Designing and making springs with music wire

[elect_eng]

I would like to make some custom made extension springs out of music wire. Can anyone provide the formula's to determine the wire thickness, coil diameter and coil length to achieve a desired spring constant?

The motivation is that I need some small springs in my research and it is difficult to find existing springs that match what I need. I've been told that it is relatively easy to wind your own springs with music wire. I'm sure I could derive the formula's, or just experiment, but this will take time. If some ready made formulas and material properties for music wire are available, I wouldn't mind saving some time.

Thank you in advance!

 

[berkeman]

I would like to make some custom made extension springs out of music wire. Can anyone provide the formula's to determine the wire thickness, coil diameter and coil length to achieve a desired spring constant?

The motivation is that I need some small springs in my research and it is difficult to find existing springs that match what I need. I've been told that it is relatively easy to wind your own springs with music wire. I'm sure I could derive the formula's, or just experiment, but this will take time. If some ready made formulas and material properties for music wire are available, I wouldn't mind saving some time.

Thank you in advance!

I googled spring design, and got lots of good hits. Here's one of the first hits, and it appears to have a spring calculator link at the left:

http://www.efunda.com/DesignStandards/springs/spring_introduction.cfm

.

 

[Cyrus]

Get a copy of Shingleys Mechanical Engineering Design. Its in there.

 

[Q_Goest]

Try shopping a catalog like Lee Spring:
http://www.leespring.com/

If you want to design your own, try these equations:
http://www.efunda.com/DesignStandards/springs/calc_comp_designer_eqn.cfm

 

[elect_eng]

Thank you all for your very good help! This puts me in good shape.

I still will welcome any additional suggestions. In particular, if anyone has experience winding small diameter (1 cm) but long length (50 cm) extension springs with music wire, I'm sure I could benefit from a quick roadmap of the tricks and pitfalls.

I also, have a question about something I learned today about extension springs. Apparently, many extension springs are made preloaded in a way that a positive offset force (i.e. greater than zero) must be applied before the spring will start to open it's coils. I can't help but wonder how these are made. Does anyone know the basic principle involved with forming an extension spring, having constant diameter, with a preloaded compression force built in?

 

[Cyrus]

They twist the wire as they coil it. This gives it a preload.

Note: You really don't want to try and make these yourself. I was told they are very, very hard to make.

 

[elect_eng]

They twist the wire as they coil it. This gives it a preload.

Note: You really don't want to try and make these yourself. I was told they are very, very hard to make.

Thanks for the explanation. That one had me scratching my head, but that makes sense now.

Do you know what the issues were in making the springs? I'm not looking for high quality and I don't need to have a pre-load built in, but my requirements are unusual and I haven't found any commercial springs that come close. If I don't succeed in making my own, I'll have to have them custom made by a spring company. This creates a long delay and considerable expense. The work is important enough to me so that I will go down that road if I'm forced to.

Oddly enough, I can get the spring constant I want with elastic rubber bands, but the rubber has too much damping for my application. So if anyone knows a material similar to rubber, but with low damping, I can consider that too.

 

[Q_Goest]

Hi elect_eng. What can you tell us about the spring you want designed? I'm assuming a conventional wound compression spring is all you're looking for. If you provide the following I'll tell you if it's possible, how many coils, etc...
Must pass over diameter:
Must fit into diameter:
Free Length:
Spring constant needed:
Maximum Load:
Is this a cyclic load (yes/no) explain:
If you have a specific wire you want to use, what is it (material type and diameter):
Any unusual temperature range? Fluids it may be in contact with?
If it's not a conventional compression spring, I probably won't be able to help.

 

[elect_eng]

Hi elect_eng. What can you tell us about the spring you want designed? I'm assuming a conventional wound compression spring is all you're looking for. If you provide the following I'll tell you if it's possible, how many coils, etc...
Must pass over diameter:
Must fit into diameter:
Free Length:
Spring constant needed:
Maximum Load:
Is this a cyclic load (yes/no) explain:
If you have a specific wire you want to use, what is it (material type and diameter):
Any unusual temperature range? Fluids it may be in contact with?
If it's not a conventional compression spring, I probably won't be able to help.

Thank you so much. I appreciate your very kind offer. Actually, I need extension springs in my application. I think I can provide the information you listed above, but I will understand if extension springs are different enough from compression springs to generate too much additional time investment for you to justify (for a stranger ). I will list the things I know off the top of my head just in case it helps you (or anyone else) provide some basic pointers for me.

