First Serve: Should Elbow Should Nearly Straight At Contact?

[Better_Call_Raul]

First serve: Had always thought of elbow being straight/very nearly straight at contact.
That is what Nate explains here.

But this study shows significant elbow bend of 35° at impact... Although not clear if the serve being studied was first or second.


High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

 

[5263]

First serve: Had always thought of elbow being straight/very nearly straight at contact.
That is what Nate explains here.

But this study shows significant elbow bend of 35° at impact... Although not clear if the serve being studied was first or second.


High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

is the elbow bent or is it a bend or angle at the wrist...this video is pretty poor...

 

[Better_Call_Raul]

is the elbow bent or is it a bend or angle at the wrist...

At impact, the elbow is straight/very nearly straight. Very little bend of the elbow. That's how I think of it. But study shows significant 35° bend,
At impact, the wrist is neither flexed nor extended. The wrist will be in a roughly neutral position at impact.
Video lesson says the contact should be right above the right shoulder.

 

[5263]

At impact, the elbow is straight/very nearly straight. Very little bend of the elbow. That's how I think of it. But study shows significant 35° bend,
At impact, the wrist is neither flexed nor extended. The wrist will be in a roughly neutral position at impact.
Video lesson says the contact should be right above the right shoulder.

It should be in line with the shoulder when view from behind the server, but in line with the swing direction.....but even though in line, the contact should be into the court in front of the shoulder.... not truly above it... are should be very close to straight with a big server and finish getting straight for the 4 Ms contact thru the ball...

 

[Better_Call_Raul]

It should be in line with the shoulder when view from behind the server, but in line with the swing direction.....but even though in line, the contact should be into the court in front of the shoulder.... not truly above it... are should be very close to straight with a big server and finish getting straight for the 4 Ms contact thru the ball...

Yes, contact should be into the court. This dimension is not shown in the rear view.

From rear view, and assuming one follows the tip of making contact exactly in line with the right shoulder, does that determine the clock position?
Or can one still adjust the clock position by manipulating the wrist?

Is the contact in this pic considered contacting at 12 o'clock?

 

[Dragy]

At impact, the elbow is straight/very nearly straight. Very little bend of the elbow. That's how I think of it. But study shows significant 35° bend,
At impact, the wrist is neither flexed nor extended. The wrist will be in a roughly neutral position at impact.
Video lesson says the contact should be right above the right shoulder.

Which study, lol

 

[nyta2]

i feel like my elbow is slightly bent at contact...
it's mostly straight if looking from the outside, but i definitely have a slight bend
a long time ago i used to almost lock my elbow, and hurt my elbow (medial epi?) that way

 

[Better_Call_Raul]

Which study, lol

Biomechanics of The Elbow Joint in Tenni PDF | PDF | Elbow | Anatomical Terms Of Motion

Scribd is the world's largest social reading and publishing site.

www.scribd.com

High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

 

[Better_Call_Raul]

i feel like my elbow is slightly bent at contact...
it's mostly straight if looking from the outside, but i definitely have a slight bend
a long time ago i used to almost lock my elbow, and hurt my elbow (medial epi?) that way

Roughly how bent of elbow at contact? Does degree of elbow bend differ from first serve 12'o'clock contact vs first serve 1 o'clock contact?


Does elbow bend at contact differ from flat and kick? (I know that wrist position is more deviated at contact for kick. not sure about elbow bend).

35 degrees of elbow bend at contact sounds too high... will need to check if it was a club player study.

High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

 

[Dragy]

Biomechanics of The Elbow Joint in Tenni PDF | PDF | Elbow | Anatomical Terms Of Motion

Scribd is the world's largest social reading and publishing site.

www.scribd.com

High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

So the article refers to a study of 1994, medical one:

Kibler WB
. Clinical biomechanics of the elbow in tennis: implications for evaluation and diagnosis.
Med Sci Sports Exerc
1994

It’s very possible they didn’t study pro players with proper mechanics, but those who suffered tennis elbow. This would explain these figures.

You need to do sanity checks before you trust any statement.

