Diy Tuning

[rollex]

We have 4 alpha testers at the moment using it in their workshops/at home doing bug testing for us. The whole "undefined parameters" section we will limit to workshops and on a per user basis most likely as we want to limit the risk we expose ourselves to if someone does blow something up and tries to sue.

 

Yes we can pull SCT/HPT tunes out. We are working on an auto detect for corrupt OSID as well, though a lot of people know how to do this manually already.

 

Once we get to beta we will be releasing to public. Though we still need to finish the website front page and payment system. We will be expanding the number of alpha testers slowly to not swamp ourselves with things that need fixing/improving.

 

I imagine realistically maybe 6 months away. Could be earlier.

 

On 26/03/2017 at 0:12 PM, MattyP said:

Are you saying that the existing software is somewhat lacking in its parameters? 

A lot of the parameters are unused/zerod but a large number are not. They may be completely uninteresting and they didn't include them for a reason. Though I suspect quite a few were just missing out of the original A2L files everyone used. I can't see any good reason why they wouldn't have included exhaust cam tables other than they simply didn't know the addresses.

 

Anyway this makes them all available, but it leaves it up to the user to figure out what the hell they do. If we or users do figure them out we will comment as we go. Lets the community chip in if they are looking for something specific and feel like having a play. Obviously you could bin your motor/car if you change something though so it is not without risk. Eg I'm sure there are items in there that could jam an injector open flooding a motor/bending a rod, or make the throttle go wide open during cruise and put your car into a wall. So lots of caution needs to be exercised.

[DarrylC]

15 minutes ago, MattyP said:

Yeah sorry for the confusion, I meant hpt, sct, etc are lacking. Looks to be coming along nicely  

Are you planning to eventually release it to the general population and is it going to be able to pull the existing tunes out?

The existing tunes are extracted like HPT and Sniper do. The software has the capability of determining the calibration even if the calibration has been "corrupted" but this may be limited to the Workshop version.

Edited 26/03/17 02:04 AM by DarrylC

[discostig]

If only there was some sort of controlled test car that you could connect and control with a real ECU to go nuts with and check what the unknown tables do (if anything), but it shuts down if it's about to blow up.  Like a motor/throttle/fuel system on a test bench with some sort of attached retarder system to produce load as it tries to rev,.  I'm sure it would be more complex than that though if you wanted meaningful parameters/results.  An accurately computer simulated car engine alone would be a whole project by itself that you'd need a team of software/physics/engineering experts to put together!  Plus you'd need to know what the outputs do in the first place so you are back to the start hahaha.

 

Probably easier if you are willing to risk blowing engines to set up a logging car on a permanent hub dyno and stockpile a bunch of cheap wrecker I6s so they can be easily swapped out if they grenade...could get expensive fast if you were guessing a lot though, depending how cheap you can find the donor engines, and of course then there's the labour component but that's not much different to what you already do volunteering your time.

[rollex]

Bang for buck we are better off analysing the ASM to get a good educated guess that way. If the axis are load and RPM that helps, then see if its a multipler/adder and log what subroutine its called from. Then you get an idea if its related to spark if its called in the vicinity of the other spark tables.

 

Then once you are confident it is spark you can simply fudge it and datalog to see what it does.

[DarrylC]

Much easier to read the asm and work it out. You get traction after reading a few million lines

 

We are identifying all the routines  we can and using them as a hint for where the parameter is used.

[HI PSI]

Great work guys, it is always an  interesting read, when revisiting this thread.  

 

I foresee that there will be limited data sharing with the new tables and their effects. As this will be the new intellectual property component, that tuners will utilise to the benefit of themselves and their clients. As these will determine  the manners and power capabilites of the engines. After all, this is what already what separates Tuners from Tuna's...

Edited 26/03/17 04:27 AM by HI PSI

[rollex]

Yes that will obviously be up to tuners if they want to keep it secret or not. Personally we will push for people to make them publicly accessible for all but if they want to keep it to themselves for a few months to get extra business then that is up to them. Ultimately if you are a good tuner people will be lining up regardless.

[IH8TOADS]

On 26/03/2017 at 9:38 AM, DarrylC said:

If you mean existing software like SCT, Sniper or HPT then I am not sure. I don't think thing SCT or Sniper community think it is lacking but the BA/BF community on HPT are not ecstatic. 

 

If you mean our software then not at all, people have reported back that the "other products" do not have them either. In fact we have things like the exhaust cam retard that no one else has. 

 

What is being demonstrated here is that there are a lot of parameters we have found and the software is helping getting them defined for the customers.  

 

Our aim is to make the product flexible enough to handle the "missing parameters" and automate the population of them across all the calibrations.

 

Not sure what other products have these features though

 

Does anyone know if SCT has the boost sensor mapped yet or if HPT have put it into the BA?

 

We don't have the boost sensor in SCT and the t-MAP is mapped for a limited number of calibrations. My code A3YB doesn't have either boost or t-MAP in either SCT or HP.

 

HPT is still missing a lot. The guys with HPT only will ALWAYS tell you otherwise, because they simply don't know what they are missing.  SCT is missing parameters too and as mentioned, it's bizarre some of the things they have missed. For example the max camshaft overlap, airflow correction for camshaft position, camshaft position modifier for boost error etc are in SCT but the desired camshaft overlap table isn't there to edit. Since 2006 when Ford introduced TiVCT no tuner I am aware of has had the ability to independently control the camshafts....it's a major flaw in the software.  Even access to the sensor calibrations is a reasonably recent addition.  This new software might prove to be some healthy competition...probably needed from an end user point of view. 

[galapogos01]

Open defs are a good thing for the DIY tuner!

[rollex]

5 hours ago, IH8TOADS said:

ince 2006 when Ford introduced TiVCT no tuner I am aware of has had the ability to independently control the camshafts....it's a major flaw in the software.  Even access to the sensor calibrations is a reasonably recent addition.  This new software might prove to be some healthy competition...probably needed from an end user point of view. 

 

It took Darryl a few weeks max to find it, then it took our algorithm a few seconds to map it to literally every other strategy that exists.

 

I suspect SCT don't do much reverse engineering and simply buy the maps out the back door from ford directly. The maps they buy most likely don't have these tables. I know Eric from HPT does actual reverse engineering though I suspect he does it all manually by hand. This explains how they got the TMAP stuff, though it also explains why when they find something new it doesn't get added to everything in one go.

 

Doing it from first principles and tracing the ASM is very time consuming. I can understand from a commercial aspect why certain companies would spend money on creating relationships with OEMs to just buy the data, if you can find the right person it is going to be cheaper though it means you haven't developed the capability to find specific customer requests.

Edited 31/03/17 05:36 AM by rollex