[Crumpler]

Because I like to make things difficult I spend a few days working on cranking fuel settings only to find out the battery was weak. Tested normal here and at store initially, then I test it after first crank and the cca was 50% less.
Then I spent several days on VE mapping only to find a vacc leak and essentially started over.
It’s coming together now.
Peep’s tune was pretty close to mine on low end. Haven’t explored higher loads yet, winter storm and such.
Knock sensor wired in to ecu.

[Crumpler]

Knock control a bit slippery.
A lot of moving parts.

Initially you are calibrating ambient noise, under a 2000-2500 rpm load.
Choosing knock window.
Changing gain on each cylinder, based on distance since one knock sensor in this set up.
Changing band pass frequency based on piston bore.
Changing time constant and noise range to get knock value average to equal reference average.

This is to calibrate for knock sampling, next step is to differentiate noise from knock.

[Crumpler]

Snow is cleared, short cruise today.
Not to shabby.
General tendency toward lean.
There is a couple a ways to modify VE in VEMS, the automated function exists but VEMS tune will crash every time. Greg I don’t know if you have had this experience?
What I have found easiest is to log drive and then use the analyzer on data. VEMS will come back with optimal number for each cell.

[worf]

what are your A/F targets?

[Hey_Allen]

If that's actually correcting by 50% on average, it sounds like there is something seriously off in the baseline air fuel tables...

[Crumpler]

what are your A/F targets?

Yeah, let's discuss that if we have a moment.
Curious what other SC people have done here.
It's a low boost situation (5-7 psi), and I am prepared to push a little into max power over max safety territory.
With VEMS reading (hopefully) valid knock events as some insurance.
I've read a couple of "rich ain't always safer" articles too.
My initial plan was 11.8-12.1 at max boost.
Subject to change, estimating 6psi is 150ish MAP sensor load (kPa) would put it closer to 12.5.

If that's actually correcting by 50% on average, it sounds like there is something seriously off in the baseline air fuel tables...

It's a volumetric efficiency model so a lot of variables in the algorithm that I am only pretending to be proficient at. But 50% off is being too rough on Peep. I have found cells somewhat close, but also some 30% off so far.
Also, the SC is pushing a lot of air even with lower loads, so I am assuming that why if need to put more fuel in?

[worf]

what are your A/F targets?

Yeah, let's discuss that if we have a moment.
Curious what other SC people have done here.
It's a low boost situation (5-7 psi), and I am prepared to push a little into max power over max safety territory.
With VEMS reading (hopefully) valid knock events as some insurance.
I've read a couple of "rich ain't always safer" articles too.
My initial plan was 11.8-12.1 at max boost.
Subject to change, estimating 6psi is 150ish MAP sensor load (kPa) would put it closer to 12.5.

When doing the chips for D.R.'s twin screw my targets were 14.7 for unboosted operation and 12.6 for boosted. In practice, the range I was comfortable with was 14.5 to 14.9 for low load unboosted, 14.2 to 14.7 for high-load unboosted and 11.6 to 12.6 for boosted. So, stoichiometric for manifold pressure of 0 or lower and 12.6 for pressure above 0.

This was however a "street" application where unboosted represents 99+% of operation.

Max power is 12.6. But, the power curve is fairly "flat" there. However, combustion temp goes up from there and if you can't hit 12.6 you'd rather be rich than lean in my book for decreasing knock sensitivity.

At 12.6 you're pouring fuel out the tail pipe for purposes of faster combustion, cooler temp, and decreased knock sensitivity. Richer than that doesn't do any good AFAIK. It's just that you'd rather be richer than 12.6 rather than leaner if you can't kit 12.6 on the nose.

Right or wrong and subject to debate and more learning, that's how I did it 12 years ago.

And of course there's timing changes which will change a good fuel map into one that needs revision.

[Crumpler]

Thank you, that helps.

