pvenuti Posted October 3, 2009 Author Share Posted October 3, 2009 Thank you both for the responses. I will try the overnight disconnect. Testing at the ECU to see what is coming back from the sensors makes perfect sense I can definitely do that once I figure out what I should expect to see. Link to comment Share on other sites More sharing options...
pvenuti Posted October 14, 2009 Author Share Posted October 14, 2009 Ok so I reset the ECU overnight after replacing the knock sensor and other than a 12 code for starter input(mines not connected) I have been error free for 200 miles or so. Thanks for the advice I’m hoping that my sensor, injector problems are behind me now. Finally got the gauges and the leak down test done and the percent of leakage seems to be nonexistent. I performed this test because two of the 4 cylinders only have 150 lbs while the other two have 200(tested them all over four times on different days(once by a third party same results each time). Any Ideas why or how I can have such different compression readings from one side of the engine than the other yet the leak down test shows no problem. Still having a cold stat problem but the symptoms have changed a bit. The car is starting better now but not great. When it does catch it chokes a bit low idle around 200 rpm then works up the courage and idles high at 1500-2000 RPM and stays that way until the engine is well on the way to warming up. Any ideas on this one? Once it’s warmed up runs reasonably smooth not perfect but reasonably and it gives a small sputter or puff about every 4 or 5 seconds easily detected when you hold your hand behind the tail pipe to feel the flow. Perhaps this is related to the rough start tuff idle issue? Link to comment Share on other sites More sharing options...
lmdew Posted October 14, 2009 Share Posted October 14, 2009 My bet is the cam timing is off a tooth or two! I've had this happen to me, took me forever to troubleshoot it. When I put the timing back where it need to be (and replace the tensioner) its been fine for months. Link to comment Share on other sites More sharing options...
89Ru Posted October 14, 2009 Share Posted October 14, 2009 c'mere horsie sorry for the re-beating For a second opinion, forget the tbelt marks...if you haven't already you should count teeth (twice spaced by an overnight, and once every 10 posts hereafter ) right side cam to crank is 44 teeth left side cam to crank is 40.5 teeth Don't take my word on these values, doublecheck the endwrench guide. Link to comment Share on other sites More sharing options...
OB99W Posted October 14, 2009 Share Posted October 14, 2009 Finally got the gauges and the leak down test done and the percent of leakage seems to be nonexistent. I performed this test because two of the 4 cylinders only have 150 lbs while the other two have 200(tested them all over four times on different days(once by a third party same results each time). Any Ideas why or how I can have such different compression readings from one side of the engine than the other yet the leak down test shows no problem. It depends what you mean by ''nonexistent'' leakage -- I've never seen an engine that didn't show at least a few percent loss. However, if the leak-down percentage is very low and the same on both sides, that indicates good valve and ring sealing. Leak-down is a ''static'' test, while compression is ''dynamic'' (being dependent on when and to what degree the valves open, besides how well they close). As I and others have said, timing error is what usually causes compression differences when the engine is otherwise healthy. Other than that, significantly misadjusted valves (not opening sufficiently) could cause lowered compression readings, but that would typically be obvious in other respects. Link to comment Share on other sites More sharing options...
pvenuti Posted October 16, 2009 Author Share Posted October 16, 2009 So by nonexistent I mean like 2 to 5 % barely off the 0% once setup. I knew that timing made sense the piston must be on the up stroke before the valves are closed so it cannot create as much compression. Just not what I wanted to hear as I have to pull the engine mounts, muffler, etc to get in there again. I guess it could be worse I could have an internal problem. Link to comment Share on other sites More sharing options...
OB99W Posted October 16, 2009 Share Posted October 16, 2009 I don't remember any discussion of who did the engine rebuild, so what I'm about to mention may or may not be a possibility. However, it's something that could explain compression differences from one side of the engine to the other, assuming timing is okay. There have been a few different heads for the EJ22. Is it possible that one head on your '92 engine needed work, and someone replaced it with another? While the pre-'97 N/A EJ22 heads are 9.5:1, those for the EJ22T (turbo model) have a lowered compression ratio of 8.0:1. I'm just throwing this out there, in case nothing more likely is the cause. Link to comment Share on other sites More sharing options...
pvenuti Posted October 16, 2009 Author Share Posted October 16, 2009 What a cheery thought, mismatched heads. I think I will take a look at that now. I assume it will be fairly obvious from how they are stamped. Link to comment Share on other sites More sharing options...
