Tutorial: Easily measuring your string tension

[MarcR]

Hi folks,

I gathered some information on the process of measuring string-tension
of the strings on tennis rackets. I post here a tutorial on how to
measure the string-tension on your racket.

First some introductory information :

1. one can measure the real physical tension (in kg) of a string
via measuring the base-frequency of the vibrating strings.
(i.e. if you hit the string-plane on your racket you can hear
a sound of certain frequency composition=pitch which depends
on the tension of the strings).
Let us call this the Frequency-method of measuring string-tension.

So one only needs to measure the base-frequency f0, plug it into
a formula and get the string-tension.
Here's the formula :

A : racket-head area
f0 : base-frequency of vibrating strings
mu : mass-density of strings

2. you string your racket with say 27 kg of pull tension, but due
to friction,stress-relaxation of the strings and other factors
the real tension after the stringing process is about ~35%
lower than your pull tension, i.e. only 17.5 kg. I mention this
because the frequency-method will measure this (lower) physical
tension on your strings, it will not the yield kg-values that you
used to string your racket with, because of the tension loss. So
don't be surprised when you later use the frequency-method an get
such low tension-values. Of course you can convert the measured
real-tension values in pull-tension values you're used to, by
"adding" the lost 35% of tension to the measured tension-value and
get a pull-tension value (that'll be a rough approximation).
Personally, I do this by stringing my racket at lets say x kg
and measure real-tension directly after stringing, lets say I
measure y kg afterwards. So the loss of tension was (x-y) kg = z.
This z-value is specific to the string I used and to my personal
stringing manner. Later when I measure the tension on my strings
I'll get the lower real tension, and to get the corresponding
pull-tension values I just add the z-value. This adjustment is
done because you're more used to the kg-values in the pull-tension
range [20,30] kg. Because you put 27 kg on your racket-strings
and when you play with your racket you play only with real 17.5 kg
on the strings, but you associate the feeling of these 17.5-kg-strings
to be 27-kg-strings, as that was the pull-tension you put on the
strings during stringing.
If you just want to measure differential tension loss you can
of course skip this conversion. In this case your not interested
in absolute kg-values, but you measure directly after stringing
a real string-tension of 18 kg and after 2 weeks you measure
only 16 kg, so you know you lost 2 kg of tension.

Now here's the tutorial as a PDF-file. I hope the reader can understand
the process of the frequency-method.

http://www.tennis-altensteig.de/marc/freqmess-eng.pdf

Have fun and I hope to see some feedback =)

Greetings Marc

Reference:
- Cross, R., & Bower, R. (2001). Measurements of String Tension in a Tennis
Racket. Sports. Eng, 4, 65-175

 

[Gaines Hillix]

Marc, sounds interesting. A few questions;

How do you measure the frequency of the string vibration?

Does one have to have a consistent method of "plucking" the strings in a certain way to reproduce consistent frequency?

Don't string lengths effect the frequency and is that factored in by using the area of the racquet head in the formula?

How do you know what the mass-density of the string is?

 

[Marius_Hancu]

How do you measure the frequency of the string vibration?

Did you check his PDF tutorial?
He's doing a spectrum analysis there.

 

[MarcR]

@Gaines Hillix:

How do you measure the frequency of the string vibration?

- Hitting with the heel of your hand on the string plane
- bring the vibrating string-plane near a microphone that
is connected to a PC and record the sound of the
vibrating strings
- analyze the recorded sound with Audacity's spectrum analyzer

=>this can be done easily at home within 1 minute on your PC
using the free Audacity recording software

Does one have to have a consistent method of "plucking" the strings in a certain way to reproduce consistent frequency?

No, I'm using this method for some time now, and I try to hit the middle of
the string plane. Of course you not always perfectly hit the center, but I
always got the same base frequency. You just have to make the strings
vibrate.

Don't string lengths effect the frequency and is that factored in by using the area of the racquet head in the formula?

You're right, string length is important. It is factored into the racket-head
area. The approximation by Cross and Bower is that A = square_root(L)
where L is the string length.

How do you know what the mass-density of the string is?

Just take the string set you want to put on your racket, determine its
weight and divide it by its length, that's the mass-density of the string.
Or just use as approximation 2 g/m for polys and 1.7 g/m for nylons.
With these approximations you can still perfectly measure differential
tension loss (which I think is the most important application, i.e. you want
to know when you lost 4 kg and then you restring).