Good to hear about the low load numbers being 14's because that's where VEMS is putting it (lambda 0.97 to 1.01).

I can at least see the path forward now.
Things were grim mid-January Dave.

[grepin]

Hi Dave. My auto tune function is reasonably stable. I make sure I burn updates regularly. It can fall over a couple of times a drive but I thought it was the computer more as the issue. I haven't done a lot of tuning by statistics. I have had a play with it and like it. I just haven't really worked out how to implement it properly. Those changes in the 50% range don't show a 50% total change it could have been at 55% prior to the change. The step before shows the actual change value. Think you know that Dave but someone else mentioned it.

[Crumpler]

Hi Dave. My auto tune function is reasonably stable. I make sure I burn updates regularly. It can fall over a couple of times a drive but I thought it was the computer more as the issue. I haven't done a lot of tuning by statistics. I have had a play with it and like it. I just haven't really worked out how to implement it properly. Those changes in the 50% range don't show a 50% total change it could have been at 55% prior to the change. The step before shows the actual change value. Think you know that Dave but someone else mentioned it.

Ok, thank you for checking in.
Hoping life is good down under.
Appreciate the feedback. Yes, correct on what those numbers represent :)
Yes, the stats also tell you percent difference of original number to new optimal number, in a separate pulldown. Which is interesting to me. I'm still assuming my differences are from the SC effect even at low loads but who knows.

[hans14914]

I was thinking about your post this morning for some reason. What kind of fuel are you using, and do you have an ethanol content sensor installed?

Stoich for standard pump gas is no longer 14.7 as it contains 10% ethanol. Target should be closer to 14.1 for the E10 fuel that most stations have. As the ethanol content is so darn unpredictable anymore, having a fuel composition sensor is really almost a must-have. However, if you are using normally available top-tier 93octane pump stuff, you are probably better using the 14.1 as the new target.

That may account for the table shift.

[worf]

Stoich for standard pump gas is no longer 14.7 as it contains 10% ethanol. Target should be closer to 14.1 for the E10 fuel that most stations have.

Very good point.

What’s max power target?

How do “modern” engine management systems trim the fuel/air ratio. Do they have a content sensor? The variation in ratio between 14.1 and 14.7 is well in excess of the required control limits.

[hans14914]

All modern cars use a fuel composition sensor.

The VE table and target AFR table is done in units of Lambda instead of AFR.

The Lambda value references the stoich value given from the ethanol content sensor, so that target table never has to change.

Fuel composition (flex fuel) sensors also contain an internal temperature sensor. This value and the ethanol content come back on a single digital input. The output varies the duty cycle and frequency to get two signals on a single input.

The fuel temperature is used to correct for fuel density.

None of this is accomplished in the older systems. Most just rely on long term trim from the slow narrow band sensors. As Worf suggests, if the trim table only allows for 10% correction values, most of the overall compensation is already eaten by the modern fuel.

[worf]

Nice @hans14914 !

VEMs can take and use composition sensor output?

[hans14914]

I am by no means a VEMS expert. It's been years since I have really worked with it. I moved over to Adaptronic, and then now settled on MaxxECU for all my projects (including my 928 control package). Principals are the same across all systems, but some are just easier to interpret. I personally have always found VEMS to be a bit of a struggle to understand and work with.

There is documentation here on how to get the sensor working on their hardware. I have not done it personally, so can't comment on how difficult it is. On newer platforms, it's really easy, as it was designed in as a base feature in the fuel models.

http://www.vems.hu/wiki/index.php?page= ... 2FFlexFuel

In the future, I may share all the documentation related to my MaxxECU implementation on the 928. I would never say that VEMS is a poor choice, hardware/dollar it's still the most cost-effective option. I would say that as far as time/results though, you will find a much better value with slightly more expensive but newer platforms. For the 928, the budget leader would be the EcuMaster Black and for a couple more bucks MaxxECU Race and AEM 508. I personally find the MaxxECU solution to be the best intersection of hardware/software/value for an 8cyl application with any advanced functionality.

