10 mpg 3.2 - can’t find where the fuel is going

[Discseven]

Originally Posted by proporsche

sounds good without problem..
Still ,you go through so much testing to find out the fuel los --crazy...still thinking it is something missing.If i understand 10 gallon difference -between freeway and city right...?

Ian... Highway mpg = 27. City = 10. So for 20 gallons of city-driven fuel... engine's doing 200 miles to a tank (before reserve.) That's roughly half of what I used to get---so that's why I say 10 gallons per tank is "missing", or going somewhere unexpected. This amount, a leak would seem very sensible. But there's no leak. So this fuel must be being consumed by the engine. Efficiently so according to the O2 readings.

After the spark plug change from W6DPO to Iridiums that you and Sal proposed, city went to 11-12 as I recall. I still call it 10. My city-drive style has become far more mpg conscious so am probably inching up the city number. Will fill tank over the weekend and post current status.

Indeed, many tests. All good. Will keep at it.




Exhaust gas analyzer

Unit fell through. I looked at new unit prices. Thousands. Can't justify that. Used on eBay... I read some op manuals for systems available. They have multiple parts that need to be included, and system has to have been maintained in certain manner to work accurately. Too risky to get this gear used. Shops that do this? Not any more. Florida did away with this testing years ago. Maybe a dyno shop. Still looking... Would be interesting to see what the exhaust composition is.
.

[Showdown]

Karl,

Since building my digital gauge which calculates MPG base on injector pulse width, rpm, time and distance moved (it's all a very complicated math equation but that's how it's done) I am able to see my MPG for various states: total, current trip, and current 5 second rolling average (I tried 1 second, 3, 5, 7 and 10 and found that 5 was optimal for human readability).

The point here is that while I'm driving (or later with a data log) I can see what my MPG is at any given point and plot that against RPM/speed, GPS location, etc...

Since the cam change (stock 2.7 CIS to M1) I have been very cognizant that my low speed, low gear MPG is not great- 10-12MPG. The amount of fuel that the cam demands at low speed to get the car moving is significantly more than the stock CIS cam. Mid range and low load cruising like at highway, it's really quite good; high 20's, even approaching 30. Up top at WOT it goes way back down again as expected.

So, on any drive in the city where I have to go from 0 up to 30, then back down to 10, then to 30, then to 40, then 0 then 30, etc... the cam and by extension, car is consuming a lot more fuel than when I'm up to 70 and cruising. Sound familiar?

I'm fortunate that I can control fuel delivery and timing with my ECU so I can set my highway cruise lean with a lot of advance to maximize MPG (that coupled with a taller 5th) means getting from home to the fun zone is pretty efficient. 29MPG in this car is better than my old Nissan and often rivals my new Audi.

Once I get the the destination, goodbye fuel, but hello fun.

I asked about timing but realize that given you're using not using an aftermarket ECU, you might not be able to get that data, or even if you had it, make any changes. I suspect, as I mentioned earlier (a few times) that the cam is just more fuel hungry or that the stock 3.2 timing might not be optimal for the 964 cam... Others have chimed in to share their cam-change MPG-dropping tales as well.

I guess my point is to add another anecdote of what happens with a cam change to this puzzle. I'm fortunate to have more levers (ECU) at my disposal than you but even still, city MPG did take a hit.

Best of luck and I'll be following along hoping you do find the smoking gun.

[76FJ55]

would you happen to have a GPS watch or cycling computer or other method of measuring distance driven? I know you said you checked your ODO based on highway markers, but I wonder if there is some way that it could be inaccurate at the lower speeds while still accurate when on the highway?

[Discseven]

Originally Posted by Showdown

Karl,

Since building my digital gauge which calculates MPG base on injector pulse width, rpm, time and distance moved (it's all a very complicated math equation but that's how it's done) I am able to see my MPG for various states: total, current trip, and current 5 second rolling average (I tried 1 second, 3, 5, 7 and 10 and found that 5 was optimal for human readability).