 

[nyta2]

Roughly how bent of elbow at contact? Does degree of elbow bend differ from first serve 12'o'clock contact vs first serve 1 o'clock contact?


Does elbow bend at contact differ from flat and kick? (I know that wrist position is more deviated at contact for kick. not sure about elbow bend).

35 degrees of elbow bend at contact sounds too high... will need to check if it was a club player study.

High‐speed video analysis studies from Kibler et al5 have demonstrated that during the service the elbow moves from 116° to 20° of flexion within 0.21 s,
with ball impact occurring at approximately 35° of flexion.

i don't know.. but i'm guessing like throwing a punch, where you don't lock out your elbow, i'm probably doing something similar (eg. tiny bend)
[edit]
actually now that shadowed it... 35 at contact, and 20 max "straightening" sounds right

 

[Better_Call_Raul]

actually now that shadowed it... 35 at contact, and 20 max "straightening" sounds right

That degree of elbow bend at contact sounds extreme.The pros seem to be straight/nearly straight at contact.
And according to Elliott study, elbow should be nearly straight just prior to contact for peak ISR to occur.

 

[Dragy]

actually now that shadowed it... 35 at contact, and 20 max "straightening" sounds right

Don’t even let this BS into your brain

 

[nyta2]

Don’t even let this BS into your brain

thx for the pics...
for flat at contact, feels like say 10?
but definitely never straightens completely locked.
was thinking my spin serves are slightly more bent... dunno probably need to improve, like everything else

 

[Dragy]

but definitely never straightens completely locked.

Completely locked is over-extended actually, unless person has limited range... If you hang off the bar, it's not same as "locked" elbow, for example.

I don't know whether science would determine it as 0 or 10 degrees, but best servers wouldn't hold arm straightening as the mass of accelerated racquet gets the best out of it. However, you will find some examples of clear bend at contact, like Felix AA. Not more than 10-15 degrees still, as far as I can judge.

 

[SystemicAnomaly]

That degree of elbow bend at contact sounds extreme.The pros seem to be straight/nearly straight at contact.
And according to Elliott study, elbow should be nearly straight just prior to contact for peak ISR to occur.

I don’t recall. Does Elliott et al actually say that? It is possible to achieve considerable ISR with the elbow bent — even bent 90°.

Nevertheless, for elite servers, the arm is straight or nearly straight (a few degrees of flexion perhaps) at contact.

 

[Better_Call_Raul]

I don’t recall. Does Elliott et al actually say that? It is possible to achieve considerable ISR with the elbow bent — even bent 90°.

Nevertheless, for elite servers, the arm is straight or nearly straight (a few degrees of flexion perhaps) at contact.

Elliott study was of "high performance players".
The rotation of the humeroulnar joint (elbow) was measured in radians per second. Elbow extension was measured as an angular velocity.

Around the time of peak of ISR, the elbow extension is very low. Around 4 radians per second during peak ISR.
Close to zero radians per second elbow extension measurement at 1/100 second before impact.
This indicates that the elbow has almost straightened at impact and is nearly straight when peak ISR occurs.

a few degrees of flexion perhaps) at contact.

If I understand the graph correctly there is no elbow flexion near impact. Elbow is extending near impact.
If elbow was bending (flexing) approaching impact there would be negative values measured for elbow extension.
Elbow flexion does occurs post-impact as seen by negative values. See the "blank triangles" that represent elbow extension (straightening).

 

[SystemicAnomaly]

Elliott study was of "high performance players".
The rotation of the humeroulnar joint (elbow) was measured in radians per second. Elbow extension was measured as an angular velocity.

Around the time of peak of ISR, the elbow extension is very low. Around 4 radians per second during peak ISR.
Close to zero radians per second elbow extension measurement at 1/100 second before impact.
This indicates that the elbow has almost straightened at impact and is nearly straight when peak ISR occurs.


If I understand the graph correctly there is no elbow flexion near impact. Elbow is extending near impact.
If elbow was bending (flexing) approaching impact there would be negative values measured for elbow extension.
Elbow flexion does occurs post-impact as seen by negative values. See the "blank triangles" that represent elbow extension (straightening).

