Mechanical Shock

It depends where you measure the acceleration. All components will see a different acceleration depending on their mounting stiffness. The mass of the glass will affect results greatly.

See Instron.com for shock testing equipment

Are you able to say what is the failure mode? And which direction of shock is causing the failure - it could just be that the additional weight of the thicker glass is just too great for a particular direction of shock.

If the shock is only in a direction where this is not significant, it could be that you are saved by plastic deformation.

Or there may be an internal component that just happened to be sensitive at the resonant frequency of the version with the glass plate.

If the saving grace is plastic deformation, the problem may reappear at low temperatures.

Correct me if I am wrong. I am no expert in this field but i remember studying something about energy and crack. for crack to propergate, you need to input a sum of energy great enough to break the link of atoms. for a soft or more elastic material which can easily be deformed, the energy from the impact loading is absort to perform the deformation first in the elastic region. Hence there is not much energy left to propagate the crack. That may be your case because plastic certainly absorb more energy than glass.

Thought experiment:

Drop a glass marble and a plastic ball of the same material (Lexan polycarbonate?) onto a hard surface (ceramic, steel, etc.)

Which one will bounce higher?

I am sure everyone will say glass, because of experience. Plastics, even very crystalline polymers, just are not as stiff and resilient as glass, and so absorb some of the energy rather than returning. If you can find Co-efficients of Restitution for the material of the balls colliding with the material of a hard surface (as above) you will see their relative performances. I am not sure what the effect of thickness would be when the diameter is many times the thickness, as in this case. My thinking is that the difference of 8 mm versus 4 mm would not be significant. You could try stacking two plastic discs together (provided that parallelism was good enough to insure good contact across the entire surface) and test that. You might also compare the Modulus of Elasticity (Young's Modulus) of the two materials.

First to be clear, the direction of shock that failed is in the z-direction, that means a knock (mechanical shock) through the glass+metal, alum stand then to the player. The test is to make sure that when reading a disc at track 1 and 20 during playback, the screen should not freeze after the knock. Another observation, the freeze does not occurred immediately after the knock, it occurs a few milliseconds later.

Physicist, does that means if we reduce the weight of the glass, we should be able to meet 5G/3ms? Talking about resonant frequency, where can I find this info on different materials for reference?

zichau, I have the same thinking that the plastic absorb the energy from the mechanical shock better than glass.

STL Engineer, we have tried stacking 2 plastic discs together and the result is the same, means it also passed the 5G/3ms so 1 plastic disc can do the job.

Now, without replacing the glass, is there a way to improve the result to above 5G/3ms? That is always the product people wants because glass looks more presentable than plastic.

I'm assuming that the shock is applied via the mounting, not directly to the mechanism. In that case, the mass of the glass is not (of itself) the dominant factor - it is more likely to be its rigidity. There's clearly also a resonance effect, for which the increased damping of the plastic will be useful. There are presumably asymmetries in the structure that convert the vertical shock to lateral (which could upset the control system for the head). So I think that we've arrived at the point where we need a diagram of the mounting if we are to say anything more sensible.

fyz

"First to be clear, the direction of shock that failed is in the z-direction, that means a knock (mechanical shock) through the glass+metal, alum stand then to the player. The test is to make sure that when reading a disc at track 1 and 20 during playback, the screen should not freeze after the knock. Another observation, the freeze does not occurred immediately after the knock, it occurs a few milliseconds later."

Yes, one would expect the shock test to be through the stand. The object being that vibrations below a certain expected level transmitted through the floor to the stand and up (z-direction) will not cause problems. As Physicist observed, there must be a translation of the z-direction shock into a lateral direction that would affect the tracking of the laser reading mechanism, unless the player has a fail-safe device that locks up the laser if the z-direction distance becomes too small, since the laser usually "floats" over the disc, and contact with the disc would be harmful to both. The observation that the freeze does not occur until a few milliseconds later is irrelevant, this is a natural delay between read time and display time due to processing time of the data by the electronics. There may even be buffering involved, where the data is intentionally held longer.

"Physicist, does that means if we reduce the weight of the glass, we should be able to meet 5G/3ms? Talking about resonant frequency, where can I find this info on different materials for reference?"

Let me give my $.02 worth here. I do not believe reducing the weight of the glass would help. My guess is that it would actually cause more problems, since there is less mass to absorb vibrations. Possibly, increasing the mass of the glass could help, but not much, unless you increased it dramatically, say, by an order of magnitude or so. It is not for no reason that old-style LP turntables were often mounted in heavy bases, sometimes even a rock, like granite. For the same reason, industrial inspection fixtures are often mounted on a heavy iron or granite bed, to reduce vibration that would interefere with the sensitive measurements. Also, resident frequency is more a function of the dimensions of the object, as well as the materials, and NOT materials alone. Think of a glockenspiel or similar instrument that uses tuned steel blocks that ring when struck. Each block has the same material, but a diferent size that will resonate to produce different tones (audio frequencies).

"zichau, I have the same thinking that the plastic absorb the energy from the mechanical shock better than glass."

Hey, that's what I said!

"STL Engineer, we have tried stacking 2 plastic discs together and the result is the same, means it also passed the 5G/3ms so 1 plastic disc can do the job."

Kinda what I thought.

"Now, without replacing the glass, is there a way to improve the result to above 5G/3ms? That is always the product people wants because glass looks more presentable than plastic."

If you can get a way with it esthetically, I would suggest a thin film (sheet) of some softer polymer between the glass and the metal plate. It could be clear or colored to match the color of the metal plate, then it would hardly be visible. Possible materials could include vinyl (PVC), Polypropylene, or some of the Thermoplastic Rubber (TPR) materials. TPR's are especially flexible (both physically and feature-wise), allowing blending of colors and various durometers (hardness). Think of rubber mats which replaced felt on old phonographs, helping to absorb vibration as well as keep a better grip on the record. This should provide sufficent absorption of vibration to pass your test.

