A/F measurement with high overlap cams | FerrariChat

A/F measurement with high overlap cams

Discussion in 'Technical Q&A' started by pma1010, Apr 11, 2004.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, Skimlinks, and others.

  1. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    I spent some time with the Tod at Continental on Friday, tuning the carbs on my P6 cam'ed (290 duration, high overlap) 308. This car is also fitted with a wideband A/F sensor (fitted post header, pre-muffler). The AF sensor has been calibrated to within 200 feet of the altitude here in Chicago so I am confident of its accuracy. Being a 77, it has no cats.

    After sync'ing the carbs, dialing in the mixture, driving the car to get it good and hot, and then re-doing the mixture (and re-checking the air flow), all per "the book", we are finding it wants to idle best and pulls best low down while recording a very rich mixture (11.5 or 12:1 on the A/F readout).

    Leaning out the mixture to the "normal" range (say 14:1) causes additional drivability issues.

    For any given A/F mixture input, high overlap cams will, other things being equal, lead to a richer mixture being recorded by an A/F sensor due to the greater presence of unburned fuel in the exhaust. A smaller overlap reduces the unburned fuel/leads to more complete combustion of the fuel and hence a cleaner exhaust. This clearly was one (the?) impetus for the "softening" of the cam profiles of the later carb'd cars as they were "tamed" to meet emissions requirements.

    My question is therefore two-fold:
    1. I assume the effect of the cam overlap on recorded A/F mixture will be greatest at low rpm/idle.
    2. At mid to higher RPM (say 3000 up), with the engine acting more like a pump, the A/F measurement versus a "normal" target (like 12.8 or 13.2, for example) will have more salience than at idle.

    Any thoughts?
     
  2. Steve King

    Steve King F1 Rookie

    Feb 15, 2001
    4,367
    NY
    pma good work here. I do not have the A/F instrumentation so my tuning is basicly with the air balance and then using my ear for the rich/lean adjustment. I also find that my 77 runs best at low end with a rich mixture.I have gone to a hotter plug to off set the potential of fouling during start up and idle. As it warms up the idle is dead on at 1100rpm (want the extra 100rpm for the AC) . The only drawback is when you do the Ferrari red line tune up you blow some black smoke out of the pipes. It reminds me of my early 327/340HP corvette motors. That big Holley alwas ran rich until you would light it up , then it would fly. So I opted for the rich mixture and also notice it runs a bit cooler for street use.
     
  3. chrismorse

    chrismorse Formula 3

    Feb 16, 2004
    2,150
    way north california
    Full Name:
    chris morse
    Russ,

    1) going to a high duration/overlap cam, particularly with big venturis yields a very weak signal at the venturi, (since there is almost no venturi effect). We compensate for this weak signal by fattening up the idle. Further, because of the lower velocity through the carbs we have very poor air-fuel mixing, lots of fuel droplest sticking to the walls, ect. We fix this by fattening up the idle some more. Clearly we are not going to get a perfectly smooth idle unless we raise the idle speed - any misfire whether lean or rich will allow unburned fuel to go into the pipes and tell us we are too fat.

    2) Following up with the high overlap allowing unburned air/fuel to go through the combustion chamber and out into the exhaust, I speculate that somewhere around peak torque you would have the greatest cylinder filling and possibly the greatest % volume of air/fuel getting past the overlap - thus showing an overly Rich condition.

    Some further speculation. It seems fairly clear that bigger cams and venturis are likely to give us higher readings but if we can record some benchmarks for stock and "improved" engines, we may be better able to use the tool and shared database to come close to optimal power. This will require some carefull records of modifications, ambient conditions, A/F ratios and dyno verification.
    Once we have established max power, we could deal with driveability issues.
    Since it is ulnlikely that any two engines will have the same mods and be operating under the identical conditions, a certain amount of interpolation will be needed. What a great time saver that would be, (not to mention vastly reducing the amount of wa$ted dyno hours).
    I have one of waynes meters, but haven't had a chance to try it, (owned the car for 5 weeks - got the cig lighter changed, bungs welded in the new header pipes, exhaust changed, K&N installed rejetted to 140 and 200 and started on the suspension bebuild - frightfull knocking and banging).

    looking forward to seeing this tools usefullness explored.

