What's wrong with idle now then? O2 sensors shouldn't need any running in.

It's a bit better at idle +/- 625 rpm. O2 sensors are generating sensible values, but I still have the warm restart issue, it did it to me this evening. I'm only doing relatively short trips at the moment, which is not helping it, I suppose.
I am trying to spot on Nano what is going on on the occasions it fails the warm restart. My throttle position always used to be 0.625V. Should it read accurately at ignition point II - it seems to move around a bit between .58 and .64 between key turns. what does a failing TPS do?

Like most pots, they tend to get dead spots on the track so don't give a nice progressive change but jump every so often. If the closed throttle voltage is wandering, that suggests it might be showing signs of wear but I can't see how that would affect starting, idle speed yes, but not starting.

I agree with Gilbert but wonder if that remains true if attempting to start with a closed throttle? Perhaps normally during cranking with a closed throttle the IAV opens a bit?

But more likely I would expect it's still fuel trims affecting starting as per my earlier thread with OL1 and OL2,

The only other things that come immediately to mind are ECT reading lower coolant temp than actual so the engine gets cold start enrichment when detrimental rather than beneficial, ignition system that starts breaking down when warm, poor compression. Compression leakage has more effect at low rpm (so especially when cranking) that at high rpm... but would have to be over several cylinders to make much difference to a V8.

Did poor starting start after just changing lambda probes?

No, the poor re-starting pre-dated the new lambda probes/ exhaust (the probes were baked into the old exhaust and the thread stripped when they came out, so when I changed the exhaust I effectively had to change the sensors). I really hadn't done much too it - I'd changed the coolant completely (follow up to HG change earlier in August). Following the HG rebuild I cleaned the MAF, and replaced the air filter, but those were both a month before the restart issue (pretty sure I reset adaptive values at that point). I think the theory that the warm engine is being given the full cold start mix is probably right - the exhaust smells v rich when it fails to start - but it doesn't do it at every warm re-start, which is a bit confusing. I'll check the coolant temp again (there are 2 in a GEMS, right? not sure which one I need to be worried about) - other than Nano, is there a good way to test them? Would the fuel temp sensor have an impact ? Are there any other sensors which might mis-read and cause the ECU to use the cold start fuelling trim. The lambda screens on nano read +39% fuel trim at a cold start, currently, I didn't screen shot the same screen at warm start (yet).

romanrob wrote:

(there are 2 in a GEMS, right? not sure which one I need to be worried about)

There's a green one and a grey one. The green one drives the gauge on the dash, the grey one supplies temperature information to the ECU. If you use the nano to show the coolant temperature, it will be taking the reading from the ECU, so the one that matters.

Gilbertd wrote:

There's a green one and a grey one. The green one drives the gauge on the dash, the grey one supplies temperature information to the ECU. If you use the nano to show the coolant temperature, it will be taking the reading from the ECU, so the one that matters.

If it's the grey one at fault, it'll be under reading. Watch it warm up from cold on the Nano. I think mine was stopping around the 75 mark, if I remember, then throwing the fault up.

New one runs up to around 85/7, depending on ambient.

Thanks George, I'll do that. Just waiting for the rains to subside here in the UK.

ok, so re temp sensors. Temp rises steadily to 72, and then the dash gauge gets to the first white line. By 85 the temp is approaching the middle of the gauge. It then continues to rise (all it idle, with a few revs) to c.95 degrees, toggling back and forth between 94 and 96. Seems pretty normal.
I keep coming back to the IAT sensor, there is a default strategy, if the ECT fails, to use the IAT (see GEMS by Poole pdf). If the ECM is ever relying on the IAT (mine's reading way too low) then I suppose it will retrieve the wrong fuelling mapping.

