Afm Spring Adjustment - Bothersome Thoughts

[Captain Obvious]

Last time I talked to Superlen, he had completed his home built flow bench and had poked around with a couple AFMs. I betcha he knows.

 

And if we're lucky, he'll be able to come up for air sometime soon!

[Zed Head]

The flow bench would probably tell the story.  Three runs; stock setting, loose spring, tight spring, with voltage readings.  He must need a diversion by now.   

[Zed Head]

Pondering how to make stone softer...

[superlen]

Hi everyone! Excellent discussion by all. I think everyone ended up on the same page, but I'll throw in my notes from a quick read through.

 

1. The 4500 is "arbitrary" in that it just happens to be where the AFM Vane maxes out with a stock Z engine. By maxs out, we mean any more airflow does not move the vane (& the corresponding voltage read by ecu) any appreiable amount.

 

2. The ECU does not stretch the pulse ANY based on RPM value. Don't confuse this with more fuel delivered with more rpms. That happens because you get two squirts/cycle...more cycles, more squirts, more fuel, not because the ECU reads rpm and decides we need more fuel.  I think we all know, but I'll reiterate there aren't any tables in the stock ECU, it's done with multi-stable vibrators, comparators, & op amps. There is zero circuitry in the ECU to modify pulse width based on RPM.

 

3. If the load (read airflow) continues to increase, after the AFM has maxed out,  a stock Z will begin to lean.

 

4. Read #3 again. At first It seems wrong or counter intuitive, but that is the way the ECU operates. Actually many ECUs do this regardless of what load sensing device (LSD) they use. If the LSD runs out of headroom, the ECU pulse width stops increasing. Remember LOAD drives fuel required, not rpms,

 

5. WTH, that can't be right. This sounds like a bad scenario. What gives? Well, the short answer is we don't care. Why? because the overall fuel system is designed to do this at the theoretical max our engine can perform. Its at the extreme top end of load so we don't need to measure any more than that. You may worry that leaning out at Full load is really *bad* and it is....well kinda. Being LEAN (AFR above 14.7) at full load is VERY VERY bad. Leaning out at full load (meaning the AFR in increasing due to stalled pulse width - NOT really a lean condition), is not bad. Fuel delivery systems are design to run a lower AFR at full load/WOT by default. Our Z's are no different. Digital ECUs do this with tables and our ECU does it with the transfer function of AFM vane position vs. pulse width and a WOT input as a kicker. So when our AFM Vane maxes out, we are probably at 10-12:1 AFM. If we continue to draw more air (we've robbed a beer truck and are fleeing from the local POPO), the AFM Vane doesn't increase any more, the ECU doesn't strech the pulse any more, and the AFR begins to go leaner, 10.5, 10.6. 11, 11.1, 12.1, 12.4, 13, 13.1, 13.2, 13.2, 13.2. (Actual numbers will vary of course depending on your AFM setting/engine combo) What saves us is the system is designed so we can't get into that 15,16,17:1 ratio at full load. I haven't measured the Zs AFR at this condition, but I would guess that 10:1 isnt too far out of line as the target starting point. It is desirable to run a fully loaded engine rich for a few reasons. First, the farther away from the lean kaboom condition, the better insurance you have against murphy's law. Second, excess fuel being dumped into the engine will help cool the combustion chamber. The only determinent is that you lose a *tad* bit of power, but not much & not worth the risk. Even the latest/greatest ECUs with the best sensors around won't try to run the engine near 14.7 at full load. Perhaps, in some race engines.

 

6. The stock AFM measures out on the flow bench to handle about 190hp before maxing out so more than enough to operate the stock Z. Take this number for now with a grain of salt as I'm not sure what constitues a stock AFM. I have 20 or so AFMs and they are all different. Rob, you can take it verbatum, as I did the math when your AFM was on the bench. One thing to note however is that it's not *too* much greater than stock. This is the reason that when you add a cam and headers, the stock system starts acting up. You now can max the AFM easier and the ECU doesn't have any way to know this. Typical fixes have been larger injectors, tweaking the clockspring, or rising rate regulators, but all of these fall down in particular areas. The Stock ECU was great for the stock system, it just isn't very forgiving of changes. The ability to make it robust enough to handle unknown conditions just doesn't exist in its analog circuitry.