I believe that "flexion" and "extension" can refer either to position or to action. Since the Y-axis is angular velocity, positive & negative values do not (necessarily) reflect position. As I understand it, a positive value reflects an extension action whereas a negative value reflects a flexion action. This seems reasonable to my mind but I could be mistaken.

 

[SystemicAnomaly]

^ The terms (above) might normally refer to movements (actions) rather than position. For position, we might say full extension (or even hyperextension) and full flexion. We might also refer to 90° of elbow flexion. Would it be proper to also call that position 90° of extension?

 

[Better_Call_Raul]

I believe that "flexion" and "extension" can refer either to position or to action. Since the Y-axis is angular velocity, positive & negative values do not (necessarily) reflect position. As I understand it, a positive value reflects an extension action whereas a negative value reflects a flexion action. This seems reasonable to my mind but I could be mistaken.

It is confusing. Can definitely see where elbow would bend (flex) post-impact. That makes sense.

But the study is measuring elbow extension as an angular velocity not as a linear velocity. Note that both approaches are valid.
Even with a straight arm at/near impact, the elbow can continue to rotate angularly in the same direction post-contact, via ISR.
Do not see why elbow extension angular velocity, measured around the humeroulnar joint, would suddenly go negative post-impact.


In the high-velocity tennis serve, the contributions that the upper limb segments' anatomical rotations make to racket head speed at impact depend on both their angular velocity and the instantaneous position of the racket with respect to the segments' axes of rotation. Eleven high-performance tennis players were filmed at a nominal rate of 200 Hz by three Photosonics cameras while hitting a high-velocity serve. The three-dimensional (3-D) displacement histories of 11 selected landmarks were then calculated using the direct linear transformation approach, and 3-D individual segment rotations for the upper limb were calculated using vector equations (Sprigings, Marshall, Elliott, & Jennings, 1994).

The major contributors to the mean linear velocity of the center of the racket head of 31.0 m · s-1 at impact were internal rotation of the upper arm (54.2%), flexion of the hand (31.0%), horizontal flexion and abduction of the upper arm (12.9%), and racket shoulder linear velocity (9.7%).

Forearm extension at the elbow joint played a negative role (-14.4%) and reduced the forward velocity of the center of the racket at impact.

 

[SystemicAnomaly]

It is confusing. Can definitely see where elbow would bend (flex) post-impact. That makes sense.

But the study is measuring elbow extension as an angular velocity not as a linear velocity. Note that both approaches are valid.
Even with a straight arm at/near impact, the elbow can continue to rotate angularly in the same direction post-contact, via ISR.
Do not see why elbow extension angular velocity, measured around the humeroulnar joint, would suddenly go negative post-impact.


In the high-velocity tennis serve, the contributions that the upper limb segments' anatomical rotations make to racket head speed at impact depend on both their angular velocity and the instantaneous position of the racket with respect to the segments' axes of rotation. Eleven high-performance tennis players were filmed at a nominal rate of 200 Hz by three Photosonics cameras while hitting a high-velocity serve. The three-dimensional (3-D) displacement histories of 11 selected landmarks were then calculated using the direct linear transformation approach, and 3-D individual segment rotations for the upper limb were calculated using vector equations (Sprigings, Marshall, Elliott, & Jennings, 1994).

The major contributors to the mean linear velocity of the center of the racket head of 31.0 m · s-1 at impact were internal rotation of the upper arm (54.2%), flexion of the hand (31.0%), horizontal flexion and abduction of the upper arm (12.9%), and racket shoulder linear velocity (9.7%).

Forearm extension at the elbow joint played a negative role (-14.4%) and reduced the forward velocity of the center of the racket at impact.

It makes sense that elbow extension (& flexion) are measured as an angular velocity rather than as a linear velocity. The whole forearm moves wrt to the upper arm as a unit, at certain angular v at any given moment. But different parts of the forearm are moving at different linear velocities -- forearm close to the wrist is moving much faster than the forearm close to the elbow. Yet they are moving at the same angular velocity. So angular velocity makes sense where linear velocity does not.

