12P MAP switching and knock control ideas and questions.

[vlad01]

For a long time I wanted a way to retard timing based on MAT only and not the charge temp which happens to be the only option in 12P

Charge temp I found to not be that useful as I found the occurrence of knock to be highly correlated with ambient air temps and somewhat manifold heat soak, and almost not with coolant temp.

The issue with using charge temp is that most of the number comes from the coolant contribution and less from air temps and the charge temp numbers are too broad to be able to accurately and insensitive to air temp be a clean trigger for the charge temp spark subtraction map. I guess this is why the table is blank from factory, it's there but not used.


Driving in Melbourne the other day and thinking about the fuel prices and knock as knock only happened in two places period, both in Melbourne where traffic conditions and lack of air flow and low gearing would cause the air temps to soar. The fuel prices reminded a post I someone posted a few days about about revisiting map switching for a low oct 91 due the the fuel crisis.

Then an idea dawned when I was driving.

What if the flex table could be looped back into the ECU to do a map switch?

My question is, having never used the map A/B switching feature, is it real time, ie switches over while the car is running? And if yes, how fast does it happen?

I was looking at some of the 12P release info and general tuning guides and stuff, I couldn't find this info specifically, I may have missed it, or it's listed somewhere else. So if anyone knows for sure, that would be helpful in looking into this further.


My idea is if those assumptions are true, I could use a flex table to use MAP and MAT axis to output back to pin B9 which could switch to map B with a reduced spark up top, when the condition goes away the map should drop back to A.

This idea went further, if one used an aftermarket knock module, it could also trigger a map switch. Only issue here is speed of the switch if using an external module.

With the flex table,you can set the table to trigger at lower maps and MAT temps before knock actually happens, giving a lead on pulling timing as proactive measure rather than an reactive one in the case of an external module.

[Charlescrown]

Going off what antus said in reply to my post yes you can switch on the run. The rest abt using flex tables is at this stage beyond me. I have just finished a map A and map B program for my Starion pulling 3 degrees of timing (guessing point to start) so I can run on 91 and have connected a wire to ECU pin B9 but need to dump the program into the memcal and wire up the grounding switch before I can say it was a success but I'm sure it will be with all the effort put into 12P. Doing some bodywork tam and waiting on some clips from Japan before I put the wipers back on and can go for a drive. I'm sure the boys in blue won't be impressed with no wipers on a classic rego car.

[vlad01]

Ah yes, it was you who mentioned the map switching. If it works on the fly, it should work with the flex driving the switching as the flex tables are running both A and B as they are shared. Its only the dedicated A or B only settings that switch over. You can configure the flex output to be digital on/off for the likes of driving relays, or a PWM for other applications. There is also a flag to invert the input, so a low can be a high instead. So you should be able to go active low to switch the map to B. If not directly, a buffer or a simple relay can do the job, but ideally directly as a clicking relay would be annoying.

I think this is def worth pursuing as it's a good potential work around to the shortcomings of the delco knock system.

I wonder what the total time is from input assuming grounding to actual execution of the cal data in map B is, it maybe fast enough with an external aftermarket programmable knock module. Yeah it won't prevent it, but might be the next best thing to a modern knock retard system in newer cars. Many of those aftermarket knock units have outputs for feeding into an ECU input.

[Charlescrown]

I'll follow your project with interest because the flex tables is something I know nothing about.

[antus]

Yes it could be done. But charge temp is supposed to be the temp of, well, the charge and it is supposed to be used for what you describe. If its not getting it right for your conversion there is a scaler to adjust the percentage CLT versus MAT to calibrate heat soak. You can also change that to get better charge temp readings.

[vlad01]

That's pretty much what I ma trying to get at, the charge temp isn't a reliable indicator of knock condition but air temps are. Changing the coolant contribution table will just mess up fueling stability, the table for the factory cam is pretty spot on.

I have also ruled out coolant temp as the major factory physically by running a colder 82c stat and adjusting the fan setting to suit. I found that this did not change the knock tendency for the condition that it occurred in.

So I need a pure intake temp vs map table that allows spark offset. This doesn't exist in 12P, but I know for sure there is one in the MAF PCMs.


But knowing that a map switch is a feasible work around. My goal is next summer to drive in Melbourne to the two locations that it occurs and log it.

I tried to simulate the condition 2 years ago by blocking off the radiator with cardboard and stuffing all the opening with newspaper. The air temps went up heaps, but the coolant temps were also up heaps. No knock occurred no matter how much I tried.

