I just got a call from a client who said that he has a broken tooth and an axially displaced bearing on one of his big gearboxes.

The broken tooth is in the high speed pinion (herringbone) as is the displaced bearing.

More information just came in about the bearing. Evidently just the inner race moved axially. The bearing is a New Departure A5228TS.

This is driven by a 700 hp dc motor with roller bearings which is connected by a large gear coupling and is turned on and off by the operator at his whim. (One time I asked the operator when he was going to load the mill again and he flipped the switch like it was the overhead lights. I told him I needed the mill to be loaded to get the readings and he just turned it back off.)

My questions are:

1. How is the magnetic center set on a large dc motor with rolling element bearings?

2. Is it possible that (if the motor is not set properly on its magnetic center) the motor shaft could surge axially and either push or pull the high speed shaft, causing the results we have seen?

I have been covering this drive for about 10 years and there has been very little change in the vibration spectrum or the time waveform.

I am not about to start making calls based upon some little bitty sidebands around a little bitty gearmesh frequency without some corroberation from a timewaveform, or PeakVue and it wasn't there. Anybody see anything here that would have made them shut this down and look for a broken tooth?

_1banbury.doc (90 Kb, 56 downloads)

Hi Danny,

Just a few questions about the spectrum you posted.

If you would like, send me the machine's data in a small database and I'll look at it in the manner in which I normally look at data and see if I would have called anything.

This message has been edited. Last edited by: Ralph Stewart,

Danny,

Any PeakVue time waveforms?

Danny,

The motor should have ball bearings and the rotor should be located by the drive end bearing housing. The free end housing should have room for axial thermal expansion so that as the shaft grows the majority of the growth is absorbed towards the free end. The only axial force the motor should be able to put on the pinion shaft would be from an installation issue or maybe a coupling problem. If the unit has been in service for some time I doubt that the motor was responsible for this failure.

Is this unit mixing rubber? The banbury's I've monitored over the years have all been, I've not seen them in any other industry. The problem that I always fight when collecting data is that during normal operation the speed and load varies almost constantly. When the batch of rubber is DROPPED into the mixer the gearbox is immediately loaded and as the mixture softens the load decreases and most of the time speed changes. The fact that different compounds are harder or softer than others adds to the problem. I have found it very tough to get trendable data because of the variables of the process.

I don't see anything in the data you posted that would have caused me to call broken tooth, or for that matter any significant problem. If the inner race moved on the shaft and the outer race, and rollers, stayed in place I can see where it wouldn't take long to fail the bearing or gear. It could very well have been fine when you checked it June 1st and failed today. I would think that if the bearing geometry would have been so out of whack when you collected data you would have seen all sorts of impacting. Some failures don't happen slowly.

Ralph,

Do you have Version 5.0?

LOR is unchanged. The 2 mounds you refer to are 1 and 2 x gmf and the sidebands are spaced at mill speed. Good question about the sudden appearance of the sidebands. I have no answer right now, but will look into it.

Don't you hate trying to analyze data in a .doc file? I'll try to put this into a database and send it to you but right now my wife is not so patiently waiting on me.

Bill,

Of course I have PeakVue twf's I don't see anything in them, but will post some later.

Joe,

This Banbury mixes plastics and the load behaves just like you describe. The plastic is added in many different forms, but most are smaller particles rather than a big lump. I record the product mixes and will compare them for relative stiffness, but like you said, it depends not only on the mix, but when in the duration of the mix that I get the readings.

I am checking into the history of the motor because my theory is that the recently installed drive end bearing (within the last 18 months) put the motor out of position and the frequent switching on and off was tolerated through May, cracked a tooth between May and June and then it broke yesterday. I completely agree that if that bearing had been in that state on June 1st, I couldn't have missed it even if I was working in my sleep.

Thanks for your help guys. We may never really know the answer here, but this is my best client and he is not prone to blaming the vibration guy for everything. (Not everyone at the plant is so open-minded He is genuine in his concern to find out what went wrong and always avoids finger-pointing. I wish I could copy him and put him in every plant in the world.

Danny,
Are you sure about the bearings? That 5228 sounds like a double row angular contact bearing which would take the axial thrust you speak of. I have never seen a straight roller put in both ends of a motor and would think it very unusual to do so. Normally if a roller bearing is in the drive end, then a ball bearing would be in the opposite end. And this would be more for belt driven applications and not direct coupled.

Ronnie

Ronnie,

The A5228ts is in the gearbox and is a New Departure bearing that I believe to be a cylindrical roller bearing, although I am not entirely sure of that. Does anybody have any information on NDH bearings? I have fault frequency data, but no confirmation about the type of bearing. The motor bearings are 6224 ball bearings.

With herringbone gearing the thrust generated by one row is opposed by the other sdo you don't need any axial load bearing capability.

quote:

Originally posted by Danny Harvey:
With herringbone gearing the thrust generated by one row is opposed by the other sdo you don't need any axial load bearing capability.

