1st Gen Haltech ECU Discussion

I'm at 2.600 ms @ 14 volts at all loads/rpms and... I think I started to smoke them on that pull. There was a smell afterwards and I need to inspect them. I may need to dial it back to 2.400 ms. Granted, I had just done a similar pull a few moments earlier, so it was a back to back 4th gear balls out pull.

 

="{"&A161&", "&IFERROR((LEFT(D161,1)*8)+NUMBERVALUE(MID(D161,FIND(":",D161)+1,1)),LEFT(D161,1)*8)&", "&IFERROR(IF(FIND("-",D161)>0,16),IFERROR(IF(FIND(":",D161)>0,1),8))&", "&IF(LOWER(E161)="unsigned",0,1)&", 0, "&J161&", "&VLOOKUP(H161,'Base Unit Indicies'!$A$1:$B$45,2,FALSE)&", 0},"&" //"&F161


Haltech has not made their CAN data the easiest to turn into adaptable code lol... or the intern rather.
Regardless I should have 95% of all possible outputs compatible with my display driver soon, most of them entirely superfluous for us Taco people.

 

You can certainly tell this is the first time someone has posted it for general use, I have a feeling the internal whitepaper that companies have been using so far is even more arcane though.

I started fine tweaking boost control again, dunno why I picked cold weather to do that but regardless that's what I'm up to. MAP prediction has taken 99% of the little gripes I had and smoothed them out. No more little jerky starts after a shift, no more fluctuating AFR at low throttle input, no more intermittent stuttering on takeoffs in cold weather with a cold engine. It just works, really feels like a stock ECU tune level of polish now.

 

Little do they know......

I need to get back into mine and do more tuning now. Seems this has really helped refine the transient throttle tuning.

 

Yup. Definitely need to get back in there, after reading this post. Transient throttle has been such a hassle for me.

 

Absolutely, Haltech created a little bit of magic with this update. I'm not really sure how the learning occurs but I can see that values do adjust and I haven't had a single hiccup since rolling the version forward. Once you're on the newest firmware it is easy to disable the classic transient throttle and enable map predication in its stead from the Fuel Tuning tab. Other than changing your engine details to reflect if you're NA/Turbocharged/Supercharged that's all that needs to be done.

 

RE This... when you upgrade your current map using NSP it seems to generate very accurate MAP Prediction/Fuel Film maps of its own based off of your tune (minus transient). I wouldn't leave both transient on and map prediction, you're just going to feed the MAP prediction garbage information and with its scale turned down it would think it isn't affecting anything and continue to trim until it was useless when you finally scaled it properly. Also I'm fairly confident this feature couldn't smoothly start and stop if it was divorced from it's fuel film modeling. It would have a delay like opening the secondaries of a carb with too small of jetting if the fuel film was gone, but I don't think they intend you to mess with it much unless there's a specific issue you need to hunt down.

 

Damn! Is this code from Haltech?

Is this operating on the raw CAN message data? Looks like it might be messing with eight bytes of message frame data, but good lord, that’s tough to parse for a dumbass like me.

 

To keep the bus moving fast they're sending data that needs final modification from the receiving device to turn it into the units you wish to see. This appears to be his translation code for changing units of something

 

Awesome!

 

Haha no that's just my excel formula to automate each dataset into a usable array in C. It's very annoying: some data are single bits, most are words (double bytes), some are single bytes, and some are partial bytes (positioned bits).

Because of this, I'm having to save the starting position and length of each, within a particular CAN ID, then parse it. And since there are so many, I can't save a "last data" for all of them like I was doing before due to memory limits, so I'm reconstructing the whole program to revolve around a Display ID and assign (or allow the user to assign) a sensor ID to each display index. Furthermore, I decided to also move to a shift register to change the Chip selects so tentatively I'm shooting for a max display count of 32, rather than 8 (before I was limited to physical pins). Speed wise it'll be limited to how fast I can clock the shift register to update all displays.

 

Switching to a shift register was a good move, I actually figured you were limited on the speed/RAM of the micro controller not physical outputs.

 

The limit has always been memory, and I won't know if I can even stuff all the new data titles into SRAM until I finish restructuring... It's gonna be a challenge. But previously the thing was damn fast since you don't really have to add any delays between turning a pin on or off. That won't be the case now with the SR - I think it needs something like a 600us rise/fall time for clock and data, which immediately reduces a full display set refresh cycle to (1/(0.0012))/displays Hz (1/0.0024 for the cycle start), give or take. Kinda hoping I read the unit wrong

I actually implemented an artificial delay originally because eye couldn't track the update frequency on things like RPM

 

I guess for a smaller set of concurrent displays you can eliminate the shift register and run in essentially high performance mode eh?

 

Ya I'll most likely have that option eventually. Right now the plan is just to leave the Analog pins open for when I'm ready to code in adding new sensors to backfeed to the HT, but there shouldn't be any reason they cant be multipurpose

 

For anyone who's curious about map prediction here's the tech article now

https://support.haltech.com/portal/en/kb/articles/fuel-film-model-map-prediction

 

A bit bummed the Elite series won't be getting this feature but I don't really mind having a spare laptop in the truck anyway, I hardly ever have to change anything in my tune anymore really.

https://www.haltech.com/news-events/coming-soon-haltech-connect-app/

 

I may try this out this weekend.

 

I am ready to fire up a new motor (NA 3RZ) with a Haltech Elite 750. Going through the set-up and it all looks good except it is reading the air intake temp at 750f! Engine not running and ambi temp is about 65f. I assume it’s a sensor calibration or input type but I can’t seem to figure it out. Any thoughts? It’s a 2-wire AX1 sensor for what it’s worth.

 

Do you have the pull-up enabled?