Once the elbow is fully extended (or hyper-extended), that angular v goes to zero (shortly after impact on the graph). The elbow turning due to ISR before, during and after contact, has nothing to do with elbow extension. The axis of rotation for the elbow has nothing to do with the axis of rotation of the shoulder. Separate actions.

When the angular v of the elbow (extension) goes negative, shortly after contact, it means that the elbow is now flexing rather than extending.

What do you mean by forearm extension? Are you referring to elbow extension?

 

[Better_Call_Raul]

It makes sense that elbow extension (& flexion) are measured as an angular velocity rather than as a linear velocity.

The other table in the study measures elbow extension as a linear velocity. Not clear which approach is the common practice.
Can this linear velocity approach also approximate the contribution to racket head speed just as the other table did using angular velocity approach.

Also not clear how in the world ISR in the below table can be measured as a linear velocity?!...
But apparently ISR peaks at 19 meters per second over the course of 0.1 seconds.

Once the elbow is fully extended (or hyper-extended), that angular v goes to zero (shortly after impact on the graph). The elbow turning due to ISR before, during and after contact, has nothing to do with elbow extension. The axis of rotation for the elbow has nothing to do with the axis of rotation of the shoulder. Separate actions.

This is true but not clear whether this can be clearly separated. Maybe so.
But visually, the elbow angular movement in space (whether it be via elbow extension or ISR) looks to very similar, and in the same angular direction,
both approaching contact, at contact and immediately post-contact. Do not see why angular velocity should suddenly drop to negative value... But you could be right and these sensors are able to separate the motions.

What do you mean by forearm extension? Are you referring to elbow extension?

That is the Elliott quote. Assuming forearm extension refers to elbow extension.

"Forearm extension at the elbow joint played a negative role (-14.4%) and reduced the forward velocity of the center of the racket at impact."

 

[SystemicAnomaly]

The other table in the study measures elbow extension as a linear velocity. Not clear which approach is the common practice.
Can this linear velocity approach also approximate the contribution to racket head speed just as the other table did using angular velocity approach.

Also not clear how in the world ISR in the below table can be measured as a linear velocity?!...
But apparently ISR peaks at 19 meters per second over the course of 0.1 seconds.

This is true but not clear whether this can be clearly separated. Maybe so.
But visually, the elbow angular movement in space (whether it be via elbow extension or ISR) looks to very similar, and in the same angular direction,
both approaching contact, at contact and immediately post-contact. Do not see why angular velocity should suddenly drop to negative value... But you could be right and these sensors are able to separate the motions.



That is the Elliott quote. Assuming forearm extension refers to elbow.

"Forearm extension at the elbow joint played a negative role (-14.4%) and reduced the forward velocity of the center of the racket at impact."

Do you have a link handy for the rest of the study? I don't recall, offhand, where I posted it.

The elbow angular movement due to extension does NOT look at all like the rotation of the elbow due to ISR. They are NOT in the same direction -- the 2 angular movements are orthogonal to each other.

 

[Better_Call_Raul]

Do you have a link handy for the rest of the study? I don't recall, offhand, where I posted it.

The elbow angular movement due to extension does NOT look at all like the rotation of the elbow due to ISR. They are NOT in the same direction -- the 2 angular movements are orthogonal to each other.

Think this is the entire Elliott study. Must download 10 page pdf.

(PDF) Long-axis rotation: The missing link in proximal-to-distal segmental sequencing

PDF | Most assessments of segmental sequencing in throwing, striking or kicking have indicated a proximal-to-distal sequencing of end-point linear... | Find, read and cite all the research you need on ResearchGate

www.researchgate.net

 

[SystemicAnomaly]

Think this is the entire Elliott study. Must download 10 page pdf.

(PDF) Long-axis rotation: The missing link in proximal-to-distal segmental sequencing

PDF | Most assessments of segmental sequencing in throwing, striking or kicking have indicated a proximal-to-distal sequencing of end-point linear... | Find, read and cite all the research you need on ResearchGate

www.researchgate.net

Ouch. Not sure my eyes or phone are up to the task.

