Following a previous thread on this board, I've recently had a number of magnetic stainless steel discs manufactured which I'm currently mounting on assorted non-magnetic machines (mostly motors) to assist data collection. The discs are parted off from 30mm dia ferritic stainless steel and are fitted using Devcon Plastic Welder. I've found that wet wipes are an excellent method of dirt removal prior to applying the adhesive, easily cleaning off grease and solidified paper stock. Also handy for wiping your hands afterwards.

Does anyone know how this particular adhesive reacts to water/heat/chemicals/etc? We have a fairly wet environment and I'm concerned about the discs dropping off. I currently have a couple of old bearing races stuck together and sitting in a bucket of water to see if the bond deteriorates.

Also has anyone experienced any problems with the motor fins becoming blocked up with stock due to the discs, and causing overheating? It's a concern of our electical guys.

Thanks Alan.

Haven't used Devcon so I can't speak to its effectiveness, but I can't see where blocking one or two fins is going to cause the motor to overheat. Assuming no stock on the motor, they shouldn't affect its performance at all. Assuming a lot of stock on the motor--well, that little disk is probably the least of your problems.

Patrick,
I used it a good bit when I was at Weyerhauser in Sav Ga. I would usually use a spot facing tool in a cordless drill (during outage due to rain forest effect). This left the surface free of paint and rust. I would then apply the Devcon and the disc.

I would be wary of attatching a disc to the painted surface as it may become unattached. The failure would not be the devcon to paint attachment but the paint to the metal attacehment. Are you following me. In other words the heat and wet did not seem to effect Devcon at all.

Patrick,
A word of caution with the devcon. It has a shelf life and once it passes that shelf life it is about as useful as bubblegum on the bottom of your shoe.

If it is yellowish in color coming out of the tube throw it away because it will not harden. It should be whitish in color.

I have gotten Devcon from grainger and other suppliers that was yellow the day I got it.

You are best of gettin it from a supplier that gets the shipments directly from the manufacture to you.

CTC is one such supplier 800-999-5290

Perhaps I should have said, "I haven't used Devcon because I like Loctite." It has the same limitations as far as shelf life, surface prep, etc.

Loctite Depend has worked well for me, but I echo the comments regarding shelf life! Practically the day after the marked shelf life, Depend is useless.

Also good comment about painted surfaces. The adhesion to the paint is one issue...but a bigger one is quality of vibration data. The best case is as close to metal-metal contact as possible, with the thinnist possible glue line. A thick glue line and/or many layers of paint actually act as a spring and change the response characteristics of the accelerometer.

Jon
Spintelligent Labs

I've just taken a look at my first set of data since fitting the magnetic discs. I've noticed a significant rise in the spike energy level on all machines which have a disc fitted. Is this because the glued discs actually provide a better vibration path than holding the sensor in position by hand, or is something else going on? The velocity readings have all remained normal.

You've increased the mounted resonance so you're getting a better response at the higher frequencies. A flat magnet on those flat discs is better (as far as mounted resonances go) than a dual rail magnet. If you stud mount them your only limitation is the accel itself. The better the mount, the more 'oneness' there is between the accel and machine--in a zenlike sort of way.

I wouldn't expect the velocity readings to look any different if there's nothing significant going on around Fmax; but, you can check by taking an acceleration reading and comparing it to readings before the disc. You can probably go out to 5 or 6kHz now and trust the data. Or if you have a dual channel box, put one accel on the disc and hold the other one next to it and look at the difference.

I mounted a number of magnetic stainless "disks" today with Depend adhesive and just have a couple of observations.

If you are going to be all over the plant, an 18V cordless 4.5" grinder with a "stringer bead" wire wheel works very well. For tight places, try the Dremel cordless with the Lithium-Ion battery and a little "flap wheel".

When using Depend, even if you hold the disk firmly for 60 secs as recommended, it still may slide on a vertical surface. A big piece of duct tape will hold it in place until it sets up. If the glue doesn't spread out on the disk easily, clean the disk more thoroughly. (I used the Dremel flap wheel). At first I thought the tube of Depend was bad, but it wasn't.

Just some tips....

I am plagued with the little thin disk people have glued on to take vibration readings so I will knock them off or avoid them. I don't want a little thin disk in an enviroment where corrosion will make for looseness. Generally I don't have a problem with a magnet mount 90# pull and only use a probe where heat is a problem for the most part.

I'm not a huge "disk" person, but my customer wanted these.... I've trained him to take his own readings so he can determine on his own when to clean fans.

As I said, they are magnetic stainless, so I don't believe corrosion will be a problem. Typically, the disk is going to give you a better surface than is available on the machine itself. Properly done, you will get more consistent data using the disks and a good flat magnet.

