Twu swing weight calculator

Aretium

Hall of Fame
Hi everyone, now I know a lot has been posted about this method. I don't really care about how accurate it is, just want to match a few rackets.

On the instructions it says to swing from the head. Shouldn't it be from the grip? As thats how we swing the racket?

How would I readjust the calculation for this?
 

Slowtwitcher

Hall of Fame
Hi everyone, now I know a lot has been posted about this method. I don't really care about how accurate it is, just want to match a few rackets.

On the instructions it says to swing from the head. Shouldn't it be from the grip? As thats how we swing the racket?

How would I readjust the calculation for this?
No. Follow the instructions and it will work.

 

El_Yotamo

Hall of Fame
Hi everyone, now I know a lot has been posted about this method. I don't really care about how accurate it is, just want to match a few rackets.

On the instructions it says to swing from the head. Shouldn't it be from the grip? As thats how we swing the racket?

How would I readjust the calculation for this?
When you swing from the head you're essentially measuring the inertia there. As slowtwitcher mentioned, the TWU formula works thanks to the parallel axes theorem which gives you the inertia at any axis given the inertia at another axis parallel to it, the mass, the center of mass (balance point), and the distance of each of the axes from the COM. So don't worry, they know what they're doing.
 

anarosevoli

Rookie
I initially thought the same but was convinced after putting weight at the head and buttcap for testing: weight at the buttcap influences amplitude but not time. It works very well. I always let it swing 50 times instead of 10 in order to minimize timekeeping error.
 

HitMoreBHs

Semi-Pro
I initially thought the same but was convinced after putting weight at the head and buttcap for testing: weight at the buttcap influences amplitude but not time. It works very well. I always let it swing 50 times instead of 10 in order to minimize timekeeping error.
Better still, save yourself the stopwatch errors and use the rather excellent Swing Tool app. Even if one isn’t convinced by its accuracy, its precision is excellent.

Unfortunately, one needs at least partial worship at the altar of Steve Jobs to use that app. (Android? You’re outta luck!)
 
Last edited:

AA7

Hall of Fame
yes. TWU method works just fine. like @anarosevoli said, try and take 50 swings or so and divide the total by 5 (to enter into time box in the swing weight calculator)... basically you want as big a sample as possible... it works out fine...
 

anarosevoli

Rookie
Better still, save yourself the stopwatch errors and use the rather excellent Swing Tool app. Even if one isn’t convinced by its accuracy, its precision is excellent.

Unfortunately, one needs at least partial worship at the altar of Steve Jobs to use that app. (Android? You’re outta luck!)
Will try, I have that app but only use it for tension.
 

El_Yotamo

Hall of Fame
Better still, save yourself the stopwatch errors and use the rather excellent Swing Tool app. Even if one isn’t convinced by its accuracy, its precision is excellent.

Unfortunately, one needs at least partial worship at the altar of Steve Jobs to use that app. (Android? You’re outta luck!)
I'm a very DYI-oriented kinda guy, what I like to do is film the swing and then put the video through a frame by frame software to try and get the time period as accurate as possible. I also calculate the SW myself rather than on the TWU site, but it's really the same formula so not very important just a matter of preference for me.
 

El_Yotamo

Hall of Fame
yes. TWU method works just fine. like @anarosevoli said, try and take 50 swings or so and divide the total by 5 (to enter into time box in the swing weight calculator)... basically you want as big a sample as possible... it works out fine...
Well you want a big sample but keep in mind that as the swings get smaller in amplitude due to air resistance it's not as easy to get an accurate read. 20-30 swings is a good amount IME, never tried 50 so not sure.
 

travlerajm

G.O.A.T.
I initially thought the same but was convinced after putting weight at the head and buttcap for testing: weight at the buttcap influences amplitude but not time. It works very well. I always let it swing 50 times instead of 10 in order to minimize timekeeping error.
I’ve gotten into the habit now of counting 100 swings, but using the lap timer every 10 swings to keep it easier to count without worrying about losing count.
With this way it’s really easy to measure to within 0.01 seconds per 10 swings and know that it’s accurate.

