X64v Posted November 15, 2007 Share Posted November 15, 2007 As many people do, I suffer from a rough idle after anywhere from five minutes to one hour after running the car, depending on ambient air temp. I fixed this by just richening up my idle, but this gave me about an 11.3:1 afr during a normal idle. My engine is happiest at about 12.3-12.5:1 at idle, but this brings back the hot idle situation. To try to analyze the problem, I took two data logs, one of a perfect idle after a drive, the other of the idle about 20 minutes after that drive, as soon as the engine started. The two logs below: Smooth idle: Click here for a full size version. Hot start: Click here for a full size version. Both logs are attached in the zip file. For the smooth idle graph, go all the way to the very end of "woo 11-12-07" What you can't see in the graphs: Coolant was 171°F during both tests (no, warm up wasn't on, I run a 160°F T-Stat) Battery voltage was 13.4V during both tests Ambient air temps were within a few degrees of each other Overall engine temps were very similar (hood was kept closed, so not a lot of heat was lost) What you see in the hot start graphs will go on for about a minute while driving (won't idle at stoplights), and a few minutes at just idle. Everyone's thinking "Vapor lock, hot fuel," right? I'm not so sure. I'm running a non-metal fuel rail, and cycling the pump on five or six times before starting (basically getting the pump to pump cool fuel through the rail for 10 or 12 seconds prior to starting the engine) does not affect this. Running a 1.5ms pulse width at idle totally fixes this problem, it'll start up and idle at 900rpm with no throttle on a 100+ degree day after sitting in the sun for half an hour, but my afr's would be, like I said, in the low 11's. A few people were talking about having their IAT sensors heat soaking and causing lean conditions, but my IAT readings were within 2 degrees of each other during the tests. I have not ruled out hot fuel as a possibility, but I'm not convinced it is, either. One possible fix is to just run in closed loop at idle. With 12.5:1 as my wideband target, it idles alright, though it will hunt a little at times (switching between 1.4ms and 1.5ms). Any thoughts either way? Quote Link to comment Share on other sites More sharing options...
mario_82_ZXT Posted November 15, 2007 Share Posted November 15, 2007 Have you tried leaning out your idle the other way that was recently discussed? Do this, open up your TB a little bit by adjusting your idle screw, lean out the ratio, and retard your timing a pretty large amount to bring it back down. This way I have a buttery smooth idle at 850 rpm with a 14.5:1 A/F ratio. This is also a good way to lean out your idle with big injectors. Mario Quote Link to comment Share on other sites More sharing options...
big-phil Posted November 15, 2007 Share Posted November 15, 2007 Thats interesting. The heat soak rough idle thing bothers me to. Quote Link to comment Share on other sites More sharing options...
X64v Posted November 15, 2007 Author Share Posted November 15, 2007 Mario - You're on a stock turbo cam, right? What's your vacuum reading with it set like that? I'm pulling 16-18 in.hg. below ambient, or about 34kpa absolute. I remember you telling me about that for big injectors, I may have to do that anyways when I move to 500cc's. Quote Link to comment Share on other sites More sharing options...
mario_82_ZXT Posted November 15, 2007 Share Posted November 15, 2007 Mario - You're on a stock turbo cam, right? What's your vacuum reading with it set like that? I'm pulling 16-18 in.hg. below ambient, or about 34kpa absolute. I remember you telling me about that for big injectors, I may have to do that anyways when I move to 500cc's. Sorry I can't remember and I can't check it now. The car has a dead hole again, hoping it's just the fuel injector because that plug is super clean/white, not drench from not combusting, and it's not burning coolant... I want my injectors to work dammit... Mario Quote Link to comment Share on other sites More sharing options...