That said, I don't think they are worth the cost to install them for most applications. I am a big fan of the Pareto Principle (80/20 rule), so I'll generalize and say that probably 20% of all measurement points would benefit from the use of the disks, and for the other 80% a 2-rail magnet or short stinger will work just fine (again, properly done).

Environment: the disk aren't the problem, the base and 'glue' is. The bonding agent will deteriorate or so it is here. I come across many of these little disk and they seem secure until you make an effort to knock them off - they come off easy for the most part. And it is appaarent that they are only held secure in 'spots'.

I have the following questions for the board:

1. What is the minimum radius (if any) of the surface a disk should be glued onto?

2. Once disk is installed, does flat magnet improve accel data in comparison with a dual rail magnet?

Thanks.

The adhesive used definitely makes a difference. "Superglue" works great for a short time (weeks), but can be very handy when trying to attach a mounting pad to a wet location. Just wipe off as much water as possible and stick it down! It will be ready to use in a couple minutes.

For permanent installations, we've had excellent results with Loctite Depend. Anywhere it's not hot, it will last for years and provide reliable service if the initial surface prep is well done. Other two-part structural adhesives also work well, although I don't have any specific recommendations. Lord Versilock was always an excellent option, unless you happened to look at the MSDS and planned to live more than a few more months.

A flat mounting pad will work fine on round surfaces with a radius of more than a couple inches if a gap-filling adhesive is used. Superglue is not a good choice for this, but Loctite Depend works pretty well.

Jon
http://www.spintelligentlabs.com

Commercial grade dental cement works very well on non-flat surfaces and it's an excellent conductor of high frequency vibration. Since it is very weak in shear, it's easy to remove a pad by striking it from the side. At one time, the Metrix Instrument Company sold it for use with their acoustic emission devices which, as far as I know, they no longer market.

Dental cement? I'll ask my dentist on my next visit. That got me thinking, though--when I had a crown put on a broken tooth, he used a UV light to cure the cement. I wonder if something like this would speed installation and help avoid the discs sliding off.

I agree with Jon that surface prep is critical. I take the machine surface down to bare metal, knock off any burrs or high spots, and "rough up" the adhesive side of the disk. I use a spray degreaser, wipe it good with a towel, then 'contact cleaner' (electrical), and then the adhesive. I put Depend on the pad,spray the surface with the 'activator', and then spray just a little activator directly on the adhesive. I hold it tight for 60 secs., then slap a piece of Duct tape over it, pressing it down good all around.

I don't have a huge number of these in the field, but some I've had in place for over 7 years with no failures.

Just because you can knock or twist a pad off with ease, doesn't mean you aren't getting good (enough) contact, and thus, data. Where I want to remove a pad, I use 2-part strain gauge cement. By design, it will give you very good frequency response, yet it's shear strength is pretty low as you can knock it off or twist it off rather easily. (I learned about strain gauge cement from Art Crawford)

But if the pads fall off, or come off when inadvertently bumped, then the installation is sub-standard and the data suspect.

The time required to properly install a pad, in typical field conditions, is what makes their use expensive, and - IMO - only suitable for special situations. An exception might be an "in house" PdM program, where you sometimes have more people than work.

"I've noticed a significant rise in the spike energy level on all machines which have a disc fitted." - Big Al

I've wondered about this myself.... my question is, does the high frequency response, especially as seen in the acceleration waveform, have any thing to do with the 'amount' of surface area in contact with the machine. Might this account for differences seen with a 'stinger', a 2-rail magnet, a flat magnet, a 1" flat magnet, a 1.5" flat magnet, etc. Anyone ever looked into this?

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

Speaking of pads falling off, I was once asked to guaranty that the pads would never fall off under any conditions!

Was the customer being unreasonable? Probably not since the I was working on a bridge crane in a NASA clean room...where they were considering going up to get the Hubble telescope mirror to repair it! The headline "Mirror shattered by falling test equipment" wouldn't be the best way to get publicity!

Pads with safety tethers....now that gets expensive.

Jon
http://www.spintelligentlabs.com

quote:

I would usually use a spot facing tool in a cordless drill

KellyB.... good idea, but why would you then glue on a disk? I have the spot facing tools, and I could set the pilot depth very shallow and just 'machine' a flat surface pretty easily in most places. This would be a lot faster than attaching pads.

Why wouldn't I want to do it this way??

Why does the base of the transducer have as much mass as it has? How does the little thin disk compare?

Here they were installed on 5000 HP motors and gearboxes; Horz above the split and randomly mounted by someone not knowing where to take readings. On CW machines they would be on the right side and CCW on the left or where ever and sometimes correct.

So with such mass and obviously plenty of room with plenty of exposed surfaces well worthy of data collection; why would you use a washer as a medium? Am I missing something here? But I also use acceleration.