How does factory determine redline?
 

I've always just thought of it as a line on the tach that I ought not pass. But who decides where that line is? And more importantly how do they make that determination?

I would assume that the early cars that had the rare (if ever equipped) factory option of titanium rods would handle a higher redline than stock. But if you paid the extra $1K for Ti rods, did you get a different tach than the rest of the Ts, Es, and Ses? I doubt it.

So how does an engine builder determine what an engine's redline is?? Or is it simply an estimate???

It's where they stopped throwing pieces 99% of the time.

Just a Wag.

its a math question and Ill get it started..the piston is always accelerating away from the fastest pt 1/2 way between tdc and bdc and as I understand it the engeers use this as the loading facter more than anything else for the generated forces mass and stuff.
Corky Bell" max boost" has a good simple ans in his book and is very good reading

often the valve train is the issue.

The Engineers set the redline at the factory by determining that if it was an Italian engine where would it blow up. They then added 15% and called it good.

The idea that Ti rods were optional on street cars is, IMHO, a myth that the <b><i>Car and Driver</b></i> crew decided to start just for laughs ... much like their promoting the '64 GTO 'ringer' that the Pontiac engineers prepped for bouts on Woodward Ave. ... as a world-beater 'production' car even though the test numbers were never duplicated by any other magazine, and certainly not by any stock cars in reviews or revisits many decades later!!! The fact that Patrick Bedard decided to promote the myth for 2.4 engines is laughable, considering that Ti rods weren't used in 2.8 or 3.0 RSR engines!

Redlines are generally dictated by the camshaft pofile and hp/torque curves exhibited by a particular engine ... not by destruction tests! Uncounterweighted 911T engines don't blow up at 7000 rpm, so quit worrying that some demon from the Black Forest is going to rear its' head and bite you if you rev past the redline on the tach on a pre-84 engine!

The 3.2, 3.3, and 3.6 engine with stock 9 mm rod bolts are a special case, and their redlines should be observed carefully!

I'm not certain about that. I have been told that the stock numbers for all reciprocating parts on a 84-89 Carrera are identical to those on the same years turbo. The turbo has a 7200 rpm redline versue the Carrera's 6200????

Warren, would you mind please clarifying the "doom" part of your above post? What happens if you violate redline in a 3.2?

TIA,

Dan

Per Frere, the redline on Turbos was 6950 rpm, and 3.2 normal engines had a chip-set redline of 6520 rpm, Club Sports were 200 rpm higher!

Many stock 3.2 engines have self-destructed in a rather dramatic way at DEs after extensive sustained speeds above 6700 rpm. The problem has been documented many times in BA's technical column in <b><i>Excellence</B></i> and at PCA tech sessions! The 'standard' recommended precaution to take with 3.2, 3.3, or 3.6 engines to be tracked seriously at high-rpm is to install ARP or Raceware rod bolts ...

With afew mods, you are supposed to be able to spin these engines at 8,000

Ti keepers, better springs & the famous rod bolt fix.

I suppose the factory could have used "failure analysis" ala HAL9000 and that would have revealed both at what RPM the failure occurred and what part failed. But I would think that it would be more involved than that. You could blow up a lot of engines before you uncovered a consistent pattern.

This is purely a theoretical question for me. Not something I need to know, but a question I've always had, the answer to which might teach me a bit more about engines - how they work and why they fail.

Aside from that, I'm bummed to think that the Ti rod option is a myth. Of course it doesn't make sense, but it was cool enough sounding for me to buy into. I wonder what other Porsche myths I believe???