The point here is that while I'm driving (or later with a data log) I can see what my MPG is at any given point and plot that against RPM/speed, GPS location, etc...

Since the cam change (stock 2.7 CIS to M1) I have been very cognizant that my low speed, low gear MPG is not great- 10-12MPG. The amount of fuel that the cam demands at low speed to get the car moving is significantly more than the stock CIS cam. Mid range and low load cruising like at highway, it's really quite good; high 20's, even approaching 30. Up top at WOT it goes way back down again as expected.

So, on any drive in the city where I have to go from 0 up to 30, then back down to 10, then to 30, then to 40, then 0 then 30, etc... the cam and by extension, car is consuming a lot more fuel than when I'm up to 70 and cruising. Sound familiar?

I'm fortunate that I can control fuel delivery and timing with my ECU so I can set my highway cruise lean with a lot of advance to maximize MPG (that coupled with a taller 5th) means getting from home to the fun zone is pretty efficient. 29MPG in this car is better than my old Nissan and often rivals my new Audi.

Once I get the the destination, goodbye fuel, but hello fun.

I asked about timing but realize that given you're using not using an aftermarket ECU, you might not be able to get that data, or even if you had it, make any changes. I suspect, as I mentioned earlier (a few times) that the cam is just more fuel hungry or that the stock 3.2 timing might not be optimal for the 964 cam... Others have chimed in to share their cam-change MPG-dropping tales as well.

I guess my point is to add another anecdote of what happens with a cam change to this puzzle. I'm fortunate to have more levers (ECU) at my disposal than you but even still, city MPG did take a hit.

Best of luck and I'll be following along hoping you do find the smoking gun.

Julian... fine digital build you did for rolling mpg calculation. Arduino based? I could sure use a rolling calculator I genuinely have minimal cam experience so am proceeding off input like yours and others. Unfortunately for me the input is not one sided. I spoke with Webcam who I'd consider the end all since their cams are installed. Their word is their 964 40/40 cams would not be suspect. Flip side is they may be biased---not saying they are. I hang my hat on their word unless thinking about yours---and I know you have mentioned it earlier. The mystery here is such that I'm open to whatever the truth bears out. Finding it has proven difficult.

Let's say these 964 cams are drawing nearly twice the fuel in repeated speed changes. Given the engine moving the same weight car as it was when doing 18 mpg city, I suspect the O2 city readings should read rich. Reason, there's no need for the extra energy expenditure to move the same weight the same distance. If pulling a trailer, makes sense the engine would consume more energy to move the greater weight---the ECU compensating for the load. I may have the principles all wrong here. This is my attempting to make sense of the extra energy that is now being burned---I can't get to a logical conclusion.

Can a 3.2's combustion cycle---no matter the cams---process nearly twice the fuel it was previously accustomed to?

Does a stock ECU with stock chips "know" there is nearly twice the fuel flow and make up for this with more air flow?

Originally Posted by 76FJ55

would you happen to have a GPS watch or cycling computer or other method of measuring distance driven? I know you said you checked your ODO based on highway markers, but I wonder if there is some way that it could be inaccurate at the lower speeds while still accurate when on the highway?

76... Interesting idea. Something could have happened to the odo system between when the engine came out for the rebuild and when it went back in. Not a clue what would have caused that. Can't rule this out without proof. Will give it go. Trying the Stride app to track milage.
.

[Showdown]

Hey Karl,

ESP32-based. Gotta have access to CAN BUS form an ECU to use it unfortunately in your case...

Take my cam talk with a grain of salt as I'm far less knowledgeable that others here with respect to cams. I'm just trying to Occam's Razor this problem.

The "tell" for me is that your AFR readings are pretty good, if not exactly where they should be; you're never super rich which tells me that the fuel being delivered by the ECU/injectors/system is all being burnt and that fuel isn't "escaping" combustion and ending up wasted through the exhaust.

