how much crankshaft twist? 1

[jstein]

We race a roots supercharged motor. After dynoing we noticed a small amount of timing(1 degree)made a huge difference in power. Now with the ignition signal triggered from the front of the crank and the load at the back of the crank and a firing order of 1-8-4-3-6-5-7-2, what cylinders would receive the greatest amount of twist and about how much? With individual cylinder timing there should be some power there(?) How is this twist related to the rpm? Thanks

 

[Metalguy]

Maybe that's why Chevy drove the dist. from the rear, huh?

After cyl. 2 and 1 fire, you can assume the timing for cyl 8 will be advanced with your setup. Think about it.

 

[jstein]

thought about it, and putting it at the un-driven end of the cam is a poor location. I would think that the ignition signal and distibutor(old school)should be the same unit, on the crank. Not seperated by gears, pulleys, bushings,and components that twist under load. Yes/No?

 

[ivymike]

the crank twists too...

 

[patprimmer]

The original SBC distributor is driven of the back of the cam, but the cam is driven of the front of the crank, so the dissy only sees cam twist, but the cylinders see crank twist.

No matter where you drive it from, if there is say, 4 degrees twist in the crank then there will be 4 degrees variation in timeing wether the front is advanced or the rear is retarded.

Crank twist will change with torque output.

Cam twist will change with rpm.

You can eliminate losses from chain stretch, gear backlash, end floats tec by running a crank trigger, but you still get the twist, and it is variable.

The only accurate way would be to have 4 cam drives directly off the crank and 4 crank triggers for a cuniform crank on a V8, one trigger and drive at each big end journal.

I know, it cant be done without a complete redesign with this taken into consideration. i imagine the design would be very much like 4 V2s or 2 V4s strapped together.

Not exactly a viable solution



Regards
pat

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[enginesrus]

Well that is what harmonic or dynamic balancers are for.
I have an old chart for a small block chev, that shows
a 12 order harmonic at about 1800 rpm that is at .2 degrees double amplitude. And a 4th order harmonic at about 5300 rpm that is at 2.0 degrees double amplitude. Both those figures are with out a damper. Now with a damper the 1800 rpm harmonic is about .1 degree double amplitude. And the 5300 rpm harmonic is at about .3 degrees double amplitude.
Which shows that if properly tuned you can reduce alot of that twist.

 

[GregLocock]

Nice observations. You are aware that camshafts twist as well, aren't you? This can cause a measurable problem in the car.

I'm trying to get a handle on the actual crank wind-up seen, but failing dismally.

I'd /guess/ of the order of some degrees. Sadly the only way to work it out accurately is to build an FE model of a crank and then apply the loads. An upper bound estimate of the torsional stiffness of the crank would be a cylindrical shaft of the same diameter as the mains, but that would be very high.



Cheers

Greg Locock

 

[Metalguy]

Interesting thread-I forgot about the cam drive being at the front. So, let's see about somehow triggering the spark from the flywheel itself. Looking at the problem purely from a crank torque-wrapup point of view, the worst case would then be cyl.1 getting its spark a bit late, following the power stroke of cyl. 2. I suppose a purist could set up a system where the front cyls. were fired from the front of the crank, and the rears from the flywheel. Hmmm, wonder if I can patent that. <g>

 

[GregLocock]

It would be very easy to build a table in your EEC to compensate the spark for each cylinder, for a given throttle position, to account for crank twist.

Thinking about it again, we rarely see torsional vibrations of more than one degree peak on a crank, so I guess THAT is a better guess for wind-up.


Cheers

Greg Locock

 

[ivymike]

we rarely see torsional vibrations of more than one degree peak on a crank, so I guess THAT is a better guess for wind-up.

Sounds like a pretty good guess to me, even if it's nose-(ideal nose) instead of nose-tail (tail is close to a node anyway).

 

[Joest]

Many modern digital ignition systems allow setting timing of each cylinder individually. OEMs likely have the resources need to determine crank twist under various loading conditions. With this knowledge it seem that if the payoff was significant (i.e. racing engines) ignition timing could compensate for any mechanical shortcomings of a motor.

-Joest

 

[ivymike]

of course you'd probably need to worry more about things like cylinder-cylinder charging variation than about the effects of crank flex on piston position...

 

[Metalguy]

I'll bet that Jstein's crankshaft twists more with that blower than most other engines, esp. if it's a SB. Someone with a dyno and a bunch of instrumentation could measure what happens with high output engines-esp. on the front cyls.

It could probably be calculated using BMEP and the crank stiffness. I used to run a big tensile testing machine that could have been used to load and measure the front throw movement while the rear was fixed, but that was a long time ago.

 

[SMOKEY44211]

It's interesting to note that the GM 2.2 Liter 4cyl locates the crank trigger pickup in the middle of the engine behind #3 main journal.This particular engine does not make use of a harmonic balancer. I built a distributor for a small block chev that allowed me to adjust individual cylinder timing. I think one of the Jacobs systems has a provision to do this electronically. Normally aspirated I picked up about 15 hp with this method. Duplicated under boost conditions should yield even greater results.

