MadDog

Well, its a beautiful day to tune in the land of road runners and chili! 50 degrees and not a cloud in the sky!

I've been risking my neck all morning tuning this beast on the desolate (and, of course, private) roads around here. I've made some great progress. This fuel management strategy is so easy to tune. I'm not done yet, but I wanted to go ahead and share something in case someone else (jeff) is tuning today.

I made a change to the fuel MAP this morning that improved the throttle response almost as much as relocating the wastegate signal. Before this change, approaching boost, letting off the throttle, and then getting back on the throttle would result in lean spikes in the AFR and a kind of bumpy ride. I suspected this was because the fuel added was turning on and off abruptly at 0psi as I twekaed the throttle around +/- 0psi. (going straight into boost wasn't a problem).

I was right. I added a line at -5kPa and added a tiny amount of fuel to the 0psi line in the MAP. This fixes this problem. Now, there is a tiny, but important addition of fuel as you are near boost. I also changed the VTEC settings so the temp dongle comes on at 0psi. I doubt this even matters, since the LTFT is pretty much eliminated by this whole strategy. But, what the hay.

I'd reccomend this addition to anyone running the eManage - even if you haven't yet migrated to the new strategy. It really smoothes things out nicely. This, in combination with Jeffs solution #2, have made a huge difference in the around-town drivability.

Enclosed is a pic of my latest 490cc injector map. You can see the addition of 2% duty cycle at 0psi before 6kRPM.

I''ve also decided to refine the MAP around the opening of the intake ports at ~4250 RPM. Since the MAP is turning out to be very nearly a linear interpolation across both boost and RPM, I am going to sacrifice one of my RPM columns to specifically attack that lean spike. I'll post the resulting MAP later today. I might even convert it for you 390cc types (insert 'Jeff') if I'm feeling generous!

MazdaManiac

iTrader: (3)

Awesome. I was never satisfied with the acceleration map as the only cure for transitional lean spikes.
Did you get a chance to look at what I did to the MAP with regards to minimizing the effect of +85% duty cycle? I am realizing now that it won't work like I want it to. I need to find out what the ACTUAL duty cycles are of the secondaries at each of the load points and then add a percentage that is equal in time to the original intent of the P2 sub-injector map.

Only 50°? It was 75° in Phoenix. Of course, it is only 48° and raining in DC.
I won't get a chance to play with the car again until February 27th.
Hopefully by then I'll be able to source a set of secondary injectors by then.

Please please please! I was never good at math!
I think I can figure this one out. All you have to work on is the "0" line? I might be able to handle that one.

How is this:

two rotors

Hey MM you don't have to suck up anymore,new 490cc injectors are a LOT cheaper than you thought!

MazdaManiac

iTrader: (3)

OK, I'll bite. Who has 'em? Mazda wants almost $500 for each and MazMart doesn't have any in stock.

two rotors

check you PM

Greddyturbo1

Hey two rotors, why not post for everyone, where to buy...thanks

two rotors

Oh sorry about that,but I was merely pointing out that the list price for fuel injectors from Mazda dealers(in Canada) is $198.65 each(part no N3H2-13-250). All 3 injector types,ie
N3H1-13-250A,N3H2-13-250 and N3H3-13-250A are the same price.
So if you divide by 1.13,that will give you the $US price,and you should be able to negotiate a discount.
As of 02 Feb06 there were 49 injectors in MNAO warehouses(or more correctly in Catapillar Logistics warehouses).

MadDog

Thought I'd post some progress on the tuning. I know this picture looks like crap, but its the best I can do right now. I did make a rediculous attempt at drawing-in the lines so you could see them a little better. Its boost, RPM, and AFR as a function of time.

You can plainly see the two lean spikes that occur when the ports open. You can also see that I have tuned the AFR a little richer around those spikes to compensate. Just to give you points of reference, boost starts at 9psi and is at 8psi in the middle. AFR is pretty constant at about 11.3-11.7, except at the lean spikes where it reaches 12. The plot starts at 2600 RPM. I'm at full boost at 3200 RPM. The lean spikes occur at 4360RPM and 6230RPM.

One of the things that makes tuning these spikes out with any precision such a bitch is that they occur at different RPMs (as a function of load, I suppose). I'm trying to keep the rich areas to a minimum RPM width. I'd like to see a AFR trace from someone with an Interceptor. I'm sure the spikes are there, too. It would be neat to see how they are dealing with them. One thing to keep in mind is that I have my LC-1 ouputing data at the maximum rate. Most dyno tailpipe sniffer 02 sensors don't have the temporal resolution needed to really show these sharp spikes. My dyno shop didn't show them, but my LC-1 sure does. Keep that in mind when you have your cars tuned, fellas. Eventhough you are paying a "professional" to tune it, you should still have the ability to keep an eye on things yourself.

I'm liking this tune so far. Power is nice and smooth all the way to redline.

MazdaManiac

iTrader: (3)

When tuning those spikes out, make sure that you are starting your fuel-add about 500 RPM before the spike.

dannobre

iTrader: (3)

One of the difficulties of the E-manage for tuning out spikes is the way it "interprets" the vaues between the cells. One way I found to do it is to group 3 RPM pionts close together around the spikes and then use the center one to drop the spike from the values on either side. That way it doesn't spread the averaging over a larger RPM range. The only problem is we don't have enough RPM cells to do this and handle the rest of the tuning range. It might be worth it to try on the APV spike at 6250. If the fuel needs are relatively linear you might be able to use less cells for that and have enough left over to handle the spikes.