So either your AFR sensor or controller is way off (probably not) or I don't know...

The engine is using all the fuel it's being given/asking for and that just happens to result in a delta of 8MPG since the cam change/engine rebuild.

I don't have any other way to explain it other than it's doing exactly what it should.

It'll be interesting to see what the oil analysis yields. Over a 10 mile drive, could 1/2 gallon really end up in the oil? That would make the oil more gas than oil by the end of a days' drive...

[LJ851]

Karl, did you change anything on the inlet side of the engine at the same time as the cams like an open air box or aftermarket air filter?

Is your oil temperature the same as it used to be in the around town driving situation?

[Discseven]

Originally Posted by Showdown

Hey Karl,

ESP32-based. Gotta have access to CAN BUS form an ECU to use it unfortunately in your case...

Take my cam talk with a grain of salt as I'm far less knowledgeable that others here with respect to cams. I'm just trying to Occam's Razor this problem.

The "tell" for me is that your AFR readings are pretty good, if not exactly where they should be; you're never super rich which tells me that the fuel being delivered by the ECU/injectors/system is all being burnt and that fuel isn't "escaping" combustion and ending up wasted through the exhaust.

So either your AFR sensor or controller is way off (probably not) or I don't know...

The engine is using all the fuel it's being given/asking for and that just happens to result in a delta of 8MPG since the cam change/engine rebuild.

I don't have any other way to explain it other than it's doing exactly what it should.

It'll be interesting to see what the oil analysis yields. Over a 10 mile drive, could 1/2 gallon really end up in the oil? That would make the oil more gas than oil by the end of a days' drive...

Julian... Are you saying it's possible for a combustion chamber to efficiently burn nearly twice as much fuel to move the same weight the same distance at the same speed?

Agreed on gas in the oil. Math doesn't make sense. But then neither do other things here. So we'll see what's in the oil. I've posted new thread asking for comments on Blackstone who is resource I'm thinking to use for this.

Originally Posted by LJ851

Karl, did you change anything on the inlet side of the engine at the same time as the cams like an open air box or aftermarket air filter?

Is your oil temperature the same as it used to be in the around town driving situation?

LJ... No change on inlet side. Oil temp is the same. Runs around 210. There's a Mocal cooler with twin fans in right front fender.
.

[Showdown]

I don’t know Karl, Thant’s above my pay grade.

Step back for a moment, are you sure you were getting 18mpg before the rebuild? This whole exercise is predicated on that idea and yet there isn’t any data to support that (at least not in this thread).

Is it possible that your odometer is off, tires/wheels not stock, or just that your 40 year old car isn’t as accurate as you think.


Sent from my iPhone using Tapatalk

[PeteKz]

I use Blackstone for oil samples on all my cars, HD motorcycle, and truck, and have been for 10 years. I am very satisfied with their analysis and dialogue. There are other vendors, but once you pick one, stick with that so you get consistent results across multiple samples (baseline). Unless you are running really long oil change intervals, don't bother spending for the TBN. It almost never gets depleted in normal use.

[Discseven]

Originally Posted by Showdown

I don’t know Karl, Thant’s above my pay grade.

Step back for a moment, are you sure you were getting 18mpg before the rebuild? This whole exercise is predicated on that idea and yet there isn’t any data to support that (at least not in this thread).

Is it possible that your odometer is off, tires/wheels not stock, or just that your 40 year old car isn’t as accurate as you think.

Julian... fine thinking on the "step back." Indeed the 18 being accurate as reference is key here. I checked this quite often (prior rebuild) to generally monitor the car's status, and driving costs. Has been consistent over many years. Less frequently I also checked the odo by default as I do see them being united for accuracy. Always also good according to state mile markers. Milage, fuel gauge, pumped gas at station... those parts were always consistent to the end math of getting 18 city mpg. Never questioned this figure as I know it to be common for 3.2s.

Question of whether my odo is accurate for city driven miles is currently on the table. Am driving tests today to check this. Also filling the tank to get a current mpg. Post this evening or in the morning on these results.

