Originally Posted By: AdamC
ISO - What would you be looking for down the microscope, and what measurements would you take and compare? My hope is to have some basic instructions so the next time a box is broken I can take a look at the remains.
OK Mate
Let’s do some Field Level Gear Analysis for beginners, Part 1- the theory and preparation
* Not knowing your knowledge level I’m going to start at zero base and build up. Understand that this is built to a standard those in the gear world have to follow so the average guy may not want to go to every level and do everything but if I’m gonna train you it’s going to be done the right way.
Prerequisites
I would recommend googling “Falk Failure Analysis Gears-Shafts-Bearings-Seals” doc.# 108-010 (free) Read this and study it.
You want the “final authority” go to AGMA and purchase the ANSI/AGMA 110.04 Gear Tooth Failure Modes (That’s what I have to reference all my findings to)
You need a little background on what you are looking at and for along with the standard terminology. First time out, lots of pictures help a great deal.
Here’s what you need the microscope for… (A good surface finish gauge will be a great help too)
When I say microscope I mean something like a dyno lite (highest mag you can get) that is removable and you can surface scan with.
We will be looking at metal. Metal is unique in that it will hold the last failure state. (Assuming the spin down doesn’t grind it all away- this does happen)
With that said, you will be looking at the grain pattern, contact angles, type and direction of spalling, angle and depth of pitting, Plastic deformation, differentiating between abrasive and adhesive wear and quantification of Hertzian fatigue. (To start with)
The naked eye just cannot do this.
Background Info
(I’ll go into much more detail after you read that stuff and we move to part 2. We really need to be on the same sheet of music)
If you can’t get all this we can wing it but the more you know- the less you have to determine. (This is the short list)
A- Base Material: it’s important to know if this gear was born from a casting, stamping, machined from a drawn or forged billet. This gives you an idea of the stresses the gear was designed to handle. (I say this because I can almost guarantee you they will never release their design and load standards so you have to “guess”) If you can get the actual alloy type that’s beautiful too but not really critical at this point as much as knowing how the billet was formed.
B- Nominal Gear Dimensions: They will usually give you that and you really need to know where to start and reference to because the amount of wear is as important as the pattern
C- Run Factors:
If available the load direction (CW, CCW, BI) If you are not the one taking it apart someone needs to index that gear in terms of which way it turns relative to the input shaft.
Also need to know if this is a constant contact gear or intermittent contact
Is it machined on the shaft, keyed, interference fit (no key) and is it dog driven
Need to know if it is bottom or top lubed
The basics- the RPM, Load, Lubricant and if applicable any root or backlash clearances
D- Heat Treatment: For the moment just need to know if the gear was surface or through hardened and then was it normalized and stress relieved ( cold treated) afterward.
E- If possible, good pictures of the gear train from every angle (especially any gear or dog that touches the one in question)
That will get us going anyway
Tooling
(No human being can accomplish a proper gear failure analysis without the correct tooling so assuming you are the level of the average mechanic and the biggest gear you would ever see is about 6” diameter.) If you can’t get all of it we can wing it (to a degree) or find a mate who has them
The scope
Finish gauge
Caliper (with attachments)
Micrometer (the diameter of the gear) with attachments
Surface Block (if it had a dial indicator & stand that would be wonderful but not necessary)
Gear tooth gauge (if you can’t get this then a machinists angle gauge)
White inspection light with magnifier
Advanced (if you really want to get down and dirty)
Get a low end NDT dye kit from someone like Grainger.
Go to a hobby store and get a good quality plastic casting resin. (You won’t always need this but having an inverse 3D casting to work with can be a life saver)
Assumptions
You have to just assume the gear was designed and built properly- if you legitimately suspect this then you are beyond failure analysis and conducting an engineering analysis. Honestly I have never seen or heard of one that was not built properly. (Which is different that the application engineering being wrong when you put a properly made gear in an application beyond its design)
Critical Indicators
You asked in your OP how to differentiate between lubrication and stress related failures. That’s the most basal question in the gear industry and the first one out of every client’s mouth. That’s what we are going to accomplish and it comes from a proper analysis of the critical mechanisms of failure as they relate to the failure mode.
Initially we will examine:
Loading- will tell us deflection, alignment (running) stress vectors and working contact path and LOA
Alignment- all degrees of freedom and with shafting, other gears etc.
Distress/Stressors- Dynamic Forces
Surface Wear- qualifying and quantifying all the pitting, spalling, micro, macro, and a whole lot of stuff
This will tell us where we need to focus.
Let me know when you are ready to start the process.
It would be good to get an old gear just to practice on.
