My theory why stock MAFs fail. (Long)

[RUQWKNF]

Ok, here's my theory I've been pondering for quiet sometime now.

The stock 86-87 MAF was designed to work with a TBO348 turbocharger, which can flow roughly 550-575 cfm, depending on boost levels and level of modifications. The factory system does not incorporate a compressor bypass valve or re-circulation valve. When we go from WOT, even with 12-15psi stockish boost levels, and suddenly let off the gas, all of that compressed air reverses back on itself. Because now the throttle blade has suddenly slammed shut and this pressure has no where to go but backwards. Back through the intercooler, back through the turbo, through the MAF pipe and through the MAF sensor out the filter. Now, the factory MAF sensor has a heating element wire that is used to measure the amount of current drop based on grams per second of air flow that passes over it. To the non engineer types out there, (like myself) a similar effect would be when you blow onto a toasters' heating elements and their intensity/brightness drops due to the sudden blast of air. All is good until you factor in years of use, heat cycles, abuse caused by vibration and constantly being subjected to reversed air flow, as well as the thin film of oil that will get blown all over the inside of the MAF and onto the circuit board from a turbo whose compressor oil seal is leaking.

My theory is that over time, this heating element becomes worn out, and or will eventually fracture from the circuit board. This would also explain why when you do the famous "Tap Test" the elements connection would be severed and no signal would be sent to the ECM, causing the engine to stuble or even shut off. Now, if you factor in putting on a much larger turbo, say one that will support 800 cfm worth of flow, such as a TA-49, you have a lot more airflow that will reverse back through the MAF sensor, further exacerbating the issue. The main reason I believe that the Buick engineers did not incorporate a bypass valve into our system could have been due to the fact that our cars came equipped with an automatic transmission. Typically, you will find compressor bypass valves, or re-circulation valves used on manual transmission equipped vehicles, so that spool up can be maintained between shifts. It also helps with the CHRA rotating assemblies transient response.

Ok, with that said, what do we do about it? Back in the day, when there were no such things as LT1 or LS1 MAF sensors, some people would incorporate a bypass valve into their setup. Either mounting it directly to the up pipe just before the t-body, or onto one of the intercooler pipes. The logical location would be to mount it just before the throttle body for the best response and least chance for airflow to back up and slam the compressor wheel backwards.
Then route the bypassed air from the valve, back into the MAF pipe directed towards the compressor wheel. The vacuum hose would need to be tapped into one of the vacuum lines that come off the vacuum distribution block on top of the throttle body, so that the valve would get a true pressure to vacuum reading.
I personally have only seen this setup done like this on two other cars. One belonged to my good friend Dave Fiscus, back in the late 90s on a WE4 that he owned. With our closed loop MAF setup, you would need to run a recirculation valve/bypass valve instead of a standard Blow off valve. Blow off valves are designed to vent the excess airflow into the atmosphere instead of recirculating it back into the system. Running a blow off valve on our cars can cause stumbling and stalling when transitioning from WOT to idle. Trust me, I ran an HKS blow off valve for a very short time with the factory setup and did not recirculate the pressure back into the system and it would always stumble or shut off during decel. This would not apply to Speed Density converted setups that eliminate the MAF sensor.

This theory can also apply to why journal bearing turbos wear out faster. Years of the rotating assembly being slammed backwards under deceleration. Especially when you factor in a poor oiling system from the factory. The factory oil system sends unfiltered oil to the turbo at all times. Trash in the oil system goes to the bearings in your turbo, combined with coking of the oil that developes from years of hot shut downs will take a toll on the bearings and oil seals. Not too mention the leaching effect of trash getting stuck in the factory oil cooler working it's way back into the oil system.

Ok, what do you guys and gals think? Discussions and thoughts are welcomed as I by no means know it all. I'm still learning stuff everyday about these cars, and have been messing with them since '87. It's a never ending struggle.
Thanks for your time and sorry so long.

Patrick

 

[BlackMetal]

I have a HKS knock-off BOV on my car (not recirculated) and it handles decel just fine.

