Pongbot Nova S Pro owners' review and discussion thread

[plunder]

• @smee0815 Chrome webclient: https://smee.github.io/nova-s-custom-drills/
• @olanga Chrome webapp: https://olanga.github.io/nova/
• @nova_fan Chrome webapp: https://nova-premium.pages.dev/
• @inshane has a Nova Drill Control PLUS webapp v3.0.0: https://shimuranmonk.github.io/nova/
• @plunder all spinny balls reference: drop point = -6 (last updated: 2025-11-03🫠), enjoy!
• Programmed Drills — Basic (modified) : #149
• @Leaseback has worked on a proof of concept webapp
• @cysigno has a working beta of Pongbot Bro webapp
• @micadot has NovaKontrol webapp at https://github.com/cadot-eu/nova-s-pro 🤷‍♂️
  

Shoppers still on the fence may post their pre-buy questions in the other thread ( Anyone try Pongbot Nova S Pro? ) where you can find/get more general/vague answers/info from @leottsense and @pongbotstore marketing reps (who are btw not club players themselves). This thread is for sharing reviews and details (sometimes accompanied by pics or footage) about particular aspects of interest which the authors (incl. plunder myself) deem relevant to bring up. In the sense of "review", this will be an on-going thread since the product is new and complex, with many interesting or critical things to talk about. Authors doht need to wait for a surfer's question, i will, for example, bump/update the thread whenever i feel like posting about the Nova, unasked, unsolicited. Shoppers or owners, who might find the posted info/stories interesting to read, may feel free to Watch The Thread.

I, myself, am the owner of 2 newly bought robots which i call robi1 (=Powerpong Omega DOM2024) and robi2 (=Pongbot Nova S Pro DOM2024). In this thread my focus be on robi2, with occasional reference/comparison to robi1 because i find it interesting myself whether i will end up using more robi1 or robi2 in the long run. Training and learning are actually different things to do with a robi: one can use a robi for a session to get some satisfying/fulfilling training in, or to learn and practice a new technique/shot. I already learned serve-receive thx to robi2 and will continue practicing it (non-aggressive vs aggressive receives), i yet have to get full continuous training sessions in with robi2, to see how that goes.

One theoretical question to myself would be:
Would robi2 (in conjunction with my PONGORI rolling catch net) be accetably sufficient to fulfill my robot usage needs for now and in future, and replace robi1 eventually? It's kinda unfair to compare robi2 against the robi1 (or the Omni S Pro) because the two complement each other in functionality but let's do it anyway to understand their differences (e.g. in functionality, in build quality). Fortunately i personally doht have to decide between the two robis or answer the question, since by owning either robi i eliminated the need to go after that question.

My posts in this thread will be edited, updated, expanded regularly, i.e. whenever needed/possible, so watch out for 'Last edited' date stamps.

Also expect me to open a similar discussion/review thread for the retrofitted Powerpong Omega (as the better variant of the Butterfly Amicus Prime) soon too. In all fairness and honesty.

EDIT: hello Kian mppb!

 

[plunder]

The above post was an intro. This is now the first real post in the review thread, awkward way of starting a review, let's start with what i really doht like, things which i miss, things which should be different because robi1 does it better too. Hopefully the firmware maker and the app developer are tuned in to this post and are listening too:

