Let’s see.
We’re trying to do some science, in an industry that has too little of it — so yes, it’s supposed to start out with a hypothesis, then an experiment design, gathering data, analysis plus a conclusion. But it’s 2024…so we’ll cut straight to a TLDR — which is that our results show that
If that’s all you’re gonna remember that’s good enough — but if you want more than a tweet’s worth of information — here’s what you’ll get in the rest of the article.
If you’re reading this article you probably don’t need to be told how popular Pickleball is — but you might not realise how popular Raw Carbon Fiber is as a hitting surface for Pickleball paddles. While the Joola Hyperion CFS 16 paddle we used for our testing wasn’t the first Raw Carbon Fiber paddle on the market — it certainly ushered in a new era of popularity for the surface — with Raw Carbon Fiber being the dominant surface form on performance pickleball paddles today.
So let’s take a moment to clarify what RAW Carbon Fiber means — aka Pickleball paddle construction 101.
Raw Carbon Fiber paddles have a central core (for example Polypropylene Honeycomb) - attached to the core are multiple layers of Unidirectional Carbon Fiber.
In this picture you can see the white plastic (the core) and the black sheets attached to it. The directional strips you see are exactly what you think they are - unidirectional carbon fiber strands.
On a Joola Hyperion CFS 16 — there are 3 layers of Unidirectional Carbon Fiber (in a 0–90–0 configuration — which means the middle layer runs perpendicular to the other layers)
It is important to note that actual carbon fibers are soft and flexible filaments that become rigid and strong when they are cured inside a resin matrix.
Despite their impressive physical properties — it’s worth remembering that when a ball is hitting a Raw Carbon Fiber pickleball paddle — it is hitting RESIN, it is NOT hitting carbon fibers.
Let's try that again:
Because when you buy a USB Microscope on Aliexpress and look at the surface of a pickleball paddle you will see something like this:
And you might conclude that you are seeing a woven carbon fiber surface.
And you’d be WRONG.
You are indeed seeing the results of a woven surface — but it’s the woven surface IMPRINTED/EMBOSSED in the resin layer as the paddle is curing. (This woven surface is made by a thin non-stick sheet called a peel-ply but that’s a subject for another article)
But just cause if you’re reading this far — you just maybe might be a teeny weenie bit of a paddle nerd — here you go:
If it looks shiny it's because it is.
It’s a giveaway that it’s still attached. Now watch what happens when you remove it.
It looks dull now - doesn't it?
And that's pretty much it. This process has been used for years in composite construction to add texture to materials to promote bonding. The unusual thing about pickleball is that it's being used to make an actual finished surface.
(The analogy that might be helpful here is that if the paddle surface was a chocolate layer cake — the carbon fibers would become the LAYERS — and the resin becomes the filling/icing. Then the texture pattern of the peel-ply is like embossing a pattern into the top layer of icing)
It is this textured resin surface that is helping grab the ball during contact and it’s a major factor to influencing the ability to spin. (if you need to know why spin is important — you’re probably not a sufficiently advanced player yet — let’s just say it’s a vital direct or indirect contributor to power, control and placement of the ball)
Ok back to why we did this.
After my first few times playing Pickleball, I found myself thinking, “This is fun let’s get a good paddle” and I went to test a bunch and the Hyperion easily stood out above the rest.
$220 seemed steep — but I did it anyway. I loved the touch, I loved the spin etc etc…
And then I started playing almost every day.
And as a few weeks went by I thought “That’s weird I can’t hit as much topspin” and I thought there might be something wrong with my paddle.
And there was.
It became apparent that my surface was wearing out. This was very strange to me because I was less than a month into playing with my new paddle.
I started talking to other players and discovered a surprising fact — EVERYONE has this problem. It turns out that while it’s not commonly discussed by Manufacturers — Some high level pros swap out their paddles after every GAME!!!
Of course most manufacturers know this (they supply the paddles to their pros after all) — but it’s clearly not in their interest to discuss it openly.
Well to us — this seemed ridiculous — a premium paddle that seemed noticeably degraded after a month of play.
And with that— we decided to solve this problem.
