Increased Vibration with Hardened 1040 Steel Shaft

[Sewing_Engineer]

Hi,

I recently had a problem with our machines shafts where the vibration increases after heat treating our shafts. The initial test was done with a 1040 steel shaft that was not heat treated for the first prototype and had a hardened chrome plating. The production sample Shaft was done with a 1040 steel shaft that was heat treated to 700HV and had a black oxide finish the shaft was designed to only be heat treated 500HV but I thought it wouldn't be a problem with having the material harder then specified but looks like it is but I still want to figure out what is going on. After measuring the shafts total runout they are both with in spec. The only part changed in the machines was the shafts all the same parts were used and when running the machines overall machine vibrations increased. All of my shaft design calculations I did show there will be no problem with the shaft deflecting in any significant amount considering the loads on the shaft are relatively small. With increasing the shaft hardness I assumed their would be little change in the vibration since heat treating has only a marginal affect on 1040 steel modulus of elasticity.

I am wondering if anyone can help me with this problem I have been debating three questions I think could cause this;

1. How the harmonic speed of a shaft will be affected in changing just the material hardness.
2. If the softer steel would have a higher dampening value even though the modulus of elasticity doesn't change much and could be absorbing some of the shock load.
3. If the hardened steel center of mass could be off slightly with having the shaft so hard (I don't think this is very likely or could be causing a significant problem).

 

[Tmoose]

More detail please.
Drawings and pictures go a lot further than words alone ever will for technical communications.

I'm hoping there is MUCH more detailed information available about the "vibration." for your sake as well as mine.
Without detailed info about the frequency of the vibration I feel completely naked and afraid and helpless to resolve vibration (or noise) problems. As a start, if the predominant or problem vibration is not at the 1X rotating frequency of the shaft, the assignment of guilt may be misplaced.

The function of the shaft(s) in unclear to me. For instance, Do needle bearings run directly on the shaft? Does the shaft support a stack of gears and pulley or other components ? Are the ends of the shaft coupled to other shafts? If so the method of alignment and the design of the couplings can make even good shafts appear to behave badly.

 

[Sewing_Engineer]

The Vibration increases in both displacement and velocity when measuring it with an accelerometer in both horizontal and vertical directions. At 500 rpm's is where the displacement vibration is the worst with both shafts with the harder one vibrating more, but they run from 90 rpm's to 2000 rpm's the displacement of the entire machine is 1.0mm with the harder shaft and .7 with the softer one. When measuring the velocity with the harder shaft in the horizontal direction at 1800 rpm's the max vibration is about 120 mm/s where with the softer one it is 25 mm/s. The shaft is 12mm in diameter and is 639mm long. The shaft runs on 3 ball bearings with a load cantilevered out the front and a load from a belt between two bearings, see image below.

 

[GregLocock]

Have you done a Gage R&R study on your vibration measurements? You may find you have an assembly issue, for example rather than a component issue.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Shafiee.M]

Did you balance the shaft?
I sugest you to calculate the natural frequency of the shaft and then compare with the bearings frequencies and suport natural frequency.

 

[BrianE22]

Looks like it would be easy enough to strike the shaft with a hammer and measure the acceleration response on a scope. See if the hardened shaft takes longer to damp out.

Maybe all those set screw connections slip and slide and absorb some energy with the soft shaft.

 

[Sewing_Engineer]

I haven't done an R&R Study we are a small company and I sometimes overlook those things. I will have some people set up and run the tests again on some more machines. I have currently ran the test on just three different machines with similar results.

I have balanced the shaft a couple of times. first initially in my calculations, and second with the first prototype after measuring vibration and making some small modifications to the secondary counterweight. But I can go through all that and look at it again to see if there is something I did wrong or that I missed.

I calculated the natural frequency of the shaft which for the first mode was 72Hz and the support calculated was 60.19Hz. This was calculated with material data for the shaft assuming 1040 steel at 500Hv, but changing the hardness shouldn't change the frequency calculations (I don't think).

I will try to measure the acceleration response of the shaft on a scope sounds like a fun study.

 

[electricpete]

Setscrews may have a tendency to push things off-center when an assembly clearance is present.
I'm not sure, but maybe (?) this effect is less pronounced for a soft shaft (where setscrew embeds into shaft) than a hard shaft (assuming no change to setscrew or tightening torque).