I just wanted to get that out there incase other people are looking at standalone. I do have plug-in modules to convert the factory harness over to any other computer, so if anyone wanted to explore a plugin option, its pretty easy to use any computer you want. I do not do a plug-in, as I find the condition of the original harnesses to be suspect, and its just easier to install a new harness than track down all those gremlins.

[worf]

… now settled on MaxxECU for all my projects (including my 928 control package).

Ok. Good stuff. Assign my question to the MaxxECU.

You know why.

[hans14914]

I dont want to get this VEMS dedicated thread too far off the rails, but as it relates to the Maxx, all you have to do is connect a sensor and check a couple boxes in the config and its ready to go. Its preloaded with a generic ethanol fuel density table (which can be modified) and because you are using a target Lamda map, it does everything for you on the back end. There are some super clever things like 4-dimension tuning that allow you to have a fully developed model for 100% gasoline, and a fully developed model for E85 and then you can blend the values between these table as referenced to a Z-axis (such as ethanol content).

I have a dyno here at the office that I need to get reinstalled and will be doing some pretty in-depth VE modeling for the factory manifold, and a couple other configurations. After I have all that data, I will consider starting a general thread about modern engine management strategy, and how it applies to the 928. At that point I will have a full map of stock engine with flappy open and closed and we can look at ideal uses of the resonance valve, as well as the effect of short runner common manifold.

Now back to VEMS and helping Crumpler get his ride rolling right.

[Crumpler]

Hey you guys, feel free to bring any pertinent info into the thread, always.
I'm wanting to get as much information (and choices) as possible out there for the next guy.

Am I crazy again or is my thread missing posts? I posted about the starvation events and Greg, Hans and Dave all helped me?!
Hans, I'm using 93 pump gas.


While VE table checked out I was getting basket case readings shifting. Very lean, I established the AE was nil and adjusted way up. Overrun and fuel cut values looked fine. Whoops that was separate thread;)

That helped but I was still seeing lean episodes during the shift sweep, as it went into low rpm/20kPa cells -- and also super lean carrying anything over 2000 rpm into a shift with clutch in and off tps which looks like what was causing starvation events.

I did two things: jacked up fuel in those VE cells and went back and lowered the IAC dc references. The lower IAC meant the car would not start, so i had to open the throttle screw a little and then adjust some rec fuel again. I don't know if that is best practice but it did work. No more shut down events.
This is the situation now: better but not pretty. For reference second bar red is lambda which still spikes at 26 afr briefly during hard shift. Second image is coming off throttle with clutch in, initial rich then lean. Green bar is IAC activation.

[Crumpler]

So with increase in AE I can see proper trend of initial decrease in afr and gradual climb over several seconds to VE cell designation with throttle.
What came next in terms of rookie mistakes is the EGO (closed loop) was hardly ever active. Essentially, it would activate in it's set parameters but was turning off with any AE. I had set AE too much too early, so any slight fluctuation in throttle would activate enrichment and kick it out of closed loop.
Vems has max throttle setting for EGO but no min throttle.
Also interesting for those that come after, AE in VEMS is calculated by dTPS AE percentage and then multiplied by the rpm scaling percentage to get actual AE. Both values are set independently.

[Crumpler]

Still fiddling with VE.
The VEMS analyzer doesn’t seem to be bullet proof even on longer logged drives. Just started doing it by hand.
Only notable event was that I forgot my fuel cut was set at 120 kPa and another pucker moment on the interstate. Stupid.

Getting some belt slip on SC again so I bought the treated non slip pulley version from 928 Motorsports. Worth nothing it shipped next day.
Not standard offset however, had to hand cut a keyed spacer.

[worf]

… so I bought the treated non slip pulley version from 928 Motorsports. Worth nothing it shipped next day.
Not standard offset however, had to hand cut a keyed spacer.