I understand the theory behind why diverter valves are "better" but in some ways it doesn't make much sense. If you're slamming the throttle blade shut, the air isn't going into your engine anyway, so the car is still going to dump fuel for the airflow it measured, but the air won't be in the engine, and it'll run rich for a split second. So when people say a BOV is bad because it lets the air out to the atmosphere, I don't see why it matters, because as I said, BOV or not, all the air that the MAF measured isn't going into the engine anyway.

Anyways I'm not sure about your reasoning for the compressor surge causing MAF failure. The entire point of the MAF sensor is for air to come rushing past it, so I don't think it matters which direction the air is coming from. Vibration, oil contamination, and simple old age seem like more likely candidates to me.

And yes it's pretty well-believed that running a BOV/recirc valve will prolong the life of your turbo/bearings because it won't be forced to spin in the wrong direction anymore.

So I agree about the bearings, but I don't think longevity of a MAF sensor is a valid reason for wanting to run a valve.

 

[ChrisCairns]

Here's what I believe. The stock factory MAF was bolted to a bracket. Once we started to upgrade we neglected to bolt it down and it vibrated to death.

Someone suggested this years ago on the mailing list at GNTTYPE as everyone was replacing MAFs regularly. (They were readily available then.) Once most of us bolted our MAFs back up to something the general consensus was the MAFs were back to being reliable.

 

[BlackMetal]

Here's what I believe. The stock factory MAF was bolted to a bracket. Once we started to upgrade we neglected to bolt it down and it vibrated to death.

Makes sense. I've also heard that removing the factory flex pipe and installing metal MAF piping transfers more vibration to the sensor. Myth or not it makes sense.

 

[jhanson1750]

I would question the validity of the fueling issue. I think that GM addresed this issue with the defueling "map" in the processor. I agree with the reversion of the turbine being hard on them.When I wasa shop forman for an 18wheeler shop there was never as issue with turbos @ the mileage that the buicks loose them. With a diesl you have no "surge" and they use filtered oil through them and @ 3-400,000 miles no falures unless they ingest something and damage the compressor wheel. I think if you conceder the "duty cycles" of a turbo as in spool-up and surge the life of a given turbo would more than double if surge were eliminated and all other things were left the same. JMHO Jon Hanson

 

[jhanson1750]

Don't know how I did that(up) just tried to edit...

 

[bison]

Ok, with that said, what do we do about it?

Patrick

Convert to an aftermarket ecu with speed density.:tongue: I dont think the failure rate was any higher on these cars vs. other cars. Most other 20 year old cars are already scrapped and melted down by now. Not too many TR's suffered that fate or will be. The cars are typically preserved with their old solid state electronics which just fail for no good reason sometimes. All we can do is speculate. Ive seen MAFS last years and years in the vibration mode and ive seen nearly stock 20k mile cars have them fail. No reason as to why from what ive seen.

 

[Turbo1dr]

Maybe that thin plastic element get's vibrating up and down when there's lot's of air passing over it under boost. Over time this would be like bending a piece of metal back and forth until it finally breaks.

Ok, with that said, what do we do about it? Back in the day, when there were no such things as LT1 or LS1 MAF sensors, some people would incorporate a bypass valve into their setup. Either mounting it directly to the up pipe just before the t-body, or onto one of the intercooler pipes. The logical location would be to mount it just before the throttle body for the best response and least chance for airflow to back up and slam the compressor wheel backwards.
Then route the bypassed air from the valve, back into the MAF pipe directed towards the compressor wheel. The vacuum hose would need to be tapped into one of the vacuum lines that come off the vacuum distribution block on top of the throttle body, so that the valve would get a true pressure to vacuum reading.
I personally have only seen this setup done like this on two other cars. One belonged to my good friend Dave Fiscus, back in the late 90s on a WE4 that he owned.
Patrick

Patrick, I also ran a system exactly like you described back in the mid 90's. It worked very well.

