Nistune Forums

Well I have been stuffing around a bit trying to tune up the old girl with a newer turbo on the car, what I am finding when I back off the throttle whilst on boost, like say back off from 5psi to 0psi, the car will surge and jerk like a mofo.

Unsure at first what was going on, I was watching the the AFR guage that has a bar graph to see everytime I backed off the throttle, the AFR would go rich then recover.

Of course if you back off the throttle fully the idle switch contacts and the ecu will do a fuel cut so no surging/jerking.

Because its a big compressor you can hear the obvious flutter/compressor surge back out of the airbox, so I figured early on the air was going back past the air flow meter causing a rich AFR.

What i didn't thick was the air flow meter voltage was actually maxing out at 5.1V under these throttle back off conditions, no wonder it was bucking and carrying on.

The AFM voltage was actually dropping to zero volts as the airflow technically stopped then reversed back out the airflow meter where the rush of air was enough to max a Z32 afm, then stabilized back to normal.

I thought about using the maxtp values to prevent it from calculating a high tp value, but this would never work because if I was actually under boost at this rpm, the tp would need to be higher anyway and if I set a max value it would make it lean.

So long story short, looks like I have to fit a recirc blow off valve to fix this drivability issue.

Regards
Mick

Seems like your only way around it....

The fuel cut / recover maps only get used when the TPS idle is enabled (throttle completely off) but if you are not letting it go all the way then it will measure what ever the reverse airflow is and inject that amount (ECU has no way of knowing otherwise)

The TTP min isnt going to be of any use since you are getting a high load reading at that time anyway and TTP max needs to be used for normal driving otherwise

If we had a variable voltage TPS then you would be able to detect partial backing off on the throttle and maybe use those fuel cut/recover tables sooner but that isnt an option on these older ECUs unless I changed the code and used something like the diagnosis potentiometer as a TPS variable input

Mick do you have a log of this, I may be able to offer some suggestions.

G'Day mick,

This is exactly what we are going through now too. In the past we have just retro fitted reserc blow off valves but now are looking for other options.

Isn't this just an inevitable symptom of compressor surge in conjunction with a hot-wire air-flow meter? Other than a recirc I can't think of a sure-fire way round this.

The 'deceleration air' setting on an SAFCII forces the ECU to ignore AFM voltage whilst the TPS is between two preset throttle positions. It will ignore AFM output when the airflow is greater than that at the larger opening (i.e. in the case of surge or compressor 'blowback'). If you're seeing max voltage during surge, this may work I guess, although it's only meant for preventing engine stall due to large turbo/short inlet between AFM & compressor.

Surge & engine stall on my CA have all but disappeared since moving from a top-mounted T70 with short inlet piping and single 45deg bend to a bottom-mount TD06-20G with much longer piping and two 90deg bends and a further 45deg bend.

We are looking at the same thing but an external electronic setup.

It will only work with switch type TPS's where under any engine condition when TPS is on (idle) the AFM curcit will be fed with a pre determined AFM voltage (like 1V)

The Map can be tuned for idle and it will bring the voltage back to normal under decel.

Sorry Guys,

I forgot I put this post up here.

Here are the log pics below that prove the problem.

I have shrunk it all down so it’s easy to see.


In this pic you see the purple line (afr) has been full cut because of the tps being fully shut, so no surging or jerking present, note the green line afm voltage, air has actually stopped flowing in (0.2v), stalled, then its about to go backwards and be measured again. tp is not increasing because the throttle is fully shut, in later pics where the throttle is only partly closed you can see it calculate a high tp also.



In this pic you can see the throttle is fully closed/fuel cut, but you can clearly see the air flow going up and down as it surges back past the air flow meter.



Much the same as earlier pics, air flow has stalled 0volts then surged back past the airflow meter and measured 5volts+



Can see again as the throttle is shut, the airflow that goes backwards maxes the air flow meter voltage.



5.1V here, maximum reading from a Z32 MAF



This is the real one to look at because in this pic you can see I have just backed off the throttle a little bit, not completely and it has caused the airflow meter to register max voltage and causes the car to go rich (accesses highest load cells) and buck/surge and carry on.



Car needs a recirc bov to make it drive nice during part throttle back off, if anyone is interested my research shows the bosch bov valve to get is part no below, this is the high pressure version, most ideally suited for our cars.

Bosch 0 280 142 110
or
Bosch 0 280 142 114
or
Audi part #06A 145 710N
or
Saab dealer #4441895

I will have results once I get some bungs welded and the recirc bov fitted.

Regards
Mick

Hey Mick

Thanks for the info. Just out of curiosity, what K value and latency did you end up using for the GTR injector / Z32 AFM combination?

I would like to know what adjustment offset to make to the AFM changeover routine for RB30/VG30 engines


Do you still have cold start running issues with the GTR injectors?

Also you will need to put in a zener diode to clamp at 5.1 volts from the AFM as the Z32 seems to go over this and can damage the A/D converter chip. The latest software manual has instructions for R31/Z31/HR31 included

thanks
Matt

RB30-POWER, the logs look very similar to mine.

seems increasing the volume between the afm and the turbine works - however, for this method to work ideally, the volume of the pre/post turbine piping would have to be the same. - very difficult with volume of intercooler etc.

I'm going to try the bypass/recirc blow off valve trick as soon as I get the car back from the transmission shop. I'll be using a Turbosmart "type 3" 38mm recirculating blow off valve, with a nice long hose going back into the turbine intake pipe.