Standard 928 Motorsports product: final product engineering required.

[Crumpler]

Yep.

I also saw that they now sell Raptor SC units without the kit .
Before I bought the Procharger, I remember trying to buy a Raptor from Carl without the kit attached and he was having none of it.
As far as I know they are the only North American authorized vendor.
Hell I called Australia a couple of times, got stone walled on direct shipment too.

[Crumpler]

One thing that I'm not making headway on is the closed loop function.
VEMS uses a term called EGO correction. When this is activated (by setting parameters), it drags to target lambda.
So I'm calling that closed loop function.
The issue is that it is automatically disabled with any enrichment.
So driving I will see small fluctuations in AE even at steady throttle. I've tried turning down the tip in AE but it seems to be more about sensitivity of the TPS.
Sooo, basically EGO is activated every once and a while driving and always when at idle.

[grepin]

Hi Crumpler

Looks like you have very quickly caught up and possibly surpassed my skills. How are you manually tuning. I find it hard on the road trying to keep up with it. What I am trying to do is review the log after a run and make changes. In the example below my VE is at 54% EGO correction is saying 59% so I should adjust my VE to 59% to add more fuel in to match target. Now from here once VE settings have been changed it seems you can save a config file and upload. Not sure if I am doing it the hard way. In the car when running you are jumping in and out of reviewing the log file and running live. I have been wondering if open loop may be a better option. Anyone feel free to correct me.
Have you discovered a way to see the route on the VE table post a run without going to log review. Is there a setting for that. It would be handy. My current issue ATM is a fuel cut/ break down sensation at 5200 rpm that auto tune is not getting on top of. As you know I am chasing a fuel upgrade / revamp before I go any further. If I can find anyone to sell me something.
Also did you get your TPS to read correctly. My car hates throttle pos enhancement during warmup. It likes pressure based. I was wondering if that could be it.
Also what after market knock sensor are you using. My car never came with one and I feel it would be a valuable addition.

Cheers hope all is well.

image.png

See below my 5200 rpm issue. Its second gear pull from start. 32Kpa region.

image.png

[Crumpler]

Hey Grepin, I can't say I'm getting more proficient at VEMS! ;)

AS to what I'm doing:

I will log a drive. Then I come home and analyze the log. I look at where the afr/lambda is high or low, and as long as it is not associated with shift events or abrupt tps events I will document and adjust that VE cell before the next drive and log.
I have not relied on the VEMS tune, mostly because the tuner does not correlate well my log notes. I am assuming VEMS tune is looking at variance and recommending the middle number -- I'm not sure anymore man.
I am essentially tuning in open loop, which is not my first choice, but I need a certain amount of AE to not go lean with throttle (I guess with the forced induction). Any AE kicks vems into open loop again. I will attach a picture, the third cell down shows the dTPS artifacts and you can see no EGO activity. That is the source of my oscillations. I have tried all variations or dtps and AE rpm scaling. I have tried pressure AE option and it didn't go well, huge lean spikes.
If you are looking at your lambda of 1.07 on the 66 map 4545 rpm pic, I would adjust AE first and not necessarily VE cell, that may be causing that tip in artifact?

I had installed the Link G4 knock lamp kit prior to the VEMS conversion, it uses the standard Bosch doughnut knock sensor however. I'm not seeing knock but I have not advanced timing from Peep's initial map yet. It looks like you did from the 5200 rpm run picture being 44 spark advance?

I'm taking mine to a pro tuner in two weeks for final tune and dyno, I'm hoping he can tell me what to worry about and what not to.