The logic was simple. Since we wanted to see what happens to the hitting surface of a paddle when we hit a bunch of balls — we obviously needed to
(a) SCAN the paddle first
(b) hit a bunch of balls
(c) SCAN the paddle again
(d) hit a bunch of balls
And keep repeating this until we were done.
And then we could look at the results.
Figuring out the pickleball equipment was easy — we already owned a ball machine and a bunch of balls — and selecting the Hyperion made sense because it was the most popular paddle around (and acknowledged as one of the best at the time)
Paddle: Joola Hyperion CFS 16
(with Hesacore Tour Grip)
Ball Machine: Pickleball Tutor Plus
Balls: Franklin X-40
Net: Swiftnet 2.1
But deciding how to scan it required a little more thinking.
Fortunately we had guidance from USA Pickleball’s (USAP) Equipment validation and testing group — the contact surface profilometer they use is the current gold standard in the industry (the Starrett SR-160) so that was the easy choice.
USAP informed us they were intending to move to a higher quality optical measuring system — but at the time we started our test they had not finalised their choice.
After some research were able to test and evaluate a very high end system by Keyence — (the Keyence VK-X300). We performed all our tests on the Starrett and the Keyence VK.
USAP eventually chose an optical system by Filmetrics (the ProFilm 3D) so for completeness we tested our hero paddle in the Profilm 3D and the results were consistent.
Side-note — although we were not aware of any instances of disbonding with a Joola Hyperion — we did test for disbonding/delamination using the Olympus Bondmaster 600 that USAP also uses — but we found no instances of either during our test.
Contact analysis tool: Starrett SR-160
Optical analysis tool: Keyence VK-X3000
Additional validation: Filmetrics ProFilm 3D
Disbonding test: Olympus Bondmaster 600
So the stage was set — we would do 30 sets of ball hits to represent 30 days of play — and we’d scan every day and explore the data.
One last but important thing — what did we decide to define as a day:
Why 540? Well the answer is fairly simple. First of all we counted ball contacts in a Pro Mixed-Doubles match (specifically we counted Ben John’s ball contacts in a match and averaged them out — and then estimated 2.5 hours of play per day)
Clearly if you think you hit way more (or less) per day than this — that’s fine — you’ll just have to adjust the data as necessary.
Secondly — we had 180 balls ;) — so instead of having to count them every time — it was much easier to just hit 3 complete sets of balls — and - literally - call it a day.
We only used/tested one side of the paddle (to prevent us from messing up while spinning the grip we put a piece of masking tape on the other side)
All players were 4.5+
Oh wait one more thing.
You might be wondering why we were equating surface roughness with the ability of the paddle to impart spin. It is of course that this not the only feature of the ball strike that influences spin. In point of fact there are many including existing ball spin, paddle head angle, stroke direction/follow-through, wrist snap, skill of player, speed of stroke, softness of paddle — but for a given player/paddle/stroke these will stay fairly consistent — and therefore the dominant effect will be the paddle surface.
Imagine this.
But multiply by 16000
Yes the data. This is why we're all here. And this is why we did it.
I should state that we expected to prove that pickleball paddles degrade quickly and we did that.
The bottom line is that the hypothesis was confirmed (in fact it was exceeded).
We originally expected that the paddle might lose about 20% of its surface roughness in 30 days of play — but the answer was more like 45–50%.
Let’s talk about this data for a bit.
The vertical axis is roughness in microns (μm) — Rt is max roughness (peak to trough measurement), Rz is average roughness. If any of you reading is a serious materials science specialist — there is some nuance in that industry as to how these terms are defined — but for clarity — we are using the values from the Starrett and the USAP equipment guidelines.
(Eagle eyed nerds might know that USAP limits for Rz = 30μm and for Rt = 40μm and might notice that the first few days of this data place the Hyperion paddle we used outside of these limits. This is true.
However — when starting the experiment we went through a significant number of paddles (of multiple brands) in a retail store with our calibrated brand new Starrett meter in hand and were unable to find a single retail paddle that was within USAP guidelines.
