Megasquirt Tuning

[rx7locost]

MV8, I just saw your pics for the tables. I created a similar blip to the existing AFR on my 13B build. In my case, I started about 500 and ended by 1200 to 1500 RPM as I recall. I did this intentionally to resolve an off-idle stumble. It worked on my car and one other car that I know of. I have also suggested it to a Corvette owner who is trying to solve the same off-idle stumble. Unfortunately, his car is put up for the Winter. It worked similar to MAP Acceleration Enrichment settings but only worked around idle RPM's. I'm not saying that this should be done, but there are times when such an abnormality in the AFE table is justified. Beyond that, I agree that the tables should typically have nice and smooth transitions.

[Tundra 7]

MV8, I just saw your pics for the tables. I created a similar blip to the existing AFR on my 13B build. In my case, I started about 500 and ended by 1200 to 1500 RPM as I recall. I did this intentionally to resolve an off-idle stumble. It worked on my car and one other car that I know of. I have also suggested it to a Corvette owner who is trying to solve the same off-idle stumble. Unfortunately, his car is put up for the Winter. It worked similar to MAP Acceleration Enrichment settings but only worked around idle RPM's. I'm not saying that this should be done, but there are times when such an abnormality in the AFE table is justified. Beyond that, I agree that the tables should typically have nice and smooth transitions.

Chuck I had just a little bit of a idle hunting and that map truck fixed it perfect. It’s in the TS guide I believe to do that.

[Tundra 7]

Posting this so Chuck can see what you have and you can see what I am talking about. The general idea. Not an expert on this.

If you have all the timing around idle to comp for oscillations, try taking out the timing and tweaking the pid gains down until it stops, find where it just starts to oscillate and bring it back a scosh.

Didn't look at everything.

MV8, I was kinda curious, how did you come up with the maps? I am going to run them but just wondering.

[benny_toe]

Ouch! To both the 6 week delay in 1st payday and what Scott wanted. Makes me wonder what changed. He never once mentioned money. He sent me at least 5 MSQs narrowing in on a decent tune. He moved from just north of Seattle to North Dakota and perhaps saw this as extra income.

What I have now isn't perfect but I am running acceptably for street tune. Be glad to share (no charge ;<} ), Won't be an exact match but perhaps something to play with. LMK

[Miatav8,MstrASE,A&P,F]

Posting this so Chuck can see what you have and you can see what I am talking about. The general idea. Not an expert on this.

If you have all the timing around idle to comp for oscillations, try taking out the timing and tweaking the pid gains down until it stops, find where it just starts to oscillate and bring it back a scosh.

Didn't look at everything.

MV8, I was kinda curious, how did you come up with the maps? I am going to run them but just wondering.

I suggest waiting for me to post the zipped msq. Knowing carb engine tuning helped, I consulted my paper library tuning and miata specific references, there are some decent ms tuning vids on youtube, I compared turbo tunes for the non-boosted areas, etc. So many different ways to do anything with this system. It is clear to me bentwrenches method, while totally different from the way I am approaching this, makes a lot of sense, but I'm working on incorporating everything I have absorbed right now and should be able to zip it and post tonight or in the morning.

[Miatav8,MstrASE,A&P,F]

The wot timing and afr are based on oem afr and Norm Garrett's chart of best timing for a stock, late 90's miata 1.8L BP determined with dyno testing. I leaned the WOT a bit from stock. If you have a cat you may want to richen it up with the stock specs. Both maps are smoothed as much as possible but they are lumpy from Garrett.
The extra lines around 1000rpm idle are for control but if you have a better way, use the extra line elsewhere.

If it helps, great. If it is total crap, I want to know that more. A shot in the dark at a base to autotune.
Good luck!

[Tundra 7]

The wot timing and afr are based on oem afr and Norm Garrett's chart of best timing for a stock, late 90's miata 1.8L BP determined with dyno testing. I leaned the WOT a bit from stock. If you have a cat you may want to richen it up with the stock specs. Both maps are smoothed as much as possible but they are lumpy from Garrett.
The extra lines around 1000rpm idle are for control but if you have a better way, use the extra line elsewhere.

If it helps, great. If it is total crap, I want to know that more. A shot in the dark at a base to autotune.
Good luck!

No kidding!, from a miata guru!. For the hell of it I had the AFR tool run a generator make a chart and it runs out to 12.6 on the top end. However I did run into a couple of guys that ran dynos and said the miata runs pretty consistent at 13.1 WOT. Well cool, Thank You!. I'll fire it tomorrow and set the auto-tune for a new VE. I had to charge the battery tonight when I got home.

[Miatav8,MstrASE,A&P,F]

FYI, no need to enter values by hand.
The zip file (MV8 NA BP SD BASE) is at the top of the pic. It unzips to an msq file.

[Tundra 7]

FYI, no need to enter values by hand.
The zip file (MV8 NA BP SD BASE) is at the top of the pic. It unzips to an msq file.

I thank you for that. I redid my entire idle sequence so no driving today just garage time. So far from what I can tell is that off idle its a lot snappier. 48 degrees tomorrow so after I drive it I may be able to to some accel enrichment too. It sounds different when I blip the throttle. Sounds raspier.

[Miatav8,MstrASE,A&P,F]

Thanks for the update.