I'm a little skeptical, though - as i'm not sure that this will resolve the issue of the air running back past the air flow meter, it might just meen you get one 'whoosh' of air going backwards, rather than a 'flutter'...

basically aren't we trying to get a large volume of air that is after the turbine, and 'move' it to before the turbine, where there may not be enough room for it :p ?

the mitsubishi recirc bov off the rb20det and rb25det and wrx would be a good valve to use, they are good, but the flange is to large for where i want to mount it, so i have to get a bosch and get some bungs welded.

no cold start issues ever since that second cold start map was found, which is identical to the first one, i just change them together and its as good as standard starts.

k value i resized as per the theory of 260cc to 444cc, then for the airflow meter adjustment, i had to actually lower the k value to get it near correct afr wise and cell accessing with load.

from like 320 down to 290

in theory you have to increase the k to maintain pulsewidth with bigger airflow meter, but it doesn't hold true i have found, the standard bosch airflow meter and z32 calibration and ve maps, just must be way different characteristics.

the only way i suppose i could increase the k with the z32 would be to resize from 260cc to 800cc, which should make it lean, then make up the difference by increasing the k, but then i think it will access the higher load points to soon.

all the theory is bs, you just have to set it so it accesses the cells and the afr are close to standard, the standard map is up the shit anyway.

would love to know how nissan do it from factory, like how they get their base figures etc.

taken me so long and much fiddling to get it all running reasonable, changed almost every bloody value in the maps and software, apart from the latency values.

still not entirely happy, there is a problem which i cant track down, its like a big miss and i think it may have something to do with the injectors changing from group to simultaneous mode, i wish i knew more how this worked, as i can't find the problem. suppose i need a dual channel scope to confirm if this is the problem or not to watch both injector channels at once.

there is a dyno day coming up in july, so see what happens there.

I'm going to have to have a go at this with the R32 AFM I have lying around here on my car

I dont get how it should be different, using the Z32 AFM, Z32 VQ maps but there must be something else in these ECUs changing the calculations since it requires dragging the K constant down instead of upwards

As you mention, by doing this, shifts the TP higher and so then on low load you will be accessing the middle of your map

Discussing last night with Dynowog on the phone, and this would require rescaling the TP maps then and shifting them higher, but then he found idle was affected

Are you able to shed some light on what you've changed in your maps at all? I'm going to try it further from this end

No one with a Z31 has ever reported these issues (yet) when changing to Z32 AFMs. Wondering whats different here??

Matt,

What does Dynowog have to typically do when he fits the Z32 afm's, scale the k up or down?

Well this is the first time hes done this. I've modified a VLT ECU for him (fitted with NIStune board and diode modification)

He has then got the Z32 AFM plugged in, but finds that regardless of K value its running too rich (even K = 0) which doesnt sound right to me, so I'm getting him to try out a second ECU

Hes running standard injectors, just the Z32 VQ map and then mucking around with K constant. Voltage looks as it should be expected

I ran my car on a R33 AFM when first doing this, and that is when I noticed the K constant just wasnt calculating right. I assumed the voltage drop difference may attribute to this somehow?

Just done the same mod on a HR31 and having similar issues with hitting the end of the map too quick (likely becuase TP is now in the middle during idle). Interesting to note late model ECCS HR31 have use a standard 0-5 volt AFM but earlier (NICS) ones dont (2-7 volt instead)

I'm grabbing a Z32 AFM and connector from Pete this week and going to try it out in my car and see how it goes. I can put a dual channel scope on it also to monitor the signals

all he should have to do, is drop the k value and it should trim up rather nice, if he didn't do injectors at the same time.

did he fit the 390ohm resistor across the pins?

nothing wrong with your guide at all.

something else is wrong if the k value has no effect i think.

should have some more results for you early next week, back to the drags this week if the weather holds out and dyno day on saturday, provided if i'm not there to late, because i have to work sat mornings.

Sorry,

In continuation with what I was saying earlier, what I think is happening in reality, is the Z32 afm reads slightly more voltage at low airflow input, i.e idle and cruise and when you get to the higher voltage 4-5V, it is putting out less voltage increase for more airflow increase.

Well I know that sounds, der! of course it does its a bigger AFM, but what I'm trying to say is that to get the idle and low cruise afr correct, you have to drop the k value because the voltage from the z32 into the a/d convertor is higher then the equivalent a/d opamp voltage would have been with the Bosch factory AFM during idle and cruise?

It has to be, or else the K value would need to be increased.

I know the VQ map is meant to fix this as its not linear, but it doesn't seem to do what it should IMHO, the maps are to similar to begin with.

What is happening i think is, when you get to the higher calculated tps, it doesn't calculate enough tp value and leans out because the k value was decreased to begin with.

I have sort of confirmed that when I overboosted the car and the car leaned out!

It should just keep calculating and increasing tp and afr stays the same. (with this example, i don't think the afm was maxed out or injectors maxed, not enough power being produced)

But you only notice this at very high boost/power because the factory maps are so rich to begin with. 4.2V+

So you have to expand your tp map, so that at around 4.2V+ of air, the fuelling is increased in those corresponding tp cells, more so then the cells below it that are around 4V and under, which seem to calculate fuel in a reasonable fashion, even with the lower k value that was set.

Hope some of that makes sense.

It's what i'm thinking anyway.