Crank Twist: How Much?

[MadBill]

As the title asks, anybody have any actual data (or even informed speculation) relative to normally aspirated V-8s? I've read that TF/FC engines wind up their cranks as much as twenty degrees (!), which if true, would translate to ~ 2 degrees for an 800 HP engine with a TF crank. Presumably lesser specimens are somewhat more rubbery. Just to be clear, I'm not talking torsional vibration, I mean the crank is a torsion bar and will twist in accordance with its moment of inertia and the location and magnitude of the applied torques.
As a back up plan (or maybe regardless, just because enquiring minds want to know), I will twirl one with a big torque wrench...

[SchmidtMotorWorks]

As a back up plan (or maybe regardless, just because enquiring minds want to know), I will twirl one with a big torque wrench...

That's been on my list of things to do when I have spare time for about 15 years, please let us know what you find.

[MadBill]

Maybe I'll set up a poll when I'm ready to go, see how close everyone is...

[bill jones]

-why couldn't someone with a dyno paint a TDC scribe line for #1 TDC on flywheel to a pointer hanging from the rear of the block #1.
-Then use a common timing light and check both ends at some low rpm where the engine is unloaded----then get there and both ends during a good loaded dyno pull.
--------------------------------------
-I was involved with a dyno project in 1980 and we tested every cylinder on a common 350 SBC oval track engine and there was enough to measure---not over 4 degrees but 2 degrees and some change---between the 4 crank throws.

[MadBill]

Yes, we need to do this! (as well as the static test, just for comparison)With all we've got planned, I think we could have this engine about worn out before it ever hits the track.
Hmmm.. perhaps a couple of crank triggers or angle encoders would be better than sending in someone with a timing light to cuddle up to the mill at 8,000 RPM.... (Seems like something that must have been done before, but...)

[Greenlight]

You can't check crankshaft twist on a dyno without having a lot of special equipment. There is a lot of information available on crankshaft twist. I kind of doubt a Top Fuel crank has 20 degrees (4 or 5 sounds more reasonable). If it did it would only last one or two runs before it would break from "reverse bending", like a paper clip does when you repeatedly bend it back and forth.

I found a few articles on the net that have good information. There is a lot more out there.


This article has a graph of harmonics and twist.
http://www.popularhotrodding.com/engine ... _windsor2/


This test has a Formula 1 engine showing crankshaft twist.
http://www.vold.com/PDFfolder/Twista.pdf


This article states that if there is more than 1/2 degree of twist "this" crank will fail.
http://www.dalemfg.com/harmonicbalancer_008.htm


Another article showing less than 1 degree twist.
http://www.powersportaviation.com/Home/ ... esting.htm

[MadBill]

Nice data mining, Greenlight!
Good info, but as you go through the details, they are really all about torsional vibration, which occurs at high frequency. The torque-induced twist I am talking about will be relatively constant at WOT, and torque reversal will occur only once per 1/4 mile run for a TF car or automatic multi speed transmission vehicle, so the 1/2 degree fatigue limit isn't relevant.
I'm still going to check this out myself!

[Brian B]

As a back up plan (or maybe regardless, just because enquiring minds want to know), I will twirl one with a big torque wrench...[/quote]

Don't waste your time>Did it 9 years ago!
First I would like to say,,you know that can of worms
you have setting to open on a rainny day? Dont
do it!

I tried to block this month out of my mined but,,,
I built a lot of 2bbl motors stock type parts always looking
for more.
The big torque wrench deal is easy to do its the info you
get that is the killer!
All you do is take a block bolt it to a table or something
put your main bearings and crank in "lock" the rear
of the crank and go to you local truck or CAT shop
get their Largest torque wrench. Find or make a reducer
to fit your crank socket and have a couple of guys help
you. Just have the front rod journal so you put a dial
indacator on it and read it. At 300 to 350 foot pounds
it was like .020-.030 on a cast SBC crank. That is from
locking the rear flang to the #1 rod journal. We did check
each journal but I dont remember #

So then what do you do with the info?
Change the "swing" of the crank?
Have the cam lobes "time" to each cylinder?
Change the "phasing" of dist?
How about ALL of the above!