How complex would it really be to build an engineering simulator to figure out what the failure point of an engine would be? I mean, ultimately, it comes down to the polar moment of a variety of small bits and pieces, spring rates and momentums for valves, error tolerances on a variety of rotational components ... wait, this sounds _way_ more complicated than I'm likely to figure out over my morning latte. Nevermind,

Dan

djm - simpler engine simulator programs do exist to predict engine capability (GT-Power), but none point out the direct failure point. That would be a much more complicated simulator that requires dynamic motion simulation combined with nonlinear structural/thermal analysis and maybe even CFD. Big, big computer running for a long time - OEMS don't spend that much time in one area. Simulations get you to a point, but eventually it's time to build and break it - and that's the fun part :)

Redline is typically where UNCONTROLED valvetrain motion takes over where the changes of catastrophic engine failure are much higher. High RPM CONTROLLED valvetrain motion is an art/science in itself with many, many variables.

Lighter valvetrain components and cam profiles are just where you start to bump up redline.

but there isn't any advantage at doing so since i doubt there are any cams available to give you the right power beyond redline
even S-cams loose effectiveness beyond 7000 RPM

the only advantage for readlining an S, is to get in the better RPM range when upshifting http://forums.pelicanparts.com/uploa...1089913050.jpg

at least that's how i understand it...

Holy smokes!.....

Even if you could figure out the theoretical failure point...you then need to consider the variability of the individual parts...from a statistical variability point of view.

Simple example: Take *identical* paper clips...made with the same material , same wire diameter, same quality control...same *everything* that is humanly possible to control...and do an experiment. Bend each one back and forth exactly 180 degrees and count how many flexes you have until failure. Big surprise ( to some of you anyway :) )--- there will be a statistical variation in the results...most likely in the shape of the classical "bell-shaped" curve. Meaning? A very few will fail very early...and a very few will fail very late...but most will fail in the middle/peak of the curve.

Upshot? Absolute prediction is impossible....can only be described as a a statistical probablility.

---Wil

So the answer I think I'm getting here is that the manufacture's redline point is probably based on something like 10% below the leading edge of the bell curve of failure points based on a simulation -with the likely failure points being the valve train and the rod fasteners. It does seem like the valve train would fail first.

For some reason I thought the answer would be more complicated.

That's just it..it IS complicated..

Lessee...how many failures...from a probability standpoint, will the manufacturer tolerate knowing that "X" percent will be claimed under warranty? Zero?...probably not...much too conservative. 5% (?)...even that may eat his lunch, and may be too much.

And all this assumes your "predictive model" is correct on all counts..that you can mathematically model the entire design/ construct process....

We haven't even begun to talk about the trade-off of performance and reliability....

It can't be very simple...IMHO...

---Wil

This is kinda off the subject....but how do those Formula 1 engines turn over 18,000 rpm's! Unbelievable!

somebody is out there who can tell you, but they'de probably be under orders from somebody else to kill if they actually would tell somebody.

well F1 engines as I understand are large bore and extremely short stroke motors...so they move alot of air with a small stroke length.

plus it is usually valvetrain issues that blow engines...F1's solution...pneumatic valve train so there are not chains and springs which can lead to "float" situations

I really doubt Porsche is redlining engines till they blow up...doing this 100 times and coming up with a redline...I am pretty sure they's have a materials specialist do some calculations and figure out the failure point of some of the components...then use a good % of saftey margin for the consumer. (plus I am sure they had real data of cars that did blow up during test runs and racing to help determine more accurate redlines)

Now a days it is most likely a comupter simulation model that calculates a theoretical failure point...then throw in some testing data...engines that were run hard and torn down inspected for wear and damage...voila your consumer redline

MJ

In my car Redline is just a suggestion ;)

Stijn - a race car builder will have cam profiles ground to their spec.

I'm surprised no one mentioned the Honda S-2000 - that's a high rpm motor.

Ford developed the GT-40 F- killer by running a bunch of engines until things broke - beefing that up and then running it some more, then beef up that part. Over & over again. R&T or someone had an article about how they unfairly stuck it to Ferr&*&i by dong things that way.

More recent R&T articles have focused on how to get rid of the valve train (solenoids, etc.) which is a major problem in getting high revs - as well as being a very crude way to control mix inputs to teh comnbustion chamber.

that i understand, but with our little engines, are there much folks that actually go far above 8K revvs? even with custom cams?

