Joe Perez

iTrader: (8)

Mmmmm, data. Engineers like data.


I found a nice, secluded back road this evening and did a couple of back-to-back pulls. The road was at an upwards incline, however I was able to match speed almost exactly and use two obvious landmarks as marker points in the log.

The pulls were both done with the engine fully warmed up, and about two minutes apart from one another. I ran in third gear, holding approximately 40 MPH as indicated on the speedo. The only thing which I changed between the two pulls was the actual spark map itself, obtained by exporting a VEX from both my current MSQ and my starting MSQ, and switching between them. Ignore the listed Min and Max values in the logs, as the raw data also included accelerating up to speed, stopping and turning around at the end of the hill, etc.


The first pull, representing the "original" spark map which is common to both the MS3 MSQ I got from a user here as well as the DIYAutoTune basemap for a '90-'03 MSPnP2:




And the second pull, using my current spark map which is advanced roughly 8-10° in all areas relative to the first, and is by no means optimized:




I then took the raw log data, imported it into Microsoft Excel, trimmed it down to just the marker area, and performed an arithmetic average on all parameters of interest. In both cases, I used the last 7 seconds of data prior to the ending marker, which represented the most even section of the hill. In addition to the mean, I also computed the delta on all parameters, and the variance between min and max was approximately equal for all parameters when compared between the two logs.

For the original map, the averages are as follows:

RPM: 3250
TPS: 26.5%
MAP: 90.6 kPa
Spark: 26°
Fuel PW: 8.7ms

For my current spark map, the averages are:

RPM: 3309
TPS: 18.4%
MAP: 78.8 kPa
Spark: 37.6°
Fuel PW: 7.3ms


Pretty compelling evidence. In the second test I was actually going 1.8% faster, and yet all of the meaningful indicators are lower. TPS was 30.5% lower, MAP was 13% lower, and the most important variable, which is fuel PW, was 16.1% lower.

16% less fuel to go the same speed up the same hill, just by bumping the spark advance from 26° to 37.6°.


So the question remains: why is everyone else so retarded? (pun intended.)

18psi

iTrader: (76)

Joe, I don't think these base maps you keep referencing are any sort of "standard". Most stock maps I've seen (on many different cars) run 40-45 and even 50* in cruise. Why should our little turds (that happen to love timing) be any different?

We're all retarded because we're too retarded to explore advance. LOL

18psi

iTrader: (76)

Just looked at all my maps from the 2000:

I ran 40* like a boss. So maybe I wasn't that retarded

Gryff

iTrader: (6)

Thats it, im playing with my spark map as soon as I get back from RA this weekend

18psi

iTrader: (76)

Just for ***** and giggles, stock subaru map:

Joe Perez

iTrader: (8)

I noticed your little stealth edit there, Vlad.

Wherever I refer to a "stock" map, I am generally referencing the MSQ files which DIYAutoTune publishes for use on their DIYPnP, MSPnP2, etc. Typically, these are the files which we point people towards when they ask "where can I find a base map?"

I have judged the ignition maps in these files to be similar to the ignition maps in the MSQ files which I have seen shared by Braineack and other users here.

The mention of FM also made me curious, and so I went and downloaded their published maps for the old Link ECU (1.6, stock injectors, naturally-aspirated.) After converting them into a more useful format, I found their spark advance map to be similar to DIYs- not an exact match, but in the same general neighborhood.

Given that that you don't hear much about Miata owners re-shaping their combustion chambers, the optimal spark map (in the non-boosted area) ought to be pretty much the same for all B6 engines regardless of whatever else has been done to them. And since spark advance is quite difficult to optimize (as compared to VE, idle, etc.,), I have just always assumed that the maps provided by DIY were already optimized for a typical B6 engine in the non-boosted area. I can understand being at a bit conservative, but we're already at more than +10 degrees and still no knock, on 87 octane gasoline.

Regarding the Subaru map, that's a bit closer to what I'd have expected. I'd love to know how to convert between the MAF-based scaling of that table and something like a MAP-based scaling, but it sort of validates me theory.

I'd *really* love to see the advance table for an OEM Miata ECU. I've been searching through the infamous "Socketing" thread at Miata.net, but I have yet to find anything thus far which is in human-readable form (eg: something other than a raw hex dump with no translation.)

hustler

iTrader: (6)

Joe,
What size injector are you running for that PW?

I cruise at 3700rpm at 64kpa and 5.1-5.3ms of PW on NB injectors at 80mph.

Ignore the whacked-out TPS signal, that's fixed now:


Any more advance means detonation without fuel, and adding fuel increases PW without reducing MAP. So, I stopped here.

Joe Perez

iTrader: (8)

As I said, this engine is bone stock. Stock 1.6 injectors, cleaned and flowtested last year.



