Why are con rods failing in modified Honda CRF450 engines?

[Camron201]

Hi, I am an amtuer engine builder working on offroad race motorcycles. On the Honda Crf450 engines I have been working on, I have been having a large number off con rods failing in numerous ways.
I have had the the con rod snap right below the piston with no sign of bearing failure. I've had the con rod shatter into over 10 peices with no signs of bearing failure. I've also had the bottom of the con rod break as if the lower bearing has seized, but do bearing damage.
I have been told that with other engine mods I have performed that there is too much oil pressure causing a hydraulic lock of my lower bearing at sudden increases of rpm , thus destroying the con rod with no bearing damage. So I tried a ''relieved con rod'' so oil could escape more easily, but have had the same issues. There are no signs of detonation on the piston. I am running a piston with a lot of clearence to avoid thermal piston growth and seize. And the piston is not hitting the head or crank flywheels. Engines have blown at all different levels of rpm.
The engine is running peak 13,700 rpms, Bore and stroke is 96mm x 62.1, Comp ratio is 13.5:1, With my modifications I am getting about 20% more HP than stock. Could this just be too much power for these con rods? Any ideas for solutions to my problem would be greatly appreciated.

 

[94JZA80]

what are the specs of the rods that have been failing? are they stock rods, slightly modified stock rods, or some sort of aftermarket rod? what material are they made of? what's its tensile/torsional/compression strength? without knowing those things, its tough to say whether a 20% increase in power over stock is enough to toast the rods. you may need to go with forged or billet rods to keep them from breaking up. given that you've seen no signs of detonation, and given that the rod bearings and wrist pins appear undamaged following rod failure, that's where i'd start. i mean i'd do some analysis first before dropping hundreds of $ on new rods, but i'd start with finding out all the properties of the failing rods. i don't know all the forumlae off the top of my head, but if you know the stroke and the rev limit, then in theory you can calculate the maximum acceleration of the rods, and along with each rod's mass you can therefore calculate the maximum forces acting on the rods using F = ma. knowing the materials properties of the rods would help determine whether they can handle the forces on them at the rev limit.

 

[Ranger Mike]

13700 RPM with 13.5 to one comp ratio...wow no wonder the rod gives
1. what is con rod material
2. is rod I beam or H beam construction
3. who made the con rod? Is it stock?

 

[mender]

Next set of questions:
1. Have you confirmed the engine output on a dyno?
2. What rpm is the peak hp at?
3. Why are you spinning it 2000 rpm past the design redline and 5000 rpm past the stock hp peak?

Rpm kills parts a lot sooner than an increase in power. Since you're not seeing bearing failure, I'd say you're simply pulling the rod apart due to the excessive rpm.

 

[94JZA80]

13700 RPM with 13.5 to one comp ratio...wow no wonder the rod gives

well i wouldn't make any assumptions just yet. after all, modern day sport bikes (at least the 600cc's) rev to ~17K rpm without con rod failure on a regular basis. and they're typically high compression motors. one thing we should keep in mind though is that, unlike the high compression, high revving 4-stroke motors of modern sport bikes, the OP's motor is a 2-stroke...i don't know how that might affect things.

i suppose i also should have asked what your specific engine mods are thus far. OP, have you bumped up the static compression, or is 13.5:1 the stock CR? if so, how much did you bump it up? also, is 13,700rpm the stock rev limiter, or did you increase that as well? and if so, by how many rpm?

*EDIT* - i see mender sort of answered my question about the rev limiter (red line is a close enough approximation).

 

[mender]

From the info I've seen, it's a four stroke, stock CR of 12:1 and cited max hp is at 8500 rpm with a peak rpm of 11,450. That may not be the max design rpm but that was the only number I found with a quick search.

11,500 rpm with 2.44" stroke gives a piston speed of 4700 ft/minute, max accel (the critical number for rods) is 5900 Gs. At 13,700 piston speed goes to 5500 and accel goes to 8100 Gs.

The extra load on the rod goes up as the square of the rpm increase, so going from 11,500 rpm to 13,700 is a 42% increase in the forces trying to pull the rod apart (tension). Sounds like it's working!

As a side note, it's the stroke that has no compression load on the piston (exhaust/intake) that usually kills the rod. Building more power (higher BMEP in the cylinder) actually helps the rod by pushing the piston down instead of having the rod pulling it down. Two strokes have a power stroke every time so the rods have an easier time of it.

Quick rule of thumb: too much cylinder pressure (usually from detonation) kills bearings, over-revving kills rods. Sometimes it's a little hard to sort out which happened first from what's left.

 

[94JZA80]

my mistake - i just googled "crf 450" quickly the other day and noticed a 2-stroke forum devoted to the crf 450 oddly enough. apparently i need to follow my own advice about jumping to conclusions lol...

 

[Camron201]

It's a four stroke 4-valve single over head cam. Ignition is stock, but in motocross/supeross there are plenty of times when the wheel leaves the ground allowing max rpm. There are also times when you can't shift right before a jump, so you just have to wind it out (over rev) to still make the jump.
Mods are: Port and polish
7 angle valve seat cut with newen machine
Cosworth piston, bumped up to 13.5:1 (12.1 is stock)
ASF process to transmission and moving parts
Aftermarket exhaust
Carb mods
Falicon con rod (used on a couple engines)

Everything else is stock. Stock rod is forged, I beam. I've used a stock rod with oil reliefs mod also. Falicon is forged, knife beam. On an inertia Dyno, stock HP was 51.4 at roughly 9,000rpm. And mod Hp was 62.1 at roughly 10,000 rpm. But the Hp curve is very flat on top, so there is still plenty of power past peak HP. I will get back to you guys with more rod specs.