I'll keep hunting, but I need to find a sensibly priced ERR2946 (IAT sensor) in the meantime

Fuel pressure reads 26 psi at initial idle, and after removal of the vacuum line goes up to 32, and then holds, even after replacement of the vacuum hose (and still running, obv). Those psi values sound moderately low, I will change the fuel filter anyway, and see what happens, but is the low pressure something I can defer until later or can it be causing this issue?
The other thing is the idle is about 625 - whereas it should be more like 700 - I wonder if I need to adjust the air bypass to get the IACV below 30 steps - it smells rich at idle, so perhaps this would make a difference ie more air, less smell, faster idle. I pulled the IACV and cleaned it, seemed fine.
i'm also getting some kind of fumes (not necessarily fuel, but something volatile/ irritating) coming into the a/c for the last few trips, and so I'll crawl under the car tomorrow and check that the exhaust fitter tightened everything up properly.

Other thoughts/ ideas?

Replaced the IAT with LAT033 which is an NNK 10001. It makes no difference, in that the ECM "reads" -1 degrees or so when it's more like 14. Connectors are nice and shiny (at IAT and at the ECM). Botht the old and new sensors perform similarly a. in the fridge, and b. in the oven. Are the sensors actually calibratable in the ECM? Not in Nano, it seems, but are they in Testbook/ Faultmate? Is there a definitive table of what the resistance values should be at a given temperature?
Not sure if this is a red herring to my hot start issue, possibly. Idle is back on track ie smooth, but still operating about 625. Should I care that it's not +/- 700 like the bible says?

I checked my exhaust, and it's all good an tight, so If I have fumes through the vents then I guess that means heater matrix? I get nice fresh air on Lo. A few weeks back I refreshed the coolant - so if there had been any radweld holding the heater matrix together it may have been weakened

IAT makes no difference unless it reports temperature above 55 degrees when it retards the ignition timing. Every one I've ever checked has always read low, mine always shows about 10-12 degrees below ambient. Unlike Super4 in Spain who was suffering from a lack of power in hot weather and it turned out that his IAT was reading high so it was backing off the timing (he's now goit a 1k Ohm resistor in place or the IAT which makes it think it is cooler). Low idle could be down to a rich mixture, although normally that will be corrected by the IACV opening up to get the idle correct by adding more air. What is shown by the Nano as the idle figure as 625 rings a bell with me and not 700 like RAVE says it should be. When idling, have a look at the IACV opening. It should be around 20-30 when warm. At idle what does the fuelling display tell you, is it showing CL by Lambda or Open Loop?

Thanks Richard
-Re IAT, i read that it is used along with the MAF to calculate air flow in kg/ hr - cold air being denser than hot air/ cold air cooling the MAF wire quicker...so the IAT wd seem to have a secondary impact in supporting the MAF ? ie it can survive on a default temp, but will mis-calculate (under-estimate ?) the mass of air flowing if it doesn't know the temperature of the air

• Both lambda sensors are closed loop at idle, they've settled down now, after being first fitted, with no issues - giving values of less than +/- 10%. At cold start/ open loop they are both showing 39%. Since the 625 idle is smooth now then idle seems less of a concern, and presumably if the ECM didn't like the 625 it would up-regulate.
• I still have the hot start issue - did it to me again just an hour ago. My IACV at warm idle is more like 40 - 50 steps ( ie 30 long plus 20 short) which was why I was thinking about backing off the air bypass via the blanking plug - but i'm not sure if the 30 step target in long plus short, or just long.
• Can the air bypass be blocked, perhaps?
• As yet my ECM still only has a zero FMFR value, presumably it needs a bit more learning.
• I cleaned the MAF with contact cleaner previously - in theory it should have worked (ie evaporates without residue), but I could redo it with some MAF cleaner if it was going to make a difference, and have another crack at cleaning the port for the IACV, but i don't suppose it needs to be spotless.
• The car's run rich (smelly idle) since i bought it, although i've done both HGs in that time, and so had the inlet manifold off and cleaned it up etc, but I can always take another look. I have replaced the air filter too.
  There was no obvious trigger to this hot start issue - I replaced the coolant with fresh a few weeks before, and the new exhaust & lambda sensors came later, partially in response to this

I have a replacement ECT sensor too (black) the temperatures I reported above looked quite sensible, is it worth replacing anyway, do you thinK? This trail on landyzone has a guy who also thought his ECT was working, but replaced it and his (cold start) problem disappeared. I don't want to just change stuff for the hell of it, but it's cheap obviously

Confused me for a moment there, talking about lambdas being at 10%, then realised you are talking about the short term trims. Lambdas should be flicking between 0 and 5V (although in reality they will often go up as high as 6V). 5V is a weak mixture, 0V is rich, the opposite way round to a 0-1V Zirconia sensor as fitted to the majority of cars. So at idle both banks should be showing Closed Loop with the lambdas flicking between 0 and 5V.