With the table you mentioned, what body part is referenced for linear velocity data? Was it the hand, some part of the forearm or something else?

 

[Better_Call_Raul]

Ouch. Not sure my eyes or phone are up to the task.

With the table you mentioned, what body part is referenced for linear velocity data? Was it the hand, some part of the forearm or something else?

This table indicates an average value of ISR of 16.8 meters per second.
Not sure how ISR can be measured with linear velocity. ISR should be measured with angular velocity.

MarshallElliottLongAxis (3).pdf

Shared with Jumpshare

jumpshare.com

 

[SystemicAnomaly]

This table indicates an average value of ISR of 16.8 meters per second.
Not sure how ISR can be measured with linear velocity. ISR should be measured with angular velocity.

MarshallElliottLongAxis (3).pdf

Shared with Jumpshare

jumpshare.com

K, I think I see what's going on here.

16.8 m/s is not the (mean) linear speed of ISR. 16.8 m/s relates to the linear speed of the centre of the racquet head at impact. The total RHS here is 31.0 m/s. 16.8 m/s is merely how much ISR contributes to that total. It is not meant to suggest that is a value for ISR itself.

Likewise, elbow extension appears to have a negative contribution to total RHS at impact. But this is not saying that -4.4 m/s is the value of elbow extension itself. If you add up all the contributions listed, they should add up to 31.0 m/s

Most, if not all, of the movements of these various segments would be measured as angular velocities. However, their contributions to total RHS are listed as linear velocities.

This table shows that IRS contributes some 54.2% to RHS at impact while forearm pronation only contributes 5.2% at impact. But I don't think that means that pronation doesn't provide a greater contribution earlier in the upward swing (prior to the onset of ISR).

It is interesting to note that wrist actions (flexion & ulnar deviation) contribute about 31.2% at impact.

 

[Better_Call_Raul]

Likewise, elbow extension appears to have a negative contribution to total RHS at impact. But this is not saying that -4.4 m/s is the value of elbow extension itself. If you add up all the contributions listed, they should add up to 31.0 m/s

So this graph displays the speed of the centre of the racquet head, in forward direction perpendicular to baseline, at any instant.

The graph also displays the contributions from 7 movements: Leg and torso, ISR, upper arm abduction, pronation, elbow extension, wrist flexion, wrist ulnar flexion.

If one chooses t=-0.15 seconds before impact (point of full racquet drop where racquet is not moving forward)
these 7 velocities will add up to 0 meters per second RHS.

If one chooses t=0.0 seconds (point of impact) all of these 7 velocities will add up to 31.0 meters per second RHS.

This graph will tell you that elbow extension contributes negative 4.4 meters per second to RHS at impact.

If one wants to know the value of elbow extension at impact, he can look at the other graph, which displays angular velocity at impact of elbow extension.

We can see that elbow extension value is still still slightly positive at impact. This means the elbow is still extending not fully straight at impact.
We can confirm from server pic at t=0.0 that elbow is nearly straight.
Elbow extension value is negative only after !1/100 second after impact. This means that server does not start to bend the elbow till after impact.

 

[SystemicAnomaly]

So this graph displays the speed of the centre of the racquet head, in forward direction perpendicular to baseline, at any instant.

The graph also displays the contributions from 7 movements: Leg and torso, ISR, upper arm abduction, pronation, elbow extension, wrist flexion, wrist ulnar flexion.

If one chooses t=-0.15 seconds before impact (point of full racquet drop where racquet is not moving forward)
these 7 velocities will add up to 0 meters per second RHS.

If one chooses t=0.0 seconds (point of impact) all of these 7 velocities will add up to 31.0 meters per second RHS.

This graph will tell you that elbow extension contributes negative 4.4 meters per second to RHS at impact.

If one wants to know the value of elbow extension at impact, he can look at the other graph, which displays angular velocity at impact of elbow extension.

We can see that elbow extension value is still still slightly positive at impact. This means the elbow is still extending not fully straight at impact.
We can confirm from server pic at t=0.0 that elbow is nearly straight.
Elbow extension value is negative only after !1/100 second after impact. This means that server does not start to bend the elbow till after impact.