    regards,
    chris
     
  4. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Chris
    Useful thoughts, thanks. I am going to suggest bigger idle jets (we are 12 - 16, 1/2 turns out on the euro [tapered] mixture screws) and a smaller main and will see how drivability is impacted.
    Philip
     
  5. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Going to a 65 idle (from 60) improved drivability significantly.
    70 idle, even more so (44 DCNF/36mm chokes/P6 cams).
    Will post A/F data when I have the car back in the next couple of days.
    Philip
     
  6. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    I don't know the number (any?) of you that are interested in this topic, but here's an extract from a Motec tuning manual:

    “Engines with high overlap cams running at low speed may pump air through the engine resulting in a false lean reading, therefore the meter may need to read richer than would otherwise be expected”

    Confirms the going in expectation and the results from our idle jet changes.
    Philip
     
  7. jmillard308

    jmillard308 F1 Veteran
    Owner Silver Subscribed

    May 29, 2003
    6,568
    Perth West Oz
    Full Name:
    John Millard
    Philip
    Yes - me!
    Keep up the good work - I'm loving it :)
    John
     
  8. fivebob

    fivebob Formula Junior

    Jan 31, 2004
    254
    Tauranga,New Zealand
    Full Name:
    Callum
    Oh well I suppose Motec can't get it right all the time ;) That's from the AFM-1 manual which is incorrect, the PLM manual correctly states "Engines with high overlap camshafts running at low speed may pump air through the engine resulting in a false lean reading, therefore the meter may need to read leaner than would otherwise be expected."

    If you work logically through it you'll see why the first quote is wrong, if there is excess air in the exhaust, i.e. air that bypassed the combustion process, then the meter will read leaner than the actual value e.g. it would read 14 when the correct value was 13, if the desired ratio 14 then the meter would have to read approx 15. This is obviously not the case with your engine which needs to be richer.

    Others have already covered the reduced Volumetric efficiency that you get with high overlap cams which means less mixture filling the cylinders and poor fuel atomisation, resulting in poor idle and transition. To get around this ignition timing at idle needs to be advanced, but without the benefit of Electronics this is not an easy task.

    The fatter idle probably helps transition because there is more fuel available to cope with the sudden influx of air, but as I haven't played with carbs for ten years I've forgotten most of what I learnt, so I'm not sure about this.
     
  9. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Fivebob,

    I'd used a different logic and agree with the statement made about the AFM -- A/F sensor reads rich at idle (or any low RPM point) as, given the high overlap, a lot of unburned fuel gets dumped into the exhaust. Result, engine is lean but reads rich, necessitating richening up the low end mixture. This is also why high overlap cams are a challenge for cars subject to emissions tests.

    Either way, clearly there is very little vacuum at idle, I am sure there is poor AF mixing and it runs a lot better with big (70) idle jets.

    I do have electronic ignition and will look at the low end ignition advance again. Typically I run a bit more down low than is the case with distribs.

    Two questions:
    - do you have an online reference to the PLM manual - I'd like to read and be better educated
    - do you have a sense of a mid-range and up A/F target and when the "cam overlap effect" is lessened sufficiently to produce a "normal" signal from the A/F sensor (and therefore providing me with the ability to shoot for an otherwise "normal" (12- 13) AF ratio target)?

    Thanks for the inputs
    Philip
     
  10. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Fivebob, I found the PLM manual. It does indeed contradict the AFM manual. I'll see if MoTeC can provide some clarification.
    Philip
     
  11. fivebob

    fivebob Formula Junior

    Jan 31, 2004
    254
    Tauranga,New Zealand
    Full Name:
    Callum
    All Motec manuals can be downloaded from http://www.motec.com/support/manuals.htm. I've sent a query off to Motec Support asking which of the statements is correct, but I agree with your analysis of what is happening re the unburnt Hydrocarbons and emissions. However I can't agree with your logic of rich/lean readings at idle, the Lambda sensor reads available O2 and any unburnt fuel will not affect it, which is why I believe the PLM manual is correct.

    Volumetric efficiency varies over the rev range, typically you will see two or three peaks in VE between idle and redline caused by intake ramming effects. Where they occur, and their strength, is a function of cam timing, intake runner length/diameter, plenum volume, primary intake length/diameter and to a lesser extent the exhaust system.