 

[jstein]

I wonder now because of distribution variations from cyl. to cyl. (boost, timing, fuel)perhaps we should make more dyno pulls measuring the actual cylinder combustion pressure and adjust the timing to its peaks.This should make the twist a non-issue. Very interesting comments, thanks. I wonder if a balancer at both front and back would help things even more. Does the torque converter act as a balancer?

 

[patprimmer]

I wonder if the triggers in ignition systems are manufactured to maintain 1 deg accuracy. Production tollerances might add up to more than crank wind up.

I dont see the point of moveing the pick up point to another position on the crank as it just changes which cylinder is accurate.

for example if we got 4 deg consistent wind up, and we trigger of the middle of the crank, the front would be 2 deg adv and the rear 2 deg retarded, but this would detonate the front, so we would retard it untill the front stopped detonating, then the back is still 4 deg retarded.

You would need to individually time the cylinders by each crank journal location.

I expect how you could measure this is by inserting tranducers for a crank trigger system at each counterweight then reading on an oscilliscope under various conditions.

This would only measure wind up at the point where the trigger is triggering? and not account for torsional vibrations or variations at various points of rotation through the 720 deg cycle. Sorry if that statement was unclear, but I am having a bit of trouble clearly and concisely expressing my thoughts on this.

I like Gregs idea of programing it into the ECU for a road car. This might have benifits re emmissions and economy.

On further thought, if say 8 transducers were at 45 deg intervals on every counterweight, we could also get an idea about torsional vibrations.

I was a little involved in the promotion of plastic manifolds, and one major advantage is that plastics injection moulding is a more reproducable and precise process than aluminium sand casting, so plastic manifolds if designed correctly, have the potential to reduce the cylinder to cylinder charge variations.

Too bad about the ports and exhaust manifold though. I don't see to many manufacturers going to 5 axis cnc ports at this stage, nor precission jigged mandrel bent exhausts with uniform fireing order distribution in the collectors, at least for V8's. On a V8 this requires a flat plane crank or a few crossover pipes.

I think I have already done to much to hijack this thread, so I will butt out now. End of rave

Regards
pat

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[GregLocock]

I'm not quite sure, but had a feeling that we optimise spark advance for each cylinder, in production. There is definitely some sort of adaptive timing strategy, that takes the spark up to the point of knock and then retards it a bit, I'm just not sure whether that works on the timing for the whole engine, or each cylinder. Of course, I may have seen it on a proto that never made it into production.

This would only apply at part throttle, I'm pretty sure we just use a lookup table for full throttle.

Cheers

Greg Locock

 

[evelrod]

On the current crop of the vintage BMC-Leyland &quot;A&quot; type engines (three main/ 5 port head @ 1310cc) with optimised timing of each cylinder. Variations (frequently several degrees) are calculated and then verified on the dyno. Currently we are getting near the 160hp mark, corrected to norms. THAT is a lot for this engine. As to crank twist---well, if I can stop laughing---!!!

Rod

 

[patprimmer]

Rod

Re crankshaft twist, yes we did get a bit accademic and pompous, for something that is insignificant, and probably unmeasurable.

Greg

As I understand it, the purpose of a knock sensor is to allow the ignition to be pushed to just into knock then retard a smidgen. This process is repeated at regular intervals to keep it at optimum.

I expect that one cylinder would always be the first to knock, so the entire engine would be optimised to the results from the worst cylinder, unless the ECU only advances one cylinder at a time up to knock, then does the test on the next etc.

Even though Rods experience is with the &quot;A&quot; series BMC and it's well known inherent design problems, Rods is a race prepared engine, so many production variations will be reduced or eliminated, appart from those designed in like the siame ports.

Several degrees of timeing can have a significant effect on power output, and I would think would be worth chasing on a modern production engine.

Regards
pat

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[patprimmer]

jstein

Re crank twist and roots blowers.

Just say, the motor is a SBC producing 1000 HP

To do this the parasytic losses from the blower might be 300 HP

Gross output 1300 HP, 300 being drawn from the front, 1000 being drawn from the back. journal 1/2 would be slightly retarded, 3/4 would be neutral, 5/6 slightly retarded and 7/8 more retarded.

The extra torque will produce more twist, but drawing some power from each end will at least partly compensate.

With the helix design of Roots blowers, the cylinder to cylinder variations in charge density and fuel distribution will be monumental vs the probably unmeasurable and insignificant effects of crank twist.

If you have an ignition system with individual tuneability from cyl to cyl, from what Ros says, I expect you could pick up some worthwhile HP by optimising on the dyno.

If you are running a points triggered maggy, then I expect this type of tuneing would be exceptionally difficult, and could only be achieved by running 8 single cyl maggys or making points cams with various timeing built in, then continually changing cams. Not what I would call a doable exercise

Regards
pat

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.