 

7. I don't remember who mentioned it, but a minor point is that Speed Density systems (using a MAP sensor) do NOT work well with radically cammed engine. The valve overlap makes a really nasty manifold pressure signal, its' not a lot of vacuum, it's fluctuating all over the space, & there are multiple resonances happening. All of this makes the MAP signal a bit annoying/impossible to use, also MAF sensors in these conditions can also suffer some of the same problems as the pulsing waves cause some airflow inversions in the intake track. In these cases, designers fall back to Alpha-N (TPS), or a combination of Alpha-N and MAF.  A correctly tuned Alpha-N will work great with an engine, but it takes a bit of tuning to get it set up. In theory, the throttle response of an Alpha-N will be a bit better than AFM, MAP, or MAF. In practice, I think it's a moot point as engines that need to use Alpha-N are already pretty responsive as they are gulping air at idle & ready to take off when you crack the throttle.

 

Well, that's all for now. Again, this is a great thread on the AFM. It's a pretty misunderstood, yet critical piece of our cars.

 

Len

[FastWoman]

Lenny, consider the following hypothetical:  Let's say airflow is pretty much maxed out (or has approached some asymptote) at some RPM at WOT, owing to Reynolds numbers.  And then RPM increases considerably beyond that point.  As long as fuel delivery isn't maxed out (e.g. injector pulses not overlapping), then fuel delivery will increase (because of more injection pulses per minute), while air delivery will not increase.  Or looking at it from a different angle -- what goes on during one combustion cycle -- the same volume of air will have to be shared with many more combustion cycles, hence less air per cycle, while the fuel delivery per cycle remains the same.  Either way you explain it, that would create a richer running condition -- again, if airflow is limited, but fuel delivery is not.

 

At some point our systems just can't spray any more fuel, because the injector pulses would overlap.  About what RPMs would generate that overlap with the AFM pegged out?  (I'm sure you've tested it.)

[superlen]

S,

 

I'm trying to follow what you typed....(not your description, my weak mind) What you are describing is a drop in the VE (volumetric efficiency) of the engine due to intake manifold design (lets say a really poor one), and yes if the VE tanked while you still had the same amount of fuel being sprayed you would begin to richen up. However, in practice, I don't think you can get there. The reason is the feedback. The power source causing the increased RPMS is derived from the air flow. If your VE (and thus your airflow) tanks then your power delivered drops and the RPMS then drop accordingly which raises back up your VE. If your VE went totally to zero, NO air would flow and 0 rpms would result.  

 

Now that I think about it, I'm pretty sure this is exactly what happens when you just run out of power at top end during a race or snappy visit to the supermaket, the engine simply can't breathe anymore air due to shrinking VE -> which creates less power -> decreases RPMS -> raises VE back up -> raises RPMS...rinse repeat.

 

As for the injector overlay, I don't remember the number offhand, but I'll run the HellFire GUI in the morning at work and take a screen snapshot for you. I have a simulator in there that shows the injector pulse width and where things break down. With the stock 180cc injectors & two sprays per cycle, I seem to recall around 9000rpm as the theoretical limit before overlap. Some degree of fudging is in that as I don't have an accurate VE map for the stock intake/cam/head. Depending on what state I left the GUI in, I can email you a copy so you can play with the simulator as well. Ping me on my HF email if you are interested in that.