 

[Marius_Hancu]

No, I'm using this method for some time now, and I try to hit the middle of
the string plane. Of course you not always perfectly hit the center, but I
always got the same base frequency. You just have to make the strings
vibrate.

That is correct.

The fundamental/base frequency is a characteristic of the mechanical system under question.

You just need to provide some excitation/input to the system in order to get a spectral response in which the first component is the base or fundamental frequency.

Where you provide that excitation or how hard it is should be irrelevant.

 

[Gaines Hillix]

That is correct.

The fundamental/base frequency is a characteristic of the mechanical system under question.

You just need to provide some excitation/input to the system in order to get a spectral response in which the first component is the base or fundamental frequency.

Where you provide that excitation or how hard it is should be irrelevant.

Marc/Marius,

How about string gauge? Occurs to me that a 15g string is going to be a "bass string" and a 19g string is going to have a higher frequency?

 

[fedex27]

[scratches head] adding characters so i can post

 

[equinox]

This method is too complex and hard to reproduce with accuracy and consistency.

 

[MarcR]

@equinox

This method is too complex and hard to reproduce with accuracy and consistency.

Have you even tried the method ? I measure string-tension for my racquets
easily within 1 minute. By the way the commercial devices for string-tension
measurement work the same. You put the device on the string plane, excite
the strings a read-out the displayed tension. But in order to save some bucks,
I prefer to use my computer and some software to measure tension.

What makes you say that one can't reproduce results with consistency,
the only human-step is the excitation of the strings, rest is FFT-analysis
by your computer.

Note: The whole conversion step, that seems to be puzzling for some people
might be completely skipped if you just want to measure differential string-tension loss à la StringMeter.



Where have all the MacGyvers gone ?

 

[MarcR]

@Gaines Hillix

How about string gauge? Occurs to me that a 15g string is going to be a "bass string" and a 19g string is going to have a higher frequency?

String Gauge is factored into the mass-density term, as the mass-density
term mu can be re-written as mu = sigma * pi * (d/2)^2 where d is the
gauge of the string and sigma the (volume) density.

 

[gmlasam]

Marc/Marius,

How about string gauge? Occurs to me that a 15g string is going to be a "bass string" and a 19g string is going to have a higher frequency?

I'm not sure if this is relevant, but I'm a guitar player. Gianes you are correct that a thicker gauge string does produce a lower tone, but can have the same pitch. I think all musical tunning standards are based on pitch of A440.

A bass guitar is tuned to E, A, D, G (E being the lower pitch string and the thicker one and G is the higher gauge string and thinner. The E string is usually tensioned lower compared to the G. All strings on the Bass can play all notes in the octave by lengthening and shortening the strings by fretting the strings between the bridge(located on the body) and the nut(located at the head stock)

So my question is if basing the tension of the string on frequncy, which is related to a muscial note= a pitch would the lower tension string and higher tension string be read as the same tension if both are tuned to a note of A? An "A" musical note can be played at many octaves in a musical staff even if the strings are tensioned differently.

 

[MarcR]

@gmlasam

Yeah, but string-length is constant on your racket. I hope so ...

Vibration-frequence of any string depends on :
- length of string
- gauge of string
- tension of string
- mass of string

 

[gmlasam]

Also, dosen't the type of mike used can also effect the reading? A condesner mike can pickup many background noises to make the reading erratic. When I tune my guitars/bass, I plug the cord directly to my electronic tuner to prevent background noises.

 

[Marius_Hancu]

Also, dosen't the type of mike used can also effect the reading?

won't move the fundamental/base frequency.

 

[MarcR]

@gmlasam

You can simply use any microphone. I used a simple mic that came with
my old cassette recorder (the red one in the tutorial) it worked fine, you can
even use the mic of your internet telephony head-set. And by the way,
Audacity has a quite nice noise-filtering mechanism, that allows you to
remove recording noise in post-processing step ( + 20 secs ). But despite
the possible recording noise you can clearly see the peak of the base
frequency. Just try it ....

 

[gmlasam]

@gmlasam

Yeah, but string-length is constant on your racket. I hope so ...

