SpeedTalk Store - Opinion Columns

Is Taylor a useful reference for racing engine design?

General engine tech -- Drag Racing to Circle Track

Moderator: Team

Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 2:50 am

Some people take issue with my dismissal of old books like Taylor as a useful reference for racing engine design.

Let's examine some of the contents and see what can be learned from it.

Kevin has provided links to some pages on crankshafts here:

http://www.crank-scrapers.com/Taylor%20MIT/

I quickly retyped some of it to make it easier to read, (the copies were difficult to read) and to highlight some points I want to observe.

It begins:
Page 492

Loads
Most designers plot diagrams of crankpin and main journal loads based on indicator diagrams and calculated inertia forces using rigid-body assumptions. The results are translated into nominal working stresses, which are held to limits dictated by experience with similar designs. As previously indicated such procedures are basically an application of the theory of similitude since accurate computation of loads and stresses is not possible in such a complex system.


Today we can accurately simulate the flexible bodies with dynamic loads (not useless static rigid bodies). This is a critical difference as this is what is required to deal with harmonic vibrations within a crankshaft as this is what causes them to break. The author could have learned more in a few days with a modern CAD and FEA software than what is printed in this book.

Hardness
250-275 Brinell (that’s about 24-28 Rockwell C)


Bad advice, a racing crank is more like a minimum of 38 C, I have made them much higher.

Crankshaft Stresses
Qualitatively speaking, the loads on a crankshaft result in stresses due to bending, torsion, and shear throughout its entire length. The complex geometry involved would make accurate stress computations impossible even if the loads were accurately known. In spite of these difficulties, however, much has been done toward rationalization of crankshaft design, largely by means of experimental stress analysis.


Again, we have very accurate and convenient tools to do this now. It seems that our present times tech was on the authors wish list.

Results of Stress Analysis

The most useful information in regards to crankshaft structural design comes from experimental studies of stress distribution in typical crankshaft samples or models subjected to arbitrary loads under laboratory conditions. Figures 11-35a shows the enormous effect of fillet radius on the maximum stress in a straight shaft and a simple crank. Given reasonably well-chosen ratios of major dimensions, fillet radius is the most important factor in crankshaft strength. The larger the fillet radius the better, provided adequate space is available for the required bearing lengths. Where bearing space is at a premium, undercut fillet radius are better than inadequate conventional ones. Fillets with non-circular contours are slightly better than fillets with a fixed radius but are not usually practicable except for large crankshafts.


Obvious to a 1st year vocational school machinist getting better grades than C.

An image shows the imagined stress concentration in a crankshaft, they are not even remotely close to reality.


Some highlights:

the stress was measured with an extensometer


This is a device that is intended to be used to measure stretching of objects, it was really only useful/accurate to use on flat surfaces, it was not a reliable way to measure something shaped like a crankshaft. Besides it tells nothing about the internal stresses. That's why the hand drawn stress images are so far from reality, they were just someones imagination.


the stress can be several times higher than that calculated by the simple beam and shaft assumptions. Fillet stresses up to five times the nominal torsional stress and three times the nominal bending stress.


To make informed decisions about the design of a stressed object, you need to accurately simulate the loads and see where the stresses are and how high they are and how they are distributed. The author is explaining that the best estimate they have is within multiples of reality. Today we can come within a few %, closer if you have the need and time.

The bevel shown in figure b was probably detrimental to strength at least in the all of the bored shafts,,,


This is the typical taper in a rod throw seen from the side of a crank, it might be concave, convex and have contoured shape today. Today we tune the stiffness of the crank with varying shapes in that area depending on many variables of the application.

Where stresses are not considered to be high, crank cheeks are allowed to remain in the as-forged or as-cast condition.


OK, good luck selling that. If China turned out a part with unfinished cheeks they would be ridiculed.

In order to reduce wear crank pins and journals are hardened by local heat treating of the surfaces either by flame or by electric induction.


Terrible advice.

Crankshaft design ratios:
Where it is required that the connecting rod be removable through the cylinder bore crank pin to bore ratio is limited to a value of 0.6


Different era.

Main journal to bore ratios 0.6 to 1.0


NASCAR ~4.2 x 2.0 = 4.76 (well outside the recommended range)

It is remarkable that the range of ratios is about the same for all engines listed.
It is also remarkable that these tend to be rather independent of specific output.


That might have been true back then when engines were at 1/2 HP pr cubic inch and operated at low speeds. We have a lot wider variety now and different needs.

Crank pin diameter should be at least 0.6 times the bore
Main journal diameter should be large than crankpin diameters.


So a NASCAR engine should have a 2.508 rod pin diameter?
And the main 3.0" :lol: :lol: :lol: :lol: :lol:
They run 45mm rod which is 1.77, way off.