Most GT class PCA racers do go to about 8000 RPM as did the factory racers. To do this generally means adding Ti retainers and heavier valve springs. Although I got some 935 heads with stock retainers and stock springs. They may not have been original type springs though. The bottom end of a 911 up untill the 3.2l cars is usually fine until 8000 rpm. Many people add rod bolts and or aftermarket rods for an extra margin of safety. This way if you miss a shift the bottom end usually stays together. The valves just float and get bent.

Your stock 3.0 or smaller engine can all hold together till the 7200 rpm "S" redline and will probbaly hold together to 8000 rpm. The redline on these motors is based on the cam profile. There is no point reving these motors passed the stock redline because they "run out of cam". There is no HP above the redline.

I have heard of a few guys running to 9000 rpm. This always involves Ti valves at about $120 a pop. One of Elgin's (a cam grinder) customers did it, he had to change valve springs every season. The key to high RPMS is a light valvetrain. It is easier to change the direction of a lighter valve. Most of today's high revving engines use 4 valves per head. That means for a given application each valve is smaller. You will notice the latest RSR is pushing past the 8000 rpm limit a little.

I suspect that the factory went through the following process for setting the red-line.

1) Chose the rev range for the engine. For a given capacity, a higher rev range will result in more HP.

2) Select the appropriate bore and stroke to keep the piston acceleration within a reasonable rate at peak rev's. If possible use existing pieces to save tooling costs. At the same time design the bearings to ensure adequate bearing surface area without making them so large as to induce excess drag.

3) Select the appropriate cam profile, CR, chamber design and induction system to provide as much as possible the ideal air flow and combustion charactoristics. Doing this may result in step 2 being modified or possibly multiple configurations might be reviewed in parallel. Based on the cam's valve acceleration, valve train weight, etc springs will be chosen which keep the valves from floating but are not so stiff that they cause additional drag or long term wear issues.

4) Build and test the prototypes and try to identify issues. In some cases portion of the engine may be tested as a single cylinder engine. This may result in some changes or updates to the oil gallaries in the crank or case to address oiling issues. Wear issues in various sub-assemblies or pollution issues may result in the design also being tweaked.

Of course, in a company the size of Porsche, they will always being trying to use as many existing pieces or concepts as possible to minimize the tooling investments. My favorite example of this is the continous spectrum of development from 571 - 911 - 906 - 908 - 917 engines.

When it is all said and done, you will have an engine which is designed for a certain rev range (and red line) from top to bottom. Increasing the peak rev's will start to push different sub-systems beyone their ideal range of operation. The order that this happens will be different for each engine, but you can get a sense of it by what Porsche changed from engine to engine.

Funny ... 40 years later they do it the exact same way - ahh progress. :rolleyes:

Some CAE simulation was done as well, but only at the component level (usually after the part breaks the 1st/2nd time).

You haven't seen/heard anything like a V8 pushrod engine seeing 11,000 RPM and living to tell about it :eek:

Ford did CAE in the 1960's? Considering the state of the art in computers back then, that must have been some effort. I used finite element models for heat exchange modelling in the late 1970s and- was the first person in my field to do so. The comp. ctr. got mad and tried to find out who was choking up the Amdahl.... This was all in FartRAN and I had to write it all myself. A real PITA (of course if it was easy somebody else woulda beat me to it).

Only a huge company could afford the expense in the 60's - and off the shelf numeric analysis progarams were not available until the 1980s so they would have to have paid programmers also. Its all over the place now - I'll bet people will buy it from Toys R Us soon to give the kids something to play with...

I read it twice and come up withy this conclusion, he was refering to the way they drove it til it broke then made it stronger, just like 40 years ago. his post indicated that on the new model they did some CAE also, but made no indication that they did CAE 40 years ago. Is my impression correct?

Hmmm. well what we call CAE today (which uses finite element analysis, where a grid of points covers the thing and you define certain sinks and sources, flows, into and out of each point with a set of equations for each) might not have been what they called it long ago (as in the 1960's).