Measuring the balance carefully with the edge of my metal ruler is also important - easier to do this to within 0.01” when the balance is just a mm or two away from the throat bridge as a reference point. When that’s not the case (e.g., in cases where the balance is about an inch south of the bridge) then I attach a object of known weight to the tip do that I can move the balance to near the bridge. Then calculate the balance without the weight.
 

anarosevoli

Rookie
I’ve gotten into the habit now of counting 100 swings, but using the lap timer every 10 swings to keep it easier to count without worrying about losing count.
With this way it’s really easy to measure to within 0.01 seconds per 10 swings and know that it’s accurate.

Measuring the balance carefully with the edge of my metal ruler is also important - easier to do this to within 0.01” when the balance is just a mm or two away from the throat bridge as a reference point. When that’s not the case (e.g., in cases where the balance is about an inch south of the bridge) then I attach a object of known weight to the tip do that I can move the balance to near the bridge. Then calculate the balance without the weight.
Your playing with fire... Shifting the balance point with known weights lets you find out about the whole real weight distribution if you make enough measurements (I suggest 27 at least for a standard length racquet, better 68); I'm so tempted but I aready waste too much time...
 

travlerajm

G.O.A.T.
Your playing with fire... Shifting the balance point with known weights lets you find out about the whole real weight distribution if you make enough measurements (I suggest 27 at least for a standard length racquet, better 68); I'm so tempted but I aready waste too much time...
No. It doesn’t.
 

Tregix

Rookie
I don't really care about how accurate it is, just want to match a few rackets.
It is very accurate IF you measure the balance and the hang point accurately (within 1mm).

The Swing Tool App makes things much easier and faster but the principle is exactly the same.
 

esgee48

G.O.A.T.
The SW app uses the same data. Only thing it does better is measuring the time to 10 oscillations with the stopwatch function.
 
  • Like
Reactions: AA7

jmnk

Hall of Fame
I’ve gotten into the habit now of counting 100 swings, but using the lap timer every 10 swings to keep it easier to count without worrying about losing count.
With this way it’s really easy to measure to within 0.01 seconds per 10 swings and know that it’s accurate.

Measuring the balance carefully with the edge of my metal ruler is also important - easier to do this to within 0.01” when the balance is just a mm or two away from the throat bridge as a reference point. When that’s not the case (e.g., in cases where the balance is about an inch south of the bridge) then I attach a object of known weight to the tip do that I can move the balance to near the bridge. Then calculate the balance without the weight.
If I may ask - what exactly do you mean by "counting 100 swings, but using the lap timer every 10 swings"? Are you able to have the racket oscillate 100 times on a single 'push'?
 

travlerajm

G.O.A.T.
I set it up on two thin steel nails. The nail/string friction is minimal and the racquet will oscillate for a long time.
I usually measure when the peak-to-neutral amplitude is between 1/2” and 1”. In this range, there is hardly any change in amplitude over 100 swings, so the period is extremely stable.

At larger amplitude, there is noticeable energy loss due to air drag. And at very small amplitude, the friction to inertia ratio is no longer negligible. At least this is how it seems to me based on doing it many hundreds of times.
 

jmnk

Hall of Fame
It will oscillate more than 1000 times easy.
I set it up on two thin steel nails. The nail/string friction is minimal and the racquet will oscillate for a long time.
wait, are you both saying that you:
  • hang the racket by a string near the top of the hoop,
  • you gently push the handle so the racket moves as a pendulum, with an amplitude of no more than 1 inch,
  • you leave that racket be,
  • and the racket will easily do 1000 (a thousand) full swings with amplitude still big enough that you can see it crossing the neutral point with a naked eye
??
 

Slowtwitcher

Hall of Fame
wait, are you both saying that you:
  • hang the racket by a string near the top of the hoop,
  • you gently push the handle so the racket moves as a pendulum, with an amplitude of no more than 1 inch,
  • you leave that racket be,
  • and the racket will easily do 1000 (a thousand) full swings with amplitude still big enough that you can see it crossing the neutral point with a naked eye
??
Traveljam was the one that mentioned 1000 swings. That would be 22min of swinging for the last racquets I measured, which seems a little too much. But it's cool to see how long the system oscillates.
 

travlerajm

G.O.A.T.
wait, are you both saying that you:
  • hang the racket by a string near the top of the hoop,
  • you gently push the handle so the racket moves as a pendulum, with an amplitude of no more than 1 inch,
  • you leave that racket be,
  • and the racket will easily do 1000 (a thousand) full swings with amplitude still big enough that you can see it crossing the neutral point with a naked eye
??
Yes. I’m pretty sure that’s true. Since if I start with an amplitude of about 3/4”, it will still be about 5/8” 100 swings later.