Best, Dave

[grepin]

You may have something regards AE. I sort of have ignored it and just checked mine. It seems high compared to original as arrived from Peep. It may be poring in too much fuel. Do you know what triggers the enrichment. Is it RPM or throttle based. I still run pressure based enrichment instead of throttle based. Only as its a pig at start up with throttle based. I give it a little throttle and it stalls and pushes liquid out the back. Pressure it seems happier. Pic of my enrichment table. Sorry its all about me at the moment.
I look forward to hearing what you learn from the dyno tune. I want to do the same however not until I have replaced the injectors and confirmed my fuel pressures etc etc. That also relies on any of the vendors selling me stuff. Hard at the moment to get anything.

image.png

[Crumpler]

In regards to AE, my understanding so far is if VEMS is set to throttle based, then it will use both dTPS and current rpm scale and come up with value for total AE.
Your dtps values are similar to mine, and that first dtps box seems to have the most control over tip in for me. Your rpm values are a little higher then mine.

You may have seen this already but I just found my CLT table was pulling ignition at higher temps, just set table to zero.

[Crumpler]

Here’s the real question: would either one of us use VEMS again?

[grepin]

For me the benefit of plug and play and a stock car and no desire to rewire sold me. Plus with no WBo2 originally on my car etc it was a neat way to step into a tuneable option. Lets see once a proper tuner gets a hold of it. So I have no regrets. With boosted applications and maybe a race car I can understand maybe not.

[Crumpler]

I’m getting a trigger error now at approximately 5000 rpm, so you are catching me at a pessimistic time. It’s a “less” trigger error which may or may not be significant or related to CPS. I don’t know because there is no real ecu support.
That’s the only thing I would change, the tech and software support.

[grepin]

Yes the support is lacking however if you find a decent forum there is a lot that can be gleamed from it. Unfortunately not for the 928 I can find.

Did I read somewhere that you dialled in your injector req by modifying it until it read 1 lambda. Was it during warm up? Can you explain what the process was. I am just starting to think ahead for when the new injectors go in.

[Don Smith]

Hi Gents:
A lot of guys on the Audi forums( specifically the AAN motor, 2.2 litre, 5 pot) are using Vems, with great results. In particular, EFI Express in N.Hampshire, does tuning. Maybe reach out to him. Or check the Quattroworld URS/4/6 forums.
G.L.
Don

[Crumpler]

Yes the support is lacking however if you find a decent forum there is a lot that can be gleamed from it. Unfortunately not for the 928 I can find.

Did I read somewhere that you dialled in your injector req by modifying it until it read 1 lambda. Was it during warm up? Can you explain what the process was. I am just starting to think ahead for when the new injectors go in.

My method was to warm up fully, so no warm up enrichment.
Then I disable closed loop, usually set EGO control min to 6700 rpm so it does not activate.
Adjust to desired idle speed.
Adjust Req fuel to 1.0 lambda.

Then of course, your back at square one with VE table.

[grepin]

Cool I will do that from the get go with the new injectors. Think its better than using the formula. I know it doesnt really matter if close but why not if back at square one. Pressure checks to be done first when I find an adapter.

[Crumpler]

Ok, so additional data, for what’s it’s worth.
Lucky pro tuner has some VEMS experience so we could jump in this morning.
He reset ignition table, I haven’t gone in logs yet, but I know he devoted more cells to higher kPa range. The original table stops at 130.
They seemed to think I was pretty close to the mark in general on the rest.
Initial dyno runs 364 rwhp, 370 torque, about seven pounds of boost.
What was obvious quickly is that the 24 pound injectors aren’t sufficient. The runs we’re leaning out AFR over 5000 rpm, to about 13.4 at redline.

We kinda tabled the rest until I can come back with bigger injectors. I’m shooting for a general goal of 400 rwhp.

[worf]

What was obvious quickly is that the 24 pound injectors aren’t sufficient. The runs we’re leaning out AFR over 5000 rpm, to about 13.4 at redline.

Remind me, what fuel pressure are you running?

[Crumpler]

At some point I converted back to the 2.5 bar 86 FPR so 36 psi.
I think because everything was pig rich at the time.
So 267cc/min flow give or take.
With new hp I’m getting a higher requirement

[worf]

At some point I converted back to the 2.5 bar 86 FPR so 36 psi.