Same tires were on car prior to rebuild when it was doing 18 city. Just recently checked toe. All good. Tires have always, and continue to wear on the insides more than the outside. Soon time to swap them side-to-side.

Car being older... am thinking "friction." Where is it if it exists? Worn wheel bearings were possible. Tested the wheels. Fronts spin equally to each other but of course less easily than rears. Rears spin equally in and out of gear (with rear end off the ground.) No crunching/grinding feel or noise from any wheel. No rock on any of them. CV were new from a few years prior to rebuild. Prior to rebuild they were repacked. Friction related during rebuild: Crank journals polished and all new case bearings. Crank spun after closing case = silky smooth. Rods: new small end bushings, large ends rounded, both done by Porsche pro shop---and I measured each afterwards being precisely on. New wrist pin. Also measures spot on. In spec pistons and cylinders were recycled. New Goetz rings sequenced & oriented correctly. New cams fit neatly. 1 new rocker. New chains (on old sprockets.) Heads fully restored by Xtreme---new valves, grinds, new guides, new springs correctly oriented. 2,000 rpm break in for 20 minutes. Break in oil & filter changed. New clutch disk and pressure plate. Part of current diagnosis: Compression tested good and sufficiently equal among cylinders. Without plugs, crank was spun by hand with tightened fan belt---smooth. Fresh gear oil remains full---just checked level. Fan/alternator spins easily. Brakes check, all OK. Where else could friction be?

Got more into doc you sent. Impressive reference. Will be using it. Thank you.


.

Originally Posted by PeteKz

I use Blackstone for oil samples on all my cars, HD motorcycle, and truck, and have been for 10 years. I am very satisfied with their analysis and dialogue. There are other vendors, but once you pick one, stick with that so you get consistent results across multiple samples (baseline). Unless you are running really long oil change intervals, don't bother spending for the TBN. It almost never gets depleted in normal use.

Pete... Saw your post in Blackstone thread---thank you. Am moving forward with samples to them.
.

[Discseven]

Now recalling discussion that led to my going from stock 3.2 config to installing: 964 cams, 1.45 overlap, SSI exhaust, Dansk muffler… Among other selling points, it was the “engine would breathe easier—more air flow.”

…Some more recall. 930s I’ve owned: 9 mpg city.

Based on mpg alone, and assuming air flow through an engine is an mpg factor, I raise the question of whether my or ANY normally aspirated 3.2—with the mods noted—can possibly achieve nearly the same air flow, and so nearly the same fuel demand as a turbo charged 3.3 does?
.

[Showdown]

More air requires more fuel.


Sent from my iPhone using Tapatalk

[Discseven]

Odometer test

Was done with Google maps. Stride app was not run.
Route followed had 2 legs.

Combined route total per Google = 12.3 miles.
Trip odo reading = 12.4 miles.

Main odo reads same as trip.
Reference below.

Google Maps, routes 1 & 2.





Start and end mileage for Google route followed.



Current city mpg = 11.55

Fill at Westar. Premium. Done at same station & pump as last time. Fill process is also the same… nozzle position in filler neck, low trigger setting, allowing pump to cut off and no topping off.

Improvement from original 10... Could be any of following: my driving habits being honed for economy, change of spark plugs (W6DPO to Iridiums), change of gas station (Shell to Westar.) At fill early on in testing, 9 and a fraction mpg was managed. Was with slightly heavier foot than now,W6DPO plugs, and filling at Shell. All mpg readings following rebuild are without AC compressor installed.





Gallons of fuel loaded and driven miles. ("KM" refers to speedo, not trip odo.)





Gauge before and after fill. (Needle gets to 4/4 if tank is manually topped off. Am not topping off for mpg tests.)
.

[Discseven]

Originally Posted by Showdown

More air requires more fuel.

... Indeed.

