# The ultimate tension loss measurement

Discussion in 'Strings' started by stoneage, Sep 25, 2011.

1. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
The title is exaggerating a little, but I have tried to exactly measure how the tension in a racquet changes over time, including the effects of playing. I have therefore built a specially designed load cell that sits in the racquet (more below). I strung the racquet with a Kirschbaum Proline II 1.25 mm (polyester) string at 23 kg. I played with it after 5 days and then again after 12 days. The tension over the five weeks until it broke looks like this:

Some things to note: The stringing took almost one hour, since I was measuring during the stringing as well (see this thread). The tension directly after stringing was therefore a little lower and the subsequent tension drop a little less than it would have been after a more normal 15-30 min stringing time. The racquet has been kept at around 18C during this period so the tension loss rate would have been a little higher in a warmer environment. An interesting fact is the the first match acted as a "fast-aging" of the strings, the tension dropped dramatically, but then stayed almost constant for a week.

One of the main reasons for doing this test was to see how well my racquetTune app was able to monitor the tension drop. Since I am measuring two different things (two strings vs an overall figure) the absolute value would probably be different, but the trend would hopefully be the same. So taking the same data as the figure above and including the racquetTune results I got the figure below. In this case the x-axis is the time in log-scale, which makes the beginning more clear and it also makes it easier to compare the trends.

I am very satisfied with the result and I think that racquetTune can be relied on to show the right trend of the racquet. Another interesting observation from the log diagram is that the drop (before the first match) is close to linear and can thus be described by a fairly simple exponential or logarithmic function.

Set up
The load was measured with specially designed load cell that sits in the racquet. It is made of high grade aluminum and has four strain gauges, two on each side, coupled as a bridge. This enabled me to detect a tension variation of 0.01 kg. It was calibrated in the racquet with another load cell/scale with a +- 20 g accuracy.

This set up has been discussed in this thread

/Sten
______________________________________
racquetTune, swingTool and netHeight, tennis apps for the iPhone.

Last edited: Dec 2, 2011
2. ### TonyBHall of Fame

Joined:
Sep 6, 2006
Messages:
2,780
That's good stuff. Looks like the experiment was well thought out. Thanks for sharing.

3. ### fgsHall of Fame

Joined:
Jan 29, 2007
Messages:
3,499
Location:
Vienna, Austria
very interesting results indeed.

i would be interested in finding out if you are going to go on playing with the stick and the load cell in place. reason for my asking is that with some strings, and the pro line II in particular, i have experienced a second tension drop around the 6 hours mark - i mean hitting hours. so far, the graph depicts just two hitting hours, the first giving a larger drop than the second.

my asking is particularly about this 6 hitting hours point because while other strings go dead, which is not a matter of tension but rather a matter of material fatigue, the kirschbaum started to spray and behave somehow much more lively, not offering any clue of having gone dead. this is the reason why i rather think that it was a second tension drop, as opposed to other strings that simply have no more resiliency left and you need to seriously overhit in order to produce a deeper ball.

4. ### KozRookie

Joined:
Feb 20, 2004
Messages:
344
Very interesting little experiment! I remember when I had to solder a strain gauge on a piece of metal in a lab class; it was a disaster! (Unsteady hands + soldering on a tiny strain gauge). Well-done!

5. ### GlenKProfessional

Joined:
Mar 20, 2011
Messages:
842
Location:
St. Augustine, FL
I am very comfortable that RT measures tension loss accurately. It is very consistent.

6. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
I will keep it for another week or so. Then I will try some alternative. Playing more often, or testing other strings.

/Sten

7. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
I have updated the diagrams in the first post so they now cover the first three weeks and four matches. No dramatic changes.

/Sten

8. ### fgsHall of Fame

Joined:
Jan 29, 2007
Messages:
3,499
Location:
Vienna, Austria
thank you for the update.

if my memory serves me right, the "second" tension drop i experienced when playing with this string was around the 6-7th hitting hour. that's when it got sort of a rocketlauncher. would be interesting, if you have the patience for this, to see what happens when you fo into the 7/8th hitting hour.

thank you for your efforts and the very informative topic.

9. ### AndresG.O.A.T.