Urgent (realistic) wishlist

• [P01] First of all, i find it annoying🤯 that you need a WiFi signal (WLAN) or Mobile Internet (SIM card) to be able to use robi2 with your android phone/tablet. Worse still, you cannot open let alone use the android app away from the robot. Clearly, if you just want to check, program or edit a series of Custom Drills, you need to be in the vicinity of a powered-on robi2 AND have a working WiFi signal or Mobile Internet connection, ofc in addition to activated GPS Location and Bluetooth signals. Did you listen and completely comprehend what i've just said?? For example, it means that you cannot lie in bed with your phone, browse peacefully thru the cryptic😵‍💫 list of Preprogrammed (or Custom) Drills, and plan your next session ... unless you've placed the Nova next to your bed and it's powered on wth. Where's my freedom, convenience, independence to open an app whenever i want whereever i want and program some ideas for Custom Drills which have spontaneously come into my mind (e.g. during club play, or after a league play)? — In contrast, operating robi1's app is independent from WiFi, Mobile Internet, Bluetooth, presence of robi1.
• [P02] Typically a Custom Drill consists of a multiball drill (only up to 5 different balls!, let's call them <a b c d e>, where each of the 5 balls can be repeated several times, e.g. aaaaabbbbbccccdddee). However there is no way of testing only 1 specific ball, i.e. sampling 1 specific ball for 1 time. The button "Combo Test" starts the entire Drill and never ends. Instead, we would want a button called "Sample Ball" which samples just the 'active ball' AND once or twice or thrice only. That's been one reason why it takes ages to setup a Custom Drill, because the sampling/testing takes too long. Imho this situation is very unacceptable, if it stayed like that. — robi1 has such a "Sample Ball" button.
• [P03] Once a drill is running, the pause between balls is max 2.0sec (nominal, i haven't measured), never longer. This is not acceptable nor sustainable. If you learn'n practice a new shot (e.g. a strawberry flick, or an aggressive serve-receive), more recovery time between the repetitions is urgently needed. The workaround is to program an additional simple ball for a breather, yet you'd still be in the 2.0sec-rhythm, i.e. without a proper pause for recovery. — robi1 has a more sophisticated implementation of timing, namely an interactive overall "bpm" (balls per minute) which is then overlayed by the individual ball delays/accelerations.
• [P04] I did several hours of serve-receive practice with spin always set to maximum, mostly with broken_in balls, but also with a few brand_new balls mixed in. Only after 2 hours in, there was neg-effective hard plastic buildup, deposited firmly on the wheels' rubber surface; due to that, the balls were thrown out with considerably less spin. So i had to open the robi2's head and figure out a way of uninstalling the 2 wheels for cleaning with an effective cleaning method; just a warm damp cloth wouldn't cut it. After another 2 hours in, i had to repeat the disassembly and cleaning. I will continue observing the (frequency of the) white buildup on the wheels, hopefully it's an issue with brand_new balls only. Either way the company should imho publish detailed documentation (in text and video form) on how to do that kind of wheels maintenance. I'll post about it in more detail with photos, but the company should do so too. The good news is, it's rather easy and straightforward to uninstall and clean the wheels. — No such issues of buildup, cleaning, disassembling with robi1 after countless of hours (i am counting!). In some future the robi1's wheels will need to be adjusted (tools are included) because they wear down very slowly, whereas robi2's wheels will need to be replaced (they caht be adjusted afaik).

2025-01-10: so far, the above 4 problems are the most memorable ones (re function) off the top of my head and could be addressed by the company without hardware work, just thru work at the office PC. The hardware, the build quality, is what it is, and i'll talk about it in a different post. If i find more unacceptable-ish issues in the firmware or the appware, i'll come back here and extend this post.

EDIT 2025-02-01: Re [P01]: in order to get a working WiFi-signal to my (non-SIM) Android tablet, i setup a WiFi hotspot on my (SIM) smartphone and established a Mobile Internet connection, so that my tablet could access the internet thru my phone. The pongbot app consumes about 6-8MB mobile data in 2-3h, very little. This works more or less stable but drains the phone's battery. So i decided, for Nova-at-home in the basement (weak WiFi signal), to buy a WiFi repeater; this setup works really well. For Nova-at-club, i'd still need to find acceptable setup solutions (weak WiFi signal? cabled power supply?😵‍💫).

 

[Saiph]

I'm interested in buying the Nova in the future so definitely watching this thread! Looking forward for your review first

 

[schrodingerskoalabear]

Looking forward to the reviews! I like what I see from Nova S Pro generally and it feels like a good price-to-function ratio for me. However, I have been hesitating for two reasons which are interrelated

(i) newness of the robot in the market so we don't have a full picture regarding the robot's longevity, or any potential time-activated issues

(ii) 90% of the reviews (video or postings) I came across can be put into a marketing-lite category with filler language. These are generally overwhelmingly positive without an ounce of criticality or weighing of pros and cons. Online posters of such text reviews, to put it lightly, also just seem too new.

