Quagmire Posted August 3, 2011 Share Posted August 3, 2011 Long time no post from me, just started a new job and their internet policy is not exactly forum-friendly Anyways, here's a lengthy one to make up for it... Background is this: Rover 3.5, 9,35:1 comp, 10 bolt heads, running SU's, and Megasquirt controlling ignition. Runs on LPG 99% of the time, fed by mixer rings before each carb, this is a closed loop system, that goes open loop at idle. Previously the 90 was running Megajolt with no issues, and the same can be said for the squirt too. The ignition map I have been using on squirt was basically a copy of my MJ map, but scaled up slightly to fit into the new bigger tables. By this I mean I inserted some extra points and just interpolated the figures to fill the new rows/columns, i.e overall advance should be basically the same as before. I rebuilt the vapouriser just about two weeks ago (first time in around 20,000 miles) and have been slowly dialling the mixture back in (well trying anyway), so luckily had the laptop plugged in and datalogging. For the purpose of discussion I always talk in petrol AFR terms (i.e stoich =14.7) whichever fuel I am on, as lambda is lambda, but I can never remember the figures. Problem arose the other day cruising up the A3 past Tolworth, maintaining a constant 50mph the car started misfiring for a second, then returning to normal, then misfiring again giving a kind of surging effect. Logs showed that the AFR's were bouncing between 12 and 20! I figured that the closed loop control was doing something wacky, either because it was duff itself, or it was getting bad data from the wideband so I went out at lunch and disabled it. The drive home was much smoother. When I got home I pulled the wideband and checked the reading in air, sure enough a rock solid 20.8% oxygen, not even a hint of movement on that reading. I therefore discounted the sensor. I have since checked the reading in the exhaust just now after engine off for some hours, and its reading a steady 31.4:1 AFR. Because of this I think the sensor is probably not faulty... I have since been tuning the LPG without the closed loop control being used, and I have got some pretty good fuel economy and driveability again (relatively anyway), but I have been getting what sounds like pinking under high load. I put this down to being too lean, and so did some more logging. This is what I saw when cruising at 60mph- my AFR is still bobbing about like nobodies business. The wideband is an innovate LC1 and is set to output at intervals of 1/6s - as I understand it having it on instant response is usually what creates a graph like this as the LC1 will literally see every "puff" of exhaust as it passes by: 1x zoom Sorry- different time code here, but same thing @ 2x zoom So if it's not the sensor, what am I seeing? Presumably some kind of misfire? Have checked and leads are all seated nicely. May pull the plugs next and check them. Does the collective have any further ideas? Thanks! Jamie Quote Link to comment Share on other sites More sharing options...
FridgeFreezer Posted August 3, 2011 Share Posted August 3, 2011 Without thinking too deeply about it - what happens if you set the LC1 to output at 1/2sec intervals? Quote Link to comment Share on other sites More sharing options...
zoltan Posted August 3, 2011 Share Posted August 3, 2011 Might be worth checking/replacing the gas filter on the solenoid valve upstream of the vapouriser if you haven't already done that, mine played up badly after running fine for a similar mileage IIRC lambda one for LPG is 15.5 AFR for what it is worth Quote Link to comment Share on other sites More sharing options...
Quagmire Posted August 3, 2011 Author Share Posted August 3, 2011 Have tried changing intervals to the slowest, which is 1/3rd of second that had no impact- still a lumpy readout... And I changed the LPG filter upstream of the vap unit when I rebuilt Will try a short log in similar conditions tomorrow on Petrol, see what I get and report back. That will at least tell me if its fuel dependant, or something else I guess. Quote Link to comment Share on other sites More sharing options...
FridgeFreezer Posted August 3, 2011 Share Posted August 3, 2011 Well I'm not sure how it relates to engine speed, by my (probably wrong) calculation: 2700rpm = 45 revs per second = 22ms per rev Whereas you seem to have a ~1Hz sine wave going on there. Would be useful to see if it changes frequency with RPM or interval. Can you add a few other sensors into your log picture - TPS, MAT and CLT would be a useful start. Quote Link to comment Share on other sites More sharing options...
Quagmire Posted August 4, 2011 Author Share Posted August 4, 2011 Fridge, unfortunately I am a lazy bugger and still running around with no TPS (running SU's ), MAT or CLT sensors. That may change this weekend as I am being evicted from my own home whilst my better half has her friends down from up North so I will have some tinkering time on my hands. Was hoping to try a petrol run on way into work this morning, but was running a bit late so no time to hook up the laptop etc. Will def have to get to the bottom of this though, as no point in fitting EFI and tuning if AFR's are going to be bouncing around then too! Cheers Jamie Quote Link to comment Share on other sites More sharing options...
Quagmire Posted April 6, 2012 Author Share Posted April 6, 2012 Update on this one- a few weeks ago I finally ordered a new wideband sensor. Fitted it and still getting error code 8. Did some trawling of innovate's forums and found that the 1.1 version firmware on the lc1 was prone to this. I found a copy of the (beta version) 1.2 firmware, which incorporates some code changes to address this, and flashed it to the unit. Since then no problems. Back running the old sensor with the new one in a box somewhere in the back bedroom as a spare. 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.
Note: Your post will require moderator approval before it will be visible.