Does anybody knows how to calculate the gear mesh on a worm shaft gear box?

Number of flights/teeth on worm times it shaft speed.

Walt

Its simple if we consider the driven gear,i.e. output rpm x no of teeth of driven gear.

If we consider the driving gear/worm,its the no of flights as has been pointed out.pl find an extract i had saved from csi sie.

-----------------------------------------------
Worm gears are often confusing because there is sometimes a question as to how many teeth are on the input worm drive. In the case of a worm gear, it is not the number of teeth that is of concern (often a worm drive only has one tooth) but the number of flights on the worm gear. The flights refer to the number of teeth that mesh with the driven gear during one revolution of the worm drive. This can be readily identified if the output gear speed, the number of teeth on the output gear and the input shaft speed are known. In this example an output drive gear with 24 teeth turning at 10 Hz is driven by a worm gear turning at 29.5 Hz. The number of flights (#F) on the input gear can be determined as follows.

(#T)out x (TS)out = GMF
(24) x (10) = 240 Hz

next,
GMF / (TS)in = (#F)in
240/29.5 = 8.13
--------------------------------------------
Another thing i want to know is whether anybody has detected worm /wheel damage ( input flight or output rpm sidebands around GMF in spectrum /waveform).Mostly its the softer worm wheel that gets damaged.


CSP

In my experience, I have seen little to no gearmesh vibration on these types of gearing devices. I suspect thats because the gear loads are fairly constant (unlike other type designs that tend to have more impacting loads when a defect exists on a tooth).
Has anyone else seen gearmesh frequencies for worm gears???

ACTT - If you suspect vibration at these frequencies for this type gear, please post. It would be very informative

Jim Powers

Thank you guys for your replay. I'm new to this forum and I greatly appreciate all the information.
The fact is, I have problems with a servo motor directly coupled to a single reduction,gear box with
worm shaft and worm wheel. I see what appears to be a coupling or mialignment problem, but I didn't know how
to rule out the possible gear mesh frecuency. Here I attached spectrums of a test running the motor at 1000 and
2000 rpm. The high 3X peak seen at 1000 rpm dramatically reduces its amplitude. Any ideas why this happened?

3X_vibration_on_servo.doc (130 Kb, 59 downloads)

The plots all seem to be from the motor and you have expressed concerns about the worm gear?

Do you have any data from the gear (gear number of teeth, flights, etc.)? Also, what type of coupling connects this motor to the worm gear?

John from PA

I thought that gear mesh frequency for a worm was the worm speed times the number of leads in the worm (usually 3 or less).

The vibration levels are fairly low.
Normally, high amplitude harmonics is a cause for concern when the levels are 10 times what I'm seeing here.....When levels are this low, it is quite possible this is normal.
Makes me wonder what your specific concern is

Have you noticed increasing trend?
Can you compare to another unit?
Do you have data from the gearbox? and do you see the same characteristics?
Is the unit causing problems?
etc. etc.

From what little I see here, I doubt you have a gearmesh problem.

Jim P

quote:

I thought that gear mesh frequency for a worm was the worm speed times the number of leads in the worm (usually 3 or less).

Yes, I agree with Danny.

We have a number of worm boxes, with either 6 or 8 starts (leads) on the worm.

If the lead-in on the wheel teeth has worn away, or the worm has been incorectly set to the wheel, then the gearmesh levels get quite high.

Ian

Dear IanS
quote:
---------------------------------------------
Yes, I agree with Danny.

We have a number of worm boxes, with either 6 or 8 starts (leads) on the worm.

If the lead-in on the wheel teeth has worn away, or the worm has been incorectly set to the wheel, then the gearmesh levels get quite high.
------------------------------------------------
1)I have never seen high vibration with sidebands in worm GB(also I dont monitor much of this type of GB).Can you pl attatch spectrum with worm gear mesh( worm rpm X no of leads)

2)Just to confirm whether Gear Mesh is input speed x no of lead/start and not input shaft x no of flight.Have you ever confirmed Input speed x no of lead/starts is same as output speed x no of teeth in driven gear.

I will be glad if my doubt is cleared.This theory that GMF =no of flight x Input rpm that I posted earlier is from one paper from CSI


CSP

Jim
I was asked to inspect on the motor and the gear box after two different motor shafts had been shear cut by
torsional strees within a few months, the motor running at 2000 cpm. The motor was replaced with a new one and again
broken shaft on the motor occurrred.
After the that, the motor was replaced one more time and then I was asked to inspect the condition of the alignment and
any other possible issue. I agree, the amplitudes are very low, and I do not have any similar machine to compare with.
I suspect there was some misalignment or mounting loosenees that was corrected before I was involved. This servo motor
moves a carrier on a stacker, and travels 15 feet in 4 or 5 seconds. The original ramp acceleration timeand decceleration
time were never changed for the last 4 years previous to this problems.The data herein attached was taken from the unbelted unit(unloaded), since it is not possible for me to collect data in 4 to 54 seconds (gear box ratio is 10:1).
The only thing which I had no explanation for is why the relatively (to other peaks in the spectrum)high 3X that appears
in the motor and dominate the spectrum with the motor running at 1000 cpm more or less disappear at 2000cpm. The coupling
is a shim pack. Thanks again to all for your input. Merry Christmas.