The open secret of the Pickleball industry is that there are wide variances in paddle surface roughness due to multiple conditions at the factories - a combination of Quality Control in supply chain, thickness of PP core, uniformity of peel-ply, tolerances in pre-preg carbon fiber resin percentage, manufacturing precision, amount of pressure in the press, levelling of the press, temperature and of course human error - e.g. uniformity of removing peel-ply.
Based on our small sample analysis we estimate as much 20-25% variance in surface roughness in a retail batch of paddles.
The data in the graph above was all from the Starrett — but we validated with the Keyence (and later on with the Filmetrics)
The Keyence (and the Filmetrics) provide surface level data (so called S values) as opposed to the Starrett that gives you line level data (so called R values). It is possible to pick lines on the surface in the Keyence/Filmetrics software to generate R values.
Here’s an example showing Keyence data — with a 25% surface roughness loss in 10 days of play.
Both the Keyence and the Filmetrics produced a MASSIVE amount of data — and at times it might be difficult to correlate or understand every variable they produce — so instead we focused on simple trends — specifically trends in the Sa value.
Here are the key summary points:
Starrett — (Rt/Rz) loss = 28%
Keyence — (Sa) loss = 25%
Starrett — (Rt/Rz) loss = 49%
Keyence — (Sa) loss = 46%
Filmetrics — (Sa) loss = 47%
A note on data deviation/best fit data. You can see that the Starrett data varies in a surprising direction in some cases (e.g. the surface gets less rough over time). The explanation for this is fairly straightforward.
Humans are NOT robots.
We cannot hit the ball the same way each time and in the same place. We are fairly consistent so we will hit the sweet spot most of the time and over time the data will smooth out — but it’s important to understand that the Starrett sampling probe is a mere 5μm in diameter — it is measuring roughness over an extremely small area.
This means that we would take the USAP recommended 6 measurements (up/down/left/right and 2 diagonals) knowing we couldn’t take them in the exact same spot each time (at the micron level).
Well, imagine if you hit the ball 10 times — it’s highly likely there are some areas of the paddle that have never been touched by a ball. Now imagine you hit a ball 540 times — while you might think it’s not possible that you missed any spots on the paddle — remember the tip of the Starrett is 5μm — that is extremely small.
Point being — while we tried to focus all aspects of our test on the sweet spot of the paddle— the data will definitely get more accurate over time as the surface gets more and more uniformly eroded.
Regardless of the deviation in the Starrett data — the overall trend is undeniable — not least of all because it is backed up by the more accurate and more widely sampled surface data.
For the science nerds among you — it’s of course theoretically possible to design an experiment using a laser guided cannon and a robot arm to accurately hit the same spot on the paddle time after time so as to precisely measure point surface erosion over x number of hits — but even if you did that — what’s the point of your data — nobody can hit the ball that way.
We embraced the slight imprecision of our data gathering methods as being indicative of real-world play.
Maybe you shouldn’t.
If you want to focus on enjoying Pickleball and don’t want to worry too much about optimising the performance of your equipment that’s totally fine.
Or if you have an infinite bank balance and you don’t mind dropping $200+ on a pickleball paddle every few weeks — that’s fine also. (perhaps you’re a top ten player and your sponsor sends you 15-20 paddles a month — Congratulations!!!)
Let’s go back to our TLDR for a second:
If you’re playing/training for 5.0/Pro Level Singles your paddle surface will degrade FASTER than this.
If you’re playing 3.0 pickleball and hitting the odd drive/speed-up every now and again your paddle surface will degrade SLOWER than this.
So keep that in mind when thinking about your equipment — and your budget.
Well, without putting to fine a point on it - now RELOAD.
Once we discovered the data above — we thought there had to be a better way.
And that was part and parcel of the genesis of Reload.
A replaceable Pickleball paddle surface. So you can have your cake and eat it.
Maximum Spin whenever you want it
The feel/spin of a new paddle without the high price
A normal pickleball paddle ends up in a landfill. Reload sheets not only prolong the life of your paddle but can themselves be recycled.
Stick it and Rip it.
Of course we still had to figure out the IP, hope that USAP saw the light and considered modifying the rules and then figure out how to make the actual thing at scale — but we can leave those parts for another article.
This site requires cookies in order to provide all of its functionality.
Not actual cookies. That would be cool - but sadly - it's the other kind.