No passover diameter required
Must fit in 1 cm diameter
Range of motion: zero to 0.7 meters of extension beyond unloaded length
Unloaded length: ideally < 0.3 meters (but any length necessary can be tolerated)
No pretension necessary
Spring Constant (at extension distance 0.35 m): 3.1 Newtons/meter
Linearity of spring constant is not too critical, but hard to quantify
Static Load conditions: 0.11 kg mass hanging vertically at 0.35 m in equilibrium
Operates oscillating continuously at 1.2 sec period with max amplitude 0.35 m
Room temperature operation in air
No material requirements but low damping needed (assume music wire is appropriate)

 

[Q_Goest]

I generally use inches and pounds so I'm converting here. My understanding, you want to hang a .11 kg weight (1.08 N) on this spring, and it will oscilate such that it will extend the spring to 0.7 meters. Spring rate is 3.1 N/m, so load at an extension of 0.7 m is 1.24 N. Free length is 0.3 meters.

Assuming a spring OD of 9 mm (to stay inside 10 mm) and a wire diameter of 0.018 inches (0.457 mm) requires a total of 224 turns assuming music wire. Solid height is about 4.06 inches (0.103 meters). Free length can therefore be made at the 0.3 m length you requested. Stress should be low enough (44 ksi) so it should have infinite fatigue life.

If you get a spring manufacturer to do this I'd guess the cost will be a few hundred $. That's just a ROM cost (rough order of magnitude). Or you can buy a roll of wire from McMaster Carr for about $6 (part number: 9666K18).

 

[elect_eng]

I generally use inches and pounds so I'm converting here. My understanding, you want to hang a .11 kg weight (1.08 N) on this spring, and it will oscilate such that it will extend the spring to 0.7 meters. Spring rate is 3.1 N/m, so load at an extension of 0.7 m is 2.17 N. Free length is 0.3 meters.

Assuming a spring OD of 9 mm (to stay inside 10 mm) and a wire diameter of 0.018 inches (0.457 mm) requires a total of 224 turns assuming music wire. Solid height is about 4.06 inches (0.103 meters). Free length can therefore be made at the 0.3 m length you requested. Stress should be low enough (77 ksi) so it should have infinite fatigue life.

If you get a spring manufacturer to do this I'd guess the cost will be a few hundred $. That's just a ROM cost (rough order of magnitude). Or you can buy a roll of wire from McMaster Carr for about $6 (part number: 9666K18).

Wow, thank you so much for your help. This saves me so much time. I can basically get in the ballpark and start doing experiments to see how effectively I can make my own. If this doesn't work out, the expense of a few hundred dollars is not too bad and can be justified.

By the way, I play guitar and the 0.018 inch wire corresponds to the G string (third string) on a typical steel string guitar. So I can actually play around with this while the wire spool is on order. Clearly, the length of a guitar string is too short, but I can experiment with the winding methods and tools that are needed.

I hope a day comes when I can return the favor in some way!

 

[Cyrus]

Let's see it working first... I think its going to be harder than you think. (Hopefully not though!)

 

[elect_eng]

Let's see it working first... I think its going to be harder than you think. (Hopefully not though!)

Yes, things are always harder than you think, and I never thought it would be easy, so it could be really hard . However, it seems cheap enough to try. I'd like to say, it can't hurt to try, but actually it could be dangerous if I'm not careful! (I will be careful though) The good news is that, if I fail, the cost of custom springs from a professional may be cheaper than I originally thought.

 

[elect_eng]

Let's see it working first... I think its going to be harder than you think. (Hopefully not though!)

I thought I'd report back on the progress here. Thanks to the good advice given to me, I was able to successfully make springs today. I received the music wire this morning and by afternoon I had made springs suitable for my application.

One key to quick success was being put in the right ballpark with Q_Goest's initial design. The choice of 0.018 wire seems to be right on. I experimented with 0.014 wire too, but it was too difficult to work with those springs. The slightest over-stress and the spring deformed.

Another key was the information from Cyrus about using twisting to create pretension in the spring. This turns out to be important because if I just wind the spring without twist, the coils end up spaced out too far and the free-length of the spring is too long. A little twist produces nice tight coils and it was even possible to achieve pretension.