X64v Posted November 15, 2007 Author Share Posted November 15, 2007 Well, hot fuel it may be. I did another test today. I went for a 20 minute drive, got everything nice and warm, on boost a little but not much, etc. When I pulled into my driveway (with a perfect 900rpm idle at 12.5:1), I took my fuel rail temps while the engine was idling, via a non-contact IR thermometer. The rail was 90°F, and the injectors were 110°F. I then closed the hood and waited 20 minutes. Ambient temp was 70°F, car was parked in the shade. The temps after 20 minutes were 130°F on the rail and 150°F at the injector. I then started the car, and it proceded to try to die at idle, as expected. After a good 45 seconds to 1 minute of revving lightly out of gear, I got it to idle stably, although it was 200 rpm low, and rough (again, like always). At this state, the fuel rail had dropped back to 90°F, but the injectors had only dropped 10 degrees to 140°F (I would guess the cause for my still rough idle, as the fuel isn't moving very quickly through them at those low pulse widths). One should note that the increase in rail/injector temperature (measured in Kelvin, of course) is ~7%. The increase in pulse width needed to correct this stumble (1.4ms bumped up to 1.5ms) is ~7%. If we really wanted to be accurate with our fuel measurement, perhaps one could install a fuel temp sensor into the fuel rail, and edit the MSnS-E code to account for this... A dead head rail would actually be an advantage here, as the rail temp would match the injector temp, instead of cooling down while the injectors stay hot. Edit: It's also important to note that I have no heat shielding on my engine whatsoever. I also have a non-webbed intake, so all the heat from the exhaust manifold and turbo likes to just come creeping up. i have the original heat shielding for the n42 intake I have, I need to cut it up one of these days and get it on there. Quote Link to comment Share on other sites More sharing options...
big-phil Posted November 15, 2007 Share Posted November 15, 2007 I need a heat shield also. I bought some reflective/insulating tubing for my wire harness at the injectors. I wonder if it would help to wrap the fuel rail but still make it look good somehow? Quote Link to comment Share on other sites More sharing options...
Clifton Posted November 29, 2007 Share Posted November 29, 2007 Mine idles lean after 1/2 hr plus restarts. I ended up setting my idle fuel very rich (11.0) and then let the wideband setting pull it back to 13.0. It won't idle smooth in the 14 or even 15's like the stock ecu did. It works as a band aid fix. I'm running the flyback and want to go to resistors to rule out a problem there. It may also be an injector problem on mine but I do miss the stock ecu at times. Quote Link to comment Share on other sites More sharing options...
X64v Posted December 3, 2007 Author Share Posted December 3, 2007 Clifton, that's on your 7m, right? Quote Link to comment Share on other sites More sharing options...
Clifton Posted December 3, 2007 Share Posted December 3, 2007 Yeah. I ran some 550's in it for a few days and didn't have the hot start problem. I may have just needed more time for it to show up again or I have a bad 700c injector. Quote Link to comment Share on other sites More sharing options...
WizardBlack Posted December 4, 2007 Share Posted December 4, 2007 Well, hot fuel it may be. One should note that the increase in rail/injector temperature (measured in Kelvin, of course) is ~7%. The increase in pulse width needed to correct this stumble (1.4ms bumped up to 1.5ms) is ~7%. I know it's been a while since my thermodynamics, so how on earth did you come up with that direct relationship? My research showed a different ratio of expansion, but the injectors themselves probably have negative effects from the heat. It may actually be more; quite a bit more maybe? The Coeff. of Volumetric Expansion for gasoline (950 x10^-6 xK^-1) shows an increase of ~21% for gasoline for a 22 degree Celcius (or K) change in temp. (check my math; that doesn't feel right but it does expand 4X faster than water) A rough estimate on anticipating the change in resistance of the copper windings of an injector comes out to about an 8.7% rise (.393% * dT Celcius) in resistance. Considering these are low impedance and probably coupled with a ballast pack, that is probably only a third of the total resistance the ECU sees. This would raise the total resistance by ~2.9%. Given that voltage is pretty much constant, this would drop the current seen by the injectors. I can't really say how that will affect the pulsewidth. And before anyone decides to me, I am just throwing that out there to spur the conversation. Mark Quote Link to comment Share on other sites More sharing options...