Vibration-frequence of any string depends on :
- length of string
- gauge of string
- tension of string
- mass of string

The lenght of strings varies on different head size racquets. So if I would to hit a mid size racquet like a Wilson Pro Staff 85 and produced a musical note G at a higher pitch about three octaves higher than with hitting an oversize racquet like Prince Orginal Graphite OS, the tension reading would be the same?

Two different racquets producing the same G note, same pitch/frequncy, but are tensioned differently, midsize is higher tension, Oversize is lower tension. Both have different string lenghts.

 

[SW Stringer]

MarcR: Excellent work! Great tutorial also. You may want to show (graphic) the selected peak for those (most everyone) not familiar with using Audacity.

The underwhelming response you've gotten from your VERY helpful post reminds me of an old american axiom: You can lead a horse to water, but you can't make him drink! There's probably a similar saying in your country.

Thanks for sharing your process with us.

 

[Marius_Hancu]

Yes, very good work.

High-tech, in the end a much improved version of knocking the rackets one against the other, to listen to the sound.

I'd also like to know if Marc has made a comparison between the Stringmeter and his method, for the same racket, to follow up on the evolution of tension in time.

Just want to know how good the StringMeter is. They certainly wouldn't read the same value, but it's worthwhile knowing the relative measurements.

 

[MarcR]

@SW Stringer

Thanks a lot for your support =)


After recording you will see this in the Audacity main window

The you click on the selection button, select with the mouse the
following range

then you choose in the menu "View" the sub-item "Spectrum (Frequence Analysis) and the following spectrum window pops up

One can clearly see the peak. Now move your mouse cursor above the
peak an the frequence gets displayed.

Hope that helps.

 

[MarcR]

@Marius_Hancu

As I don't have a StringMeter in my possession, I couldn't compare the
frequency method with the StringMeter method. All I know is that the
StringMeter uses a spring to measure how much force is needed to
move two crossing strings away from each other. This force is somehow
correlated with the string tension.

@gmlasam

The lenght of strings varies on different head size racquets.

Of course it does. And you will get for an OS racket and a mid-size racket
with both under the same string-tension, a lower base-frequency for the
OS racket strings. But it's all considered in the formula (A,mu).

Two different racquets producing the same G note, same pitch/frequncy, but are tensioned differently, midsize is higher tension, Oversize is lower tension. Both have different string lenghts.

Look at the formula

the head-size of the racket is considered in the formula.
So where is your problem ?

What you're basically saying is that you measure for both rackets the same base-frequency of let's say 550 Hz.
Plugging this value into the formula for the mid-size wilson with A=548 cm^2 and mu=2 g/m
will yield 13.19 kg real tension. Now calculating the OS tension with lets say
A=690 cm^2 and mu=2 g/m will give us 16.61 kg. So same pitch on each racket but for each
racket a different tension, namely the rackets current tension.

 

[matchpoints]

Sweet, the geekier way to check tension lost. It also has the additional coolness factor. Thanks MarcR.
Cheers

 

[hangzhou]

Marc: Great work. I guess the real tension S stands for the tension embedded within the stringbed incorportaed with every parties, such as string elasticity, density, grommet friction, etc.

As you pointed out, there is a correction term 0.988, which surprised me. What kind of correction term will you use for rackets with different shape, i.e. Yonex?

Thanks

 

[MarcR]

@hangzhou

As you pointed out, there is a correction term 0.988, which surprised me. What kind of correction term will you use for rackets with different shape, i.e. Yonex?

Cross and Bower derived the basic formula for a rectangular racket-head. In additional
calculations they derive the correctional term of 0.988 to get the tension of the
strings within an elliptical racket-head (which fits the shape of most rackets quite
well). But for Yonex frames that tend to of a more rectangular shape you might
even want to use an correction-term nearer to 1, say 0.925. I'll check the paper
of Cross and Bower if they considered Yonex frame-shapes. But in my opinion
all these smaller approximation-terms are not that important if you say you want
to measure tension only with an accuracy of +- 0.25 kg.

 

[David Pavlich]

Would this program require different information for 17 gauge Rip Control, 17 guage Lightning XX, 17 gauge Timo Banger, 17 guage Gamma Infinity?