Sounds about right. You'll note that the logs I posted above were going uphill at a steady speed. Additionally, your fuel pressure is higher than mine in cruise, because you have an NB and I have an NA, so the effective flowrate is greater. So 7.3 ms for me going up a hill at 78 kPa and 5.2 ms for you at steady cruise at 64 kPa is basically the same thing.


Something is seriously wrong with your setup if you're detonation-limited to 15° at 2500 RPM at 100 kPa.

I question whether 15° in your log is actually 15° in real life.

Also, why do your log and your spark table not match? The log indicates 34° at 56 kPa at 3850 RPM. Your spark table says 31° in that region.

Reality check.

hustler

iTrader: (6)

35* of intake cam advance on 87 octane with 10:1 compression.

I only see about 93kpa since I'm in Dallas and it's a long ways from the ocean.

I did very minor tweeking and reduced spark angle a touch and cleared out all my junk logs...that's why it doesn't match. I'll try to make a new log tomorrow on the way to work. The example was a recent adjustment where I alternated between two settings that showed good results. I went with the one that suggests .2ms shorter PW.

hustler

iTrader: (6)

My 11.5:1 engine with the MSM intake cam on NA sensors (good riddance) and 93 octane:

Don't chastize me over the lazy tune on this...it's a total Barrett Jackson tune that spits fire, bucks, and genuinely drives like ****. However, it got me to the track for M@H.

Loren

I know the topic here is primarily spark advance, but since it was spurred by unsatisfactory fuel economy... sure, you can see some improvement by tweaking the advance, but...

You can absolutely run leaner than 14.7 in cruise. Since you live in a hilly area, you need to be careful in tuning this. I'd try to map what kPa you're at at each RPM in your typical cruise modes. Say, 35-80 mph in top gear on flat ground. That is true cruise.

You should be able to EASILY set those areas to 15.5. For lower speeds (say, under 60), you can probably go leaner than that. Some people claim to have run as lean as 17-18:1.

Try setting the center of your cruise area to 15.5, taper to 14.7 by maybe 70-80 kPa, and then to your preferred richness at 100 kPa. Anything BELOW cruise MAP can taper even leaner, you won't be doing anything down there, anyway. Couple that with as much advance as you can get away with in cruise... should make a HUGE improvement in fuel economy.

Also, be sure your fuel-cut is set so that you don't use fuel anytime you're off-throttle coasting. That's a trick that all modern cars use (they call it Deceleration Fuel Cut Off) that can help a lot with economy.

There is definitely some discussion on the topic of fuel economy on the Megasquirt forums if you dig for it. Also a little bit on some of the eco-geek forums (like ecomodder.com), but those guys are more into lightness, aero and driving technique than ECU tuning.

FWIW, I've seen 48 mpg in my stock 1991 Miata. (had to "hypermile" the hell out of it to get that, but I did... hoping for 50+ once I get the MS tuned) More recently, I actually achieved 43 in city driving. (also hypermiling... but not so much that it affected how long it took me to get anywhere, or held up anybody else)

Chasing fuel economy is an interesting diversion... and generally isn't likely to get me into as much trouble as chasing speed.

Joe Perez

iTrader: (8)

Ok, so that's now two highly modified engines whose tune is irrelevant to a discussion on stock basemaps.

Thanks for playing, and enjoy your free copy of our home game.


Yes, and my last car was ultimately tuned this way. I may eventually go in that direction with this one, but for the moment I need to generate a tune that will breeze through California emissions. That means oscillating around stoich in cruise.

Assume for the moment that 14.7:1 in cruise is a given. That's how the OEM ECU operates, and it was good for around 30 MPG highway, mid to high 20s in the city. If you were to look at a complete log of my daily commute, you'd wonder if I was aware that the throttle has any possible positions other than wide-open or fully-closed. And you'll note that I am aggressively advancing the timing at 100 kPa as well, a region in which I think we can all agree that running at or leaner than stoich is generally not a good idea.

If I had a dyno in my garage, I'd be posting logs of WOT operation and comparing torque rather than fuel pulsewidth. But as all I have access to is the street, I have to go about this indirectly and assume that lower MAP and PW for a constant, partial-throttle load means higher efficiency (which I can measure directly) and will tend to translate into greater available peak torque at WOT (which I cannot directly measure.)

Given this, it's obvious that a lot of us are flailing around with grossly conservative ignition maps. I want to find out two things.

1: Why?

2: How far do we need to push the spark advance map to find the sweet spot?


I appreciate your reasoning, but fuel economy was merely a symptom which spurred my curiosity. That, in combination with a recent thread in which a third-party criticized the ignition map of a fourth-party's ECU tune, gave me a sort of Dr. House moment. Screw the patient, I'm interested in the disease.

My end-goal here is knowledge, not some arbitrary mileage number.

That said, if you'd care to share your ignition map, I'd love to see it.