    At higher revs the interia of the incoming charge is much greater than at idle so the cylinder filling is much better and given that carbs, to some extent at least, self adjust the amount of fuel according to flow, you should have a more "normal" A/F reading. However if one of the aforementioned VE peaks is to sudden for the carbs to cope with then you may get some anomalies at certain points is the rev range. I know I've had problems with this when tuning fuel injected engines, but they don't self adjust like carbs.

    I think it's time I broke out my books on the subject, Prof Blair has something to say about this in his book "Design and Simulation of the Four Stroke Engine" so I'll do a bit af late night reading to make sure I fully understand what is happening.
     
  12. fivebob

    fivebob Formula Junior

    Jan 31, 2004
    254
    Tauranga,New Zealand
    Full Name:
    Callum
    Here's an example that shows the changes in VE that occur, the points where the manifold pressure graph smooths out are where VE is greatest, you will also note a small change in A/F at these points.

    The engine is a 2 litre turbo that was causing me some knock problems at 6200 rpm, right where the VE was greatest, I had to dump a lot more fuel into the engine and retard timing at this point to stop the knock, when I changed cams this point moved about 500rpm up the range. While it's not 100% relevant to the discussion at hand it does illustrate the point.
     
  13. mk e

    mk e F1 World Champ

    Oct 31, 2003
    12,899
    The twilight zone
    Full Name:
    The Butcher
    With a carb engine, anytime air is flowing, so is fuel. That means that you can not have unburned air present in the exhaust without there also being unburned fuel. With FI, that is not necessarily the case since the injectors fire for finite times which at idle are much less than the total air flow time.

    I use an O2 sensor on my car (EFI) to determine the mixture and it works great, althought I have much shorter duration cams. When a cylinder stops firing for any reason, the ECU reads the mixture as off the chart lean because the engine is pumping unburned air (and maybe fuel too) into the exhaust. The air is full of O2 so the %O2 in the exhaust shoots up and the ECU thinks is very very lean. If on the other hand I was detecting unburned hydrocarbons, the meter would read very very rich if a cylinder stopped sparking but was still injecting fuel. So, does your meter look at O2 or hydrocarbons or CO2 or maybe CO? The answers will be different.

    For bottom end and diveablity tuning, I still like the butt-o-meter the best, especially when coupled with an occasional plug read.
     
  14. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Mark

    I believe the NTK UEGO (L1H1) sensor measures both the level of oxygen (O2) and the level of hydrocarbons with the Nernst cell depending on the existence of a lean or rich condition: see:

    http://www.techedge.com.au/vehicle/wbo2/wbntk.htm

    Philip
     
  15. mk e

    mk e F1 World Champ

    Oct 31, 2003
    12,899
    The twilight zone
    Full Name:
    The Butcher
    That's a complicated little bugger.... it looks like it's using a standard narrow band O2 sensor to decide if the mixture is rich or lean. Then based on that decision, it measures how rich or how lean using the Nernst cell. The problem is that at low rpm, a large part of what it sees is unburned, so both O2 and hydrocarbons are present and the sensor can not deal with that. The O2 sensor detects O2 and sets the Nernst cell to measure a lean condition, but when it tries, the hydrocarbons (which it thinks are not there) react with the O2 spontaniously on the catylist of the nernst cell, so no voltage is require to drive reaction. No voltage tells the readout that it is rich, not lean because there is no O2 to measure, but the cell is in the lean measurement mode....

    ....so to make it much shorter the readout tells you nothing about what is happening in the cylinder because the sample is way too contaminated with unreacted gasses.
     
  16. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Mark: Data is great, isn't it - as long as you know how to interpret it. And, in this case I am not sure we do.

    Fivebob: From MoTeC (remember, I am not using their sensor, or meter, but am posting with reference to Fivebob's comments). Given how the Techedge O2 WB works, I d not believe what follows is applicable, but here it is:

    [MoTeC engineer]: "The PLM manual is correct. Keep in mind that when a cylinder misfires there are 14 parts of air to the single part of fuel. Additionally an oxygen sensor is not a fuel sensor. It reads oxygen concentration with respect to the atmosphere. Therefore even with unburned fuel blowing through the exhaust, a misfire still shows lean. Now to high overlap: The piston is moving up and beginning the compression cycle and the air/fuel mixture in the cylinder would much rather escape through the open valves than be compressed. Couple this with the exhaust pipe reverberation which may be creating a low pressure at the exhaust valve side, the incoming charge is steered out towards the exhaust pipe and goes past the sensor telling it that it's lean. You have to keep dumping fuel in because most of what needs to be burnt to idle escapes before the exhaust valve is closed. In this case there is no misfire, it's just that the fuel you intended for use in your engine got sent out the tailpipe instead. So now you have to add more than is necessary to make up for the loss. Finally the other thing to consider is sensor placement and how far it is from the end of the exhaust system. Exhaust pulses travel up and down the pipe and therefore at low engine speeds can draw fresh air up the pipe past the sensor. The sensor should be screwed into a bung welded on to the exhaust pipe a good 3 feet or more from the 'end' of the exhaust system"