The other problem with "spike" removal is that the spikes aren't always in the same spots...although with the P2's severed it would negate the factory PCM from adjusting the fuel to them to change the A/F when it wants

MadDog

Hello? Is this thing on? (tap tap tap)

As I said before, the map turns out to be very nearly a perfect 2D linear interpolation with RPM and pressure. Therefore, there is no benefit to having columns devoted to specifying additional RPM points if they are just on a straight line between the adjacent columns anyway. The emanage will do that interpolation for you.

I actually computed the error between a perfect 2D linear interpolation, and the existing RPM columns that were occuring every 500 RPM. The error terms form a 14X16 matrix, since there are 14 columns that could be removed from the total of 16 columns. I found several columns in my MAP that had maximum errors of less than 3% duty cycle. Error = (interpolated value) - (actuall value). Most of the error terms were below 5% across the board anyway. The fact that the Renesis can be tuned with a nearly linear function indicates that tuning this thing aint rocket science. I relocated these columns with the smallest error terms to increase the resolution around the first lean spike.

I chose one column from the low range and one from the high range, to increase the resolution around the lean spike that occurs between 4250 and 4500 RPM.

Here is what the current MAP looks like. See that nice little ridge in the vacinity of the first spike?

Next, I'm going to move one of those lines clustered around the first spike to tackle the second one.

MadDog

Here is my first cut at addressing the second spike.

dannobre

iTrader: (3)

HUH?

Anyway...youare the man with the charts...I only wish that I could do that in excel...if you have a template to play with...please post it. That is cool

Have you found that the opening point for the SDAIS valves is always the same...and the spikes are in the same place with your setup....cause they used to move around for me based on the load/time thing...

Also...how much do you think small spikes like you have it down too are a problem??

MadDog

Yeah, those graphs really help the visualization. But, I'm not using excel, I'm using Matlab. That's also what I use to do the 2D interpolation for changing the boost or RPM values, scaling for different injectors, etc.

The spikes definitely move around. They occur at lower RPM in the low gears, and higher RPM in the high gears.

As long as I can get the peak of the spike to around 12, then I think I'm okay. That seems like a reasonable trade-off.

dannobre

iTrader: (3)

Can you bracket the peak and richen it up a bit earlier to carry it forward over the spike? That might give you a small loss for the added safety of the whole map. Like MM said...if you start a bit earlier..it will buffer the spike and won't go as lean. Personally I don't think that the small spikes that you have will be a problem.

You have done a great job with implementing this change.. It makes the kit much more viable for those with the Greddy Kit...or for those that are happy in the HP range that can be derived from this turbo.

MadDog

That's exactly what I am trying to do. The log that I posted shows it richening up before the spike arrival. I am playing games with the RPM values that I can use to give me A) the narrowest rich zone. B) still address the fact that the spike occurs at different RPM values. C) give satisfactory AFR at the peak of the spike. Its an optimization problem.

MazdaManiac

iTrader: (3)

I don't think those spikes will contribute to a power loss particularly.
I would only worry about the possibility of a detonation event being triggered by that lean spot which could be addressd with a degree or two of retard just before those spots/

MadDog

I'm the retard here. I still haven't implemented the ignition timing control

dannobre

iTrader: (3)

Mad...have you wired up you ignition to be able to do the retard?? If I recall the Greddy kit doesn't do ignition adjustments?? Then you need to see what the ECU does to the A?F when you pull a bit of timing

Jeff: I assume you have left yours set up to be able to pull timing? What have you noticed if you did??

dannobre

iTrader: (3)

Jeff: have you logged the fuel injector pulsewith to see which injectors the ECU uses to change the A/F when you change the timing?? Hope it's the P2's

MadDog

Since taking the P2's away from the PCM, there is very little that the PCM can do with the fuel trims to really affect the AFRs any more. I'm in control of a much, much larger volume of fuel than the PCM can adjust via the fuel trims and the remaining injectors.

MazdaManiac

iTrader: (3)

One thing I found in my move to Phoenix is that the gas sucks!
Right now, I am running about 2° of timing retard for each pound of boost. I'll probably be able to bring some of that back once I narrow down to the spot where the rattles occur at WOT and full boost. Seems to be in the 4800 range, but I haven't had the time to tune it out.
In MD, this wasn't an issue and I only ran 1° per pound.
It will probably be a bigger issue when the temps start to creep above the 100°F mark.

dannobre

iTrader: (3)

I don't think the fuel trims are affecting this...the ECU seems to adjust injector pulsewidth directly as the timing changes. I never logged this and was wondering if either of you had. When you get your ignition wired....see how adjusting the timing affects your A/F maps. I found that it really changed things when i played with the e-manage NA.


Bad Fuel really Sucks.......We're lucky here..we get good fuel 94 is readilly available...and they don't **** with it as much because of the temperature

MadDog

Interesting. I though that the fuel trims were the only vehicle that the PCM had to adjust the injector duty cycle. Could it be that the injector duty cycles were the same, but that the combustion process was more incomplete with the retardation of timing, giving a higher apparent AFR? I have no idea if this could actually be the cause. Just a thought.

dannobre

iTrader: (3)

Someone would have to log the Injectors...and see what's happening. I just noticed that when you dialed in a map..and then changed the timing...esp advance...that the A/F changed accordingly. FT's are based on the closed loop adjustments and don't occur in open loop areas in real time....so they would only be carried forward after the fact....I don't know.....someone pipe in here and help us out