Car Runs Lean In Cold Weather

[Tompo]

Hi all, I'd appreciate some help as I'm having some weather dependent problems with my MS2 on my 1990 Mazda MX5. The car is currently NA, runs electronic fuel injection, cam angle sensor, batch injection, MAT sensor, variable throttle position sensor, 440cc injectors.

I'm using TunerStudio v2.6.04.

To keep things brief, I installed the megasquirt about 3 weeks ago, the car was on standard injectors (205cc). Everything was working well, I had been using VEAL for about 2 weeks and was consistently hitting my target AFRs; the car was running great!

A week ago I installed larger injectors (440cc) in preparation for a turbo. I recalculated my req_fuel value and set off using VEAL to try and improve my map. Since then I've noticed that I can go out using VEAL and I'll hit my target AFRs but when I take the car out the next day and I'm no longer hitting my AFRs (after saving the changes VEAL makes).

I've noticed this issue seems to be related to coolant temperature and I've logged some figures below:

Warm day:
Time MAP RPM AFR CLT
273 46 2628 14.6 95.7
347 33 1004 14.1 96.2
392 39 1992 14.5 96

cold day:
Time MAP RPM AFR CLT
197 53 1876 17.5 73
222 37 1036 16.9 74
277 37 2579 16.3 77

Going into a bit more detail, I only seem to hit my target AFRs when the coolant temperature is above 90degC. Once I exit WUE the car runs lean until my coolant gets to 90degC. I tried increasing WUE to run until 90degC to see if that helped but in reality I rarely left WUE so I didn't see this as a viable solution.

Can anyone explain why after using VEAL and saving the changes I hit my target AFRs but then the next time I take the car out all my AFRs are way out?

It feels like it must be something to do with the larger injectors as I don't remember noticing this with the original injectors. Are there other settings I need to change other than req_fuel?

I really can't understand why this is happening, if anyone can point me in the right direction it'd be amazing. I just want the car to run as well as it used to. My megasquirt experience has been really positive up until now and I'm sure it's down to my user error

[elutionsdesign]

What are the specifications for your new injectors vs. the old ones? Dead time, flow at low PW, flow at various voltages?

[Tompo]

Thanks for the reply. The only flow rate I know is the old ones were 205cc/min and the new ones are 440cc/min. I don't what voltage this is at though.

Dead times; old ones were set to 1.1ms, the new ones are set to 1.0ms. This isn't verified, but based on lots of internet searching (could well be rubbish I suppose!)

Unfortunately I don't know any more than this.

Something that was bugging me whilst out on a long motorway cruise today...Autotune just couldn't seem to hit my targets. I could maintain constant rpm and load and for example, my wideband (and tunerstudio) were reporting 14.0 when it should have been hitting 14.7. It made no difference how long I maintained that speed the AFR wouldn't change. I could watch VEAL in that area and see no cells changing. Is there something that would prevent VEAL hitting my target AFRs? I thought it could be down to VEAL not being able to make big enough changes so I increased the max cell value and percentage change values but still no joy. I also had no filters set in this region so it wasn't anything as simple as that I'm afraid.

I've uploaded a log from today which shows this.

Timestamps to demonstrate the above point:
Time__AFR__Target
682___14.1__14.6
1312__15.6__14.7
1362__15.8__14.7
1555__14.0__14.7

And so on and so on...

Current ideas:
My VE and AFR tables are scaled different so VEAL is having a hard time interpolating. Although, my AFR target is 14.7 in so many cells that it's rarely having to interpolate. Is it generally a good idea to have VE and AFR table scaled them same?

For some reason my MAT correction is limit the changes VEAL can make. Is this even possible???

I'm starting to fall out of love with Megasquirt now

[nathanhardy]

Fuel pressure? Could it be set to high causing you duty and pw to be too low and ms is try to catch it's tail

[Tompo]

Thanks Nathan although you might have to humour me a little and go into a bit more detail if you'd be so kind.

If my PW were too low what issues would that cause. How would I go about ensuring my pulsewidths are correct?

[nathanhardy]

http://www.megamanual.com/v22manual/minj.htm#size

So this help me quit a bit I had my fuel pressure to high with made my 42lbs injectors think they were 60lbs causing ms to lower the injector pw below 1.7 msec and I was told below that ms has a hard time tuning a stable AFR.. Do a data log with injector duty cycle and pw see where they are at idle and full throttle the duty should be around 85-90% at full throttle and pw no less then 1.7msec @ idle..most injectors are flowed at 3bar

[nathanhardy]

I'd turn everything back off like ego and retune, if you just re calculated your required fuel correctly you shouldn't have needed to change anything else.

[Tompo]

Hi Nathan,

Thanks for the link, I've had a read through that and most of it makes sense.

I doubt my fuel pressure is the problem to be honest as many people are running these injectors on stock MX5s. I have checked my logs from yesterday and looked at duty cycle and pulsewidths.

Pulsewidths at idle are constantly above 1.7msec and my injector duty cycle never went above 20%. I didn't expect my injector duty to ever be very high, reason being I'm running (relative to the standard ones) very big injectors for the stock MX5 engine so there's never enough air going into the engine that requires the injectors to flow very much. When I eventually boost the engine I am sure my injector duty will be in the 80s.

So after trawling through my logs taken on a long cruise yesterday I think I might have actually made some progress (EGO correction has been off for some while by the way). Whilst at motorway speeds (70-80mph, 5th gear) I'm almost always hitting my target AFR of 14.7. However, there was a particular condition where I would hit 15.7. This happens at a particular engine speed and MAP that lies bang in the centre of 4 cells whose values vary by 3 or 4 points. Is MS simply struggling to tune this point because of the conflicting values in the surrounding cells.

I added a couple of points into these cells and I will try to recreate the 15.7 condition and see if it's helped....maybe all is not lost!

I also intend to setup a custom VEAL filter based on MAT so that I can be sure I am always tuning in the same area of the MAT curve. Theory being that I can be sure my fuel map is good and then I should "just" have to tune the MAT curve.

Thoughts/comments/criticisms always welcome on any general tuning remarks I'm making ha! Still reasonably used to this game but I'm trying to put my engineering degree to good use and remain analytical when reviewing my logs. It all sounds so simple when put into words!

As always thanks for the feedback