Control of Spring characteristics

[engtiuser]

We have an application where a helical compression spring is used in clamping application for an industrial robot. We found that the clamping load was dropped in about 30% in just 6 months. The temperature around the spring should not be higher than 70C. We verified the spring provided by the supplier is below stated load value given in the drawing. Unfortunately, there is no requirement given in the drawing to specify the relaxation requirement. I am planning to add the relaxation requirement (e.g. 3% in 2 years) in the drawing to make sure the next batch delivery to be problem free. I would appreciate if someone can provide recommendations for what kind of test the vendor can perform to substantiate the conformance.

 

[israelkk]

To check if the relaxation of 30% over 6 month make sense, can you provide the following spring details?
1. Outside/inside diameter
2. Wire diameter
3. Number of total and active coils
4. Type of coil ends (square and ground, etc.,)
5. The compressed height of the spring when clamped
6. The minimum force the spring designed to apply at that compressed height
7. Wire material
8. Any heat treatment (if any) that was applied to the spring
9. Any other operations done to spring before/after the heat treatment
10. Any other information you can provide

Specifying guarantied relaxation of 3% in 24 month requires the manufacturer to actually put the springs in the compressed state for 24 month before delivering them and check the results. It is impractical and will be very costly. If it was for a long time development project as is common in the aerospace industry, that approach is practical where, during the development, few springs will be compressed to the desired height and be checked from time to time for the relaxation. But firstly the spring needs to be correctly designed.

 

[dgallup]

If the spring is over stressed it probably looses the clamped load very quickly, first couple of compression cycles. Have you checked if between 0 and 6 months?

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[engtiuser]

The spring supposed to be preset 2 times at solid height for 1 minute. Then, the spring will then be shot peened 100% fully saturated on surface follow by powder coating. The spring is made out of AISI 5160 H steel. The spring is pneumetic driven and the force applied should not be over as there is pressure regulator and pressure relief valves in the system. With proper process in place, I have no doubt the spring shouldn't relax that much in short half a year. My trouble is to find some sort of accelerated test to prove the part met the relaxation requirement (similar to accelerated life test for fatigue). To test the relaxation over the complete operating period seems to be impractical as there are many applications where spring force needs to be in maintain over 5 or even 10 yrs span.

 

[israelkk]

Stating "I have no doubt the spring shouldn't relax that much in short half a year" without showing calculations to back this statement, means nothing. If you want a sound answer please provide the information requested earlier. Knowing this information will allow calculations that can indicate if the relaxation is as a result of over stress or malfunction in the manufacturing process.

 

[drawoh]

engtiuser,

As I recall from college, coil springs are not hard to analyze. How about pulling out your machine design text and analyzing the thing yourself. It sounds like your spring is fatiguing, another thing explained in your machine design book. Your problem can be fixed by selecting stronger material, and/or by changing the geometry of the spring. This will help you a lot when you chat with your vendor.

--
JHG

 

[engtiuser]

Thanks for the reply to try to help me to determine if the 30% relaxation is reasonable or not. Whether the spring should relax 30% over the period is not my concern. However, my question was raised from the validation and verfication aspect. I am looking for the industrial standard test (ideally acceleratted) to determine the creep/relaxation characteristics of the spring in a short period of time.

 

[MikeHalloran]

My industrial standard test is to give israelkk the full specs of the spring and the application and my expectations, and ask him to comment, based on his wealth of knowledge and experience.

Would you be more inclined to do so if he sent you an invoice?



Mike Halloran
Pembroke Pines, FL, USA

 

[engtiuser]

I am not prepare to pay anything as this is a pure information exchange forum and I am not sure if EngTips support soliciting activities. I was thinking there should be some sort of standard industrial accelerated test to verify the creep characteristic and I don't think the full spec of my spring has anything to do with the information I am looking for. May be it is just me who has this problem...Thanks everybody.

 

[cowski]

I am not prepare to pay anything as this is a pure information exchange forum...