I‘ll try the experiment tonight (Set it swinging and see how long it goes).
 

HitMoreBHs

Semi-Pro
Yes. I’m pretty sure that’s true. Since if I start with an amplitude of about 3/4”, it will still be about 5/8” 100 swings later.

I‘ll try the experiment tonight (Set it swinging and see how long it goes).
Dude, I’m a geek and do make a lot of time for academic, scientific and technical experimentation. This one though... o_O
(Make sure you report back here with results! I insist!!)
 

travlerajm

G.O.A.T.
Dude, I’m a geek and do make a lot of time for academic, scientific and technical experimentation. This one though... o_O
(Make sure you report back here with results! I insist!!)
Just did the experiment.

I started it out at roughly 1” amplitude. By the time it got to 1000 swings, the amplitude was down to about 2mm.

It’s still going at 30 minutes now, and clearly visibly oscillating, but amplitude is now less than 1mm.

I used 2 plastic hex sided ball point pens weighted down with a 2-lb jar hanging over the end of the desk. I expect I could get it to go an hour if I used a more rigid and friction-free support fixture.
 

jmnk

Hall of Fame
Just did the experiment.

I started it out at roughly 1” amplitude. By the time it got to 1000 swings, the amplitude was down to about 2mm.

It’s still going at 30 minutes now, and clearly visibly oscillating, but amplitude is now less than 1mm.

I used 2 plastic hex sided ball point pens weighted down with a 2-lb jar hanging over the end of the desk. I expect I could get it to go an hour if I used a more rigid and friction-free support fixture.
If you can tell that somethin is 'clearly oscillating' with the amplitude of less than 1mm - bless your eyes....
I'm not sure if this actually improves the accuracy of the results anyway. I used to hang rackets on something flat (cylinder pencil, hex pencil, nail, etc) - but that resulted in a racket 'walking' on the support. So then I notched the the nails - which prevented walking but increased friction.
Taking an average of so many periods where, as you noted yourself, the period _does_ actually depend on the amplitude is not helping either. Ideally, theoretically, the period would be the same regardless of the amplitude - but that is not the case in a real world (@Tregix very nicely illustrated that). I believe that is the reason why SwingTool app takes only so many periods into consideration - the author likely determined that it makes sense to average only over just several periods as the period is going to vary less that way.
 

travlerajm

G.O.A.T.
If you can tell that somethin is 'clearly oscillating' with the amplitude of less than 1mm - bless your eyes....
I'm not sure if this actually improves the accuracy of the results anyway. I used to hang rackets on something flat (cylinder pencil, hex pencil, nail, etc) - but that resulted in a racket 'walking' on the support. So then I notched the the nails - which prevented walking but increased friction.
Taking an average of so many periods where, as you noted yourself, the period _does_ actually depend on the amplitude is not helping either. Ideally, theoretically, the period would be the same regardless of the amplitude - but that is not the case in a real world (@Tregix very nicely illustrated that). I believe that is the reason why SwingTool app takes only so many periods into consideration - the author likely determined that it makes sense to average only over just several periods as the period is going to vary less that way.
I might do another experiment to see if I can identify in what amplitude range the period is most stable. My guess is that the period is changing the most when the amplitude is highest (which seems logical because both air drag and friction could factor in more). And the period also seems to get less stable at really small amplitude (as the amplitude approaches zero, the period has to either go to infinity or zero).

In any case, I find measuring 100 consecutive periods (divided into laps of 10 swings for easier counting) to give me reproducible results to within 0.001 sec/period every time, which is good enough for my purposes.
 

