Tuning Novice & Auto Tune

[chrisgunton]

I recently has a turbo fitted to my 1995 Miata, 1.8, along with a megasquirt. It has a base tune, and is currently running off boost (wastegate open).

Can i just fire up autotune with default settings and let it do it's thing with default values? I have AFR targets set, just not tuned to.

[LT401Vette]

Opinions here will vary widely
1st you want to make sure it is working correctly on your setup,especially in these early version. Run it with "Update Controller" unchecked, make sure it isn't just taking the VE values to crazy places like 0 or maxed out. If it is there is likely something wrong, maybe with VE Analyze, maybe with your setup. A bad O2 sensor for example will of course leave VE Analyze clueless.
As a novice tuner you have a big risk of not identifying if something is not working correctly!

Otherwise it works very similar to VE Analyze in MLV, but is more adaptable to your setup and you don't need to keep switching between apps. For example this VE analyze is dual table and table switching aware. With table switching active you can have VE analyze running on both tables but it will only correct the 1 that the controller is using.

So ideally you can do just what you said if you have a ball parked starting map. I have taken base maps that could just barely run the engine, rev it around in neutral until letting it get the no load areas running good, then put it in gear and drive lightly until it gets those areas down and progressively drive it in a wider range.
You will likely want to do some hand tweaking along the way.

I have used this method a lot of times with MLV's ve analyze but you had to keep switching between the 2. In TS's VE analyze if you have Update controller checked, ever 10 seconds it will send any change to the controller.

But let's make sure the bugs are shaken out before we get too daring

[M30guy]

In the new VEA live, where is the AFR target map?

When going over VE cells with my pointer. It will say 14.0/14.7afr on ever cell.
I changed the afr target with ego corrections thinking that was it, but it still says 14.0/14.7 in the VE live map cells.

Map 1 and 3. No difference.

[LT401Vette]

It uses the regular Target AFR tables. If you are not using target AFR tables like in a Narrowband those may be disabled on you.

I probably better put a button or something to get to them even when the controller isn't actually using them.

Also the target afr values in the tooltips only get updated each time start VE Analyze. Then it will go back to the AFR table and update the values, it doesn't recognize changes to the AFR table while it is already analyzing. That is something on my todo list :0

[ausloki]

With the autotune function will you be adding the facility to utilise an external AFR map, Something we can tailor to suit our exisitng 16x16 tables or bigger ?

[LT401Vette]

With the autotune function will you be adding the facility to utilise an external AFR map, Something we can tailor to suit our exisitng 16x16 tables or bigger

If there is a reason or a demand I certainly could, so explain to me why you would want to use an AFR table different from what the controller is using? Is it to have a 1:1 ve to target afr cell count? I have always thought even 12x12 is more than needed on afr.

If you had different values in the custom table, then you would cause ecgoCorrection once you start running for real.

I could present a 16x16 interpolated view of the table, heck I could probably even make it so you could edit the AFR's in the 16x16 and adjust the 12x12 to match it.

[ausloki]

My main thought was to have a 1:1 ve to target afr reference. more for visualisation on the tuning side.
I know VE maps are commonly shared for the Mazda KL engine i have but i never see much in the afr target tables beeing shared.

To me it just seemed strange these days that we have 12x12 and 16x16 maps but still uses an 8x8 for our afr target tables.
I'm still a noob when it comes to the tuning aspect so feel free to enligten me

[LT401Vette]

MS1 Extra uses 8x8 and MS2 & 3 use 12x12 for afr. So it is or can be dissimilar in size to the VE table is, but there is reason. Tables quickly use memory in the controller, so unless there is reason to make a table bigger the firmware guys typically try to avoid doing so.
As the VE and AFR tables are interpolated, you really only need bins spaced to cover points of inflection, then the ve value will be interpolated to the next cell. VE tables are more complicated and can have reason for more points of inflection due to engine characteristics than afr tables.
Where an engine may have an interesting reaction to an airflow condition between 5000 and 6000 RPM, requiring you to make a non linear change needing a bin in there, the afr table is likely still going to target the same AFR. As the AFR is a desired value, the ve is maps out reaction to real conditions.
So typically I think the 12x12 is more than enough as usually most of the table is repetitive or somewhat linear anyway.

With that said I can see a reason for wanting to see a ve table matching size for the visual. That is actually probably pretty doable.

[M30guy]

As far as actually tuning VE to a target afr.

Does this program do a better job then MLV?
If so (or not) what's different about it?

I'd use it but I'm kinda like, with my little MS1 8x8 table.

I love how it remembers the o2 delays in the areas you want.
Would save a TON of time if it was in MLV.

[LT401Vette]

This will work very similar to MLV's VE Analyze. It was not just an integration of the MLV code, it was a full rewrite to change the few things I wanted changed in MLV. I do plan to put about half of these components back in MLV to bring MLV up to speed with TS.
The lambda delay table is obviously new. In MLV I have learned to do tables in a couple passes using different WB delay in different table regions. So now this takes care of this for you. Being in ms instead of records makes it adapt to different data rates.
Other differences:
- It supports Lambda as Lambda.
- Limits to Authority
- Supports Dual Table
- Supports Table Switching.
- It can update the table as you go, so as it comes up with the recommended improvements, it can calculate the future recommendation from the more ideal value.
- There are more constraints on the cell weighting.

In Dual table or table switching conditions you can start VE Analyze on both tables and it will only correct the table while it is active.

So this certainly is an improvement to MLV.

I still think 8X8 is fine. How much would you expect to change cells if you had more cells?

[tristanlee85]

I'm just extremely thrilled for the lambda delay table. I was dying for that in MLV. Great work. Hopefully I can play around with it this weekend.

[Black99rt]

Phil, for the defualts in the EGO delay table, can you let us know what distance the sensor is from the port on your set up, and which sensor controller you are using?

[LT401Vette]

The default table generates the X&Y axis based on your VE table min and max.
As you can imagine I have seen log files from a lot of cars . Most fall into these ranges. But for reference of what my test car is..
A 9.1 Liter (555 CID) with a 2 1/8" Primary pipe that is I believe about 30" long, the sensor id in the collector and I use an LC1.

i doubt that is a close reference, but most fall pretty well in the same range. Probably the biggest variance is along that bottom, coming off 0 TP or over run with light throttle, I have see some cars have a very slow reaction to the AFR.

[Black99rt]

Those are areas I don't mind at all cleaning up by hand anyway. Thanks for the info, sounds similar enough to leave alone. My primarys are a bit smaller than yours though, lol.