Unfortunately, you have not exchanged enough information. A well designed & manufactured spring should not lose that much force in that short a time. israelkk, a recognized spring design expert on this forum, has offered a free design review. If he finds nothing wrong with your design, you can concentrate on what is happening in the manufacture of the spring. However, if the spring is underdesigned, drawing notes and accelerated tests won't help. It seems to me that there is nothing to lose and potentially much to be gained by sharing the spring design information that he has requested... </2¢>

 

[engtiuser]

You are hitting the nail right in the head. We have been using the spring for a couple of years and the springs are all well behaved, until recently the supplier changed and we started seeing problem. The spring needs to be manufactured correctly and that's the part I want to verified. There are so many things can go wrong during manufacturing, from getting the raw material to finishing up the spring. The best way to verify the quality is through testing. Note: We did the analysis before picking the spring and it was verified through testing and application that the spring design is correct. I would appreciate your feed back for the specific information I am looking for. I do appreciate Isrealkk to offer the free design review.

Thanks.

 

[israelkk]

In my 40 years of experience I can't count how many times I heard the common statement "We did the analysis it was verified through testing and application that the design is correct".

To this statement I asked the following:

1. Do you know that a wire has a minimum and maximum allowed tensile strength according to the wire specification? Not all engineers even knew it. They didn't even bother to read the specification of the wire. The next question was:

2. What was the actual wire ultimate tensile strength of the spring that you tested and verified? All of the time the answer was we don't know. The next question was:

3. Therefore, how do you know that a spring made of a wire with the minimum allowed tensile strength will do the job? maybe the spring that was tested was at higher or even the maximum ultimate tensile strength allowed? Which means that a wire at minimum tensile strength (which can have up to 20% lower ultimate tensile strength) will do the job?

Your statement "There are so many things can go wrong during manufacturing, from getting the raw material to finishing up the spring."

The part "from getting the raw material" make me suspect that you didn't ask for a C.O.T. of the wire to make sure that the wire is OK. In aerospace this is not enough. Usually, three untreated pieces of wire and three heat treated pieces of wire are actually tested to ultimate tensile strength to verify that what is declared in the C.O.T. is genuine.

I hope my worries in your case has no basis and all my questions were checked and answered during the spring design and testing.

The correct design process should be as follows:

1. Design the spring based on minimum guarantied tensile strength
2. Test the actual strength of the wire. If it has larger ultimate tensile strength than the minimum
3. Calculate the spring properties expected from the actual wire strength
4. Test the spring and make sure it meets the calculations of step 3.
5. Calculate what will be the spring properties at minimum tensile strength and make sure that it meets your requirements.

P.S.

AISI 5160H steel is covered by ASTM A 304. This spec don't even guaranty minimum and maximum tensile strength. It only gives hardenability depths for quenching. The hardnes at various depth can greatly vary resulting in similar variations in tensile strength.

 

[engtiuser]

We specified the load verses displacement requirement in the drawing together with dimension and full material specification. Each spring we received, together with the CFC, must have a serial number together with force/displacement tested report (issue by third party independent lab). This eliminated the subquality spring when we received the part (the part must meet the dimensional, force vs displacement requirement). The unknown is the actual process (heat treatment, shot peen..) and this could affect the creep/relaxation characteristic of the spring. This, again, is the subject of this thread. As I have raised this question repeatedly and there is no answer to it. It sounds more and more like there is no way to verify the creep characteristic in short duration (accerelated). Is it only me who wants to verfy the spring that used in my design will not relax more than, say 10%, in 10 years? Do I have to clamp the spring for 10 yrs to certify it's use.

Note: the C.O.T and verification of tensile strength should be straight forward and we can put it in the PO. However, we don't seem to have problem related to tensile strength at this point. The relaxation is the only problem.

 

[MikeHalloran]

You can specify any damn thing you want to.

What we are trying to establish is whether a product meeting your specification has any chance of performing in the manner that you actually want.

... and perhaps to then suggest ways in which your goals might be achieved by means other than individual third party measurement and certification, and design verification by actual test.

... For instance, by designing the spring correctly and robustly, so that your company might survive a change of supplier.

the springs are all well behaved, until recently the supplier changed and we started seeing problem.

... suggests one obvious solution; go back go the old supplier, who knew what you wanted and how to get it.

The spring needs to be manufactured correctly and that's the part I want to verified.

Your problem is that you clearly don't know what, exactly, constitutes 'correctly', for your magical spring design, so far unrevealed.

There are so many things can go wrong during manufacturing, from getting the raw material to finishing up the spring.

Yes, that's on page 2 of any spring book.

The best way to verify the quality is through testing.