=====================================
(2B)+(2B)' ?

 

[Tmoose]

Measuring "the acceleration response of the shaft on a scope sounds like a fun study."

You did not answer if and how you are processing the measuring the output from the accelerometer, although to get results as you report in velocity or displacement you must be doing "something."

I'm hoping you can go a few steps further and extract some frequency information.
If the predominant or problem vibration is not at the 1X rotating frequency of the shaft, the assignment of guilt may be misplaced.
Unbalance of the shaft especially.

In the jpg you provided it is not clear to me what are bearings.
What is the large gray and gold component over at the left hand end of the shaft?
Are there two long keyways cut in the shaft over towards the right end ? Machining keyways like that often causes even stout shafts to bend.
Additionally, I'd not be surprised if the hardening process caused an un-keyed Ø12 mm shaft to bend and twist and reach for the sky.

*** A check for runout at multiple stations along the shaft with a dial indicator is a necessity. Straightening a part that is 700HV/60HRC can be a challenge as well. I would opt for peen straightening over press straightening.

 

[btrueblood]

Your vibrations are measured at 1x rpm, or ...? Not clear how 2000 rpm (33 Hz) speed will excite the shaft at 72 Hz, unless some other unbalanced forces occur at higher than 1x rpm? Your shaft bending frequency calcs are made with all the attached hardware modelled as well (point masses or distributed masses?

 

[Sewing_Engineer]

To measure the vibration I am just using a Extech Vibration meter which does not give me any frequency information. I haven't been able to talk my company into buying a DAQ or similar frequency analyzer. Are there any 2+ channel frequency vibration analyzers you know of you would recommend for under $5000?

In the JPG the arrows shown below point out the bearings.
The large gray and gold component over at the left hand end of the shaft is a counterbalance.
there are a couple of flat spots machined into the shaft for locating parts but all parts are attached with set screws/locating screws.

We do straighten the shafts after hardening I will check with the manufacture what straightening method we use.

The vibrations listed above are measured at the max vibration points when running the machine at fixed speeds. The first order peak is at 500 RPM's (8 Hz) but that is with the rubber damper's on the the bottom of the machine if I take them off it shifts the 1X speed to 1260 RPM (21 Hz) but it runs a little louder so I have been collecting my measurements with them on since I added them.

my shaft bending frequency calculations are made with all the attached hardware modeled as distributed masses on the shaft.

When running the machine at each speed from 90 rpm's to 2000 rpm's in 5% intervals the max vibration is at 500 rpm's with both shafts but the amplitude is higher with the harder one. The overall amplitude at every speed is higher with the harder one then the softer shaft. I think I am going to just make the shafts softer but I would like to figure out why this is happening. Maybe it is an unbalance problem or I have an additional design error also. I definitely am going to spend some more time testing and do some more research. But any feedback or direction will help.

 

[Tmoose]

Hi Sewing Engineer,

1 - Which model Extech Vibration meter?
This one looks to be about the least expensive that offers readings in velocity and displacement ( and acceleration).
I'd check with Extech if they offer a unit that has a port to output the raw vibration signal, and if you could trade-in your meter towards that one.
Then you could output the signal to a PC based FFT analyzer.
There are some freeware/share ware ones, plus some quite reasonably priced ones as well.

http://www.spectraplus.com/PricingInfo.htm

2- Is the shaft
visible during operation?
3 - Is it visibly "whipping" during periods of high vibration.
4 - Is there real estate available to "rigidly" support a 4th bearing around the middle of the shaft?
Normally more than two bearings supporting a shaft is asking ( begging) for trouble, but by inspection this long, thin shaft might not object and respond with large radial bearing loads due to imperfect colinearity.

5 - Was the (third) bearing at the extreme right, beyond the pulley, part of the original design, or added to stabilize things after some testing?

Time for some pictures/drawings of the entire machine/general arrangement.

I think the rubber mounts are "isolators" not "dampers."

 

[Tmoose]

"We do straighten the shafts after hardening I will check with the manufacture what straightening method we use."

My concern would be the straightness of the shaft as it is installed in the machine.
So I would measure it in-situ with a dial indicator, and record the results.