Joined:
Apr 21, 2005
Messages:
12,527
Location:
Mar del Plata, Argentina
Poor Soderling!!!!

10. ### Smasher08Legend

Joined:
Oct 2, 2010
Messages:
5,177
Location:
The 6
Sten, when will you be bringing your wonderful apps over to android??

11. ### benjamins_80New User

Joined:
Jul 7, 2010
Messages:
65
Location:
Kansas City
Wow, very impressive. I have been using racquettune for a few months now and really like being able to track my stringing history. I am finding it especially useful in noting useful life of strings and where the ideal tension is for what I like. I had been stringing full poly at around 63 lbs. I found that I liked it after the strings had lost some tension and instead of playing a guessing game of how much lower, I was able to measure it and know where to string in the future. Thanks for a great app.

12. ### julianHall of Fame

Joined:
Dec 20, 2006
Messages:
4,109
Location:
Bedford,Massachusetts,US
Are you sure your second drawing is OK?

Are you sure your second drawing is OK?
How did you get data for 100 hours?

13. ### julianHall of Fame

Joined:
Dec 20, 2006
Messages:
4,109
Location:
Bedford,Massachusetts,US
Tension of strings?

Is 23 kgs below 53 pounds specified on a throat of a racket?

14. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
Of course I am sure. What is strange with 100 hours other than it is a tension drop due to playing, but that is hardly surprising?
I measured it in the same way as all the other points! I

???
53*0.454=24.062
I think it is OK for Wilson to round that off to 24 kg.

PS I have updated the diagram in the first post to the final data

15. ### julianHall of Fame

Joined:
Dec 20, 2006
Messages:
4,109
Location:
Bedford,Massachusetts,US
Do you have experimental data for 1000 hours?

Do you have experimental data for 1000 hours?
Your figure implies that you do
Does 1000 hours mean 1000 hours of play?
I guesss it does NOT.

Last edited: Dec 2, 2011
16. ### taurussableProfessional

Joined:
Oct 2, 2012
Messages:
1,295
great thread! i might be most of the time playing with tension in the 40 range lol

Joined:
Nov 22, 2010
Messages:
487
It does not.

18. ### Old ChemistRookie

Joined:
Jun 6, 2012
Messages:
102
Location:
fgs

Exactly what I have observed with proline II strung at 53-57 lbs. After 5-6 hr playing -- less spin and less control -- shots start to spray. The string is still comfortable but plays differently.

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
Great stuff, stoneage.

The string only loses 1.2kg through 6-8 hours playing time and 4.6kg at rest. It also looks like it has practically stopped losing tension through playing, by the end. This seems to indicate that most of tension loss doesn't come from hitting. Does this mean the best way to compare the tension loss of two strings is the loss at rest?

Would the string have lost a lot more tension if you had done more hitting at the 120 hour stage?

I was just comparing this to the TWU data on the Proline II 1.25 mm. In the TWU data the static loss is 8lbs and the total loss is 22lbs. Why is the total loss much higher? The TWU static loss is tested sixty seconds after pulling. At the 0 hrs stage yours is 4.6lbs. Does this mean the TWU data is unreliable?

Last edited: Mar 15, 2013
20. ### mikelerModerator

Joined:
Sep 26, 2008
Messages:
21,863
Location:
Central Florida
Some strings seem to drop that initial tension and then stay relatively stable. Others keep dropping each time but still play the same. It is the strings that drop tension AND feel different each time out that I think most of us are looking to avoid.

21. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
The string loses tension over time. If you apply an extra load, e.g. by playing, this process is accelerated. So yes, if I would have started to play earlier, say 1 hour per day from day one, the tension would have dropped quicker. But my guess is that the final tension after 5-6 weeks would have been about the same.

When a string is loaded during stringing the molecules start to glide relative to each other causing a permanent deformation (and stress relaxation). This "gliding" however, becomes more difficult the more the string is deformed, which means that the tension loss slows down. This typical for most visco-elastic materials

A simple model is to view the string as two springs and one damper:

When you load the string during stringing both springs are stretched to take up all the load since the damper is rigid (a fast action). The damper will then start to glide, which means that spring E2 is relaxed. Eventually the damper has relaxed E2 completely and the system has settled to a lower tension caused by the stretching of E1 only.