HitMoreBHs

Semi-Pro
Do you need strings in the racket to do this? I just bought a new racket and want to match it.
You only need one string to hang the racquet if you want to do matching with the frame unstrung. I use:
1) length of leftover string
2) couple of plastic/acrylic round spacer beads
3) couple of brass crimp tubes
These are cheaply available from craft/jewelry making stores, e*ay or A*azon.

The string needs to be tensioned sufficiently so as not to sag when the racquet is hung from it. Otherwise, the measurement to the hang point will be be inaccurate.
 

jmnk

Hall of Fame
I might do another experiment to see if I can identify in what amplitude range the period is most stable. My guess is that the period is changing the most when the amplitude is highest (which seems logical because both air drag and friction could factor in more). And the period also seems to get less stable at really small amplitude (as the amplitude approaches zero, the period has to either go to infinity or zero).

In any case, I find measuring 100 consecutive periods (divided into laps of 10 swings for easier counting) to give me reproducible results to within 0.001 sec/period every time, which is good enough for my purposes.
The point is not that your measurements of ~100 swings will not give you reproducible results. The point is that the number you will get by taking the average of 100 periods, even if that number is consistently the same, is not any more correct than the average period you will get from just several swings
 

Slowtwitcher

Hall of Fame
The point is not that your measurements of ~100 swings will not give you reproducible results. The point is that the number you will get by taking the average of 100 periods, even if that number is consistently the same, is not any more correct than the average period you will get from just several swings
The first guy to figure this out did it in 1748, so I'm sure you guys eventually will get around the solution to this problem :cool:
 

travlerajm

G.O.A.T.
The point is not that your measurements of ~100 swings will not give you reproducible results. The point is that the number you will get by taking the average of 100 periods, even if that number is consistently the same, is not any more correct than the average period you will get from just several swings
If that’s the point, I disagree.

I’ve found that I can time a ten-swing period to within +/- 0.03 sec. timing 100 swings removes that error, which is enough error to throw a swingweight measurement out of whack (for my purposes).
 
Last edited:

junbun

New User
I think using 100 swings would reduce the amount that our timing error influences the calculation.

The average human reaction time is about 350 ms so I would guess manual timing with a stopwatch would have something like +- .7 seconds error even if we include imperfect detection of when the racquet crosses the swing line.
Which translates to roughly +-50 SW depends on your racquet and where you're taking the measurements from. In theory, 100 swings should reduce the amount of error introduced by the first and last swing measurements to +-5 SW

However, i'm not sure if other inaccuracies might be introduced by using 100 swings as others have mentioned.
lastly, if you are consistently late by your reaction speed then you're probably really close to the actual timing 1 - 2 SW for 10 swings.
 
Last edited:

junbun

New User
So i played with the tool a little bit and also used the swingweight tool from appmaker.se on my 3 matched Ultra tours
Unstrung specs with plastic on 305g , 290 SW, 9HL measured with a Babalot RDC at the shop.

Strung measurements with swingweight app:
All racquets have a slightly different string setup
Racquet 1 : 325 SW +6g overgrip, 325g, 32.6 balance
Racquet 2 : 324 SW, +6g overgrip 325g 32.4 balance
Racquet 3 : 340 SW,, +7g overgrip, +12g handle + 9g around the hoop from 4 to 10, 348g, 32 balance

When trying to use the TWU tool I always miss-timed and had readings of 260 - 280 SW which is lower than my unstrung spec lol.

Investigating the TWU tool further it seems these are the following errors that would affect your SW calculations:

Balance point : an error of 1mm = 2SW
Hang distane : an error of 1mm = .75 SW
Time to complete 10 swings : .1 seconds = 7 SW difference.

I don't think i can time it within an accuracy of .1 or even .2 seconds, you probably have to do video playback for this.
Now theoretically 100 swings would reduce the time error by a factor of 10
IF the racquet can swing 100 times without other errors introduced.

For comparison's sake, the swingweight app seemed much more accurate +- 2 SW at most (I used the mode of 5 tests).
On the third test I would usually have to restart the swing otherwise the racquet loses too much momentum and starts being inconsistent towards the end. That means roughly after 50 swings is when it starts getting unstable.

If you have an iphone i would pay the $4 for both the swingweight and racqettune app it saves a lot of time.
If you're on android you might be able to find some app that measures the time it takes some pendulum to cycle back and forth and then plug in those numbers into the TWU tool.
 