I am happy to buy a machine with limitations (and potential long-term issues even) as long as I have knowledge of what they are so I can have reasonable expectations in aligning robot use with improvement of my bang-average intermediate skill level needs.

Long story short, looking forward to hear of your experiences!

 

[plunder]

All 12 Tomahawk Serves

In this post I am sharing 12 Custom Drills for all possible Tomahawk serve variations, in their most dangerous form, for the Nova user to receive. What's the "most dangerous form" of a Tomahawk serve, please? It's when the ball has the highest kinetic energy (highest speed, plus highest rotation): the ball flies fast, low, wide, AND with the maximum amount of sidespin (side-pure, side-top, or side-under), landing near the edge of the player's table side. The shorter a Tomahawk Serve lands, the more harmless it becomes, the easier it is to control the receive. You'll also learn that with brand_new balls and superclean wheels and a regular racket, it is almost impossible to receive some of these serves because of the crazy amount of sidespin, at first. When you notice a decrease in the amount of sidespin after a short while, your wheels are prolly clogged up with [P04] buildup.

The below series of screenshots/animated GIF's with set angle ~21° in conjunction with IRL angle ~21° works really well for demo/illustration purposes but having such a big set angle makes the Programmed Drills awkward to use from the 2 corners of the table (pos9, pos7). For your final exact positioning of the Nova at the corner positions, i'd rather recommend that you adopt a lower set angle, as i did. See the EDIT at the very end of this post for my final values, i.e. set angle = ±9° in conjunction with IRL angle 15°:

---

Toma01: From pos7, Tomahawk Right

---

Toma02: From pos7, Tomahawk Right with Underspin

---

Toma03: From pos7, Tomahawk Right with Topspin

---

Toma04: From pos7, Tomahawk Left

---

Toma05: From pos7, Tomahawk Left with Topspin

---

Toma06: From pos7, Tomahawk Left with Underspin

---

Toma07: From pos9, Tomahawk Left

---

Toma08: From pos9, Tomahawk Left with Topspin

---

Toma09: From pos9, Tomahawk Left with Underspin

---

Toma10: From pos9, Tomahawk Right

---

Toma11: From pos9, Tomahawk Right with Underspin

---

Toma12: From pos9, Tomahawk Right with Topspin

---

comments:

1. I am showing in post#9, where exactly and how to setup the pos7 and pos9, to get the 15° set angle. Btw from the two corner positions the set angle shouldn't be over 21° because it'd make sense to choose the Angle such that it's max Drop Point value of |-10.0| = |10.0| coincides with a long-line throw by the robot, At 21°, the robot is already sitting at the corner and cannot drop the ball further to the side ("out!"). However, with regard to the Programmed Drills, anything between 8.0° … 18.0° would be a more practical/pragmatic choice than the firstly suggested 21.0°.
2. Each drill consists of 4 balls (as proven by the right-hand side). The left-hand side, however, shows only 1 ball, for example for Toma04 (ball#2). That's because the software miscalculated the trajectory of the other 3 balls, believing that they would fly out. You should sample my settings, and if with your Nova unit the 3 balls do fly out (i.e. doht drop on your table's side), then reduce the speed of those balls by 0.5 and done!
3. If you know your maths, the above pics show the ball settings of 48 different balls. Most of these 48 balls do land on the table, consistently, with their 2nd bounce landing near to the table's edge (= wide Serve). Because of [P02] there might be 2-3 balls whose settings aren't optimal: they sometimes land out or not very near to the table's edge. Feel free to identify those few balls and adjust their settings (typically, just the Speed needs a tweaking by ±0.5, that's all).
4. Because of [P01] and [P02] it takes forever to program and test these 48 balls, doesn't it? Please tell us how many hours it took you to complete the programming and testing task. Also [P04] starts ruining the serve-receive practice session after a few hours in, so have your toolbox handy!
5. Apart from [P04], it is possible to learn how to receive all of these 48 Serves within a single afternoon session (couple of hours), at different distances from the table, and non-aggressively vs aggressively. After that session you'll feel like you could receive any kind of Serve (big confidence booster).
6. It's true that robi1 can reproduce not all of these 48 Serves easily, or this kind of realistic dangerous quality of Tomahawk Serves. The Omni S Pro caht do them all either because of its head height and head position (table mid center). I guess one caht have it all within 1 single robot! And even if robi2 did it all, the [P03] kills it again: you and me, we are not robots, are we?