It was amazing how easy this was once I got the hang of it. I made a winding tool out of a long screwdriver and a tap-holder. Then I clamped the wire in a vise - pulled the wire straight - put some twist and tension and then started spooling the wire on the screwdriver (0.25 inch diameter shaft). The trick is to get the right number of twists which is quickly discovered by trial and error. Another trick is to figure out the spooling diameter needed to end up with the final diameter needed. The coil needed to be wound at 6 mm diameter to end up at about 9 mm diameter. Again, just trial and error.

I'm now planning to improve the quality of the springs with better tools to maintain constant tension as I form the spring. This should result in a professional looking spring but even today's springs work.

Thanks again for the good advice. Perhaps next week I'll post a picture of a spring.

 

[Cyrus]

Cool, post a picture of it! My hats off to you sir.

 

[berkeman]

Yes, very cool. I learn something new every day here on the PF.

 

[elect_eng]

Cool, post a picture of it!

I'm posting two pictures of a spring which is very close to what I need to make. With a 0.1 kg mass, it has 1.1 s period, which is close to the 1.2 s period that I need. I just need to make another that is little longer and I'm there.

This spring was made with about 25 feet of wire with 2 full twists per foot and 10 pounds of line tension. I made this one using a controlled tension, but I really didn't see much better results compared to when I just put a little tension by hand. However, pre-twisting the wire is critical to get nice tightly spaced coils.

It works fine for my application, but I find that it is very difficult to control the pretension over the length of the long coil. The beginning of the coil ends up with a lot of pretension and the end has almost none. However, since I don't need pretension, I was able to slide the spring over a 3/8 inch rod and stretch out the coils with pretension. You can visually see the non-uniformity over the length, but this is just a visual thing that does not affect the overall spring performance. You end up with a spring with low-damping, high natural frequency and an effective spring rate that can be matched to the specification.

One picture is the free spring lying on the table with a centimeter scale ruler in view. The other shows the spring extended while holding the 0.1 kg mass in equilibrium.

Thank you all again !

 

[Cyrus]

Good god that a long spring. I wouldn't want to coil all that wire! Hahah, good job though.

 

[Q_Goest]

Thanks for the pics. Maybe you should go into business making hand wound springs!

 

[elect_eng]

Good god that a long spring.

Now I can't resist showing a picture of his big brother. He has a period of 1.6 s when attached to a 0.1 kg mass.

This next spring came out a little better and I didn't use the 10 pound constant tension this time. I also used less pre-twist (1 twist per foot), which seems to be about right. I figured that if I made an extra long spring, I could cut out the section that looks best.

However, once again, there was pretension in the first part of the spring, and almost none at the end of the spring. Does anyone know the physical explanation for this effect? I figured a pre-twist would distribute itself evenly over the whole length, but it seems that the beginning of the coiling process wants to take up more of the existing pre-twist. I'm wondering if this is expected, or maybe I have to look and see if a step in my process is inducing this effect.

 

[Cyrus]

Did you calibrate your springs to make sure they are linear?

 

[elect_eng]

Did you calibrate your springs to make sure they are linear?

I have not done that yet, but I will tomorrow. Linearity is not supercritical in my application, but a measurable nonlinearity over my range of use might indicate a problem. My plan was to just hang different weights on the spring and measure the extension and do a plot. If you are curious, I'm happy to post the plot.

That is a good question because I ran across an interesting effect related to the pre-tension variation over length. If the spring is hanging with a weight in equilibrium and the pretension in greater than the force from the weight, the compressed coils are inactive. Then, when the weight starts oscillating, the number of active coils depends on the position of the weight, which is basically a spring nonlinearity. So, the pretension needs to be small enough to prevent this effect, which means that all coils need to be active (open with space) while the spring is in operation. This issue probably never arises with a professionally made spring, I assume.

 

[Q_Goest]

Why do you want any pretension at all? Why not just wind it with a small, constant gap? As you've noted, if there's a pretension on some of them but not all or even if each winding doesn't have the same pretension, the spring rate will change (decrease) as more of the windings separate and become active. If you have a small gap between each winding, that won't happen, even if the gap isn't perfectly constant.

 

[elect_eng]

Why do you want any pretension at all? Why not just wind it with a small, constant gap? As you've noted, if there's a pretension on some of them but not all or even if each winding doesn't have the same pretension, the spring rate will change (decrease) as more of the windings separate and become active. If you have a small gap between each winding, that won't happen, even if the gap isn't perfectly constant.