Ah. What I suspected.

That's why you need bigger injectors.

[grepin]

I’m getting a trigger error now at approximately 5000 rpm, so you are catching me at a pessimistic time. It’s a “less” trigger error which may or may not be significant or related to CPS. I don’t know because there is no real ecu support.
That’s the only thing I would change, the tech and software support.

Hope you feel better about VEMS for now. Keen to see the future results. Just in case you havent sen this. When I hear the car running I think surely VEMS is fine for us. Cool display for your race car.

[Crumpler]

That is a cool display!

I’m that guy that would buy it and then set the screen to analog;)

Did some work over the weekend, but in new injectors. These are 42 pounds. Green Giant variety.
Now I will be redoing VE table.
Also redoing the crank breather with dedicated lines and catch can, the grand experiment of drilling out the filler cap did not hold up to redline rpm’s — to put it in objective terms…
Oil everywhere in subjective terms.

[Crumpler]

Back in the rabbit hole with new injectors.
Chasing strange low throttle behavior. Series of lean surges. Can’t tell if it’s just VE mapping or actual injector delivery. There were no dead times given with shipment. Called vendor and got dead times they use in house.
Started with validating fuel pressure. That’s the good news, the FPR is spot on.
36 pounds on rail with fp relay jumped and 29 pounds with 18 mm hg vacc applied to FPR.
There’s a bizzare ritual within VEMS to find dead times by adjusting target lambda map. Which I guess I will do next.
This essentially blows up any track time this season. I think I forgot any driving skills I developed at this point anyway.

[Crumpler]

Well that was fun.
Ok the vendors specs were pretty darn close. At 14 volts the dead times were about 4% off.

If I can rule out fuel pressure and injector malfeasance, I feel comfortable going ahead with VE mods.
I did get the revised crank case breather fabricated.
Three lines to a catch can, vented to atmosphere, and a return line to pan.

[grepin]

Looking good. Hopefully the new injectors clear things up for you. I haven't put much mind to injector dead times. Noted for when the injector upgrade goes in. Fixing oil cooler hoses now.

[Crumpler]

Thanks Greg, I’m looking forward to seeing those new fuel rails go on your car too!

[grepin]

Yes pretty happy about about how neat it will be. Feels like a better option than using the tail adapters.
Hey no rush but could I have a look at a capture of your injector settings. Not to copy just want to see what it looks like when set.

[Crumpler]

Sure man, including capture from updated 42 pounders. Also sheet that came with 24 pounders that i had been using.

I took the car out and let vems autotune the other day, just to see what it would do.
Drove long enough to tune all relevant cells. Thought it did a pretty good job -- until i did the same drive and looked at logs. Started finding VE holes, afr at 18 at steady loads, etc.
I just can't get onboard with their software.

[grepin]

Yes I sometimes wonder the further down track we go whether there could have been better options. I am still happy with plug and play on a stock car. Lets see what happens with my new fueling setups. I assume you will be going back to the dyno. I hoped for me when I get to that point It will iron out any issues and auto tune will cease. I am accepting my skill set is lacking. Also if I had the stock ECU etc I wouldn't even know what its doing. Because we have visibility with the tuning software are we being over critical or being presented with detail to worry about. Sometimes I do feel like putting the stock stuff back in to confirm a car issue and not a tune issue.

[Crumpler]

Probably all is true. It’s a good plug and play option and we now know enough to be critical ;)

It’s a million times better then the old stock computer for me now. I also feel like the ecu itself is comparable to other units I would have access to with my entry level expertise.
The auto tune just doesn’t seem bulletproof.

[grepin]

Thanks Dave,

What was the part number for the 24/25lb injectors you were using?

[Crumpler]

This is the link:

https://www.fiveomotorsport.com/audi-bm ... injectors/