Aretium

Hall of Fame
I've been trying different ways to time. I think the best way is with a camera and breaking down the frames.

I've done the 50 times and 100. I ended up just doing loads of 10 swings and discarding the anomalies and working out an average.

Not sure how accurate the calculation is, but I can at least match my rackets closer.
 

travlerajm

G.O.A.T.
So i played with the tool a little bit and also used the swingweight tool from appmaker.se on my 3 matched Ultra tours
Unstrung specs with plastic on 305g , 290 SW, 9HL measured with a Babalot RDC at the shop.

Strung measurements with swingweight app:
All racquets have a slightly different string setup
Racquet 1 : 325 SW +6g overgrip, 325g, 32.6 balance
Racquet 2 : 324 SW, +6g overgrip 325g 32.4 balance
Racquet 3 : 340 SW,, +7g overgrip, +12g handle + 9g around the hoop from 4 to 10, 348g, 32 balance

When trying to use the TWU tool I always miss-timed and had readings of 260 - 280 SW which is lower than my unstrung spec lol.

Investigating the TWU tool further it seems these are the following errors that would affect your SW calculations:

Balance point : an error of 1mm = 2SW
Hang distane : an error of 1mm = .75 SW
Time to complete 10 swings : .1 seconds = 7 SW difference.

I don't think i can time it within an accuracy of .1 or even .2 seconds, you probably have to do video playback for this.
Now theoretically 100 swings would reduce the time error by a factor of 10
IF the racquet can swing 100 times without other errors introduced.

For comparison's sake, the swingweight app seemed much more accurate +- 2 SW at most (I used the mode of 5 tests).
On the third test I would usually have to restart the swing otherwise the racquet loses too much momentum and starts being inconsistent towards the end. That means roughly after 50 swings is when it starts getting unstable.

If you have an iphone i would pay the $4 for both the swingweight and racqettune app it saves a lot of time.
If you're on android you might be able to find some app that measures the time it takes some pendulum to cycle back and forth and then plug in those numbers into the TWU tool.
Are you sure you are timing starting at zero? Or are you starting your count at 1 and only timing 9 periods instead of 10?

When I time, I always use lap timer on my phone. I hit start before I start timing, so that I can position my hand with finger over the lap button and not move it. Then hit lap button next 10 times every 10 periods, so that 100 swings are measured on laps 2 thru 11. Guaranteed repeatable to within less than 0.01s per 10 swings.
 

junbun

New User
Are you sure you are timing starting at zero? Or are you starting your count at 1 and only timing 9 periods instead of 10?

When I time, I always use lap timer on my phone. I hit start before I start timing, so that I can position my hand with finger over the lap button and not move it. Then hit lap button next 10 times every 10 periods, so that 100 swings are measured on laps 2 thru 11. Guaranteed repeatable to within less than 0.01s per 10 swings.
I'm probably half a count off i'm guessing. Are you saying all your 10swing times are within .01 seconds of each other over the 100 swing period?
How are you hanging your racquet and what is the angle you are releasing at?

I have mine on 2 screwdriver cylinders, the types they use for glasses so they are pretty low on friction i think. I use about a 15 degree angle (The racquet will swing off-axis if the angle is too high depends on how coordinated and how good your setup is ). I wait for it to swing a few times before starting the count. The total error the app estimates is usually .002s sometimes it will give a .004s error which i don't use that result.

BTW since we have the same racquets I measured my hang distance at 64.3cm with 17 gut strings, curious if we have the same hang distance.
I always think my measurements are off.

@Aretium
Pretty much every method that gets swing measurements by hanging off the top mains uses the same parallel axis formula, the accuracy depends on how well your setup is, and measurements.
 

Aretium

Hall of Fame
I'm probably half a count off i'm guessing. Are you saying all your 10swing times are within .01 seconds of each other over the 100 swing period?
How are you hanging your racquet and what is the angle you are releasing at?