So I think it is fair to say that most of the stuff on those pages was either obvious or wrong.
Someone wanting to gain knowledge would be much better off to spend the same time looking at modern parts and learning CAD and FEA at a community college.

Main journal length can be as short as 0.30 times the journal diameter when centrifugal loads are counterbalanced.


Wow. we have come a long way.
You do not have the required permissions to view the files attached to this post.
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby Kevin Johnson » Sat Mar 09, 2013 3:42 am

Replicate the analysis of the forms with modern techniques, ceteris paribus. That would actually be a very worthwhile project. You will need to hire some assistants to gather the materials for you. When you are done, approach MIT Press.
Kevin Johnson
Guru
Guru
 
Posts: 7107
Joined: Tue Nov 22, 2005 5:41 am

Re: Is Taylor a useful reference for racing engine design?

Postby barnym17 » Sat Mar 09, 2013 7:23 am

Time marches on, at one time it was considered scientific fact that 200 mph could never be exeeded in the 1/4 mile.Those back then did the best they could with what they had to work with and subsequent generations have expanded on their knowledge.Likely armed with the technology available today they would come up with the same results as todays engineers.However we should remember we wouldn't have todays tech if not for the efforts of these now somewhat outdated scientists.
A modern passenger car has computer power that nasa could only dream of during the lunar missions.
barnym17
Guru
Guru
 
Posts: 1036
Joined: Fri May 12, 2006 3:37 pm

Re: Is Taylor a useful reference for racing engine design?

Postby twl » Sat Mar 09, 2013 7:38 am

Personally, I try to avoid ridiculing historical giants in the field.
twl
Expert
Expert
 
Posts: 814
Joined: Thu Feb 21, 2013 11:13 am

Re: Is Taylor a useful reference for racing engine design?

Postby strokersix » Sat Mar 09, 2013 9:28 am

A bit surprised by reading this.

Best approach in my opinion is to study and understand Taylor and others in the context of their time. Then use that knowledge as a foundation under new analysis techniques.

Foolish practice to rely on virtual analysis (FEA, CAD, and the like) without understanding history and theory that came before. How do you know your FEA results are any good without checking via hand calculations or comparing to prior work?
strokersix
Pro
Pro
 
Posts: 216
Joined: Sat Aug 23, 2008 6:12 am

Re: Is Taylor a useful reference for racing engine design?

Postby MrBo » Sat Mar 09, 2013 9:49 am

Pretty sure it was this guy (Sonny Leonard) who called Taylor’s books his “Bibles” or somethig similar. He builds HUGE engines for IHRA Pro Stock.
I bought the books after his comments.
(I would have scanned pages for Kevin, but the books are too thick to fit in the scanner & get a good pic.)
http://www.sonnysracingengines.com/
"I promise you Sheriff, I won't throw one more rock... Didn't say nothin' 'bout no brick!" --Ernest T Bass
User avatar
MrBo
Guru
Guru
 
Posts: 1428
Joined: Tue Jan 12, 2010 5:59 pm
Location: Ontario Canada

Re: Is Taylor a useful reference for racing engine design?

Postby Dave Koehler » Sat Mar 09, 2013 10:34 am

I don't think Jon is saying that everything in Taylor is out of date. Just that time, tools and procedures march on and the section on cranks that he and Johnson are going on about is too far removed from today.
Dave Koehler - Koehler Injection
Fuel Injection - Nitrous Charger - Balancing - Nitrous Master software
http://www.koehlerinjection.com
"Never let a race car know that you are in a hurry."
User avatar
Dave Koehler
Guru
Guru
 
Posts: 3053
Joined: Mon Oct 04, 2004 10:19 pm
Location: Urbana, IL USA

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 11:03 am

Foolish practice to rely on virtual analysis (FEA, CAD, and the like) without understanding history and theory that came before. How do you know your FEA results are any good without checking via hand calculations or comparing to prior work?


We do much better than checking them against hand calculations. There is a continual development loop validating that the simulation results accurately predict what happens in reality. When these cod4es are developed it isn't just done by some guy with a computer and a bright idea. Someone with a problem to solve and deep pockets pays for directed development to add a specific capability to the software to simulate a phenomena that they have observed and quantified.

It is so reliable that this is possible to get right on the 1st attempt using common out of the box software that anyone can buy.

Seven minutes of terror.

http://www.jpl.nasa.gov/video/index.php?id=1090
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 11:05 am

MrBo wrote:Pretty sure it was this guy (Sonny Leonard) who called Taylor’s books his “Bibles” or somethig similar. He builds HUGE engines for IHRA Pro Stock.
I bought the books after his comments.
(I would have scanned pages for Kevin, but the books are too thick to fit in the scanner & get a good pic.)
http://www.sonnysracingengines.com/


Having actually made many racing cranks for Sonny Leonard myself, I can tell you for certain he does not follow that bible when it comes to the cranks he uses.
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 11:09 am

twl wrote:Personally, I try to avoid ridiculing historical giants in the field.