David

 

[MarcR]

@David Pavlich

The only parameter in the formula that is affected by the type of string
is the mass-density mu of the string. One might argue that the microscopic
properties of the string (i.e. nylon,gut,poly) must have some influence on the
frequency of the vibrating strings. But Cross and Bower conclude in their
paper that "provided that the vibration amplitude is small, then the vibration
frequency of a steel string is the same as that of a nylon string of the same
mass density, the same length and at the same tension." So in the frequency
method primarly the mass of the string counts, not its microscopic structure.
So you only need to calculate the mass-density of the mentioned strings and
use these density-values in the formula. But these density values seem to lie
in the range of [1.7,2.0] g/m.

Anyways, you need to choose the values for mu as exactly as possible
only if you want to calculate absolute tension-values. If your interested in relative
/differential tension values, the differences in tension count and the possible
inaccuracies of the chosen parameters are eliminated via the substraction.

Saying that I start wondering how the commercial frequency-based tension
measurement devices determine the mass-density of the string to measure.
Can the user supply or choose the string he wants to measure with these devices ?
I didn't have the chance to use one of those commercial devices yet. But I
think it'll be useful if someone possesing such a device, could just compare the
tensions calculated via either method and post the results. But be careful,
these devices display tension-values that are converted into the range of
pull-tension values [20,30] kg , as these values are more meaningful to
customers as the much lower real tension. The do this conversion by
empirical determined multiplicative factors that are multiplied with the
real tensions to get corresponding pull-tensions.

 

[gmlasam]

MarcR,

So how does one obtain this software? Is it shareware, meaning users can demo the software with a time limit, or freeware?

 

[MarcR]

@gmlasam

It's Freeware (you even get the Sourcecode so its OpenSource too).
Audacity's homepage is at http://audacity.sourceforge.net/
But of course you could also use any other recording software and
FFT-analysis software.

 

[gmlasam]

@gmlasam

It's Freeware (you even get the Sourcecode so its OpenSource too).
Audacity's homepage is at http://audacity.sourceforge.net/
But of course you could also use any other recording software and
FFT-analysis software.

Ahhhhh...THANKS!!! I'll go try it out.

 

[spirit]

I've read all this with great interest, and if I were to get the time, I might even try it for fun, but I wonder why one doesn't just purchase a string meter, which I was planning to do. They are not very expensive, are they? Usually all we want to do is to measure the string tension just after the racquet is freshly strung, and then periodically as we put playing time on it, so we can use the tension loss as input in a decision when to restring. Wouldn't a string meter do this fine and be simpler?

 

[Radical Shot]

Actually, I can play "Smoke on the water" on my C10 Pro, and I can usually tell that there is a tension loss when the last "daaahh" in "Daa daa daaahh, da daa da daaaa, daa da daaaa, da da daaahh...." sounds a little flat.

On these occasions, a "Head" string dampener wedged near the side of the frame usually stiffens it up a bit and then it sounds sweet.

 

[matchpoints]

Actually, I can play "Smoke on the water" on my C10 Pro, and I can usually tell that there is a tension loss when the last "daaahh" in "Daa daa daaahh, da daa da daaaa, daa da daaaa, da da daaahh...." sounds a little flat.

On these occasions, a "Head" string dampener wedged near the side of the frame usually stiffens it up a bit and then it sounds sweet.

lol, one good laugh after stringing and before going to bed....thanks man :mrgreen:

 

[MarcR]

Anyone trying this method should remove string dampeners before
recording the sound of the strings. As these string dampeners
corrupt the sound of the strings.

 

[Gaines Hillix]

I've read all this with great interest, and if I were to get the time, I might even try it for fun, but I wonder why one doesn't just purchase a string meter, which I was planning to do. They are not very expensive, are they? Usually all we want to do is to measure the string tension just after the racquet is freshly strung, and then periodically as we put playing time on it, so we can use the tension loss as input in a decision when to restring. Wouldn't a string meter do this fine and be simpler?

Spirit, a stringmeter is fine to measure tension loss over time or from one string job the next on the same frame with the same string, but as I understand it the method being discussed here can accurately measure the actual tension of the string job right off the machine. My experience with a string meter is that it is usually off by 5-10 lbs. from reference tension.

 

[MarcR]

@Gaines Hillix, spirit

Right. With the StringMeter you can measure relative tension-loss.
The reference tension is the tension you measure right after stringing,
and you can compare all later on measured tensions with this reference
tension. The tension you measure with the StringMeter are usually off
by a certain amount of kg/lbs. But as you're measuring differential tension
loss this doesn't really matter.