No, it is not necessarily contradictory. The curve only shows the tension after the stringing was finished, so the string had already lost a couple of kg. The stringing took almost 1 hour since I was measuring on the same time and had a couple of measuring cables to handle when rotating the racquet. How much it lost during this first hour is difficult to say since there are several effects causing a change in the tension. See this thread:

Sten

___________________________________________________________
racquetTune, stringBed and swingTool racquet apps for the iPhone/iPad.

Last edited: Mar 16, 2013

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
So the 0 hour stage is 1 hour after the stringing process. The curve starts at 20.9, suggesting the string lost 2.1kg through stringing. Is this correct?

Sorry, I don't understand the tension variation thread you've linked. It seems to suggest tension loss is very high after pulling, then rises again 35 mins after. Does this mean the TWU static loss is unreliable because they measure too early?

Why is the total tension loss so high for the TWU data? In there data, the pro line II loses 22lbs, or 9.98kg. This is a lot more then the total tension loss of 5.8kg in your graph. Does this mean there testing method is too aggressive?

Is the best way to compare the tension loss of two strings to look at the static loss? In your graph, the string only loses 1.2kg through playing time and 4.6kg at rest.

23. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
Yes and no. It is correct that the the string lost 2.1 kg. But on the same time tension was added buy pulling the crosses so the loss would have been higher in a single string.

You can't compare the two. TWU measures the material data, I measured what happened in a racquet during stringing. What you see in the tension variation thread are the effect of three things: Creep/stress relaxation in the string, racquet head deformation and string elongation from crosses.

When you string the mains they lose tension because the material loses tension, but also because the the racquet head is deformed. This is what is showed in the first half of the curve. Then you start to pull the crosses. The mains get longer because of the "wavy" form of the stringed racquet. This elongation increases the tension. The crosses will also negate some of the head deformation which further increase the tension. Without pulling the crosses the tension would have continued to fall.
http://appmaker.se/?m=4&s=5

If you add an 3 kg tension increase from the crosses you end up with similar figures in both cases.

As I said in my last reply the reason for this difference is that I waited to play so I think that a combination of static and dynamic data as TWU does it is valid.

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
I still wonder, with the TWU testing procedure being quite short (I think), will strings continue to lose tension after the test. So, if two strings have a similar total tension loss on the TWU data, but one has higher static loss, one has higher dynamic loss. In the future won't the string with higher dynamic loss lose more tension? Thus the string with same total loss, but higher static loss, is better.

How do I add the 3kg tension increase from the crosses using the TWU performance database? If I bump up the reference tension 10lbs, the total loss only decreases 0.3kg, or 0.7lbs

Last edited: Mar 17, 2013
25. ### Povl CarstensenLegend

Joined:
Jun 14, 2004
Messages:
6,050
Highly interesting, thank you!
Would be great to see it for syn gut, and gut...

Last edited: Mar 16, 2013
26. ### Povl CarstensenLegend

Joined:
Jun 14, 2004
Messages:
6,050
Deleted
10chx

Last edited: Mar 16, 2013

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
Also, corners reckons the TW tension loss tests are too aggressive. Their tests make copolys lose more tension than they would in the real world. Is this possible, baring in mind there seems to be a limit to how much tension a string can lose?

Last edited: Mar 17, 2013
28. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
I don't really know enough about the TWU test to have a definite view of its validity.

But in general, creep and stress relaxation in a material is a highly non linear phenomenon, which makes it difficult to accelerate tests. A high load under a short time doesn't have the same effect as a low load under a long time. You would need several material constants to describe the material. Even the much simplified model with two springs and a damper I showed in the earlier post leads to three material constants that needs to be measured, and a fairly complex equation to calculate the tension drop. An equation that I very much doubt that 99.99% of stringers/player wan't to concern themselves with (I wouldn't).

So you can't use a single value (the TWU tension drop) to exactly predict how much tension a racquet will loose, especially since it would differ if it is a recreational player playing once a week or a college player playing 2 hours a day. But you can use it to compare strings: A string with a high TWU value will probably lose tension quicker in practice than a string with low value. But you shouldn't bother about the decimals.

Sten

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
If I bump up the reference tension 10lbs, the total loss only decreases 0.3kg, or 0.7lbs, which is still way off your chart figures.

