DIY tuning

[Sbrach]

I'm just thinking out loud in response to the countless comments in every N55 tuning thread from JB4 users who say that they could make more boost.. I have no desire to do anything with JB4 and would never take that route over flash only tuning personally. I just thought it was interesting that so many people say that when coming from JB4 and was curious what the cause is. JB4 hijacks the boost solenoid. If what the DME calls 100% WGDC isn't actually 100% then that would explain why the JB4 can achieve higher boost when both are commanding 100%. Would really need to scope the signal at 100% DC and see if it is actually steady state. Its not really relevant since, as you and I both said, a couple extra psi when outside the turbo's efficiency range doesn't really net you anything.

 

[LamboLover]

I'm just thinking out loud in response to the countless comments in every N55 tuning thread from JB4 users who say that they could make more boost.. I have no desire to do anything with JB4 and would never take that route over flash only tuning personally. I just thought it was interesting that so many people say that when coming from JB4 and was curious what the cause is. JB4 hijacks the boost solenoid. If what the DME calls 100% WGDC isn't actually 100% then that would explain why the JB4 can achieve higher boost when both are commanding 100%. Would really need to scope the signal at 100% DC and see if it is actually steady state. Its not really relevant since, as you and I both said, a couple extra psi when outside the turbo's efficiency range doesn't really net you anything.

I went here with VR38DETT in 2009 when the revealed tables did not allow 100% WGDC with good control until the code was altered to tightly control it. Because that platform had pressure operated actuators without tight springs and fairly large wastegate flappers, it did benefit from 100% WGDC at the top in terms of increased airflow and power.

On N55 I have taken a different development priority as I believe these basics are settled simply by using the factory tables and that given the modest capability of the N55 turbo it would be all too easy to overspeed the compressor, raise EGT and make a pressure cooker, unlike VR38.

If there is any real doubt about the logged wastegate duty cycle being correct, I could investigate (from disassembling the link between the logged variable and the final PWM out to hardware) but it would divert from other work that presently I think would be more productive.

 

[Sbrach]

I don't think its worth wasting your time on. I can put a scope on it if I get really curious. Thanks

 

[Sbrach]

Latest log. Started tune over with PPK base bin. Flashed 135is DCT software. Bumped up ethanol blend from 30% to 45%. Making my way up to straight E51-54 out of the pump. Flashed EKP to help the LPFP and increased Commanded HPFP at WOT. VRSF 7" HD on the way. Need to order a walbro to drop in and hopefully should have an ignition upgrade available for N55 soon because I think that is my next road block.

https://datazap.me/u/sbrach/rev39-3rd-wot?log=0&data=4-14&solo=4

 

[LamboLover]

Do you think increasing high fuel pressure is likely to help or hinder flow? I have not seen a flow vs pressure curve for these pumps. I did not think the VCV was a restrictor but a recirculator, analogous to a traditional fuel pressure regulator.

 

[Sbrach]

I don’t know. Annectdotally many people have said it helps catch the drop in HPFP in the midrange at high load if you increase HPFP in the low rpms. If we knew injection time we could probably deduce more but I believe I am flowing the same amount of fuel at higher PSI just with shorter injection time. On N54 you can run much lower pressure without misfiring due to the piezo injectors so I think most people have run out of volume before they run out of pressure on that platform. Increasing pressure doesn’t help because they are already starving the pump. On N55 I think the inverse is true.

 

[Sbrach]

Really I don’t like how they control HPFP but I guess it’s only an issue from a tuning point of view. They have a WOT rail pressure 1D table that must have an offset built into it for pressure drop expected at WOT at each RPM breakpoint to achieve the desired rail pressure. When we throw more boost than stock at it injection time is greater causing a larger pressure drop than the stock table values account for and pressure drops below 2000psi even though the pump isn’t maxed out yet.

 

[LamboLover]

Does the target rail pressure not follow the table then? I looked up the solenoid injector on N55 and think it was about half that of the N54 Piezo. I haven't investigated the available injection window yet.

