Jump to content

Recommended Posts

Hi,

 Took the car out for a couple off hours and a tank of fuel. Need to fill up before going for a drive with local TSSC Sunday.

Idling at 700-900, before was 900- 1k1.  AFR 13.5 - 14.7. Just reduced ignition advance idle control from max 10d advance to 5d. Lost more idle speed. I performed this, as getting a classic pinking sound when i dip the TPS at sub 1k revs.

My map is only 500 RPM increments for RPM cells; Going to change to 200 RPM steps, as I think it will speed up the conclusion of adaptive mapping?

I am still unconvinced that ignition timing can to some degree be related to a TPS. I have many road tests to prove otherwise. Convince me please.

I am thinking about setting all the timing points to a TR5 centrifugal degrees, and using to MAP vs RPM to set additional degrees? Fitting a BARO sensor and utilise the BARO vs RPM to allow for me to drive +/- KM's in height and still have correct timing and fuel? I have at least 4 MAP sensors to hand. Is there a way to tap into the balance (CP) pipe gaps? So I could fit two MAPs? 1/2" to 1/4" rubber tee's?

Cheers,

Iain.

PS. A lot of low TPS, 1K5 to 3K RPM adjustments registered, all minus around 2 to 11% of fuelling requested.

 

 

 

 

 

Edited by spitfire6

Share this post


Link to post
Share on other sites

Hi,

 21 L and 110 miles: Nearly 24 MPG.

The AEM is certainly paying for its self as I can see the difference on the digital fuel gauge; normally miles done plus fuel left equals around 100. Now looking like 110+.

Will see what it's like on the 40 round trip to work on mostly 60 MPH over the next few days, or maybe a whole week?

Cheers,

Iain.

PS. Still looking for a means of measuring manifold vacuum ie load.

 

 

Share this post


Link to post
Share on other sites

Reducing your rpm slots to 200rpm is a good move especially at say sub 3000rpm, but that will slow down your adaptive table learning unless you increase the allowed load site threshold. I suggest you set the load and rpm allowable threshold to 127 which will allow learning whilst you're anywhere inside a given load cell and give you a fair but coarsely adapted map, but at least will get you lots of cells populated in the adaptive table

Also allow a wide afr tolerance, say 0.5afr.

 

The way the emerald learns it's values is the ecu has to trim the mixture and then only when the mixture is within the allowed afr range will a value be learned. This means that you need to spend lots of time at any one rpm/throttle site for it to learn anything useful  even with the allowable limits set very wide.

 

Once you've got a fuel map that's trimmed up over as many sites as possible you can then reduce the afr allowed range to say 0.1afr, and the loadcell threshold to say 32 and then the ecu will refine your coarse map even further.

Share this post


Link to post
Share on other sites

As regards your wish to run MAP sensing for load .. that's fine, but TPS is a far faster way to map the ecu esepcially if you want to visit a dyno and pay for mapping hours.

As mentioned previously, every time you revisit a particular throttle position and rpm site the airflow is the same as the last time you visited it.

Smilarly every time you revisit a particular manifold vacuum and rpm site the airflow is the same as last time you visited it.

If it wasn't you'd never be able to map the fuel or ignition tables at all, the values required in each cell would be different if the "load" varied every time you revisited a particular cell. In actual fact the "load" is exactly the same each time you revisit any particular cell in the map.

 

Are you UK based? 

Share this post


Link to post
Share on other sites

Hi Obe,

 Here are some screen shots of some settings I am presently running:

 

 

Over 6K RPM, I increase step size to 500 RPM. 

"Similarly every time you revisit a particular manifold vacuum and rpm site the airflow is the same as last time you visited it." But, I have TPS as load site..

Thanks Obe, I am in NW UK, AKA GB, but wet.

Cheers,

Iain.