    All (well, at least the one other person following this!): I talked to James Patterson at Norwood. After listening carefully, James put it succinctly. It will give you a false reading at low rpm, much better at high rpm. At low rpm, just tune for drivability. He also provided some timing advice too. All helps.

    I picked the car up tonight. It certainly has more grunt this year. To protect the guilty, I'll just say the service personnel road tested the car and were all thrilled with how it performs.

    It certainly doesn't seem to run out of breadth and just keeps going high up into the rev range. WOT is 12:1 or richer, so I am going to test some different main and air cor jets next.
    Philip
     
  17. fivebob

    fivebob Formula Junior

    Jan 31, 2004
    254
    Tauranga,New Zealand
    Full Name:
    Callum
    The Techedge works the same as the Motec/Precision/FJO/.... they are just controllers for the wideband sensor, the NTK UEGO sensor that comes with the Techedge can be used with the Motec but I prefer the Bosch LSU as it has a faster reaction time.

    As I said before they are all O2 sensors, the do not as such measure HC/CO etc concentrations but rather measure O2 availability/demand. http://www.megasquirt2.com/PWC/LSU4.htm has the best write up I could find on how they work and explains it far better than I could.

    From what I know about tuning ECU's (I've tuned about ten cars) the Motec tech's explanation is indeed correct, and I'd add in the possibility of charge dilution via reversion (ie exhaust gasses being sucked back into the cylinder/intake).

    The man from Norwood's advice is probably the best I've heard in a while, some people try and complicate things by insisting everything is tuned by instrumentation, when the seat of the pants dyno is sometimes a better method, especially for driveability.

    BTW can the Techedge sensor be logged (preferably against RPM) I'd be interested to see a trace of AFR for partial throttle transition vs WOT.
     
  18. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Callum, thanks for the Megasquirt reference. Reversion is an issue at low RPM. When Continental were testing the car, Gary (SM) indicated you could see the "exhaust" from the carb trumpets flowing upwards at low RPM, then it would change direction and wham, off you go.

    From James's advice, I've upped the initial timing (to about 15 BTDC at idle) and slightly opened the throttle plate screws (it was wanting to idle about 650 - 750 rpm when hot) to get it closer to 950. This should also help draw a bit more vacuum at idle.

    James also opined that the wideband A/F sensor will be a useful tuning guide at mid-range and up (as I'd guessed/hoped).

    It is "not bad" on transition considering the cams (P6) and venturis (44 DCNF and 36mm). Low down you have to roll into the throttle otherwise it'll fall flat on its face, but about 3000 you can stomp on it. There is no hesitation at 3800 (typical cross over point from idle to main influence) with these idle jets, although there was with the 60s or 65s.

    On progression, I am running an F36 ET and have F30s in the box of jets. The Comp Dino ran F6s, but I believe the transition holes in the carb body were different so I am not sure I'd draw much conclusion from this. Gradual throttle openings it is still rich (and the F30s will lean it) but I want to focus on the main and air correctors first.

    On constant throttle, mid to upper RPM I am running rich (typ 12's:1) and have opened the air correction jets to 175s from 165. My guess is I'll end up at 185 and will also reduce the main (currently a 150).

    Stomping on it, the accel pumps seem to be well sized and it is going v rich (11.5:1). Again, I'll address this once I have the mains and air correctors more in the ballpark. I'll be at Road America in two weeks and will get some good data then.

    Thanks for the inputs.
     
  19. snj5

    snj5 F1 World Champ

    Feb 22, 2003
    10,213
    San Antonio
    Full Name:
    Russ Turner
    Interestingly, your final jetting is looking like mine.With 40 DCNFs and 36mm venturis, I'm running a wee bit rich (12s) at 145/190 with 55's and F24s (original dino ETs). Am planning on leaving it a tad rich for the new 257 (vs stock 244) intake cams due in next week, finally. the new cams have a broader ramp up and intake area, so am expecting across the range increased requirements. I have a set of 57 idles (!) already here in case. As an aside, it 'feels' happier on the lean side with 140 mains.