 

[GNVenom]

Several years ago we did some flow testing because of the MAF "dropout" that we experience between shifts. This anamoly causes an instantaneous lean detonation condition (seem like everyone sees the same thing, that huge knock count between 2nd. and 3rd. shift). Turns out, it has to do with the way the MAF works, i.e. "hot wire" element. The element uses a certain amount of regulated current to maintain its temperature and by passing air across it, causes it's temperature to drop. The circuit in the MAF then adds more current to the wire, to maintain the preset. The more air flowing, the more current needed to maintain it. The problem is, we flow too much air across the element, causing the circuitry to overload - there is not enough power available in the reg circuit to maintain that level, so the circuit momentarily drops out. You see it in your scantool; the LV8 numbers peg at 255 until the shift, then drop to maybe 230 or 240. When this happens, the ECM goes back into fuel control, dropping the fuel to match the <incorrect> air flow. Therein lies the lean "pop" that we commonly see.

So, it seems the drop in the airflow number is artificially caused by the excessive air flow and the MAF's inability to maintain its regulation because of it's limitation within it's design.

Bumping the MAF voltage up to 15.2 volts, however, minimized this anomaly. I believe the bottom line is that the MAF supply voltage gets depleted during WOT operation...we see this as a common occurence on our scan tools.

Some might say using a volt booster is a band-aid approach but I disagree. That's the one sure way of preventing system voltage drop to the MAF.

There's my 02.

 

[strikeeagle]

Damn TRs!!! Now you tell me, after I traded in my GTO!!!

Actually, I don't have a theory. I just figure mine will break someday, at which point I'll have to do *something* - not sure what yet?

strike

 

[83ttypecooled]

Hi,
I'll tell you what, my '86 Firebird with a 2.8L V6 also eats MAFs at a regular rate, about once every 75K or so. I think it's just a part that has strong limitations.

 

[VadersV6]

They go bad because the potting epoxy begins to decompose and the electronics start getting hit with constant heat and moisture...but the main issue is the strain gauge. Strain gauges are this thin amber colored film that has a given resistance to it. If the film stretches or moves, the resistance changes. This allows strain to be measured...they're commonly used in bathroom scales. GM used these in the MAF's...as far as I can see. I made a strain gauge calibrating/lapping machine on a little sony robokids robot, back in 98 or so, and learned alot about strain gauges. I was pretty surprised when I got this car and saw the strain gauge inside the MAF, but immediately understood why it was being used. Now, 20 years later, FORD is using these in their MAF's...which is really just a strip of strain gauge that drops in a tube. Anyway, the air passes over the strain gauge, deforming it, which provides a given resistance that can be matched to a given airflow. The thermode is also used as part of the measuring package. The Thermode needs a given voltage to maintain a constant temperature. More air passes over the thermode, cooling it, requiring the computer to dump more voltage at it, to heat it back up to the temp it needs to be at. This provides a voltage curve that can be matched to known airflow parameters. The strain gauges stretch with time, because of heat and improper cleaning, and just time. You stretch a strain gauge permanently, and its permanently dead. Mine is all wavy like a potato chip. This means its bad and the data it spits out cant be trusted.

 

[Grumpy]

I just think its a 20 year old part I DRIVE my junkers a LOT .. personally I have had 1 MAF die on me (well that I can remember ) ... we have a gazillion miles on all our cars .. we also had the "hardpipe" with no problems .. these were only low/mid 11 sec street cars

 

[SloGN]

Several years ago we did some flow testing because of the MAF "dropout" that we experience between shifts. This anamoly causes an instantaneous lean detonation condition (seem like everyone sees the same thing, that huge knock count between 2nd. and 3rd. shift). Turns out, it has to do with the way the MAF works, i.e. "hot wire" element. The element uses a certain amount of regulated current to maintain its temperature and by passing air across it, causes it's temperature to drop. The circuit in the MAF then adds more current to the wire, to maintain the preset. The more air flowing, the more current needed to maintain it. The problem is, we flow too much air across the element, causing the circuitry to overload - there is not enough power available in the reg circuit to maintain that level, so the circuit momentarily drops out. You see it in your scantool; the LV8 numbers peg at 255 until the shift, then drop to maybe 230 or 240. When this happens, the ECM goes back into fuel control, dropping the fuel to match the <incorrect> air flow. Therein lies the lean "pop" that we commonly see.