Strongly disagree. That's pretty much the most expensive way to verify anything, and the statistics are terrible until you increase the sample size, in which case it gets more expensive.

Note: We did the analysis before picking the spring and it was verified through testing and application that the spring design is correct.

We're kind of curious to see your analysis, or to be provided enough information to do our own analysis.
Testing verified that a small number of springs were adequate, provided that they were made by your 'old' supplier, who clearly knows more than you do about 'your' springs.

I would appreciate your feed back for the specific information I am looking for.

It should be obvious by now that creep and relaxation cannot be accelerated with any statistical significance, which is why you are not getting a simple answer to an impossible question.

I do appreciate Isrealkk to offer the free design review.

I also appreciate israelkk's generosity, and patience with you.

If you feel that your spring design is so elegant and magical that you want to keep it for yourself, or if you have concerns about some proprietary feature of your application, please do engage a local expert, with appropriate nondisclosure terms, to review your analysis and your application. I personally think that you are in this way over your head.



Mike Halloran
Pembroke Pines, FL, USA

 

[israelkk]

If everything as you posted is true that each spring is tested and verified, then the only conclusion is that the spring stress at compressed state is much too high from the relaxation resistance point of view. If heat treatment was faulty then the spring had to fail the tests and verification procedure.

Shot preening only affect thin depth on the surface of the spring and it can only assists against fatigue due to cyclic stress conditions that initiates firm the wire surface. From you posted till now it seams the spring is working in static stress condition where shot preening will not help. Anyway, relaxation indicates that the high stress the spring sees are not just superficial but goes deeper into the wire axis far beyond the shot peeling affected zone.

 

[israelkk]

Sorry for the typo errors, here is the corrected post,

If everything as you posted is true that each spring is tested and verified, then the only conclusion is that the spring stress at compressed state is much too high from the relaxation resistance point of view. If heat treatment was faulty then the spring had to fail the tests and verification procedure.

Shot peening only affects thin layer on the surface of the spring and it can only assists against fatigue due to cyclic stress conditions that initiates from the wire surface. From your posts till now it seems that the spring is working in static stress condition where shot peening will not help. Anyway, relaxation indicates that the high stress the spring sees is not just superficial but goes deeper toward the wire axis far beyond deeper than the shot peening affected layer zone.

 

[engtiuser]

If there is no accelerated test available to verify what I need, just post it and so be it. My question is straight forward and the answer should reflect that. I don't think I ever asked for the reason why the spring relax or comment for the robustness of the design. There are propriety policy/different comfort level for revealing the design details, and it needs to be respected. The answers to the forum is voluntary base and I have no intention to waste anybody's patient to answer something that I didn't ask. If I see any question which I cannot answer from the forum, I will just ignore it. At the end of the day, I did get paid to waste my time here.

 

[ColdCoil]

try having the parts cycle tested...another alternative is heat setting or "warm setting"

Chris

 

[ColdCoil]

Here is an idea of the process that I'm speaking. There is a much shorter method by just pulling the parts from the furnace and placing the spring in a setting machine to a desired height then holding for a short period of time. Ask your current supplier if this is an option available. I WORKS and it will solve the setting issues for the long term.

The Hot-Setting Process

Hot setting is usually carried out after the shot-peening

process and consists of compressing each spring to a fixed height while at an elevated temperature. This procedure causes the spring to suffer plastic set and shorten in length, thereby losing load during the process rather than in service. As a result, the hot-set spring subsequently loses load in service at a much reduced rate than a spring without the benefit of hot setting.

In order to maximize load stability in service, it is necessary to maximize the amount of plastic set in the spring during hot setting.

It is important to make a clear distinction between the plastic set and load loss that occur during the hot-set process, and the load loss that takes place in a hot-set test (sometimes called a “hot-oil test”). The hot-set test parameters and pass/fail requirements are often called out on the customer print or in a related technical specification. The purpose of the test is to verify that the finished spring has the required level of load stability necessary for correct operation in service. The test details and parameters are, however, very similar to the hot-set process itself, so it is important to be clear regarding the differences. The test usually requires the springs to be clamped to a fixed height and placed at an elevated temperature in a furnace for a period of several hours (usually one to 24 hours), and the load loss or loss in free length (i.e. the plastic set) compared against a maximum allowable value. In order to minimize the load loss during the test, it is necessary for the springmaker to maximize the load loss during the hot-set process.