There is two likely factors there, lack of atmosphere, I am 600m up, Melbourne is pretty much sea level, the other is the much higher coolant, this might be affecting the fueling in a way that side steps the condition or the condition is physical, like less overall charge density despite temps being high. I think it is a combo of higher air density + high air temps that are causing it. The higher air temps are the trigger for knock I found consistently. but it seems when the air density outside and/or inside the engine is high enough in addition to the temps, the knock actually occurs

The city tunnel is where I can 100% create knock, low air flow due to trucks moving air at the same speed as the traffic, hot air from the trucks going up hill and potentially higher that normal air pressure as the tunnels are probably below sea lever. The coolant temps are still good at a max of about 90c.


I think I had tried like 2 deg less timing, but it hurt power a little overall and still had some knock, not anywhere as obvious. I think 3 deg should be enough to stop all knock, but we are approaching towards the factory timing and significant power loss. Having a temp loss for these rare by repeatable conditions is completely acceptable for me as with any good factory knock control system, this is normal. But it would shit me having to limit the entire engine 100% of the time, where I can hopefully set this up and only have to limit it less than a fraction of 1% of my driving, but the engine will thank me for it.

98 ron worked well to prevent it, but I would like to be able to use 91.

Interestingly, E10 with its 95-96 ron makes zero difference over 91, that is surprising to me.

Enhancing AFRs within reason made no difference either, I tried this a few years ago I normally run 12.8, I have gone anywhere from 13 to 12.0 and still not real difference, going less than 12 is just a waste of fuel and loss of power. These engines make best power at 13 WOT, but 12.8 adds a little extra safety with little cost to fuel usage.

[antus]

The baro code should compensate for height from key on, and 12p has improvements to update this while the engine is running, though the factory software only did it at start, so if you drove from melbourne up to 600m high it might start running crap and need an engine re-start. 12P should be fine for that. If thats what you want to do, then yeah wire up the map a/b switch. And either switch it manually depending on your fuel or loop back a flex table like you suggested. I know we've talked about flex before, but as a reminder you use the ose flash tool to configure what you are mapping to flex, it goes in to the OS part so you need to flash the whole bin in (can use oseflash tool for that with 12p if you want, the ALDL loader is in a constant ram location to make full bin flash 12p to 12p possible), and ose flash tool will also spit out an XDF that is configured to edit the flex tables with the size/axis etc that you configured for your bin exactly. Then test it how you will and loop it back to the map switch. You should see the flex output and map switch change at the same time in the data views.

[delcowizzid]

If its knocking in traffic heat it must be well past mbt timing when its not hot

[vlad01]

Well as mentioned it never happens in any other traffic except for two places that I know of in the years and 100,000s of km driven, but it would happen without fail once the weather was hot enough, but nowhere else. The timing is about 28-29 deg max at which is pretty bag on for a Buick with pacemakers. I wouldn't run that with shorties or the logs though as they choke the engine and lower the knock limit significantly.

On the VR (now sold, but I kept the engine) it was 27 deg on the dyno for MBT, It also ran 98 exclusively, but it has 9:1 comp where as the other cars are all 8.5, hence the slightly higher timing. I also never drove the VR in the city other than the once and not in those two locations from what I recall, so I can't say if that engine would do it too, probably not since its on 98. 98 also "fixed" the issue on this car in question.

I have nailed it down to abnormally high air temps, not just a heat soaked manifold, like actual outside air temps which would probably be in the high 40s I would imagine.

As mentioned in a post up top, I have reduced timing a couple of deg and enriched a bit, but it didn't change much if anything, maybe knock was for a shorter duration?, lower thermostat also didn't help. I am sure it will stop if I take a heap of timing out, but that would be significantly less than MBT and not something I want to sacrifice permanently. It's a rare occurrence but a repeatable one, and this is the best solution I can think of that should work off a reliable data point, one that doesn't exist in isolation in 12P.


Yes Ant, I am all over it. I have already played around with setting up the flex tables in the flash tool and got one loaded up in 12P without an issue. I have set one up already as a virtual cruise control neutral safety switch since the box physically doesn't have one nor is there an easy way to physically add one. But the flex table allows an rpm and road speed constraint on the cruse control safety loop, so if I were to "accidentally" put the box in N and resumed or set the cruise it should only allow it to rev to about 3 or so K before the relay is triggered open and cuts the cruise. Saves it sitting there on the the red line, yes they do that from factory on the autos if moved into N, but mine is a manual so I also have a switch on the clutch pedal and also adding the flex for N, as it is still possible to run the cruise like that without it knowing it's revving off its face.

[Charlescrown]

Just to let you know the map a and b switching is working and like antus said it can switch on the run. I never knew how many maps a and b configurations there actually were in 12P. Many I have no idea what the do. Next think I'll look at if it's possible is switchable VE table so if I go down the path of E85 switching from 98 but at this stage I have no idea. It's called discovery learning and lots of good help from this forum.