 

 

Screen Shot 04-23-18 at 07.33 PM.PNG

Screen Shot 04-23-18 at 07.34 PM.PNG

Screen Shot 04-23-18 at 07.35 PM 001.PNG

Screen Shot 04-23-18 at 07.35 PM.PNG

Screen Shot 04-23-18 at 07.36 PM.PNG

Screen Shot 04-23-18 at 07.37 PM.PNG

Share this post


Link to post
Share on other sites

5ade2c00457c6_ScreenShot04-23-18at07_49PM.PNG.0d27e0dbf7e46ef6ab96f50b86fe757c.PNG

Here is what the speed sites are above 5k6.

5ade2c7cece7e_ScreenShot04-23-18at07_56PM.PNG.4dfe4979e9b7b12e8a9e20116f7cc6d5.PNG

So, for speed and load increase 4 fold to 127, as a first step, and increase target to 1.0.

The K/Gain is a guess by the way. Any ideas on a better value from 0.25 & lock-on time of 250ms?

 

Cheers,

Iain.

Edited by spitfire6

Share this post


Link to post
Share on other sites

Hi,

 Did not think I could do this; here is my running map. Emerald K3 ECU.

Cheers,

Iain.

22_April_2018.map

The above map is for anyone to view or edit. LOL.

You will need to install:

Click to got to download site.. Safe, honest (When first uploaded)

Emerald K3/K6 ECU, sw:  K3 v1 r0.94

If you haven't already got the SW.

Edited by spitfire6
changes

Share this post


Link to post
Share on other sites

I notice you've got your throttle percentages quite large per load site, presumably you have small butterflies? If you find you're driving around alot on cruise between throttle 0 and 1 you can usefully reduce the initial percentages to say 2 percent between load sites for the first 3 or 4 throttle sites and gain a lot of extra resolution (if your throttle butterflies need it).

You have the rpm sites at 200rpm initially, that's good for refining the map over small rpm changes, but your speed integer of 32 means that the adaption only occurs whilst the crosshair cursor is in the middle of a cell. This means you probably only have something like a 50rpm spread over which adaption is being learned, and the rpm scatter is likely to take the rpm outside those learning limits so set the speed_int to 127 and that will allow learning over the full width of each cell. You can turning on adaption for idle too rather than leaving it closed loop / not learning. Set the load_int to say 80 which allows learning though most of the throttle travel in any given cell, but not at the edges of the cell (I suggest this because f your 5 percent load site increments, that's quite a wide spread, but I'm used to large butterflies with a high rate of change of airflow per degree).

Set the target afr to 0.5 to begin with, that means that if the afr reached during fuel trimming is 0.5afr rich OR weak of the target then an adaption number will be stored. This will give a coarse learned fuel map .. you can reduce that later as you refine your map based on the adaptions learned. I send cars out with the AFR target at 0.1 afr once the fuel map is nicely refined.

I also set the after start delay to 60 seconds on the basis I dont want any adaptions occuring until I know the engine has stabilised after a startup.  Hot fuel being purged from the engine bay and other heat soak and after-start fuelling stuff can influence the learned values so you want adaption to only take place when you know the engine bay and fuel system has had cool fuel and air through it after a hot restart.

Similarly if your normal engine run temp is something like 80deg you can stop adaption until say 75deg, that way you don't have the influence of any warmup trimming taking place. You still have 7 percent of injection correction in your coolant table at 60 degrees, so best plan is to wait until youre close to thermostat temp before allowing adaption. Any fuelling changes that need to be made to the warm up tables then only occur in open loop operation and can't influence the adaptive table.

 

There's a lot to look at, including the lambda signal smoothing and position of the lambda which influence what gain you might choose to use. If you're ever on your way south near sheffield let me know and I can talk you through some thing face to face which is always easier than trying to explain stuff on the net!

Happy tuning though .. the emerald is quite intuitive really once you get to grips with it.

 

Share this post


Link to post
Share on other sites

I think he's using modified PI throttle bodies, which are flippin' huge (46mm IIRC and too big really)......... so most of the interesting stuff happens over the first few degrees.

Good information there - thanks!  Clearly the Emerald is a bit smarter than MS....

