1995 3.4L Repair/Diagnostics Journal:

My original issue was/is a 45+ mph torque converter shudder that lessened a bit with a couple 2 oz tubes of conditioner. However, the shudder still manifests, as b4, after decelerating and then a lightly feathered acceleration but only after 45 mph or greater. You punch it and its fine. It's only been driven less than 30 miles since the new injectors were installed a few months ago. O2 sensors are factory originals so no surprise they're possibly failing.
================================================== ============================================
Sorry I didn't comply with Fred's data log convention's to the letter but I did get the rpm's up to almost 4400 momentarily. I am crunching & collating data surely but slowly but please chime in if something besides the crazy trim data jumps out.
================================================== ============================================

• Fuel trim(s) looks crazy. I guess that's what Chinese crap injectors get ya. If 128 is the ECM's baseline or metric then something is rotten in Denmark.
• Is there be some other reason besides dirty or faulty injectors that the ECM is thinking the pulse width "on time" should be so long?
• I did finally get a DTC (64) "RIGHT O2 SENSOR LEAN / MAP sensor circuit signal voltage low " which, among other things, can be a sign of clogged injector(s) of course. I had a few issues with Windblows, a 24 year old lappy, and scan9495 app error pop-up's refusing to load app correctly.
• The .004v of the O2 sensors in the xls doc (line 2084) is negligible at best. There are several other .0**v readings also.
• Pretty sure 28+ "Base inj p/w (ms)" (line 1199 and more) is almost double even a "normal" wot pulse width for an injector. Dirty or bad injectors? Or, is the PCM compensating for perceived lean reading(s)?
• Left bank-2 ST & LT fuel trim (BLM) are fine and within parameters. So, as the app can't separate individual injector values, how can we know which injectors are affected on the right bank?
• Does the ECM send the same pulse mod instruction to each side/bank/injector equally? Does one bad injector requiring the ECM boost mean that the other two injectors get the same "open" time? And, that's why the extra fuel dispersed floods the engine with too much fuel?
• The normal (sea level) MAP at idle should also be around 40 kilopascal's or less at idle for the L32 I'm guessing so I don't see anything abnormal in that column.
• TPS values seem within limits but I seem to remember reading somewhere yesterday that the MAP and TPS should or may have some corollary values, no?
• The Air/Fuel ratio is ok (14.7:1) most of the time but dips rich a couple points and the right bank, coincidentally, shoots up to180. That's a whopping 40% more fuel the PCM is forcing down the right bank's throat. Ouch.
• If .45v is the norm for O2 sensors, my PCM is going nuts trying to maintain a 14.7:1 ratio which I guess is "normal".
• Interestingly, the left bank voltage shows rich with normal ST & LT trim. The right bank shows lean with but with correspondingly high ST & LT trim.
• The CAT is rusted garbage as is the entire exhaust system but if the torque converter needs replacing, the exhaust will have to wait.
• So, new MAP, O2's, & exhaust system are a good idea, given they're all 29 years old.
• Test injectors.
• Still need to scan9495 transmission tab for any data that may help. Too bad it can't give torque converter data.
• Does anyone know if the PCM can get "stuck" on a specific block set in one of the 18 "operating conditions" it corrects, while it formulated/calculated a long term solution or corrective values? Does it (PCM) use an eeprom to store data/values? I was wondering if a coded "instruction" or a function in the memory bank (block) or whatever could get corrupted and send false information to components, positive or negative. For instance wrong pulse/width open time to an injector. I read that the PCM updates calculations about 10x per second which probably lessens room for large or persistent errors. However, I thought that maybe jumping back and forth (given car speed fluctuations) between closed loop operations and not, the extremely sensitive O2 sensors could send "bad" data back to the PCM. Thereby, the PCM uses 'bad' data in calculating it's mini-algorithms, if you will, for the closed loop block learning. I guess it's only a short term issue anyway I suppose as, hopefully, the defective parts causing the larger deviations in the PCM, forcing correction, get rectified quickly.

I'm still going over data and learning a lot. If anyone can see data or variables that I should pay more attention to, chime in. I know there's only a handful of folks left in here but hopefully even a random surfer might notice something .

 

find an injector shop close to you and have injectors tested. i picked up a set of rebuilt injectors for the 3.8 from doctor injector that were good price and worked great. also had my new injectors for the lt1 tested and adjusted to flow even.

 

So, can I just get a noid kit and do it myself?

 

no they test flow on bench with flow test machine.

 

Gotcha.

 

I sent mine in through the mail. shop is a couple hour drive one way.

 

So, was there something besides the wacky trim in the scan/data that pointed to the injectors more than something else?

 

not really, I would suspect those and/or 02 sensor unhappy. I am not up on the 3.4 so any other issue that may be common I have no idea.

 

Thanks craby. I've read in a few different posts here and there that the injectors' (barring really crappy engineering or quality) failures are predominantly from clogged or dirty screens, no?

After a 13 year hiatus in the Shawnee Forest, the fuel system was trashed of course. B4 removing the injectors, I blew out everything on top of the plenum with 135 psi of air w/ emphasis around the injector seats. Four of the injector's came out attached to the rail but two of the tops crumbled in my hands, leaving the bottoms in the ports so I covered the four open ports and blew around the two stuck ones again b4 yanking them. They actually came out easily. It's that horribly designed hood seal above the plenum the leaves the injectors sitting in a lake. After getting 'em all out, I used a modified vacuum hose on the shop vac to suck out any potential debris inside each port. I also blew out the fuel lines and soaked inside the feed line with a little bit of evaporust from the rail to the filter and blew it dry with filter removed and then put in a new filter. I cleaned the fuel rail to factory new, replaced the regulator diaphragm, two filters, replaced the fuel pump, replaced a piece of the feed fuel line/coupler by the rail, replaced the Vapor Canister Purge Valve because the line coming into it was clogged with tons of charcoal, and of course the injector's. During this, I soaked the tank with 14 gallons of vinegar for 4 days (neutralized w/baking soda after), used electrolysis on/off for 2 days, pressure washed it and used a heat gun to dry it thoroughly. Lastly, I blew out the tank with the compressor to see if anything else came loose and shop vac'd as far as I could reach by hand.

Even though I'm pretty sure everything was pretty thoroughly cleaned, I couldn't have got every micron of any residual whatever that may be left in the system. I wonder if maybe just pulling the injectors, clean or replace their screens, and then putting another filter in it would work?

Of course none of this matters if the torque converter is bad.

 

have you checked the evap canister to see if its charcoal has gone to pieces and/or full of gas. it can get gas in it and cause this. mostly caused from over filling gas tank but other issues can cause it. failing evap system can cause rich condition and or mess the a/f up