I’m confused by your statements here

I didn’t come to these conclusions from inspection of the graph. The graph above displays radians/s, not meters/s. It displays angular speeds of various joint rotations / articulations. No linear speeds.

While Table 1 shows RHS contributions at impact, I am not seeing how you are extracting RHS from the graph. Can you elaborate?

Other graph? Is there another graph that displays the values in Table 1? 7 velocities? The only graph that I see displays 6 angular velocities.

I do agree with the last part of your post here. The small positive value for elbow extension at impact implies that it is still straightening (unless the elbow is already straight and it is now hyperextending).

The graph does suggest that the elbow starts to flex shortly after contact. I found this a bit unexpected. I would have expected elbow flexion a little bit than this.

With the Sampras serve, we see quite a bit of elbow flexion after contact. But with other servers, like Federer, there is little or no elbow flexion during the follow-thru.

 

[Better_Call_Raul]

i don't know.. but i'm guessing like throwing a punch, where you don't lock out your elbow, i'm probably doing something similar (eg. tiny bend)
[edit]
actually now that shadowed it... 35 at contact, and 20 max "straightening" sounds right

Pretty close. You are not too far off.
It feels terrible to contact with a straight or nearly straight arm. Feel a lot of stress in triceps. Injury risk???
Feels much more comfortable to have a nice 10-24 degree bend... Lots of leeway here!

"At ball contact, Elbow flexion (20° ± 4°)".

An 8-Stage Model for Evaluating the Tennis Serve: Implications for Performance Enhancement and Injury Prevention

The tennis serve is a complex stroke characterized by a series of segmental rotations involving the entire kinetic chain. Many overhead athletes use a basic 6-stage throwing model; however, the tennis serve does provide some differences.To support the ...

www.ncbi.nlm.nih.gov

 

[Dragy]

Pretty close. You are not too far off.
It feels terrible to contact with a straight or nearly straight arm. Feel a lot of stress in triceps. Injury risk???
Feels much more comfortable to have a nice 10-24 degree bend... Lots of leeway here!

"At ball contact, Elbow flexion (20° ± 4°)".

An 8-Stage Model for Evaluating the Tennis Serve: Implications for Performance Enhancement and Injury Prevention

The tennis serve is a complex stroke characterized by a series of segmental rotations involving the entire kinetic chain. Many overhead athletes use a basic 6-stage throwing model; however, the tennis serve does provide some differences.To support the ...

www.ncbi.nlm.nih.gov

Do you get that this picture isn’t suitable to measure elbow flexion?

 

[Better_Call_Raul]

Do you get that this picture isn’t suitable to measure elbow flexion?

I'm not using the picture to measure. Probably cannot measure accurately from that view.
I am citing the research study findings:

"At ball contact, Elbow flexion (20° ± 4°)".

 

[Dragy]

I'm not using the picture to measure. Probably cannot measure accurately from that view.
I am citing the research study findings:

"At ball contact, Elbow flexion (20° ± 4°)".

You can always make sanity check and measure yourself if this can be true. Can you pick a frame from high speed video with proper angle and measure if 20 deg is anyway close to reality?

One of these angles is perfect. To my eye, it’s not hyper-extended, but I cannot see 20 degrees here. It’s important to put lines correctly though, along the bones or whatever…

 

[Better_Call_Raul]

While Table 1 shows RHS contributions at impact, I am not seeing how you are extracting RHS from the graph. Can you elaborate?

I was looking at the linear velocity graph. RHS at any instant is the sum of the 7 motions at that instant.
IIRC, there are two graphs. One measuring linear velocity contributions and one measuring angular velocity of the individual 7 motions.
Do not believe anything can be extrapolated about RHS from the angular velocity graph; it is simply a raw angular velocity measurement of the 7 motions.

Believe both graphs and measurements must be made (angular velocity and linear velocity) in order to finally arrive at the RHS speed at impact and contributions at impact displayed in the table:

wrt elbow extension position at contact
My sanity check is the shadow swing, 15 degrees bend feels very comfortable. Which is within the 20 +/- 5 degrees cited in the Kovacs 8 stage study.