If the tension loss is similar, then by the end of the TWU procedure the tension loss would have all but stopped, right? This means I can use the total tension loss figure. Otherwise, if the racket still has a lot of tension loss to go, I should use the rate of loss (static loss deducted from total loss).

30. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
What reference tension?
In my set up I strung the racquet with 23 kg setting on the machine. Adding 3 kg from pulling crosses into the mains means that initial tension was something like 26 kg, giving a total tension drop of 9 kg. However, the extra 3 kg was just an educated guess, it could have been 2 kg or 4 kg.

I am saying this again: You can NOT compare the tension in my test with the TWU figure since they represent two very different situations.

I hope you don't think that the TWU figure gives you the tension loss in a racquet? (because it certainly doesn't)

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
I mean't tension loss in the strings, not the racquet.

For the TWU performance database, is it more useful to use rate of loss (total loss minus static loss) then total tension loss? This allows you to see how much tension a string loses post-stringing, either through hitting or at rest. So a string with more rate of loss will ultimately lose the most tension. This is what I have been doing in the past.

Last edited: Mar 20, 2013

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
bump......

33. ### tlmLegend

Joined:
Jul 21, 2004
Messages:
9,599
Sure is a lot of trouble for nothing really. I already know that my strings play the best the first day or 2 and then they go down some and stay about the same.

I only use them for a week anyway and then change them. I don't need to measure tension loss, I can see the difference in how my strings play as they lose tension. Plus I already know that nothing beats poly when it is new.

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
I use multi's. I don't break them, so I like to use them for 2-3 months. I'll buy a tension meter myself and find out how much tension strings are losing post stringing, and when they stabilize. I've been curious about this for ages.

Last edited: Mar 21, 2013

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
These two quotes contradict each other. The top quote suggests the final tension loss will be the same regardless of load. The bottom suggests the load impacts how much a string loses, and when it stabilizes.

Last edited: Mar 21, 2013
36. ### tlmLegend

Joined:
Jul 21, 2004
Messages:
9,599
Wow you can use mult's for 2-3 months thats a long time.

37. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
No, they don't.
The first one says that my guess is that they would eventually end up the same (or at least similar).

The second statement talks about about how much a string loses, that is correct. But it doesn't say what would happen eventually. It just says that how fast a string will loose tension depends on the load.

If you look at the spring and damper model again (which is actually good for understanding what happens): As I wrote "Eventually the damper has relaxed E2 completely and the system has settled to a lower tension caused by the stretching of E1 only." But if you stretch the string a little extra by playing during that process, both springs will be stretched and the load increased. That will lead to the damper being pulled a little extra. When the deformation is back to normal the damper is therefore more elongated, E2 is more relaxed and the total tension is lower. The tension loss is thus accelerated. But the end result is the same (E1 carrying all load), it is only the time to reach that state that will differ.

Sten

___________________________________________________________
racquetTune, stringBed and swingTool racquet apps for the iPhone/iPad.

Last edited: Mar 22, 2013

Joined:
Jan 12, 2012
Messages:
695
Location:
Rotherham, United Kingdom
Thanks. I think I understand the damper model. Does tension loss stabilize at a similar point, regardless of how much you accelerate the loss/increase load?

In another thread, corners stated "I believe TWU's procedure for tension loss measurement is too aggressive. Basically, their tests make the co-polys lose more tension than they would in the real world". Is this false?

Last edited: Mar 22, 2013
39. ### stoneageRookie

Joined:
Apr 8, 2010
Messages:
257
Location:
Stockholm, Sweden
The model says it does. However, it is a simplified model so real materials might behave differently. But I think it is a reasonable assumption as long as the acceleration consist of playing more or less often.

I have never used the TWU tension data so I can't really comment on how good they are in practice. I have read the paper of the methodology of the test though and I was a little curious about the relative short testing time for something that essentially is a long time phenomena. Some materials exhibit a fairly quick initial tension loss and then a slower rate whereas others are more even. And these test risk to miss that difference. But doing real long time tests under different loads and temperatures would probably cost to much.

Since there are no other tension loss data available they are certainly better than nothing as long as you don't take every kg literally. And maybe you should only compare strings within the same category.