On stringmeter.com the manufacturers of the Stringmeter say the following :

Q. I just strung my racquet at 60 lbs. The Stringmeter is reading only 40 pounds. Does the Stringmeter work or am I doing something wrong?

A. Your Stringmeter does work and you aren't doing anything wrong. This is the most common question about the Stringmeter and the most misunderstood concept about the device: The Stringmeter's numbers are only relevant to itself. While it can be used during stringing, you cannot use the Stringmeter to VERIFY a stringing job. That is not its purpose. The Stringmeter is used to MONITOR tension.

Keep testing your racquet weekly. When the Stringmeter shows 35 instead of 40, you KNOW you have lost 5 pounds of tension. This is the purpose of the Stringmeter, not comparing it between machines.

But with the frequency-method you have the capability to measure the
absolute real tension on your strings. But I think the StringMeter absolutely
fits your needs if you just want to know when you lost x kg of tension.
Go ahead and get yourself a StringMeter, it's more handy than an PC
and fits in every racket-bag.

 

[spirit]

@Gaines Hillix, spirit

Right. With the StringMeter you can measure relative tension-loss.
The reference tension is the tension you measure right after stringing,
and you can compare all later on measured tensions with this reference
tension. The tension you measure with the StringMeter are usually off
by a certain amount of kg/lbs. But as you're measuring differential tension
loss this doesn't really matter.

On stringmeter.com the manufacturers of the Stringmeter say the following :




But with the frequency-method you have the capability to measure the
absolute real tension on your strings. But I think the StringMeter absolutely
fits your needs if you just want to know when you lost x kg of tension.
Go ahead and get yourself a StringMeter, it's more handy than an PC
and fits in every racket-bag.

MarcR, Using the string meter doesn't sound much different than this computer generated method. Neither method will match the tension of a freshly strung racquet to be the tension you set the stringer at. Per your original post which I quote here:

"you string your racket with say 27 kg of pull tension, but due
to friction,stress-relaxation of the strings and other factors
the real tension after the stringing process is about ~35%
lower than your pull tension, i.e. only 17.5 kg. I mention this
because the frequency-method will measure this (lower) physical
tension on your strings, it will not the yield kg-values that you
used to string your racket with, because of the tension loss. So
don't be surprised when you later use the frequency-method an get
such low tension-values. Of course you can convert the measured
real-tension values in pull-tension values you're used to, by
"adding" the lost 35% of tension to the measured tension-value and
get a pull-tension value (that'll be a rough approximation)."

So neither the string meter nor the computer method matches the stringer set tension, but you can use either to set a reference tension right after stringing and then measure periodically as you use the racquet to guage tension loss over time, right?

I suppose the computer method is superior in that it does actually measure the physical kilograms of tension on the strings, and perhaps the string meter does not (we don't know this for sure, since you have not compared a string meters measurement of a specific string job to what you would measure via the computer method). So if one is interested, mostly because of curiosity I presume, the computer method is superior. It is certainly a lot more fun if you have the time and like physics and math.

 

[SW Stringer]

spirit says: "So neither the string meter nor the computer method matches the stringer set tension, but you can use either to set a reference tension right after stringing and then measure periodically as you use the racquet to guage tension loss over time, right?"

The stringer set (reference) tension is merely what the machine ATTEMPTS to pull on each string. If you've read the Physics and Technology of Tennis you'll understand (hopefully) that it can't pull this tension on all the strings, especially the crosses due to friction with the mains, friction in the grommets, etc. The overall stringbed tension is going to be much less than the reference tension, and the formula that MarcR quotes was developed by the racquet guru's and represents the ACTUAL attained overall stringbed tension.

As you surmised the frequency method is an exact (and very sensitive) way to measure relative tension loss, so it doesn't really matter if you accept this lower tension number given by the formula or not, the percent tension loss will be the same regardless.

 

[MarcR]

Hi,

I coded a little program that takes over the measurement of your string-tension.
You just have to prepare your microphone under XP for recording in the
Windows Audio Mixer, unless it's not already properly installed.
Then just launch the program freqmess.exe enter your racket and string
data and click Record. Then hit your strings once and bring the string-plane near
the microphone. Wait.
The tension T will then be displayed in the field T.

You can get the zip-file of the program via
http://www.tennis-altensteig.de/marc/freqmess-eng.zip

Remember the calculated tension is the real tension on your racket, not the
pull-tension that was used to put the strings on. The tension will decrease
by up to 40 percent minutes after stringing.