So, it seems the drop in the airflow number is artificially caused by the excessive air flow and the MAF's inability to maintain its regulation because of it's limitation within it's design.

Bumping the MAF voltage up to 15.2 volts, however, minimized this anomaly. I believe the bottom line is that the MAF supply voltage gets depleted during WOT operation...we see this as a common occurence on our scan tools.

Some might say using a volt booster is a band-aid approach but I disagree. That's the one sure way of preventing system voltage drop to the MAF.

There's my 02.

thats another good reason to convert to a speed density system




Patrick.----- you are correct on the surge causing maf problems over a certain amount of time. The airflow while building boost is not a suddden thing it builds slowly.But the compressor surge from the throttle blade shutting creates a very big surge and along with that surge comes resonsant energy as the air has to change direction.

I have been thinking about doing the recirculation valve for a while now.

i think that some of the guys with the ls1 sensor with big turbos and front mounts seem to have a much more common failure rate that the guys with the smaller turbo/intercooler set-up.

 

[Joe Lubrant]

Found this on a site...forgot link.

Tech Tip Corner
Mass Air Flow Sensor Failure

Premature or Repeated Mass Air Flow Sensor Failure
General Motors Remote Mounted Designs
Instances of repeated Mass Air Flow Sensor failure on General Motors applications utilizing the remote mounted Delco unit can often be traced to an engine back-fire condition. The sudden intake pres - sure buildup can induce cracking of the heated element resulting in MAF sensor failure.

Several common causes include lean misfire due to low fuel pressure or restricted fuel injectors, breakdown of the secondary ignition including internal ignition coil arcing, and a dead spot in the Throttle Position Sensor (TPS).

A lean fuel condition can be verified by using a scan tool to monitor the Block Learn memory value with the engine at a steady no load cruise RPM. A reading above the 135 to 140 range would indicate a lean fuel mixture, or an exhaust lean above the Oxygen sensor. Internal ignition coil arcing between the coil primary and secondary windings can be very difficult to trace. Checking for internal arcing requires one of the new engine analyzer/scopes.

A dead spot in the Throttle Position Sensor (TPS) can be verified by monitoring the sensor voltage as the throttle is moved from the idle position to the Wide Open Throttle position - very slowly - with an analog voltmeter or lab scope.

Joe

 

[GNRick]

Well, I was having a backfiring issue not too long ago. Up pipe blew off and took out a chunk of my new old style silk screened hood insulator . I hope my MAF lasts a while longer. As for the air flow theory, would the MAF last longer if we go back to the stock air box?

 

[VadersV6]

Backfiring would totally melt and deform the strain gauge...it would take way too long to describe what a strain gauge is, so they just say "heating element". Strain gauges were cutting edge technology back then...and look. It took over 20 years for Ford to catch on, and now they're calling it cutting edge..lol.

 

[BLACK6PACK]

knock on wood... I have 208,000 miles on my maf.

 

[bison]

Strain gauges were cutting edge technology back then...and look. It took over 20 years for Ford to catch on, and now they're calling it cutting edge..lol.

LMAO! You should email them and let them know that Buick had them back in 1983!

 

[GNVenom]

Vaders, the component you're referring to is not a strain gage, it's a resistive heater element sandwiched between two "Kapton" layers. It gets quite hot, I believe around 160 deg. F and it is airflow that cools it down...causing its regulator circuitry to add more current to maintain its regulated temperature. Kapton can handle mucho heat, but in time, with ambient temperature added to the mix, the element eventually fatigues and stretches out. Heh, just like us humans...
the fatigue eventually causes either an open condition or an air gap between the Kapton layer and the element. Air gap prevents the air temperature to saturate properly, causing a mismatch between metered air flow and actual air flow. Hence, the "rich" condition when it finally goes out.
I don't know who started the "rap" test, but since it's basically solid state, I would never recommend this as a diagnostic means.
The circuitry is placed on a metal plate with a ceramic substrate, almost spacecraft spec, and then housed in a sealing "jelly" that is never supposed to break down. It may eventually happen but the Kapton element will not outlive the sealant they used from the factory. So, it's a "heat vs. time" situation that kills those MAF's.

Convert to speed density? Great idea.