ACTT
Thanks for the recent post...it explains much more of the problem.
I cannot explain why 3X becomes less dominant at 2000 RPM (vs. 1000), but I assume you are more interested in solving the problem than understanding the reason for relative strength of 3X at different speeds.
Again, the levels are low and now that you've explained that the concern is shear loads on the motor shaft this data may not be very useful.

A few more questions:
1. You state that ramp times have not changed for four years prior to recent failures. Can you detail what has changed from the time prior to recent failures? Looking in detail here may be the key to solving problem
2. You state you took data on the motor "unloaded"....I'm not sure how this would help you understand problem. Can you explain? In retrospect, this explains the low vibration levels.
3. Can you take crude torque readings on the input shaft using a torque wrench? This may help identify if excessive loads in the reduction gear exists. I was recently involved in a problem where a torque-tender coupling (sort of a slip clutch) was tripping. Design personnel thought it was because of misalignment. Turned out it wasnt misalignment....simply that the new gear had much higher torque loads than one in service for years (their basis for selecting the torque spring settings)....duh. In any case, we were able to measure the torque using a torque wrench to confirm loads and solve the ROOT CAUSE.
As you state, they may have already solved the problem without knowing it....keep this post updated as you learn more....that way we ALL learn.

Merry Christmas
Jim P

CSP,

quote:

1)I have never seen high vibration with sidebands in worm GB(also I dont monitor much of this type of GB).Can you pl attatch spectrum with worm gear mesh( worm rpm X no of leads)

Attached are some spectrums showing a rising peak at 8x running speed. These are taken from a worm box input shaft with an 8 start worm, meshing with a 59 tooth wheel.

quote:

2)Just to confirm whether Gear Mesh is input speed x no of lead/start and not input shaft x no of flight.Have you ever confirmed Input speed x no of lead/starts is same as output speed x no of teeth in driven gear.

The attached spectrums also show that input shaft speed x no of starts = output shaft speed x no of teeth on wheel. This HAS to be the case. The gearmesh frequency is the gearmesh frequency, no matter where it is measured.

As I noted previously, worm boxes will give substantial GMF's if they are not set up correctly.

Hope that this helps.

regards,


Ian

The_data_above_comes_from_the_non_drive_end_input_shaft_on_a_worm_gearbox.doc (88 Kb, 35 downloads) Worm Box GMF

is the shim pack a 6 bolt item? maybe 3 X is coming from coupling, but i have no answer for it disappearing at 2000cpm.

ACTT,

If the motor shaft broke due to torque (conical break), then you need to measure the entire machine running event. The starting or stopping may be culprit and not the gear mesh. A more direct approach would be to measure torsional vibration (dynamic strain) on the motor shaft throughout the machine cycle.

Walt

IanS
Thanks for the data plots on GMF....
Do you know what was the cause of the high levels??
I'm trying to understand the mechanics of the GMF peaks. Were the shafts not alinged properly?

Jim,

quote:

Do you know what was the cause of the high levels

The data posted is from a gearbox that is still in service, so I am not sure about this particular example. However, in my (limited) experience, high GMF in a worm box is usually due insufficient/incorrect lead-in.

Apologies if I'm telling you something you already know; the lead-in refers to the position of the worm relative to the radial centre-line of the wheel. It allows an oil wedge to develop between the worm tooth and the wheel tooth, which is very important since the contact between a worm and wheel is all sliding motion (unlike an involute gear tooth, which has a large component of rolling contact)

If the lead-in is incorrect, there will be insufficient lubrication betwen the worm and wheel, and the initial contact between the worm tooth and wheel tooth will not be as smooth as it should. All this results in high GMF's

Trust that this makes some sense.

regards,

Ian

This message has been edited. Last edited by: IanS,

Ivibr8

I'm trying to get more details about what change prior to failures. Indeed, I'm more concerned about solving the problem and I'm assuming there's no way to collect data under real load conditions for this short duration application. I attached an additional picture of the motor-gearbox.

Walt

Please provide more insight about the torsional vibration test. How is it done. A recap: This motor
displaces the carrier for a stacker with a reocurring
cycle of about 4 to 5 seconds from start to stop, back and forth.

3X_vibration_on_servo.doc (1,152 Kb, 21 downloads)

ACTT,

I would suggest using strain gages and telemetry on this unit to see what the actual stress is on the shaft during the brief time it is running as well as the starting and stopping stresses.

Have you had the motor shafts inspected after failure to try and determine the failure mode?

ACTT,

If I were doing this project, I would use my Binsfeld strain-telemetry system to measure static and dynamic torque on the motor shaft. Strain gages can also be configured to measure shaft bending or axial thrust, as necessary. I would also use my ZonicBook 8-channel system to simultaneously measure the torque/bending strain, shaft speed, motor current, and vibrations (accelerometers) throughout several operating cycles. The frequent machine cycles are probably causing the high stress in the motor shaft. My suggestion would be to consider replacing the disk-pack coupling with a torsionally soft elastomer type. The measurements would clearly show any benefit to changing the coupling or other modifications.

Walt
w_f_strong@msn.com