With my stringing-jobs with polyester I usually loose about 10 kg of tension
(depends on how accurate one does stringing and what strings he uses), so to
convert the real tension in the range of pull-tension I just add up 10 kg
or 22.2 lbs, respectively. The proper procedure would be to use the program
to measure real tension right after finishing your stringing-job and calculate the difference
between pull-tension and measured real-tension and to add this offset to
later real-tensions that were calculated with the program.

NOTE: Before recording the string sound remove string-dampeners.

Greets Marc

 

[Gaines Hillix]

Hi,

With my stringing-jobs with polyester I usually loose about 10 kg of tension
(depends on how accurate one does stringing and what strings he uses), so to
convert the real tension in the range of pull-tension I just add up 10 kg
or 4.5 lbs, respectively. The proper procedure would be to use the program
to measure real tension right after finishing your stringing-job and calculate the difference
between pull-tension and measured real-tension and to add this offset to
later real-tensions that were calculated with the program.

NOTE: Before recording the string sound remove string-dampeners.

Greets Marc

Mark, good work! One question here....did you mean 10 lbs or 4.5kgs?

 

[MarcR]

@Gaines Hillix

You're right I meant "10 kg or 22.2 lbs". Got mixed up with
the conversion from kg to lbs.

Greets Marc

 

[POGO]

MarcR,

Thanks for your information. A laptop really comes in handy with this as you can have the laptop nearby your stringer and do the calculation. Most if not all laptops have a builtin mic that works well.

Thanks,
POGO


Hey!! Wouldn't this program be a nice feature buitin to an electronic stringer? Just imagin an electronic stringer with a an actual laptop computer builtin, the lbs will automatically display on the screen as you tension and you can also serve the web while stringing or even lookup string patterns.

 

[MarcR]

Summary of events during/after stringing that cause the big tension loss
after Stringing :

- 1 - stress relaxation : The force that acts on the string due to the tension
causes molecule bonds of the string-polymer to break. This causes a
dramatic loss of tension during and directly after stringing. Overall the
stress-relaxation causes approx. 3 kg of tension loss.The tension loss
due to stress-relaxation is higher for polys than nylons.

- 2 - friction : When stringing the crosses you loose some tension due to friction
between the mains and the crosses. Friction accounts for roughly 2.5 kg
of loss in tension.

- 3 - racket deformation : During stringing the racket-head suffers under
deformation due to the tensioning procedure.

- 4 - machine : The pull-arm of the strining-machine is not in the same plane
as the strings are. They´re displaced by approx. 15 degrees. So you
loose some tension here, as they're not co-linear. 1-2 kg are lost here.

- 5 - knot-tying : When you finish your stringing-job and tie off the knots
you'll also loose (depending on your tying-method) some tension.


So overall you may loose roughly 10 kg from your pull-tension. But no problem,
calculate your personal tension-loss value and add this offset to the values that my program gives you.

@POGO

Laptops always come in quite handy One could even make a tiny
diagnosis center out of the program. One can even frame stiffness via
frequence measurements. But one has to excite the racket in a different
manner. More to come

my plan for the future, this program on a Java-Handy with built-in microphone.
Pefectly mobile



Greets Marc

 

[MarcR]

I measured the mass-density mu of some of my strings I have
in my stock, Pro's Pro strings, mostly Polys.

I will populate this table further.
The only nylon string is the PP Flexel with 1,33 mm diameter.
I'd rate the PP Poly Nova as a role model of Poly strings.

Greets Marc

 

[Kevo]

MarcR,

I am curious to know what language you used to write the program. I would like to make something similar for the Mac, and I wonder if you would be willing for me to use your code as a starting point. I have been able to use your original instructions to do this manually, but I think this would be much quicker and easier.

 

[MarcR]

@Kevo

A Mac port is highly welcome.

- I used Python for the graphical user interface, there is a Mac version of Python available http://www.python.org/download/ and wxPython.

- Python was also used to do the FFT on the Wave file. But I don't know if wave-files are supported on the Mac.

- To convert the raw recorded data to a wave file I used the sound package sox at http://sox.sourceforge.net/

- To record the sound it mainly uses PortAudio (http://www.portaudio.com/) which has a Mac port too.