In an open car where you may not be able to hear detonation, an oem knock sensor input to the ms is a great tool that can be used to dial it in as a gauge/red light and provide correction if you want. You have to set the threshold above normal running mechanical noises. An idle test should show the normal mechanical noise level. Your ms should accept the analog input directly. A gm unit is piper thread. You can lube the threads with antisieze but don't use teflon tape or anything to insulate the sensor from the block.

If you don’t have a knock sensor, I’d listen for det rattle with a dash display of very large kpa, tach, and wb02 while also logging. After warm idle check, perform casual normal driving kpa from 1.5k to 4k, followed by wot kpa to 4k. At any point in testing, if/as soon as you encounter any rattle, quickly note kpa, tach, 02 then let off.
Select the adjacent kpa (lighter load/numerically lower) at that rpm and hi lite the row out to max rpm.
Use the minus button to remove ½ degree from the row so each cell is changed ½ individually instead of smoothing/averaging the row.

Next, hi lite from each of those cells to 100kpa and smooth with the red crayon button. Select two columns at a time to smooth and do this out to max rpm.

Retest to duplicate rattle. If no rattles, perform the same tests to 4k but now use the next higher gear (3rd when you should be in 2nd) to pull from 1.5k, duplicating a steep hill.

[Tundra 7]

Thanks for the update.

In an open car where you may not be able to hear detonation, an oem knock sensor input to the ms is a great tool that can be used to dial it in as a gauge/red light and provide correction if you want. You have to set the threshold above normal running mechanical noises. An idle test should show the normal mechanical noise level. Your ms should accept the analog input directly. A gm unit is piper thread. You can lube the threads with antisieze but don't use teflon tape or anything to insulate the sensor from the block.

If you don’t have a knock sensor, I’d listen for det rattle with a dash display of very large kpa, tach, and wb02 while also logging. After warm idle check, perform casual normal driving kpa from 1.5k to 4k, followed by wot kpa to 4k. At any point in testing, if/as soon as you encounter any rattle, quickly note kpa, tach, 02 then let off.
Select the adjacent kpa (lighter load/numerically lower) at that rpm and hi lite the row out to max rpm.
Use the minus button to remove ½ degree from the row so each cell is changed ½ individually instead of smoothing/averaging the row.

Next, hi lite from each of those cells to 100kpa and smooth with the red crayon button. Select two columns at a time to smooth and do this out to max rpm.

Retest to duplicate rattle. If no rattles, perform the same tests to 4k but now use the next higher gear (3rd when you should be in 2nd) to pull from 1.5k, duplicating a steep hill.

I looked up the way Microsquirt see knock. You have to buy a GM spark module and run the knock to it then us the SPAREADC wire so the microsquirt can see it. Interesting there isn’t a direct hook up. I’ve never looked into this before. Thanks for the instructions. Pretty thorough write up here. Would be nice to have a knock sensor with a usb on it and send it right to the computer. I doubt something that easy is available. I’ll drive it first see how it is.

[Bent Wrench]

Auto-tune works really nice. You need time to make a lot of hits in the cells you want to adjust. So the cells at 80kpa and 5-7k rpm can be problematic. That’s where the dyno comes in I guess.

Yes the dyno will get you into the"illegal" areas.

Additionally when using auto tune hold the power in each cell as long as needed to give Auto time to make corrections. That can speed up the process. I use a long hill and different speeds in different gears.

[Tundra 7]

Auto-tune works really nice. You need time to make a lot of hits in the cells you want to adjust. So the cells at 80kpa and 5-7k rpm can be problematic. That’s where the dyno comes in I guess.

Yes the dyno will get you into the"illegal" areas.

Additionally when using auto tune hold the power in each cell as long as needed to give Auto time to make corrections. That can speed up the process. I use a long hill and different speeds in different gears.

Today Ended up drizzly I wish I had a hill like that. Ours are short. My new Idle tune is nice. We’ll get the rest. Now it’s just finding pockets in the weather to do it.

[Miatav8,MstrASE,A&P,F]

I looked up the way Microsquirt see knock. You have to buy a GM spark module and run the knock to it then us the SPAREADC wire so the microsquirt can see it. Interesting there isn’t a direct hook up. I’ve never looked into this before. Thanks for the instructions. Pretty thorough write up here. Would be nice to have a knock sensor with a usb on it and send it right to the computer. I doubt something that easy is available. I’ll drive it first see how it is.

I didn't know which ms you have. I think some will take the knock sensor directly. The module is called the electronic spark control or ESC but it is basically an amplifier to boost the strength of the output. Some earlier ESC also had a vac switch with emission carb applications.

You can make your own "ESC" with an arduino and a resistor or buy one as a kit for about $60 that is standalone and includes an led but will also work with ms. The link also shows what you would setup in ms or with the standalone, they set it up for you based on cylinder size (speed across the cylinder to the plug is a factor).

https://www.arduino.cc/en/tutorial/knock

I will look for the standalone site.

[Miatav8,MstrASE,A&P,F]

They are not as cheap as I thought. I'd find someone parting an old chevy and get the connectors, esc and sensor at the same time.

The expensive (piece meal) way:
esc (used)
https://www.ebay.com/itm/Electronic-Spa ... SwEVlb6gXC

connector (2 for 1)
https://www.ebay.com/itm/Knock-Sensor-H ... SwY-NdxA9n

esc connector (good deal)
https://www.ebay.com/itm/Electronic-Spa ... SwvNVcG9Rg

sensor
https://www.ebay.com/itm/New-Knock-Deto ... SwV3hb7F86