EDIT:

I redid all 12 Tomahawk serves, using the +9° set angle setting from pos9 (likewise from pos7: -9° set angle), while using the 15° IRL angle, see post#9. I should do an online spreadsheet of the new 144 values? Here we go:

Tomahawk Cerve Receive (Nova at 15° IRL angle)

docs.google.com

(screenshot as of 2025-01-21, the red values mean that one could reduce them by -0.5, the green values mean that they are a bit on the conservative side and one could try to increase them by +0.5 to learn if the ball still lands on the table most of the time):

((comment: In future sessions i'll try to setup the hardest topspins from pos7/pos9, i.e. finding the lowest Head Height in combination with the highest topspin and then highest possible speed (or vice versa for a slow spinny loop). If that fails, i could place the robi2 behind the table on a stool. This drill setup will be for a blocking exercise)))

 

[schrodingerskoalabear]

Appreciate the hard work on writing up your findings, very very useful.

Since I am unfamiliar with the interface of the app, all the serves configuration looks really intimidating. But I can just tell that the Tomahawk serve settings is one of the most useful pieces of information for any Nova S or prospective owners.

 

[plunder]

This post be about wheel maintenance, not build quality (that'll be topic in another post).
Wheel maintenance

It's straightforward and very easy to open up the plastic encasing, 4 screws and the head is open:

The wheels are held down, in place, by loctited screws. I diy'ed handtools which allow me to hold a wheel in place while unscrewing the screw. Not much force is needed to take off the screws. The black washer is made out of plastic, interesting:

The wheels are made out of plastic and have a thin layer of rubber on top ("tyre"):

Only at max revolution rate (of the wheel) would the white balls start to "rub off" on the rubber tyres. The sliding frictional force/heat is then simply too much for the white plastic surface, such that molecules of it get separated from the ball and transferred to/deposited on the rubber. I use the max spin setting for serve-receive exercises of Tomahawk Serves, or for blocking exercises of extreme topspins:

In general, rubbing alcohol doesn't go well with rubber, or was it oil, haha? Anyway, so you got white hard plastic embedded in black hard rubber. My goto method of getting the "gunk" off (technically, the white stuff is not gunk!) is by using tape; this works reasonably well up to a certain degree:

For the remaining white spots (freckles) i use a soft eraser (if you used a hard rough eraser, you could damage the black rubber, no doubt):

As one can see, the middle section of the tyre looks shinier and degraded. That makes sense because the sliding friction of/by the white ball surface induces wear and tear to the black rubber surface. That's nothing I worry about; if i needed to replace/renew the wheels in some future, it'd be easy to do so: simply buy new ones, pop them in, and you're good to go. Today i completed hours of nonstop training with the Nova at non-max spin settings, and there was no white buildup on the tyres!

So how often does one need to do this kind of wheel maintenance? — It depends on how fast you make the buildup grow on the rubber! If you content yourself with running the "Programmed Drills" only, then expect no considerable white buildup to form on the wheels, because the pre-installed drills doht employ highish spin settings (but rather lowish 🥱). Once there is buildup on the wheels, the robot cannot throw out crazy spinny balls anymore, so the point of the exercise (e.g. blocking extreme topspin, or receiving Tomahawk Serves) becomes moot. Personally, as soon as i notice a drop in produced spin, i go and recheck the black rubber surface, and eventually do the wheel maintenance.