I have mine on 2 screwdriver cylinders, the types they use for glasses so they are pretty low on friction i think. I use about a 15 degree angle (The racquet will swing off-axis if the angle is too high depends on how coordinated and how good your setup is ). I wait for it to swing a few times before starting the count. The total error the app estimates is usually .002s sometimes it will give a .004s error which i don't use that result.

BTW since we have the same racquets I measured my hang distance at 64.3cm with 17 gut strings, curious if we have the same hang distance.
I always think my measurements are off.

@Aretium
Pretty much every method that gets swing measurements by hanging off the top mains uses the same parallel axis formula, the accuracy depends on how well your setup is, and measurements.
The software attached in the video is to 0.01 seconds. So it's basically that accurate. The angle which you swing doesn't matter. Just as long as you let it swing a little first.
 

Aretium

Hall of Fame
I'm probably half a count off i'm guessing. Are you saying all your 10swing times are within .01 seconds of each other over the 100 swing period?
How are you hanging your racquet and what is the angle you are releasing at?

I have mine on 2 screwdriver cylinders, the types they use for glasses so they are pretty low on friction i think. I use about a 15 degree angle (The racquet will swing off-axis if the angle is too high depends on how coordinated and how good your setup is ). I wait for it to swing a few times before starting the count. The total error the app estimates is usually .002s sometimes it will give a .004s error which i don't use that result.

BTW since we have the same racquets I measured my hang distance at 64.3cm with 17 gut strings, curious if we have the same hang distance.
I always think my measurements are off.

@Aretium
Pretty much every method that gets swing measurements by hanging off the top mains uses the same parallel axis formula, the accuracy depends on how well your setup is, and measurements.
Sorry didn't spot, I use 1.30mm strings and I'm getting 64.5?????
 

jmnk

Hall of Fame
[...]

For comparison's sake, the swingweight app seemed much more accurate +- 2 SW at most (I used the mode of 5 tests).
On the third test I would usually have to restart the swing otherwise the racquet loses too much momentum and starts being inconsistent towards the end. That means roughly after 50 swings is when it starts getting unstable.

[...]
@junbun
what do you mean by "the mode of 5 tests"?
what do you mean by "On the third test I would usually have to restart the swing"
 

junbun

New User
Sorry didn't spot, I use 1.30mm strings and I'm getting 64.5?????
Sorry I was referring to travlerajm and the Ultra Tours we both have.

I think your video method is probably more accurate. You can probably calculate your twist weight too .... turn the racquet 90degrees and subtract that number from your SW to get TW.
 

junbun

New User
@jmnk
"mode" meaning I take the number that shows up most often instead of the mean or median. So my last few tests came out 342 340 340 340 339
I use 240

I release the racquet once and take multiple tests without stopping the racquet. Each test on the app takes 16 swings and 1 additional in the beginning. during the third test the racquet will usually be too slow to for the app to track @ around 50 swings.
 

jmnk

Hall of Fame
@jmnk
"mode" meaning I take the number that shows up most often instead of the mean or median. So my last few tests came out 342 340 340 340 339
I use 240
I see, makes sense (if you meant 340 that is)

I release the racquet once and take multiple tests without stopping the racquet. Each test on the app takes 16 swings and 1 additional in the beginning. during the third test the racquet will usually be too slow to for the app to track @ around 50 swings.
Interesting approach. I'm not sure if that is the way to go though. If you look up the very interesting graph by @Tregix you will see that in practice the period depends on the amplitude - so the periods in your 'second' and 'third' measuring try will be less and less than the ones measured on the first one - because the amplitude gets smaller. It is actually a not insignificant amount. I would think you may be better off making sure you always start with the same initial amplitude, let the app measure over those 16 swings, and then restart the process again from the same initial amplitude.
 
Last edited:

Irvin

Talk Tennis Guru
The software attached in the video is to 0.01 seconds. So it's basically that accurate. The angle which you swing doesn't matter. Just as long as you let it swing a little first.
If you believe the angle does not matter try lifting the racket a inch farther and see if you get the same reading.
 

Irvin

Talk Tennis Guru
@Irvin

Do you know what angle is best to use ?
The smaller the better. It really depends on the pivot. You want to displace the racket far enough to get the number of periods you need and have enough displacement to measure accurately. Try to keep the displacement exactly the same each time you start.
 
Top