Alternatively, a spray on rubbery contact adhesive might work just as well. Try applying the glass to the metal with the adhesive in between, both after fully curing (drying) and before curing to see if it makes any difference.

Good Luck with your project. Let us know how it turns out and the make and model when it goes into production!

I have some doubts that a thin plastic interlayer would make a great deal of difference. I would have suggested thick plastic with glass coating (like a laminated car windscreen), except that there would be problems if the edges were exposed - so I don't think that is an option.

For MikeHoo: Does the player already exist without this base - and if so does it already suffer the same problem when not attached to this base?

I agree, a thin plastic interlayer would not make a great deal of difference, if it is a hard, rigid plastic. And perhaps my "thin layer" was misleading, I would suggest possibly 0.5 to 1.0 mm (.020 to .040 inch) or more. This should be sufficient if a material of low durometer is selected. The steel part could also accomodate a thicker "rubber" pad by counterboring one side, leaving a wall all around with a minimum visible gap.

I do not know what the steel part sits on, whether it has legs, feet, etc. or sits flat, but if the latter, it might also accomodate a "rubber pad". These are often used for vibration insulation.

Good thinking. On a related subject, I doubt that the weight of the steel plate is too helpful. So couldn't the plate be hollow with a suitable 1-mm filling? Then the most suitable material (regardless of appearance) could be used - possibly a relatively soft viscoelastic material? (I do have a possible reservation with this - placing the compliant layer further down will increase the basic amplitude of any wobble - on the other hand, the lower resonant frequency might help).

Thanks for all your comments.

To Physicist: Does the player already exist without this base - and if so does it already suffer the same problem when not attached to this base?

I have tried this experiment by removing the stand and all the bases. The result is perfect - it passed the 5G/3ms. This is another situation that I do not understand why. The player is tilted around 75deg when placed on the shock table and there is an outer aluminum rim wrapping around the plastic housing. Is it because the player is more stable with the alum rim acting as the support base and the vibration from the knock is well absorbed by the plastic housing? But then, when the stand and bases are installed, it can passed at 3G/3ms only.

To STL Engineer: The steel base has already 5 rubber pads acting as foot for the palyer. But these rubber pads doesn't help.

There may be other reasons why the rubber pads do not appear to help:

1. They may not be taking very much or any of the load. This may be particularly true if the testing machine is clamping on another part, bypassing the rubber "feet" entirely.

2. The effect of the rubber pads is probably accumulative, that is, you do not notice that they are helping, unless if you remove some of them the problem gets worse.

3. If the rubber feet are small in diammeter, they may actually be a cause of the problem, if they are taking the load, because they would allow much more lateral movement that a large diameter pad would not..

Here is a suggestion. You know the old saying, one picture is worth a thousand words? Instead of feeding us information, a bit at a time, each time we make a suggestion, that might be obviously wrong if we knew more about your problem, why not post a few photos or a drawing of the player and test set-up? That way it would be far easier to diagnose your problem.

So how to post the drawing for discussion?

Mike,

Depends on what form your drawing is in. Most CAD packages allow you to convert or export a file (about the same thing) as a JPEG or BMP image file. I believe either one can be added to a posting using the CR4 Message toolbar and clicking on the Insert/Edit Image icon that looks like a small green flash camera. This gives you two choices, either type/paste the URL if you have the drawing on a publicly accessible webpage, or browse your computer/network to upload the image file. I am not sure if CAD files will upload directly. Other files may work also, like PDF or Word, but I have not tried that. I do know that JPEG, BMP, and GIF will all work fine.

If your drawing is on paper or you cannot get the file to convert to one of the image formats there are two options. First, if it is a CAD or other graphic file that will not upload to CR4, print it on paper. Next, whether CAD print or hand drawing on paper, you can scan it into an image file with either a dedicated scanner or one of the new multipurpose machines (print/copy/scan/fax). If one of those is not available, use the highest resolution digital camera you can find and take a close-up photo of the drawing. Dowload from camera to PC, then upload to CR4. Photo will usually be in JPEG format already.

Thanks Guys. Appreciate the discussion we have. As an engineer, I would like to understand more and help to solve the problem. But at this moment, I cannot reveal too much on the product so this is the end of the discussion. I will come back again when there is another problem to solve.

Looks like another example of "tell me how to solve my problem, but I really can't give you the important information that would let you solve it".

You can't have your cake and eat it too!

In this situation, without a drawing, it is like taking shots at a target in the dark after being spun around three times, and you say, "Don't shoot in the direction of my voice!"

I now think it could easily be the mass of your new base and the compliance of the rubber feet of the player that cause the problem. Glass is generally much denser than plastic. Whatever you do, the assembly will vibrate around its centre of gravity. If that is in the player itself, the movement will be small. If the mass is dominated by the base subassembly, all the movement will be in the player. Use soft coupling such as rubber feet between the player and the base, and the movements will be large - and probably at just the wrong frequency for the control system of the player to tolerate. As a first move, I would mount the player as rigidly as possible on the glass, and provide feet under the metal base that are stiffer than the total assembly by the same amount as the assembly is heavier than the player. The further you can spread the feet, the better the results should be.
Incidentally, if I understand your test set-up correctly, it could well bear little resemblace to practical post-shock conditions
Finally, I would reiterate STL's recent point (and mine in #6) that we could be for more helpful if you could provide drawings of the assembly and the test systems (with masses defined).

If you have this shock only from time to time, type in Google: CD anti-shock buffer