Does the Mac have a C- compiler installed like the gcc. I think if
we have a gcc running on your Mac we could also make the whole
stuff in C or C++ an compile it on an Mac, to get a Mac executable.

For the current windows version I used py2exe to get executables for
the Windows platform.

Greets Marc

 

[mido]

Marc R
Very nice project and extremely useful application. Very helpful. Congratulations!

I was using the first method (Audacity) for few weeks.
Then I tried the second application.
Both run on Dell workstation on Win2000Pro using external computer mike.
Mixer device for second method was SoundMAX Digital Audio.
The results were very consistent and close: 568 Hz using first method and 571Hz using second method, which corresponds to only 0.3 lb of tension difference.

Anyway, fully agree with “spirit” and “SW Stringer” – the main reason to measure the string tension is to have a reference, to measure relative tension loss, to compare tension now and later (after few hours or days).
The actual number doesn’t matter and we’ll never know it.
The important thing is to know what’s the reading when racquet performs perfectly (eg. 29 lbs or 51 lbs or whatever, then after 5 hrs of playing it could read 2 lbs less, so we know what’s going on).

With Marc’s PC frequency method it is accurate and takes less than 60 sec.
Mechanical Stringmeter is handy and costs $20-30 http://www.stringmeter.com/index.html
Electronic Tension Tester is $40-50 http://www.eagnas.com/maxgen1/etest.html
Gamma ERT 300 Tenniscomputer is $160 http://209.166.188.140/gammasports/daughter.cfm?ID=283

 

[dufferok]

MarcR...

I downloaded your program, strung one of my HEAD LiquidMetal MP rackets at 66 lbs. in the mains and crosses. Using your program and a simple microphone that came with my webcam, the measured tension was 45.5 lbs. Add in the offset you suggested (22.2 lbs) and this gave me a tension of 67.2 lbs. Very close to the tension setting of my stringer while stringing the racket.

I also own a Stringmeter tension gage. Measuring my freshly strung racket with the StringMeter showed a reference tension of aprox. 48 lbs.

Therefore, I would say that your program is very accurate in measuring the actual tension of the racket (accounting for a 22.2 lb loss in initial strung tension). Well done! A big congrats on what I would call a huge accomplishment.

 

[Kevo]

MarcR,

Email me at kevo at gatorgraphics dot com with the necessary code and I will see if I can put a Mac port together. I have gcc installed and I can download the other bits. I think quicktime can read wav files so I believe everything should work.

Thanks

 

[MarcR]

@dufferok

Thanks for using the frequency method and posting the results you received.
Let me summarize :

You strung your racket with 66 lbs, mains and crosses. And after you finished
your stringing-job , you measured a tension loss of 66 lbs - 45.5 lbs = 20.5 lbs .
So whenever you measure the real-tension of this particular racket just add
up your 20.5 lbs to the real-tension my program calculates, and you'll get the
absolute pull-tension.

With my stringing-jobs I loose approx. 22.2 lbs. I use a
simple dropweight-machine. But this offset may be smaller when using a
electronic stringing machine (as tensioning is more accurate) or an other string
(like Nylon, which looses less tension during/after stringing). So the best
solution would be to calculate your personal tension offset for every stringing-job
and note it down for your racket and later add this value to the real-tension.



@kevo

I'll prepare a source code package and send it you via email. Thanks
for your support for a Mac version

 

[MarcR]

I prepared an OpenOffice and Excel spread-sheet to analyze
the tension of your racket-strings.

OpenOffice spreadsheet

Excel spreadsheet

Just enter pull-tension [lbs], head-size [in²], mass-density [g/m],
pull-tension [lbs] and for each measurement the frequency [Hz] or
real tension [lbs] in the corresponding fields.

 

[MarcR]

@Mido

Thanks, for using the frequency method and for posting your
experiences here.

When you use the Audacity method to measure the frequency
the maximum FFT-size you can do is @16384 sample-points.
And the sound is recorded at a sample-rate of 44100, which
means the frequency resolution of the spectrum you get
from the FFT is 44100/16384 = 2.69 Hz. My program uses
a 617400 sized-FFT to get the spectrum, which leaves us at
a frequency resolution of 44100/617400 = 0.071 Hz . So the
oversampling gives us a much higher frequency resolution.
This is the reason why Audacity and Freqmess frequencies
differ a bit, as the Freqmess FFT is much higher resolving.

Marc