 

Len

[superlen]

8300 RPM assuming 78% VE at that rpm.

 

Also, I want to revise my last thought on maxing out at top speed, its not the VE tanking at top speed, its just sheer power not enough to combat the wind resistance and power requirements to go faster. It *could* be the VE tanking if you had a really really crappy intake manifold design with too small of runners, then the VE would fall dramatically at the higher RPMS. However, you're typical cars intake is of course sized properly, which is good. We don't want it to be the deciding factor.

 

Len

Edited March 30, 2015 by superlen

[Chickenman]

 

7. I don't remember who mentioned it, but a minor point is that Speed Density systems (using a MAP sensor) do NOT work well with radically cammed engine. The valve overlap makes a really nasty manifold pressure signal, its' not a lot of vacuum, it's fluctuating all over the space, & there are multiple resonances happening. All of this makes the MAP signal a bit annoying/impossible to use, also MAF sensors in these conditions can also suffer some of the same problems as the pulsing waves cause some airflow inversions in the intake track. In these cases, designers fall back to Alpha-N (TPS), or a combination of Alpha-N and MAF.  A correctly tuned Alpha-N will work great with an engine, but it takes a bit of tuning to get it set up. In theory, the throttle response of an Alpha-N will be a bit better than AFM, MAP, or MAF. In practice, I think it's a moot point as engines that need to use Alpha-N are already pretty responsive as they are gulping air at idle & ready to take off when you crack the throttle.

 

That was me Len. With a Speed Density system and big Cam you simply run Alpha N, as you mentioned, up to an RPM where you start pulling a reasonable vacuum, then switch on the MAP feedback. Any street-able cam is usually stable enough by 2,000 to 2,500 RPM that you have a sufficiently stable vacuum signal to run the MAP. ( 2,500 RPM is on the high side ).

 

GM TPI systems with speed density have the ability to re-programmed to enable Alpha N under those conditions. A friend of mine Tunes local Road Racing/Track Day/ Autocross EFI systems ( mainly GM, Ford, Mopar )  and has several Speed Density systems working just dandy on motors with really big cams.

 

Running a DD Audi Quattro 1.8T, I'm more familiar with the BOSCH ME7 systems and tuning with Unitronics and Eurodyne. It's fairly easy to program the ME7 to run Alpha N up to a set RPM and then switch to a MAP system. Or MAF if you prefer. A lot of VW/Audi drag and Road Racing guys run MAFLESS systems based on just the MAP sensor. The " good " Tuners have all gotten wise to running Alpha N till about 2,000 to 2,300 rpm on the big cammed motors, but " off the shelf " tunes are pretty basic. 

 

Personally, I would always want to run a MAF on any Modded street driven vehicle engine. The part throttle drive-ability is so much easier to control. But there are some guys that prefer a " Clean " looking engine bay. Personally I prefer function over appearance...but that's just me.

Edited March 31, 2015 by Chickenman

[superlen]

Chickenman,

 

Thanks for the info..good stuff. I knew many of the systems blended or switched between MAF/MAP/Alpha but didn't know any of the details on the specifics. My only experience with Alpha N is on my bike. It's a Victory Vegas and the signal coming off it's MAP was quite crazy at times. The stock ECU only uses it to sample BARO at startup is my understanding. The fueling system on it is completely Alpha-N from that point on.

 

I have a MAF from an Infinity Q45 to play with on HellFire but I haven't tested it yet. I'm going to plumb it ahead of the AFM for an interesting test as I should be able to read both signals and correlate airflow/IAT and massflow. I'm a bit wonderous about how the intake routing will affect the signals. I

 

Len

[Chickenman]

Testing of MAF on the 1.8T forums have shown that they are quite sensitive to air turbulence. This is true to all MAF's in general.

 

With that in mind, an ideal bench setup would have a straight section approx 6" after the MAF, before any bends or connection to TB. At least 3" in front ( more is better ) with an air-bell entrance. Debris screens or flow straighteners on the front of the MAF should be left in place.

 

That seems to give the most consistent Testing results.

Edited April 1, 2015 by Chickenman