Imho the PDF-user's operation manual should detail this rather essential topic. Even new owners might will run into this issue right away, if they start out with Tomahawk serve-receive practice: 1-2 hours in, and then the black rubber is clogged already!

EDIT: btw the above photos stem from my 3rd session of wheels maintenance (today!) and it took quite some efforts & time with the proposed combo method (duct tape, plus pencil eraser). such a session throws you off and caht be done during a training session; do the wheels cleaning before, or after, a planned training session. I also took photos of my 1st cleaning session, that was a week ago, when i last used robi2. Click on Like, if you want to see those photos as well (they look very similar), and i'll post them here too.

Here are some pics before my very 1st cleaning session (you can see the upper wheel)...:

...(you can see the lower wheel)...:

...(before unscrewing the wheel for the very 1st time, sorry for blurry pic)...

...(with different lighting)...

Okay and now let's compare with a photo from before my 8th🤯 cleaning or so:

The central area of the rubber has been shining up and from the changing width of the white plastic traces one could infer if the rubber surface is not perfectly even/regular. At this point i am not too worried about the consumption (wear'n tear) of the black rubber surface. When the wheels in cleaned condition start becoming non-functional, i.e. need a full replacement, i would post the news here. As it is obvious and natural, both the balls and the rubber get worn down through robot use! The question is only 'to which extent' and 'how fast'.

 

[plunder]

This post is about my comments on the pre-programmed drills, which are simply named "programmed drills" in the app. As always, check back once in a while to see, if i've updated/added content in the post.

Programmed Drills

Having them is very useful and every owner should have gone thru all of them <1-ball, 2-balls, 3-balls> × <1-star, 2-stars, 3-stars>. For timing i set 1min (it's not possible to set 50sec or 75sec or etc) and then do 1 difficulty level, e.g. 1-star. Then the drill completely stops (since the app doesn't feature a timer for automatic/programmed pauses, like robi1 does have). Now i would clear the few balls which have gathered around me, and i'd walk over to the tablet and increase the difficulty level by +1, i.e. 2-stars. Hit 'Start', and i'd get another 60sec of the same drill, just "diffculter". Then the drill completely stops, i clear the balls in my way, but also have a look at the ball container and decide if i wanna collect all balls (from the ground and the PONGORI rolling catch net) already for the container refill. My ball container —I use the small fabric one not the big hard plastic one— is good for 120sec with any pair of 1-star or 2-stars drills, e.g. 1min of a 2-stars drill, then another 1min of a drill at 2-stars. Before reaching the end of the 3min-mark, the ball container would run empty, especially with any 3-stars drill.

1-star, 2-stars, 3-stars, what's the difference between these three difficulty levels? — Mostly it's the pacing between the balls, the increased "bpm" (balls-per-minute) if you will, the ball frequency. Depending on the drill, 1-star can feel a bit too slow, whereas some drills at 3-stars feel too fast paced. Any other difference? Well, more stars also mean that the balls have a bit, just a bit, more spin, i.e. more underspin or more topspin. Apart from the faster pacing and increased spin, i doht notice other potential differences such as changed head angle or changed ball speed.

With my above described 1min-routine, running thru all stars of the 1-ball folder (called "Programmed Drills (Standard) — Basic") takes about 32min total. That's a really nice, effective and fun way of warming up, since it's still up to you how you wanna react to an incoming ball, i.e. what kind of shot to apply. Continuing with the 3-balls folder (called "Complex"), it takes another 2.5h to complete that folder ((I discovered a simple bug, the drill "All Random" defaults to "Sequential" whenever one sets the stars, so one has to click on "Random" every time before starting the drill!)). If you used a ball container which is good for <1min 1-star 1min 2-stars 1min 3-stars> of any Programmed Drill, then you could get thru the 3-balls folder maybe within 2.0h.