What kind of coupling is between the motor and gearbox, and the gearbox and the load?
You are right that a properly installed herringbone gearset will not impart axial thrust itself, but it can _transmit_ axial thrust. Is it possible that a thrust from the blender/mixer/squisher thingie that this motor runs has traveled through the gearbox? Perhaps the bearing that moved just happened to be the weakest link?
One nice thing about a spur gear is that it won't transmit axial thrust. If the couplings have some axial give to them then this shouldn't be an issue, but it's a thought.

Danny,

I only have 4.9 right now. Sorry.

This message has been edited. Last edited by: Ralph Stewart,

Danny, what kind of condition was the gear coupling in? A locked coupling will not allow the pinion to float where it needs to (herringbone following the output bull gear).

Ran into a problem one time with broken teeth on a double helical. Bearing OD's and bore measurements were in spec, but for some reason when the gear box cover was bolted down tight, the bearings were held by too tight of a bore fit. This would not let the pinion float where it needed to. Possibly from some gear case twist on the base compared to the repair shop (perfect) base. What would be real interesting if possible and if time allowed would be to try and float this pininon in the bores with the top cover bolted down.

Inspect the teeth wear on this pinion closely. Does it look like one one side of the herring bone was carrying the load?

Danny, I would look closely at the "load" side of this equation. It's unlikely the motor would displace the bearing, unless it (the motor) came loose and locked the coupling up or something. But a serious load excursion will try and move everything attached... all that errant energy is trying to "go to ground" any way it can.

Don,

I don't know about the condition of the coupling but that is a very good place to look.

Rusty,

This is a double reduction gearbox. Do you think the load could transmit axially through 2 gearsets and 3 shafts?

quote:

Originally posted by Danny Harvey:
This is a double reduction gearbox. Do you think the load could transmit axially through 2 gearsets and 3 shafts?

Danny, depending how this gear box is set up. Is it like an extruder screw, you know once it loads up the output shaft will thrust in one direction. If so yes. The output shaft will thrust and load the thrust bearing. Being a herringbone, the other two shafts will follow and center themselves (teeth) to follow the bull gear shaft. Check the loaded side of the pinion teeth. If more or heavy wear is found on the side with the broken tooth, that means both sides of the herringbone were not carrying equal load. Something was preventing this shaft from floating where it needed to be for both teeth sets to carry the load. You'll be off to a good start if you find this.

You put a "shock" into the gearbox output and it's going to travel back to the high speed pinion which is the smallest and lightest link in the chain. I just can't see the motor or coupling (unless it's in really bad condition) breaking a tooth on the pinion. Of course, if the tooth broke and then went through the mesh, that'd be pretty severe too.... might even displace a bearing?

Danny,

Do you have the PV time wavforms? Just want to take a look there if possible.

Rusty,

I figured the Weakest link theory was at the heart of your and Don's idea and it makes sense if a big axial shock were to travel from the mixer rotors in reverse progression. I'll take a closer look at how the mill is built. The only way I can see it happening is through the coupling between the gearbox output shaft and the mill pinion.

Don,

I'm looking back for some photos I took of the gearing and I'll post whatever I find. I'm not sure about how much axial load the mixer rotors generate during normal operation. They have a pretty elaborate shape as I recall which could generate some. I'm pretty sure though that if a destructive blow came directly from within the mill, there would have to be obvious damage to the mill.

Mike,

Sorry, but I missed your questions. There are geartooth couplings between the motor and gearbox and gearbox and the mill gears.

Bill,

Here they are. Let me know if you see anything in there.

_1banburytwf.doc (114 Kb, 21 downloads)

Some more answers.

The mill rotors do generate considerable thrust loading but it is captured by snap rings, shaft shoulders and keeper plates so that if the shock load had come from within the mill, there would have to be obvious and catastrophic damage to the mill. Not only that, but the gearbox is connected to the "mill connecting gears" which drive the mill rotors and they are spur gears.

I have asked for inspection of both the couplings during the machining of the new pinion and shaft assembly. They also plan to replace all bearings in the drive. They will be retained for inspection.

I have some photos of the gears and I will attach them.

The first photo (in the last post) is of the high speed pinion and gear. The pinion is where hte break occurred.

The one attached here is the low speed pinion and gear. There is some pitting present and it appears that there has possibly been some axial loading. There is also some product in the picture which is hard to tell from the pitting. This gearbox has been in use here for about 8 years and was manufactured in 1983. I don't know how this photo compares to earlier ones (I don't have access to them but have asked)

Danny,

The A5228TS is a single row cylindrical roller bearing. The construction is similar to an "NU" type bearing. Dimensions are 140 X 250 X 82.6 I believe that NDH has been out of business for some time now. Torrington may still make an equal to this. Way back when, New Departune was the ball bearing end of NDH and Hyatt was the roller bearing end.

dj

Ralph,

I must apologize for making you think your eyes are going bad. You are correct, that is not 2 x gmf. It is about 55 x running speed with mill-speed related sidebands and I really don't know what it's significance is. It's mighty small, but interesting. I would guess that it is the natural frequency of one of the low speed gears which are know to have wear and regularly show up in the spectrum on the low speed side.