I'm ridiculing people that think the data is accurate or relevant.

If someone thinks it is relevant they should go make some 23 Rockwell C induction hardened cranks for 4.188 bore with 2.5 rods and 3.0 mains and beveled rod throws and see how that works out for them. :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: .
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 11:37 am

Dave Koehler wrote:I don't think Jon is saying that everything in Taylor is out of date. Just that time, tools and procedures march on and the section on cranks that he and Johnson are going on about is too far removed from today.


Yep.
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 12:03 pm

I think the same will be true for design of cam profiles.
There are hundreds of books and papers documenting methods that are obsolete or incomplete ideas but nothing that describes the methods used by the current tools that have been out there for more than 10 years.

People just don't have any reason to write books or papers explaining how to duplicate good, modern, commercial, development.
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby Erland Cox » Sat Mar 09, 2013 12:14 pm

If I look at failures of modern cars that where developed using FEA it looks like it could have been smart
to keep someone from the older generation of designers as advisors.
If I look at the knowledge that the students have that leave our technical schools today I can say that it is useless without experience.
Then they are put to design things that they don't understand and that can not be good.

Jon, I have an interesting project for you if you have some time.
Remember the jpeg to CAD to CAM thread.
All my life I have been told about separation of flow over the short turn.
To make it go away you pull the short turn back.
Is there really any separation over the short turn at all?
I believe what happens is that when the valve flows well enough its entire periphery the air stalls in the combustion chamber.
This makes the flow take only one path past the valve.
If we make the short turn flow worse the chamber will not stall but we are fixing the wrong thing and making two wrongs become a right.
You seem to have so much energy debating so I think you could use that energy to see if you can prove me right or wrong.

Erland
Erland Cox
Guru
Guru
 
Posts: 3010
Joined: Sun Jan 29, 2006 9:46 pm
Location: Lund in Sweden

Re: Is Taylor a useful reference for racing engine design?

Postby SchmidtMotorWorks » Sat Mar 09, 2013 12:21 pm

Erland Cox wrote:If I look at failures of modern cars that where developed using FEA it looks like it could have been smart
to keep someone from the older generation of designers as advisors.
If I look at the knowledge that the students have that leave our technical schools today I can say that it is useless without experience.
Then they are put to design things that they don't understand and that can not be good.

Jon, I have an interesting project for you if you have some time.
Remember the jpeg to CAD to CAM thread.
All my life I have been told about separation of flow over the short turn.
To make it go away you pull the short turn back.
Is there really any separation over the short turn at all?
I believe what happens is that when the valve flows well enough its entire periphery the air stalls in the combustion chamber.
This makes the flow take only one path past the valve.
If we make the short turn flow worse the chamber will not stall but we are fixing the wrong thing and making two wrongs become a right.
You seem to have so much energy debating so I think you could use that energy to see if you can prove me right or wrong.

Erland


What specific failures related to FEA? How do you know what analysis was done? Most things designed in CAD never get budget to have enough FEA done to them and even if it is done, accountants and manufacturing engineers often overrule the recommendations.

I agree about experience, working 6 months in a machine shop making stressed one of a kind parts will teach someone far more than reading a book. Some things you learn better by seeing, feeling and hearing the failure.

Make a sketch in paint of what you mean, I can test it.
Are you thinking lay the short side way back like 30 degrees since most of the turning happens at the seat anyhow?
http://www.ocmfg.com Aerospace Machine Work: Prototypes, Tooling, Molds.
SchmidtMotorWorks
Guru
Guru
 
Posts: 10144
Joined: Mon Apr 11, 2005 1:30 am

Re: Is Taylor a useful reference for racing engine design?

Postby Erland Cox » Sat Mar 09, 2013 12:44 pm

Just Google BMW rear en cracks.
I believe that someone designing parts for a car must have understanding about how a car works.
The Universities teaching them should have an auto shop with skilled mechanics.
Now it is like a bank clerk is asked to build a sub frame for BMW understanding only how to use a computer.
For any tests on the head to be meaningful we must use the head that was used in the previous thread.

viewtopic.php?f=15&t=30828

Just open up the chamber between the plug and the cylinder wall and see what happens.

Erland
Erland Cox
Guru
Guru
 
Posts: 3010
Joined: Sun Jan 29, 2006 9:46 pm
Location: Lund in Sweden

Next

Return to Engine Tech

Who is online

Users browsing this forum: 5speed1966, cgarb, crash, dwilliams, Google [Bot], HarleyInStereo, MikeD, more cubes, pdq67, racear2865, statsystems, Yahoo [Bot] and 37 guests