this.

 

Another update.

We swapped the EHA modules and re-ran our checks. The 1-6 bank is still misbehaving. At idle we set the AFR to 13.1 after warmed up (luckily, it wasn't far off). We still drop 2 full points (from 13.1 to 10.8~12.1) at 2500 rpm versus idle. The 7-12 bank stays exactly at the 14.2~14.5 it's at (lambda unplugged) from idle to 2500 rpm, if anything is has a lean dip to 15.5~16.0 until it stabilizes.

Another observation: if we tap the air sensor plate on the 7-12 bank, the idle bogs and pops right back up. If we do the same on the 1-6 bank, the idle bogs and takes 3-5 times longer to come back to normal. We also previously noted that if we depress the air sensor with the engine OFF, the 1-6 bank side takes longer for the plunger to presumably return to contact with the sensor arm (you can feel when it's pressing against the arm vs when it's not as it's moved).

The AFR's are pretty close to equal on both sides, so with all this I think it might be something in the FD. There was really no change from the EHA swap. So it seems those are OK.

Is there any way to ensure this diagnosis is accurate?

 

Good news!
We talked to Larry at CISflowtech, and he advised we really really should check the EHA current before thinking about servicing the FDs. So, we finally broke down and did that.

We round the EHA current to be steady at 60ma, no matter how the FD mix was set. So.. with some more advice from Larry, we went searching through connections at the computer connectors to narrow down the issue. We swapped KE ecus left/right and had the same trouble, so that eliminated the ECUs.

We then probed all the sensor connectors: potentiometers, thermotime switch, idle microswitch, coolant temperature - and bingo. Coolant temperature sensor was showing (engine hot, ~150 deg) 900ohm on one side, 900k on the 1-6 (troubled) side. I've never dealt with a dual channel CTS (NTC on the wiring diagram for the TR) so never suspected.

Swapped it this morning, set the base mix with the EHA current pegged to 10ma, and it passed smog!

We still have an issue though. At WOT, 5k rpm, the engine hits what feels like a rev limiter. We'll need to do more testing to figure it out, but even feathering the throttle it might make it to 6k before it breaks up. Not sure what to suspect yet, injectors, ignition, etc.

 

good to read that one failure has been found
bad, that there is now the next one.

meassure the ignition wires resistance first. it is the easiest way to start.