 

[Sbrach]

Edited to add more data.

Stock table and stock load it is about 5 bar below at peak load according to a couple of old logs I pulled up.

Stock table with increased load it is almost 60 bar below at peak load.
https://datazap.me/u/sbrach/rev26-e30-3rd-wot?log=0&data=14-15-34&solo=34

With increased target on my latest bin it dips to 13 bar below target at peak load but recovers fairly quickly. I may be able to increase demand a bit more just prior to the dip and smooth it out a bit more.

https://datazap.me/u/sbrach/rev39-e51-3rd-wot?log=0&data=13-14-28&mark=33

Stock

Increased target

 

[LamboLover]

Thanks for the logs. That is rail actual. If rail target isn't in there or available I can find it and get it added.

 

[Sbrach]

Rail target is not one of the available loggin channels in MHD. I’ve been going by the KL_PRAIL_H_VL table values. I suppose there could be some modifiers after that table that alter the final target. Additionally it would be very beneficial to be able to add a LPFP sensor and log it natively.

 

[LamboLover]

I'll look into it. I was assuming since you are pushing the HPFP that differences between target-actual are due to HPFP limits. But as you say many weird modifiers occur in these ECUs, so if you are able to see target fuel pressure tracking the lookup table that helps. Obviously you don't want to just log the RAM variable that has just been spit out from the table, I'll look for something better.

Blue sky idea for extra sensors/actuators is to put them on CAN as too little remaining IO or hassle adding pins to ECU plugs that requires inlet manifold removal.

 

[LamboLover]

There is a designated PID for "Raildruck Kraftstoffsystem Sollwert" = Rail fuel system setpoint, so have sent details to MHD to ask for it to be added to the next update, unless of course it has already been tested and doesn't work.

 

[Sbrach]

Awesome. Thanks.

Is injector PW something that could possibly be logged as well?

 

[LamboLover]

Yes, but it is not a standard PID, so would have to find a RAM variable to log. Or several and the available windows. Will likely get to it at some point as would be very useful.

 

[Sbrach]

Would someone mind explaining the behavior of these cold start tables? Specifically the distinction between (VVT) and (Throttled).

I assume the duplicate tables are unintentional?

Lastly, is it typically necessary to alter the cold timing table for E50? Lowest ambient temps I should see are 0C.

Thanks

 

[LamboLover]

You might be one of the first to try them, so don't know how much you will need extra if at all at 0C. The duplication is unintentional, the addresses are the same so I just deleted them from my xdf.

The VVT tables are when the ECU is using valvetronic, throttled when using the throttle. Normally valvetronic is used, but I gave both sets of addresses to Martial in case valvetronic does not work in very cold temperatures as I recall that older versions of VANOS did not. IIRC, valvetronic is an electric motor rather than using oil pressure like VANOS, so probably not a real concern.

 

[Sbrach]

Thank you. Typical morning temps here are about 0-5C. I've upped ethanol content to ~50% and the worst I get is a slightly longer crank and a couple stumbles when i creep out of the garage to warm up in the driveway. Not a big deal but wanted to experiment with the tables to see if I can improve it and its obviously easier to experiment if you understand what the variables are. Thanks again for your help.

 

[Sbrach]

So I added about 20% to the table below 30C and cold start was great this morning. It wasn’t very cold though. 47F

@Twisted Tuning if I edit the cold start table but use the Disable Cold Start flash time option is the table overwritten or does it only affect Vanos and ignition timing?

Thanks

 

[Sbrach]

Log with cold start table changes in attached screen shot. Unfortunately coolant temp was 19C so not a great test. Hopefully we get a few more cold mornings here in AZ. Still not sure these tables aren't overwritten by the Disable Cold Start flash time option. Will try without it checked tomorrow. Fuel is E54. Also, wondering if there is a DCT version of the start fueling tables. Most tables that have a trans variant have all three, I changed both for now.

https://datazap.me/u/sbrach/rev43-cold-start?log=0&data=6-9-13-14-15-16-17-20-21-34&solo=13