Just my 2 cents
Brian B
Dakota Parts Warehouse
The Ferrea Valve Guy

[MadBill]

Hey, thanks Brian! At least you have some data to share!
My first thought was to figure the engine's peak torque, get a wrench that big from a truck shop, bolt the crank down exactly as you describe and set to. On further pondering though, I realized that for a torsion bar (which is what the crank is acting as in this case) the twist depends on the length being twisted. Therefore, I calculated the torque per cylinder pair and multiplied it by the distance of each from the crank flange. Totalling up the four, I then divided by the distance from the plane in which the wrench will act to the flange (don't have my data at hand, but call it 25") This gave the torque value at the damper which would replicate the cylinder pair totals (~275 ft-lb in this case). Once I have the twist results, I will calculate the relative twist at each cylinder pair location as a function of how far they are from the flywheel flange. <Edit>(I expect my number will be a bit smaller than yours, assuming you applied all 350 ft-lb. at the damper location.)

My plan is then to set up an individual-cylinder-adjustable spark advance distributor to compensate, and do some A-B testing, hopefully increasing power and reducing the tendency for detonation.
(With a little luck, cam twist matches crank twist and keeps the #7 & 8 valve events more or less in synch with 1 & 2.)

I'm tentatively set up to try it on Saturday; will report back one way or the other....

[Grocerius Maximus]

I remember reading somewhere, back in the good old days, the early BC days (Before Computers) how they used to measure rated output and shaft power on big ships was to measure the amount of twist of the prop shaft down in the bowels of the engine room. Might be bad memory but recall figures of over 360 degrees were seen at flank speed. Any old Navy guys here?

[Rick360]

Wouldn't the amount of twist depend on which cylinder is firing?

You don't have even torque applied on all journals at the same time. When a front cylinder fires, the crank would twist more during it's power stroke since it is applying torque further from the flywheel. When a back cylinder fires it is so close to the flywheel there would be very little twist and the front cylinders would be taking power (pumping and compression etc) which would try to twist the other direction toward those cylinders. The total amount of twist would change a lot throughout 2 rev's. There could hardly be any "steady" twist, but there might be some consistent dynamic twisting that could be accounted for in timing of cam or ignition.

Rick

[SupStk]

Wouldn't the amount of twist depend on which cylinder is firing?

You don't have even torque applied on all journals at the same time. When a front cylinder fires, the crank would twist more during it's power stroke since it is applying torque further from the flywheel. When a back cylinder fires it is so close to the flywheel there would be very little twist and the front cylinders would be taking power (pumping and compression etc) which would try to twist the other direction toward those cylinders. The total amount of twist would change a lot throughout 2 rev's. There could hardly be any "steady" twist, but there might be some consistent dynamic twisting that could be accounted for in timing of cam or ignition.

Rick

Rick, That is the same thing I mentioned to one of Brian B's guys when he flew the idea past me. As far as I know the idea of grinding the crank with increasing index offsets on each pair of rods wasn't tried or tested. As Brian mentioned, if you have the crank ground how far down the string do you do go to correct the other things that are now out?

[Ken_Parkman]

Some turboprop engines use a torque measurement system where there is an unloaded reference shaft through the center of the loaded shaft, and a set of teeth between the two shafts. More torque, more twist, and the distance between the teeth changes, which can be measured to get an actual torque.

I would think that the actual load and radius would result in an actual fairly small stress on the crank, and not much twist. But the deflections if the engine is operated where one of the important engine orders is in resonance with the crank could be huge, way more than the engine torque could manage. It's a different style engine, but I've seen resonance based deflections that if I hadn't seen it I would still not believe. We're talking huge here. Taught me tremendous respect for the power of vibration. Also why the effectiveness of the damper is probably the most important part of the crank twist story.

[MadBill]

...effectiveness of the damper is probably the most important part of the crank twist story.

That's what got me interested! The measured torsional vibrations (which don't seem all that big in V-8s, e.g: http://www.atiperformanceproducts.com/p ... /index.htm ) seem to have a big effect on performance. David Vizard tested a bunch, from flyweight 1 pound aluminum hubs to 8 lb. plus lumps and even at 900 RPM/sec or more acceleration on the dyno (which should have penalized the high mass unit due to its greater inertial resistance) found it made ~10 HP more. This confirms that crank twist is a significant power robber, not just a crank breaker, and I can't see how low frequency variants wouldn't also be a problem.
To re-state: There will certainly be high frequency torsional vibrations due to inertial and firing impulses interacting with the resonant frequency of the shaft (which is of course what the damper is there to minimize) but they will be superimposed on a static twist caused by the torque input by each connecting rod

[MadBill]

Oh yes: As far as the cam is concerned, you could always put a secondary cam drive on the back of the crank, and force the back cylinder valve events to follow the lagging pistons...