Ok, I have been listening to this sound for a couple a hundred miles and am stumped.

Best described as a roaring sound coming from the back, but does not occur at all speeds or driving conditions. Specifically, only when in either 4th and 5th and only at lower rpms (say between 2800 and 3200).

Additionally, I only hear it when holding at a constant speed. I.E. if I accelerate through that rpm range - no roar. Only occurs if I let off power, then re-apply power but only to maintain a constant speed. Consequently it only seems to occur while "cruising" in 4th and 5th (at the stated rpm range).

My first fear was tranny bearings, but wouldn't I hear it at all speeds?

CV joints?

please help, so far my only solution is to either downsift or turn on the radio

Is it a whine or a rumble? Check the tranny fluid level first. Could be the intermediate bearings are going. The top one goes first usually as it is the last to get oil if the tranny fluid is low.

Geoff

With the failure that your tranny had, it may be that one of the intermediate plate bearings is going bad. Or there is some excessive backlash in the transmission.

Is it a whining noise?


We can check it next Saturday when you come over.

fluid level is good, just changed it actually.

Clay, you know the sound reverse makes? Like that but, lower in tone.

What exactly is backlash? I am curious as this noise seems too occur as I am transistioning from accelation power to steady crusing power or decelerating.

Backlash 101.

Every gear in a 914 transmission other than reverse are "helical cut" gears. The teeth are not straight but spiraled. This will make the gears run quieter. But only when properly meshed. The disadvantage to them is that the driving gear will "climb" or "decend" the driven gear if there is too much backlash. That means that the gear is allowed to move in and out on it's rotational axis. This changes the relationship between the drive and driven gear. That makes noise. The drive gear will ascend the driven gear under power, and decend the driven gear when coasting.

What you are probably hearing is a gear lash problem. It may be due to the bearing running loose when the nut backed out on you.

It should not be fatal, but we need to listen to it next Sat.

thanks for the lesson,

see ya on the 29th.

Backlash is out much clearance there is between the teeth on meshing gears. When one gear is driving another, one side of the driving teeth are pressed against one side of the driven teeth. When the loading changes (such as letting off after accelerating), this reverses, so briefly the "driven" gear may drive the "driving" gear. The teeth will tend to bounce off each other, taking up clearance on one side, then bouncing to the other side, back and forth. The more clearance, the greater the forces involved. This creates noise, and can damage the gears over time. There is always SOME backlash, and always some noise, but in a "good" gearbox, the backlash is either small, or the noise frequencies of the various bits all end up mostly cancelling out.

The forward speeds in the 914 gearbox are all "helical-cut", with the teeth cut at an angle. This gives more tooth surface area for a given width of gear, which makes the gear a bit more robust in wear, and cuts noise as the backlash is naturally less. The reverse gear, however, is straight-cut. It has to be, as this gear, unlike the other five, is engaged by meshing and unmeshing, which can't be easily done with helical-cut gears (a diagram would help here). Straight-cut gears are much noisier, as the tooth lash is much less constrained. The whine is basically the gears ringing from the teeth bouncing back and forth.

If your forward speeds are making a similar whine, then it's either caused by a chipped or worn gear or two, or by a failing bearing. Rolling element bearings make all kinds of noises as they fail: whine, rumble, roar, squeaking. It depends on exactly what's causing the failure. The fact that it's only happening in the upper gears and only at certain speeds makes me think that what you're hearing is some resonance that could be caused by several things almost anywhere in the gearbox. The sound is always there, probably changing in frequency, but only under the circumstances you describe are the frequencies "matched" sufficiently for you to hear it. In other gears and speeds, the frequences cancel each other, or when then add up are still so low that they can't be heard.

What to do about it is another matter. I had a '99 Miata for five years and 70K miles that made a similar noise mostly in 4th from new. Several cars did it, and there were enough complaints that Mazda investigated and determined it was an unfortunate resonance, and they decided not to do anything about it. I have yet to hear of an actual failure caused by this. I simply didn't use 4th all that much, so it didn't really concern me. You may pull the gearbox apart and find nothing. Simply replacing one gear pair or a bearing may alter things enough that the noise changes (perhaps better, perhaps worse).

The last bit is backlash in the final drive. Again, due to resonance effects, it's still possible that it could be this even though it only shows up in certain gears. This is the spacing of the pinion to the ring gear, and is adjustable via paper shim gaskets between the intermediate plate and the gearbox housing. Setting this is a royal pain, but well worth the effort. These parts are sufficiently stressed that too much backlash will kill the final drive eventually.

in your case the backlash is variable due to the proposed slop in the bearings. if the bearings were to spec, you probably still have some small but measurable backlash which, in my understanding, is defined as total rotational play for one gear with the other one fixed (ignoring thrust for now).

jbb

edit: lapuwali beat me to this with the non-reader's-digest version.

QUOTE (Jeff Bonanno @ Jan 20 2005, 10:58 AM)

in your case the backlash is variable due to the proposed slop in the bearings. if the bearings were to spec, you probably still have some small but measurable backlash which, in my understanding, is defined as total rotational play for one gear with the other one fixed (ignoring thrust for now). jbb edit: lapuwali beat me to this with the non-reader's-digest version.

Actually, both you and Clay pointed out what I'd neglected to (despite my windiness): that a loose bearing is another cause of actual backlash problems (as opposed to the bearing itself being noisy). I also liked Clay's climbing analogy.