Pictures are confusing and I'm not good at drawing the lines and measuring the elbow extension angle at impact...

On a possibly related note, I have trouble getting the racquet face to the inside of the forearm at impact.
It is more in line with forearm, as in below Tomaz pic, or even to the right. Tossing more towards 12 o'clock does not seem to help much.
Racquet face is still to the outside of forearm at impact.

 

[SystemicAnomaly]

I was looking at the linear velocity graph. RHS at any instant is the sum of the 7 motions at that instant.
IIRC, there are two graphs. One measuring linear velocity contributions and one measuring angular velocity of the individual 7 motions.
Do not believe anything can be extrapolated about RHS from the angular velocity graph; it is simply a raw angular velocity measurement of the 7 motions.

Believe both graphs and measurements must be made (angular velocity and linear velocity) in order to finally arrive at the RHS speed at impact and contributions at impact displayed in the table:

wrt elbow extension position at contact
My sanity check is the shadow swing, 15 degrees bend feels very comfortable. Which is within the 20 +/- 5 degrees cited in the Kovacs 8 stage study.

Pictures are confusing and I'm not good at drawing the lines and measuring the elbow extension angle at impact...

On a possibly related note, I have trouble getting the racquet face to the inside of the forearm at impact.
It is more in line with forearm, as in below Tomaz pic, or even to the right. Tossing more towards 12 o'clock does not seem to help much.
Racquet face is still to the outside of forearm at impact.

Just saw the other graph(s). I’ll have to look at this later…

 

[TennisCJC]

On a possibly related note, I have trouble getting the racquet face to the inside of the forearm at impact.
It is more in line with forearm, as in below Tomaz pic, or even to the right. Tossing more towards 12 o'clock does not seem to help much.
Racquet face is still to the outside of forearm at impact.

I would see a pro and take a few lessons get the hand outside of contact is critical to hitting a good serve without impinging the shoulder joint. the photo on the L where the shoulder, arm, hand and racket head are roughly in a straight line requires a steep angle of the arm out of the shoulder joint. That isn't good for your shoulder and it isn't good for your serve. The photo on R is much better but the contact is a bit far to the R for my liking. You could have the contact above the R shoulder and still have the hand outside of ball contact. How is your kick or topspin serve? Near impossible to hit a bit of topspin unless you can get that angle. Pictures of Federer and Stosur show an angle of near 45 degrees for their 2nd serves. This angle promotes ISR and pronation too. Try choking up, standing at the service line and hitting serves where you concentrate on nothing but getting the hand outside contact. You can start these serves from the racket drop position. Initially stop the hitting hand where it points to the sky to your R. Good conti grip is required. Or, ignore my advice and see a good pro but this is critical to serving well and your health.

 

[Better_Call_Raul]

Just the other graph(s). I’ll have to look at this later…

Also note that the graph alone does not directly indicate the position of the elbow at impact.
But it does show that elbow motion suddenly changes from extension to flexion around t = 0.0.

And the corresponding blurry server pic at t = 0.0, suggests arm is nearly neutral at impact.

Looking at the graph alone (without server pic), it is possible that elbow is severely bent at/near impact and then suddenly bends 1/100 second after impact.
But to suddenly that would be very awkward physically... Most likely the arm is within 20 degrees of neutral at impact. And then goes to neutral 1/100 second after impact before going into flexion.

 

[Better_Call_Raul]

The photo on R is much better but the contact is a bit far to the R for my liking

Tomaz from FeelTennis is a great coach but not sure why he has contact way off to the right there. Maybe he is exaggerating.
But Tomaz could have been demoed better.

I do try to drop my front shoulder to avoid shoulder impingement. And I am moving my toss towards 12 o,clock and trying to contact
above right shoulder just as Nate/PlayYourCourt demonstrates...

How is your kick or topspin serve?

Have never been able to learn kick. Struggle hugely with kick. Can only hit flat and some slice...

The "racquet to the outside at contact" problem might be wrist deviated outwards instead of inwards.