That's a good question. I agree with all of this. However, there is a slight complication as I've started to revise the specifications. I left some of this out, since I thought it might not be interesting. But, of course the details are important. Originally I said that any free-length could be tolerated. However, this was a temporary compromise I was wiling to make to get some test results. Now, with the success I've had, I can try to target the final goal right away.

For the final system, I have a limited total system length specification. I'm also finding that my needed period is a little longer than I thought, which means the equilibrium length is longer (now 0.5 m for 1.4 s period) If I allow the free state of the spring to have gaps, then I run out of space. My first springs were wound with no pre_twisting and the gaps were too large. This led me to use the pre-twisting idea to try and reduce the gap. However, it is very difficult for me to control the gap once pre-twist is used. The gap or pre-tension ends up being a function of length. So basically, I try to get just a little bit of pretension. This produces the shortest possible free length, but the gaps are all there once the spring is stretched to its equilibrium position. The key to this working is that the oscillation amplitude (0.3 m) is less than the equilibrium extension length (0.5 m). I think this last spring I made can be trimmed slightly and will work, but I still need to test to make sure.

I do have the ability to take pretension out simply by putting the spring on a 3/8 inch rod and stretching the spring. The rod tends to prevent over-stretching. This is not easy to control, but with experience I could do it well enough.

I'm sorry about leaving out some of the important information. I thought too much info might be boring to read through. But, that is a dangerous thing to do when smart people are around.

 

[elect_eng]

Did you calibrate your springs to make sure they are linear?

Here is a plot of the spring as it exists now. This shows the nonlinearity due to inactive coils when the displacement is too small. This is a very interesting effect. These results indicate that I'm almost there. In my application I need a working range of 20 to 80 cm. I can improve the problem area (20-30 cm) by identifying the compressed coils and stretching them out.

This has been a really interesting and educational exercise, and it was also useful to save the time that would be needed to have custom springs made. The other advantage is that when I do go to have custom springs made later, I will be very certain of the specifications that I need. However, it's clear to me that making a professional quality spring is not easy at all.

If anyone out there someday contemplates doing something similar, be prepared for a few days of challenges, along with some minor blood-letting and very sore hands.

 

[minger]

Nice job. That is all, haha.

 

[Cyrus]

Here is a plot of the spring as it exists now. This shows the nonlinearity due to inactive coils when the displacement is too small. This is a very interesting effect. These results indicate that I'm almost there. In my application I need a working range of 20 to 80 cm. I can improve the problem area (20-30 cm) by identifying the compressed coils and stretching them out.

This has been a really interesting and educational exercise, and it was also useful to save the time that would be needed to have custom springs made. The other advantage is that when I do go to have custom springs made later, I will be very certain of the specifications that I need. However, it's clear to me that making a professional quality spring is not easy at all.

If anyone out there someday contemplates doing something similar, be prepared for a few days of challenges, along with some minor blood-letting and very sore hands.

That graph is wrong. It shouldn't pass through zero if it has pretension.

 

[elect_eng]

That graph is wrong. It shouldn't pass through zero if it has pretension.

That's a good point. Actually, in my case the data points are not wrong, but the linear interpolation between the first two points is misleading. Since my spring has some open coils, the (0,0) point is correct and the spring deflects with even a tiny mass attached. However, I can't be sure that the initial shape is not very different from the linear shape shown. There probably should be a sharper rise before the slope flattens out.

This region is difficult for me to measure and is outside my range of interest, so I won't try to correct the plot. In my range of interest, the spring behaves very much like an ideal spring with pretension. The operation is piecwise linear, and if you trace the straight line-fit back to the y-axis, there is a force offset, as you would expect.

 

[Cyrus]

That's a good point. Actually, in my case the data points are not wrong, but the linear interpolation between the first two points is misleading. Since my spring has some open coils, the (0,0) point is correct and the spring deflects with even a tiny mass attached. However, I can't be sure that the initial shape is not very different from the linear shape shown. There probably should be a sharper rise before the slope flattens out.

This region is difficult for me to measure and is outside my range of interest, so I won't try to correct the plot. In my range of interest, the spring behaves very much like an ideal spring with pretension. The operation is piecwise linear, and if you trace the straight line-fit back to the y-axis, there is a force offset, as you would expect.

That looks and sounds reasonable. Your spring is also very linear for what you're trying to do. Kudos.