In Open Loop condition the short term trims should be 0%, default for the programmed fuel map. 39% would show adding extra fuel which may have something to do with your hot start problem. However, don't forget that once it has completed it's learning the long term trim will adjust to offset any wear or variation so the short term can move either side of 0%.

I don't follow your comments on the IACV at all. What do you mean by long and short? At idle it should be between 20 and 30, which are steps over closed, 0 being fully closed. You adjust it using the screw that is directly above it on the throttle body (that once had a blanking plug in it). Air through the passage to it can also be drawn in through the pipe from the purge valve. Now that might be relevant as the purge valve opens to dump excess petrol vapour into the inlet but only when running at speed. If the purge valve is opening when you try to start it, you could be getting a big dollop of petrol vapour being dragged in. Check that it seals and maybe try unplugging it.

It's not impossible that i'm lifting the wrong values as IACV steps - on Air screen no2 in Nano/ Gems I see: Secondary air status 4; curr run line position 20, long term adaptive idle 31, short term etc 27; idle speed 625, IACV 49%. I took the 31 and 27 to be my reference for Steps, but that's a bit of a guess on my part, and maybe it's one of these other values. The Nano guide could be clearer for this screen.

Thanks for your input, Richard, I will also investigate the purge valve - that whole EVAP system is a black box to me at the minute. Charcoal canister to soak up excess fuel.... Does that matrix ever need rebuilding btw?

Yes, you are misreading the values that you see. The Idle speed is idle speed reference so it is at 625 rpm and that is the target idle rpm that everything else adjusts to maintain it at that. The next one down is the IACV and will show a number between 0 and 200, which is the number of steps it is open and not a percentage. It is this one that you can alter using the idle screw so it is between 20 and 30 at warm idle. The others are learnt values that the ECU uses to correct for tolerance variations in components, fuel pressure, etc.

Well, to be totally clear (lazy typing) Nano says "Idle air control valve (%) " with a value of 49... so it sounds like BBS got their Menu wrong (ie added the %) , which I hadn't considered, but in which case the accompanying user guide suddenly makes sense... Let me play with this and see what happens

popped onto the BBS forum - their development team were looking into this a year ago... sounds like some major editing is taking place to remove one percentage sign

So using the "IACV percentage" value on Nano I reset the steps to c. 23 at warm idle. Still no improvement on the warm starts though. When it fails a warm start I reach for the Nano and write the IACV value up to 30 or so, from the zero or so it read prior to start-up. No idea if I should be trusting a pre start-up value, but it reluctantly comes to life that way - I guess conversely I could write a lower fuel setting.
Starting to think of just replacing the IACV and TPS just to eliminate them, and maybe the ECT sensor. Is there anyone/ anything that can dig further into the ECM than Nano - to try and work out what fuelling schedule it's using/ why? Currently it will do short term fuel trim 39% on any start, it looks like. For a cold start that's fine, but for a warm start that doesn't work.
I'm also planning to replace the fuel filter, which clearly hasn't been done for aeons (and so may get my fuel rail pressure up), since one union is frozen solid and clearly hasn't been touched in a while. I have a good 5/8" nut to grip on, and the fuel filter nut is good at 13/16". I have been dousing with WD 40 for a few days now, no joy. I haven't been able to get a flare spanner on the 5/8 end as there is not enough clearance to get one on, unless you use a 1970s bike spanner. So I haven't been able to get a crows foot on and try and rattle it free. I've banged/ tapped the spanners with a hammer, nada...
I really want to avoid the car going up in smoke (so not heating the union), and/or having to replace the filter to pump fuel pipe. Anyone got a funky way to release a rusty union of what I guess is stainless steel vs steel