Loren

I'm on a stolen ignition map at present and have done no tuning at all on my MS (just installed it a few days ago). The map I'm running is one that was dyno-tuned on a stock 1.8 making 130 rwhp, so it's worthy of consideration even though I have little experience with it. It's here: https://www.miataturbo.net/megasquir...e2/#post903093

My previous MPG-runs were done on a stock ECU (8MHz overclocked to 9MHz) and base timing set to about 16 degrees. Overclocking the ecu "stretches" the timing map, so stock timing would have come on about 10% later than stock. And I'm 6 degrees more advanced than stock... whatever that map is.

But, really, my MPG numbers are more a factor of the combination of lightness (2060#), narrow tires (175's), gearing (4.1 vs 4.3 and taller tires), aero (hard top and right mirror delete)... but mostly how I drove it. To get close to 50 mpg out of a Miata takes a lot of "momentum driving", hardcore "pulse and glide" and judicious use of engine-off neutral-coasting. Not exactly the kind of thing the typical commuter or most "normal" drivers do. I get 28-30 mpg around town without trying. (just from good habits) 30-36 on the highway depending on how fast I run.

Ending fuel economy tangent.

Good timing definitely unlocks power. My previous Miata (96R) also had an overclocked ECU, and timing advanced to 18. But, it was using a Bipes, which pulled some timing above 3500. I have a dyno chart of that setup (and comparison with the stock ECU and no Bipes) somewhere... actually, I'm sure it's in the archives of Miata.Net. I was pulling something like 118-120 hp to the wheels with a very healthy torque curve.

Ah, here is is:
Dyno - ECU Mod & Bipes - MX-5 Miata Forum

Pretty much demonstrates that bumping timing from a "typical" 14 degrees (which is already 4 degrees greater than stock) to 22 degrees can yield significant torque and power increases. The Bipes was pulling a max of 12 degrees starting at 3500.

In this instance, I'm sure some of the power increase came from decreasing the overly rich fuel mixture (12-12.5 stock) to something leaner (14ish below 4k, less than 13 until over 6500).

Based on this experimentation, if you started with a stock map, it seems that you could add 10-12 degrees below 4000, and then taper that down to near stock level from there to 7k+.

A knock sensor is on my list of things to do with the MS. I plan to do some real world "find the limits" testing and create my timing map from those results. Don't know how long it will be before I do that.

Joe Perez

iTrader: (8)

Ok, so no valid data there, either.


Yes, but there's the rub. I have utterly no idea what a "stock map" looks like under moderate load at 4,000 RPM, so I don't know what 10-12 degrees more than it would be.

That's kind of the whole point of this thread.

Oni

iTrader: (1)

Interesting Thread!. Joe how are u monitoring knock/detonation?

Cheers
Scott

Loren

23 years of Miatas and nobody's cracked the stock timing map? (no, I don't have it, either... but I'm kind of surprised that it's not out there somewhere)

I was reading a thread here on MT last night where a guy mapped the "achievable" parts of the stock timing map by monitoring it through the OBDII. That seems like it could give us a good idea of what the stock map is where it matters. (I'd do it, but no OBDII)

Here it is. It's not much, but it's something.
https://www.miataturbo.net/megasquir...93/#post713985

richyvrlimited

iTrader: (1)

Pretty sure that this is the table 18psi mentioned. The load scale is next to useless however.

Oni

iTrader: (1)

id think it would be safe to say 15 is max load and 0 min, perhaps guestimate 0 as 0kpa and 15 as 110kpa to allow for differnt elevations. im about to sleep but im sure someone can put taht in excell and adverage out the values in steps of 10kpa and then adverage each cell based on the ones arround it. like that smoothing spread sheet does thats about somewhere. its not 100% but it would get you pretty darn close id imagine.

hustler

iTrader: (6)

The VVT engine is stock and runs only 10* of cam advance in cruise.

18psi

iTrader: (76)

I think you know the answer to the 2 questions:
1) a few years ago a gt2860rs turbo was considered huge on a miata. bipes and fmu were considered satisfactory. The Link and MS1 were considered cutting edge new technology. idling and cruising @ 13:1 on pos ev1 injectors was standard. Its JUST THE WAY IT IS.
Now we got folks questioning things. Questioning tuning strategies. Setups. Etc. Figuring out more efficient ways of doing things, buying better technology.
You get the idea.
Basically: " We were too retarded to even know we were retarded "

You're among the few on here (hopefully) making strides in the right direction. Hell, I bet half the folks on here are still running FM/BEGi/DIY base maps or ones they stole/borrowed/copied from others and don't even question the values because "it runs good enough".


2) I don't think there is a set number. Esp since you're running a dinky 1.6 low comp engine. I bet you can advance it to 50* or something like that. I say test the limits til you hit det, a brick wall, or start using more fuel than normal then back off a little.

Which is most likely what you are going to do anyway.

/coffee infused morning rant

yessir. I think there's another one or two as well, but they may have been for an NB
Agreed. On all counts.

Someone like Joe can probably do that in his sleep.