What can i comment about the Programmed Drills, what are my impressions/observations good bad ugly?

[pd1] Very good thing: I ran all "Programmed Drills (Standard)" from pos9, instead of pos8 (=Standard), and it worked well! Likewise, one can use the pre-programmed drills from pos7, without the need of editing any ball settings. And if a long line ball went out too often, you could use the calibration button for that specific drill.

[pd2] Very good thing: every drill and each ball (and also the stars) has its own optimized pacing, giving the drills a more natural, non-robotic feel. Whoever designed these simple(?) 2-balls or 3-balls sequences, must have put much thought in fine-tuning the timings between the balls, and i relish the delays between some shots.

[pd3] Good/bad thing: Seriously all balls per se, from the Programmed Drills/any stars, are easy balls to me. The balls with underspin (e.g. "Push(B)" or "Loop(B)") have very little spin reaching my racket, no matter the stars: they are thrown out with some underspin but when they bounce off of my table's side, the spin gets almost nullified. The balls with topspin (e.g. "Drive(B)" or "Drive(F)") doht have much spin either, imho. And my impression is that all balls get thrown at the same head angle and that there is no variation in speed: none of the "Drive"-balls are fast, but all of them bounce off so highish that i can smash/slap them all(!) back. In short, all balls have very little quality: they bounce off highish, have little spin (sometimes zero), and fly slowish. It's fun to hit and land easy balls and make only few unforced errors but from some day on it'll get old. — No big deal you say, because you can program very challenging balls thru Custom Drills. You're absolutely right. But this post is about the pre-programmed drills.

[pd4] Bad/very bad thing: None of the balls of a Programmed Drill can be edited, like increasing the spin or the speed of a ball, or its drop point location, or the ball repetitions. Or the head angle. All you can do is switching between "Random" vs "Sequential", and do some general calibration of the drop point (such that all balls of the drill do land on your table's side). I find that hard to accept. It would be much better and more valuable if the Programmed Drills folder served as template which then could be duplicated and edited by the user (as it's possible in robi1's app).

[pd5] Not good: The "Complex" folder consists of 3-balls drills only, even though the app allows for 5-balls combinations. I wanna see some relevant 4-balls and 5-balls drills in this folder.

[pd6] Maybe it's possible for the app builders to offer a settings page —valid per drill (like the calibration-per-drill page) OR valid for all Programmed Drills— which allows the user to specify the quality of balls in 1-star, 2-stars, 3-stars Programmed Drills: head angle, ball spin, ball speed, and overall balls-per-minute. Seriously, pd3 will become a personal problem rather sooner than later: the way it is now, the balls are too easy. And it'd be unacceptable to re-create all those (otherwise nice) Programmed Drills from scratch, only for the purpose of personalizing the head angle plus the ball settings. We private end consumers doht have the time and energy to do that job. It's not our job. It's the company's job to make it as easy as possible for us to get the most out of their product! I really dislike the entire process of creating a Custom Drill on the Nova: one needs a lot of motivation, energy, patience, etc, to manage to create a new Custom Drill because of [P01] and [P02], it's a pita in the long run, not sustainable, you'll learn. What will happen with lazy users is (fair assumption): they manage to create a small handful of Custom Drills during the first two weeks after buying, and then leave it like that, doht find the energy in later months to add further Custom Drills because it is a pita.
If the app builders caht manage to provide such a settings page, then the simplest solution to my problem would be for them to revise the ball settings for every ball in every drill for all the stars. Simply, by making the balls higher quality, more challenging, less easy.

Okay, users who are energetic and motivated 365/365 doht need to deal with pre-programmed drills and their shortcomings. They can just ignore all that and add Custom Drill after Custom Drill from scratch, even if it takes hours and days. Personally, i use, like, and enjoy the pre-programmed drills (for the time being) and they do make me a better, more confident player, but I wish there was a way to change the ball settings, especially for spin.

 

[plunder]

In this post we'll treat the subject of positioning the Nova at pos9 (likewise pos7), i.e. how i do it consistently.