    Looks like I will only be running up to 30 overlap or less.

    Did you need to port your intake manifolds for the 44s? I think I need more airflow.

    great thread
     
  20. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    I port matched (new) fiber "spacers" between the intakes and the carb bodies. Interestingly, one side needed significant taper, the other little if any taper to align with the carb throat, indicating that the carb body does not sit on precisely the same bore center as the intake manifold.

    Russ, 42s or 44s might work well for you too, given the ~5% difference in displacement.

    As an aside, I think the PO ran the 44s with 160 mains, 70 idles and 190 AC, F30 ETs with 44s/P6 and using 34mm venturis. I'd guess this would produce a very rich condition through peak torque and lean out somewhat from there. Hopefully the WB will assist informed choices. More to come.
    Philip
     
  21. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Russ
    One more thing. I believe (no data, just SOTP) the 44s give a lot more "oomph" than the 40s above 4000 rpm
    Philip
     
  22. snj5

    snj5 F1 World Champ

    Feb 22, 2003
    10,213
    San Antonio
    Full Name:
    Russ Turner
    I believe it.
    On the dyno at WOT you can feel it wanting more air as the hp curve begins flattening out close to 6500 and holds it as if wanting more air. I'm hopeful cams will help that a bit. Have monitored the trumpets and have not sensed any reversion - but then I have very short overlap compared to P-6s.
    So for 44s, would have to port the spacer. What are you using for trumpets, and does the stock airbox work?

    Oomph is good. Of course with a street application, I need to keep the air velocity up for oomph below 4000 rpm - and the 36 venturi on the 40 has no ill effect at all. :)

    Many thanks
    rt
     
  23. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Russ
    I think you are feeling the cam profiles. I think even the early 308 cams start to run out of gas at 7000 rpm (my 260s did -- whereas the P6s typically produce peak power at about 8200 rpm). Dema will have perspective on the "top end" of your cams.

    On the trumpets, I polished those that came with the 44s - visually they are similar to the 40 DCNF unit.

    Ferrari UK has the spacers.

    I am running the stock air box. While most of the testing has been done without the filter, box or rear deck. The ram effect helps, although I note all the Michelotto cars had both sides feeding the air box so maybe there is something to be gained...Of course, then there is headers and a straight through exhaust to comtemplate...
     
  24. pma1010

    pma1010 F1 Rookie

    Jul 21, 2002
    2,559
    Chicago
    Full Name:
    Philip
    Most books and tuning manuals I have read indicate tuning for an A/F ratio of around 13:1 (12.8:1 to 13.2:1) being the typical range.

    The Weber tuning manual shows a graph of HP, A/F ratio and RPM indicating the mixture should be enriched towards the early and late part of the RPM range (idle jet, AC jet).

    By contrast Guy Croft's book on tuning Alfa engines describes having the engine run rich in the mid range and leaning out to close to 14:1 at the upper end of the rev range. Many of these observations are for FIAT/Lancia engines with peak torque around 5000 - 5500 and peak power around 7000 rpm. Similar to a stock 308. See image below.

    See: http://www.carobu.com/html/308gt4_euro.html for an example of the 308 dyno curve.

    By contrast, a P6 cam'd 308 will typically have peak torque at 6000+ and peak HP at about 8200 rpm.

    See: http://www.carobu.com/html/308gtb_dry_sump.html.

    Those of you with multiple dyno runs under your belt, how are you setting up your A/F ratios across the rev range? Specifically, at peak torque and at peak power?

    Philip
     
  25. snj5

    snj5 F1 World Champ

    Feb 22, 2003
    10,213
    San Antonio
    Full Name:
    Russ Turner
    Here in the first graph are some reference dyno numbers on a 3.2 qv, paper air filter, 34mm venturis, stock US headers, Euro resonator, tubi - the first set show how differing A/F from differing jetting do not make a large impact on output as long as they are close. The jetting difference is that one is with a 145 and one is with a 140 main jet.
    The second lower graph is a riched up jetting on stock cams showing variation of A/F with a larger 36mm venturi, same car, with greater output from more flow.
     

Share This Page