Nick

Share this post


Link to post
Share on other sites

Hi All,

 Very good information presented.

Yes, I have PI ITB's.

I will LOG the TPS sites with more emphasis on part throttle and decrease the resolution up to WOT. 

The car is running so much better & runs fine in 4th with OD switching in & out at 30 speed limits. I would have been using 3rd before with OD.

I have adjusted the speed & load integer from 32 to 127 and will burn the adjustments in when arrive back at work in the morning. I have limited max change to +/- 10%, but will increase the + to 15%, as saw a few sites at +10% required.

Under snap acceleration, the AFR drops to > 15 before going to 12 ish; Need to revisit acceleration enrichment table i.e. "Accelerate pump".

 

Thanks everybody for you help. Especially Obe, for his emerald knowledge and recommending the AEM Lambda that will soon pay for its self.. 

Thank You all for putting the Triumph smile back on my face.

Cheers,

Iain.

 

 

 

Screen Shot 04-24-18 at 10.19 PM.PNG

Edited by spitfire6
Added new Log TPS.

Share this post


Link to post
Share on other sites

46mm if that's what they are is good for about 50hp/cylinder so pretty big! The good news about rescaling your initial throttle percentages is that the ecu makes a calculation based on interpolation so when you rescale the throttle sites you won't completely mess up the mapping you've already done, it will work out the new numbers to put in the cells for you.. it won't be a perfect correlation against the new percentage sites because the interpolation is linear and the change of airflow is essentially a log (or inverse log?) function,  but it will be somewhere close and means you're not starting from scratch again :)  the fact the interpolation is linear is also a very good reason to stack your off-idle cells close together for best relationship of fuel v airflow.

Share this post


Link to post
Share on other sites

image.png.879525cffa92a38402e9a94fc2399e0a.pngHi, Reset the TPS settings from linear to, in my eye, log.

I think it's curved enough.

Filled up today and getting 24MPG. My speedo reads low, so checked my odometer (19.5 miles), against GPS and I have a difference of 1.15. So, MPG is 27.6. Thanks math!

Next, to be able measure engine load, ie; MAP, so i can set spark timing via load and not TPS. A couple of hypodermic needles in the two balance hoses seems a plan? Proximity of fuel rail prevents use of original PI tapping's. 

Getting brief lean on hard throttle acceleration so will re-visit accel settings.

As confidence in the system is growing exponentially, I will look at cold start running and revisit the fitting of a idle control motor.

Thanks guys.

MOT for real on Monday!

Cheers,

Iain.

 

Share this post


Link to post
Share on other sites

Hi,

First off ; Hail Craig, for getting the forum back.

Obe, I program Siemens PLC's & DCS's for a living, but mostly binary & HMI. The ability to change jets, ignition timing, accelerator pump size and response with a keyboard is fun. Getting all the parts to work together will be another long experience.

Time to read up on the Emerald data logging abilities; TPS vs Lambda on acceleration for a start. Way too much extra fuel on accel below 2500; engine nearly stalling on rapid foot pressing. LOL. 

For spark timing; I think a MAP or two is required to follow a TR6 vacuum capsule.

The AEM is the best piece of kit I have bought this year! Even better than the thermal image camera I bought! 

 

Now to get the drop links ready for MOT on Monday. & the horn fitted the correct way up so it doesn't fill up with water and die!

Cheers,

Iain.

 

Screen Shot 04-27-18 at 06.05 PM.PNG

Screen Shot 04-27-18 at 06.06 PM.PNG

Screen Shot 04-27-18 at 06.07 PM.PNG

IMG_0827.JPG

Share this post


Link to post
Share on other sites

Hi,

 "fuel acceleration" settings has been much fun. Didn't realise that excessive settings and excessive right foot was equal to low MPG; refill shows 17 MPG. LOL & LOL as it was worth it.

Going back to past settings, that showed a lean on pumping right foot, from circa 35% previously to 75%, and now change to 50%.

Happy days.

Cheers,

Iain.

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×