OB99W Posted October 16, 2009 Share Posted October 16, 2009 Sorry, please forget what I said about the heads. After further checking, it seems that the compression ratio difference is due to piston volume, not heads. That brings the area to look at back to timing. Link to comment Share on other sites More sharing options...
pvenuti Posted October 25, 2009 Author Share Posted October 25, 2009 Ok, so I pulled it all apart the other day and counted the teeth on the belt 4 or 5 times and found sure enough I was off by one tooth on the left cam. So like everyone says don’t use the marks on the belt don’t try to eyeball the notches in the plastic just count the teeth. I picked up 30lbs in both cylinders in the left bank so while they aren’t the 200psi I get from the right they are now pushing a solid 180 from each. The cold start problem is gone now she just zips to life with one crank. Now I have to ask do I try advancing that left cam one more tooth to see if I can get the 200psi so both banks match? Also I have a bit of a new issue. When the car first starts from cold it idles at 1500rpm for the first 3-5 minutes then kicks down to 900-1000 rpm. Any thoughts on why it’s such an aggressive cold idle? Thanks again everyone for the help. Link to comment Share on other sites More sharing options...
hankosolder2 Posted October 25, 2009 Share Posted October 25, 2009 I don't think that 1500 RPM on cold start is unusual. I started a '95 Legacy today and a '98 Outback (w a 2.2 conversion) and they both were in the 1200-1500 RPM range for a few moments. It seems like yours is taking too long to slow the idle down and your final idle speed seems high- Subarus generally have a pretty low warm idle- perhaps 600-800 rpm. As far as I know, idle is determined by the coolant temp (as read by the coolant temp sensor, which I think you already replaced?) Since this engine is used in a modified chassis with long runs to the front mounted radiator (yes?) , make sure that the coolant temp sensor is in fact reading the coolant temp and there's not an air pocket keeping it from getting a good reading. Also make sure that you have the correct T-stat in the car. If your Idle Air Control valve is sticky, that can cause various idle speed issues. If you have unmetered air leaks anywhere that could also be the source of your issues. Finally, any unresolved codes or disconnected sensors could possibly be an issue. That's all I have. Nathan Link to comment Share on other sites More sharing options...
OB99W Posted October 25, 2009 Share Posted October 25, 2009 Now I have to ask do I try advancing that left cam one more tooth to see if I can get the 200psi so both banks match? With proper timing and no carbon buildup on the piston crown or the head, 180-185 psi is typical (especially with the non-interference engines). I assume that both banks are now correctly timed, so perhaps one head got shaved a bit. I wouldn't suggest intentionally mistiming in an attempt to equalize the readings -- the 10% difference isn't particularly significant, and it's likely that the engine will perform better with proper timing than with it changed just to equalize compression at cranking speed. Link to comment Share on other sites More sharing options...
pvenuti Posted October 26, 2009 Author Share Posted October 26, 2009 Thank you both for your thoughts. You are correct the radiator has a long run and is up front while then engine is in the rear. I will try to setup Select Monitor to see if I can watch the coolant temp sensor real time and go from there. Link to comment Share on other sites More sharing options...
pvenuti Posted October 28, 2009 Author Share Posted October 28, 2009 I setup select monitor so I was able to confirm that the coolant temperature is sending valid info to the ECU. I can sit there and watch the car warm up etc. The tool is interesting in that it shows some good info on a lot of sensors. One thing I found confusing and need to look into more is that several of the parameters seem to bounce around a bit. Like temperature might read 180 then -58 then 0 and then back to reading a good temp say 185. Or maybe RPM might be 850 then I reeve the engine it detects the change to 1200 then has zero for a few seconds. I'm assuming that it must be an issue with how quickly my laptop is communicating with the ECU. Maybe I need to adjust the baud rate or something on the port. I will also try to run some SeaFoam through the IAC again to see if that helps. Link to comment Share on other sites More sharing options...
89Ru Posted October 29, 2009 Share Posted October 29, 2009 (edited) Hmmm...it could be a number of issues *Intermittent sensor(s), this I doubt since multiple data is affected *ECU sending bad values in between good ones *Laptop software buffers timing out and clearing themselves *Port parameters (parity, etc.) but I doubt this since it would either work or not...the ECU probably does not autobaud (detect the baud rate), i.e the baud rate should be fixed *Some other issue If you want to make this an academic exercise, for a second opinion of what the ECU is sending try downloading realterm software http://realterm.sourceforge.net/ to your laptop, this will help diagnose the actual raw serial data that is coming from the ECU. It won't be pretty, you won't know the byte positions of the rpm, temperature, etc. you'll have to decipher these by observation -troublesome since the data may be hopping around. edit: First, try changing the flow control settings on the laptop's serial port (three options: hardware, Xon/Xoff, none) Edited October 30, 2009 by 89Ru add info Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now