Zmanco Posted December 4, 2007 Share Posted December 4, 2007 I had similar issues with inconsistent idle (speed and roughness) on my NA until I: - switched to a PWM idle valve - set the min rpm value for O2 correction to 600 rpm - set the AFR target for that point on the map to 15:1 It took a lot of trial and error to get PWM control to keep a stable idle, but it works pretty well now. The idle is very consistent across a broad range of temperatures, loads (headlights/fans on/off) etc. Even if you don't have a PWM valve, you might try an experiment and lower the rpm for the O2 correction to begin working. Quote Link to comment Share on other sites More sharing options...
X64v Posted December 6, 2007 Author Share Posted December 6, 2007 WizardBlack, I did not come up with a direct relationship, just took measurements of temperature and noted the pulsewidths needed. I am aware that I made no account for thermal expansion rates or anything of the sort, which is why I simply noted those numbers. However, since you brought it up, according to nist.gov, the thermal expansion rate of gasoline is .00069/°F, making the expansion with dT=40°F 2.76%. I think I'd need someone like z-ya to come in and explain the effects of small resistance changes on fuel flow and what the ecu does/thinks it's doing. For what it's worth, I'm running low Z injectors with NO resistor pack. V3 megasquirt takes care of that for me. My guess is that it would only slightly change the opening time of the injector (and with injector opening times being variable by only 10% or greater at 1ms, I'm not sure of the impact of injector temp changes on it's function). Zmanco, I have tried running O2 correction at idle, and it does seem to work, though I was hoping to try something to fix the fuel temp problem rather than just running O2 correction. It can hunt at times with correction on, causing my idle to bounce up and down a little bit. And 15:1 afr would be way too lean for my motor at idle. It seems to lean miss starting at around 14:1, and really really likes about 12.5:1 or so (really really likes meaning highest idle speed and most vacuum). Quote Link to comment Share on other sites More sharing options...
WizardBlack Posted December 6, 2007 Share Posted December 6, 2007 WizardBlack, I did not come up with a direct relationship, just took measurements of temperature and noted the pulsewidths needed. I am aware that I made no account for thermal expansion rates or anything of the sort, which is why I simply noted those numbers. However, since you brought it up, according to nist.gov, the thermal expansion rate of gasoline is .00069/°F, making the expansion with dT=40°F 2.76%. I think I'd need someone like z-ya to come in and explain the effects of small resistance changes on fuel flow and what the ecu does/thinks it's doing. For what it's worth, I'm running low Z injectors with NO resistor pack. V3 megasquirt takes care of that for me. My guess is that it would only slightly change the opening time of the injector (and with injector opening times being variable by only 10% or greater at 1ms, I'm not sure of the impact of injector temp changes on it's function). Zmanco, I have tried running O2 correction at idle, and it does seem to work, though I was hoping to try something to fix the fuel temp problem rather than just running O2 correction. It can hunt at times with correction on, causing my idle to bounce up and down a little bit. And 15:1 afr would be way too lean for my motor at idle. It seems to lean miss starting at around 14:1, and really really likes about 12.5:1 or so (really really likes meaning highest idle speed and most vacuum). Well that expansion ratio (.000069/F) sure makes a lot more sense but I've run and re-run the ratio I came up with from sources using Kelvin. Perhaps I just ain't thinking right. Quote Link to comment Share on other sites More sharing options...
sweetleaf Posted December 6, 2007 Share Posted December 6, 2007 I think your rough idle problem when trying to run leaner on idle, is too short of a PW. That is probably why it ran ok with the 550's. The only way to correct it with large injectors, is to turn down the fuel pressure. Yeah. I ran some 550's in it for a few days and didn't have the hot start problem. I may have just needed more time for it to show up again or I have a bad 700c injector. Quote Link to comment Share on other sites More sharing options...
Zmanco Posted December 6, 2007 Share Posted December 6, 2007 It can hunt at times with correction on, causing my idle to bounce up and down a little bit.If you do decide to use O2 correction, set it to change every XX ignition events, and then adjust the value for XX. On my engine it's best when between 24 and 30. If it's hunting, try a lower number, or a much higher number. Quote Link to comment Share on other sites More sharing options...
Gearshredder Posted December 18, 2008 Share Posted December 18, 2008 do you guys happen to know a thread or a common solution to this idle problem but with a stock 79 280zx? (no megasquirt) thx Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.