If increased fuel use is a result of rebuild mods allowing so much more air flow than there was prior to the engine's rebuild, and taking into account all the tests done to date (see first post) revealing no mechanical or electrical faults that have not been corrected---notably parking brakes dragging, and brake booster failing while testing was in progress and that too fixed---then there come the wideband readings suggesting the engine running relatively on target AFRs... combo of mods do have plausibility. I can't fathom a 3.2 processing this much air/fuel but I don't know this stuff inside & out. So my reluctance in this regard is pure speculation.

...Restrictor plate(s) come to mind.

Edit:

Looking at various sources to see if restriction makes sense, there are claims that improved exhaust air flow INCREASES MPG due to the lessoning of back pressure. That makes sense. But given more air flow, that would sensibly call for more fuel.

Regarding the capacity of a 3.2's cylinders for air flow, I came upon calculations posted by Sal:

Let's keep this simple by simply just mathematically calculating max air flow in Liter/min and assuming a Volumetric Efficiency of 1.0

First we'll talk about VE and the 911 engine and some history, the 3.2L engine has a tuned intake system, the runners are tuned for increased VE at about 5800RPMs and this is the sweet spot for peak torque in these motors. Amazing is the fact that 5800RPMs these engines run at a VE of about 1.1 this means the engine is actually ingesting 10% more air than the displacement of the cyls! At the time the 3.2L was developed it was the first OEM (non-turbo) motor to exceed a VE of 1.0 and Porsche was recognized for the achievement. This is why all those online calculators don't work! Then the 911 engine internals on a per HP base are far lighter than any Muscle Car V8 on a pound per HP base. One needs to understand the that spinning 30-60lbs of metal on a crankshaft at 6000RPMs is no easy task and can easily consume well over 100HP internally alone! This is simply wasted HP that never makes it out to the wheels. The 911 engine really reduces these losses a lot compared to big American iron V8s.

So now for the math:
A 3.2L 4 stroke ingests 1/2 it's displacement in air per rev, so for each rev we ingest 3.2/2 = 1.6Liters of air per rev. Let's save that value.

1.6L / Rev

Now let's assume a red line of 7000 RPMs, 7000 revs per minute we are ingesting 11,200 Liters per minute.

So if we assume a VE of 100% we can ingest at most 11,200 L/min at 7000 RPMs it really is that simple to calc in volume.

But volume is not the same a mass of air so we can adjust for mas assuming sea level conditions at room temp of 68F then the we simply multiply by aprox. 1.2 so at sea level 68F and 7000RPMs we can take 11,200 * 1.2 = 13,440 grams/minute

So here's air flow at 7000RPMs at sea level in volume and mass:
11,200 l/min (volume)
13,440 g/min (mass)

Of course these are not exact because we assume a VE = 1.0 and that's not the case at 7000RPMs but for quick calcs to size injectors, fuel pumps and such this is a very good way to mathematically do the calcs.

In my modified 3.2L it puts down about 230HP at the wheels (265HP at the crank) and I monitor air flow very precisely during a dyno pull. The car has a lab grade MAF installed so I know exactly what the air flow in g/min is at any given time during a run. I also very accurately calculate VE in my monitoring system and log everything. I'll post more details on real runs soon, with the graphs showing what/how the VE behaves across the RPM range. It's very interesting to see this data from a real run.
__________________
Sal


https://forums.pelicanparts.com/porsche-911-technical-forum/863819-3-2-3-4-airflow-galons-per-minute-cubic-feet-per-minute.html

Am after an air/fuel equation to apply to 3,000 rpm. Working on it. Idea being to see how that fits a 3.2's expected mpg.

.

[Showdown]

Karl,

Have you by chance changed your flywheel from the stock one to a lightweight one?

I ask as the lightweight flywheel can/may have a negative effect on mpg in that more rpm’s are needed to help it spin (as it has less inertia), there’s less coasting (again less inertia) and as it spins up faster, it’ll achieve higher rpm’s and that can require more fuel

At highway speeds it should have no effect.