 

[brewnog]

Agree about engine speed here, over-reving might do it...

 

[Ranger Mike]

Mender..that was excellent and i am writing that one down...you nailed it..
Camron201 you need to get custom steel connecting rods from carillo or crower

 

[Camron201]

Thanks, I've contacted carillo, but haven't pulled the trigger yet.

 

[94JZA80]

11,500 rpm with 2.44" stroke gives a piston speed of 4700 ft/minute, max accel (the critical number for rods) is 5900 Gs. At 13,700 piston speed goes to 5500 and accel goes to 8100 Gs.

a quick question...you calculated the total distance traveled by the rod by multiplying 2 times the stroke by the rpm, but that just yields the average speed of the rod. did you use this average velocity to continue on calculating the max acceleration that resulted in a max force of 5900 G's? if so, wouldn't the max force exerted on the rod actually be greater than 5900 G's, since in actuality the rod's speed is instantaneously zero at both top-dead-center and bottom-dead-center, while its speed is actually greater than 4700 ft/s half way between TDC and BDC, regardless of whether it is traveling up or down?

 

[mender]

Yes, that is the average piston speed, so peak is higher, roughly double. The exact speed will depend on the rod to stroke length ratio, as that changes the dwell time at the top and bottom of the stroke, and obviously the acceleration rate will change as well. For the extremes, an infinitely long rod will have a sinusoidal motion, whereas a rod that has the same length as the stroke (equally impossible) will have the piston at the bottom of the cylinder for 270 of the 360 degrees.

I confess that I'm lazy and just enter the numbers into my desktop dyno so I can compare trends. However if you want to get into it, here's a page with the calcs and a good explanation:
http://ftlracing.com/tech/engine/rsratio.html

Take a close look at the graphs and you'll see that peak piston speed (and 0 accel) doesn't happen half-way between TDC and BDC and the exact location of the crossover changes with the R/S ratio. Also, note the difference in shape of the accel graph; the shorter rod length has a dip in the piston accel. As the R/S goes to 1, that dip will go to zero because of the dwell time at BDC, showing one of the extremes that I mentioned above.

ETA: there is another factor that affects the peak accel but it has a minor effect. Piston pin offset affects the rod angularity slightly and offsets the peak accel from TDC/BDC but usually isn't considered.

 

[Phrak]

The solution is obvious. CRF means cam rod failure.

 

[94JZA80]

Yes, that is the average piston speed, so peak is higher, roughly double. The exact speed will depend on the rod to stroke length ratio, as that changes the dwell time at the top and bottom of the stroke, and obviously the acceleration rate will change as well. For the extremes, an infinitely long rod will have a sinusoidal motion, whereas a rod that has the same length as the stroke (equally impossible) will have the piston at the bottom of the cylinder for 270 of the 360 degrees.

I confess that I'm lazy and just enter the numbers into my desktop dyno so I can compare trends. However if you want to get into it, here's a page with the calcs and a good explanation:
http://ftlracing.com/tech/engine/rsratio.html

Take a close look at the graphs and you'll see that peak piston speed (and 0 accel) doesn't happen half-way between TDC and BDC and the exact location of the crossover changes with the R/S ratio. Also, note the difference in shape of the accel graph; the shorter rod length has a dip in the piston accel. As the R/S goes to 1, that dip will go to zero because of the dwell time at BDC, showing one of the extremes that I mentioned above.

ETA: there is another factor that affects the peak accel but it has a minor effect. Piston pin offset affects the rod angularity slightly and offsets the peak accel from TDC/BDC but usually isn't considered.

thank you for that link with all the excellent information. as you can see in my first response, i was a bit lazy myself b/c i only explained how to apply calculus to a function describing the vertical motion of a connecting rod of specific length and stroke, but didn't actually try to find that function...at first i thought to myself that it might be sinusoidal in nature. but its interesting to note the slight deviation from a true sinusoidal function in the crossover points on the acceleration graph. in other words, it is interesting that max velocity on the downstroke slightly deviates from 90°, and that it significantly deviates from 270° on the upstroke...i kind of assumed that max velocity would simply occur at 90° and 270°, but intuition can be misleading when it comes to these things. piston and rod movement is obviously more complex than i originally thought.

 

[mender]

but its interesting to note the slight deviation from a true sinusoidal function in the crossover points on the acceleration graph. in other words, it is interesting that max velocity on the downstroke slightly deviates from 90°, and that it significantly deviates from 270° on the upstroke...i kind of assumed that max velocity would simply occur at 90° and 270°, but intuition can be misleading when it comes to these things. piston and rod movement is obviously more complex than i originally thought.

The offset from 90 and 270 will be the same amount towards TDC as long as the piston pin is centered in the piston. If you cut the graph in half at the 180 mark and flip the 180 to 360 part upside down, it should overlay the 0-180 part of the graph.

This kind of stuff is why I love engines!

 

[pete1623]

The rods have out lived there useful service life at these loads. The answer to this problem is to install an hour meter and replace the rod at a shorter service interval. I have built many of these motors. A CRF450 @ 60+ HP the dependable service life got very short. These motors are utilizing lots of handed down F1 technology.
Titanium valves are stock on these models.
Honda recommends inspection and/or replacement of pistons at ten hours.
The pistons are slipper skirt and have to be run at .0015 piston to cylinder clearance.
The stock CRF450 piston skirt is only .828" tall by1.386" wide!
Real light weight. But a very short service life.
The days of a whole dependable race season on a motor are over.
Pete