Maybe a small piece of the puzzle as this far there hasn’t been a smoking gun.


Sent from my iPhone using Tapatalk

[Discseven]

Tested:

AFM intake air temp sensor = OK
Engine temp sensor = Questionable…

Of two readings, lower engine temp ohms OK. Higher engine temp ohms is not to spec.

Tests are always repeated to confirm the results repeat consistently.


Specs applied for this are from:



Thank you Julian.



Notes & spec references:


AFM Intake air temp = 1.73 ohms. Suggests air temp being 15 - 30 degrees C / 59 - 86 F. Ambient is 75 dF so this makes sense.







Engine temp sensor

Two ohm readings were taken to test the engine temp sensor. First reading was taken while engine was warming up. Second when engine reached op temp, 210 in this engine’s case.

With dash gauge temp at hash mark just below 210: Ohms = 287.
This puts #3 cylinder temp in the 80 C / 176 F range. Makes sense relative to dash temp gauge reading. Calling this OK.

With dash temp gauge at 210: Ohms = 42.6
This doesn’t fit specs. Ohms should be 160 - 210 range. Either sensor is off or temp gauge in dash should read much higher.

Ohms climbed slowly as engine cooled. This correlates with specs.









Not sure what to make of the engine temp sensor results.
.

[Discseven]

Originally Posted by Showdown

Karl,

Have you by chance changed your flywheel from the stock one to a lightweight one?

I ask as the lightweight flywheel can/may have a negative effect on mpg in that more rpm’s are needed to help it spin (as it has less inertia), there’s less coasting (again less inertia) and as it spins up faster, it’ll achieve higher rpm’s and that can require more fuel

At highway speeds it should have no effect.

Maybe a small piece of the puzzle as this far there hasn’t been a smoking gun.

Julian... Nice thinking. Just looked back in my rebuild parts list to confirm what went in. Fly is a Sebro 93010203301. Stock. BTW, just used the doc you provided to test two temp sensors. Thank you for that info.
.

[scarceller]

Both the IAT and the CHT sensor have the same temp to ohms spec. Here's the full spec:
Temp C Ohms
-30 4700
-20 4500
-10 4300
0 4000
10 3700
20 2500
30 1825
40 1150
50 943
60 735
70 528
80 320
90 253
100 185
110 170
120 155
130 140
140 125
150 110
160 95
170 80
180 65
190 50
200 35
210 20
220 5

But once the CHT sensor ohms drop below 180 ohms the sensor no longer does anything to fueling. Meaning, if ohms are between 0-180ohms fueling change factor is simply 0%

This is why you can remove the sensor in a fully warm engine and bridge the 2 pins with piece of wire (0 ohms) and engine should run fine.

Wile I agree changing a sensor that's out of spec is a good idea, this is not your issue with poor MPG. Even if the sensor is below 60ohms it has no effect on fueling.

[Discseven]

^^^ Sal... Thanks for those details. Bridging CHT pins to keep a warm engine running... nice work-around.
.

[Discseven]

Tested:

Speed sensor - static resistance = OK
Dynamic AC voltage = OK

Reference sensor - static resistance = OK
Dynamic AC voltage = OK

Test procedure is per Specialized ECU Repair:

Disconnect harness plug from ECU. Multimeter, set on ohms, connects the following pins:

Speed sensor: connect pins 8 and 27. Ohms to be between 600 - 1600.
Reference sensor: connect pins 25 and 26. Ohms to be between 600 - 1600.




Speed


Reference


For dynamic testing: Check car’s battery has 12+ volts = OK (12.6 v)
Gearbox in neutral, parking brake ON.
Multimeter set to AC voltage.

Speed sensor: connect pins 8 and 27 with multimeter. Crank engine. Voltage should be greater than 1.0 v = OK (is 2.6 v)

Reference sensor: connect pins 25 and 26 with multimeter. Crank engine. Voltage should be less than 1.0 v = OK (is 0 v)

.