Positioning the Nova at pos9 (and pos7)

I'd wish the company would release a stencil-like jig, i.e. a flat rotary quadratic platform that can be easily aligned right at/in the corner of the table and on which the Nova is placed, to standardize the exact location of the robot in alignment no pun intended with the app. It would then be easy to physically rotate the robot's orientation, read out its angle, and enter that number into the app when setting up a Custom Drill. Without such a well-defined exact location of the robot within the ninth sudoku cell, how could the visualization in the app possibly be accurate?

Anyway, first we need to be able to reproduce an exact location (x-, y- coordinates) and exact angle of orientation (° degrees) of the Nova on the table, in a fast and easy manner. From my testing of the Programmed Drills, a perfect choice of orientation be +15.0° (or likewise -15.0° for pos7), not 21°. We'll leave the robot at this orientation for all Programmed Drills and Custom Drills initiated from pos9.

Step 1. Position the two rubber feet of the Nova's rear section as closely to the table's edge as possible. With your finger you should be able to touch/feel the alignment of rubber and edge:

Feel free to mark that touching point with a piece of tape:

Same with the other rubber:

Here a more frontal view of the tape:

The 2 tape snippets are not sufficient IRL to accurately set the 15°.

Step 2. With a cheap thin carpenter's framing L shape square, held against two rubber feet, you can see the angle of orientation forming at the table's edge. For convenience, i shifted the framing square such that its "100"-mark was right at the tip of the 'triangle in the picture':

Convenient because now you know where to place the triangle ruler. And bull's eye, the measured angle reads indeed 15°:

Since the correct angle has been confirmed, you can put the triangle ruler away, and draw a fat green line along the framing square (with a dry erase marker):

If you have the framing square handy, you'll be able to reproduce the exact position and orientation of the Nova in the ninth sudoku cell simply with the help of the framing square and the two green lines, while the two red tape snippets are just for quick rough reference to get you started:

For pos7 it's likewise. Once you have drawn the two fat green lines for pos7, it's super quick'n fun to place the Nova exactly there. Just place the robot on the table somewhere near the table corner, slide the L shape framing square "underneath" the robot, touching three 🧐 rubber feet simultaneously, then slide/shift the robot towards the corner such that the framing square perfectly aligns with the two fat green lines. First with the one green line, then with the other green line. And if you want, you could leave the framing square the whole session time there (so i finally found an IRL sensible purpose for the ultracheap 5.99€ China-made framing square which i had bought over 2 years ago hooray!):

Step 3. In the app, you can set the angle to "set angle = +15°", as one normally/logically would do. If you do so, you'll notice that the visualizations (ball flight trajectories) are quite a bit off! That's understandable because the app doesn't know where exactly you positioned the Nova within the ninth sudoku cell. By experimenting with different set angles, you'll learn that, for app visualization purposes, using the "set angle = +9°" setting is more suitable. So if you want the app visualizations to be a better matching predictor of the ball trajectories, you should go with that set angle instead of +15°, even though IRL the actual angle is +15° as proven by the photos. Then, when you work with Custom Drills, simply ignore the +9° info bit ("virtual discrepancy") in the drill settings.

Comment:
In practice, one can/could ignore the app visualizations altogether! In the end, what really counts is that the robot produces balls from the above exactly defined pos9 which have the properties and quality as desired. And that depends only on the ball settings of the drill, not on the set angle (app info bit) of the robot. And certainly not on the app visualizations! The manual even states that "3.5.5. Real-time Animation (…) This animation effect is for reference only. ( Figure 3-16 )". However i find the app visualizations immensely useful, since the ball consistency is pretty good and the trajectory predictions are pretty good too. In fact, imho the app visualizations are the strongest point of the app, so i wouldn't want to do without them!
The choice of 15° instead of 21° for the IRL angle enables us to use the Programmed Drills from pos9 without editing/modification/calibration: all balls land on the table under nice sector coverage. Understandably there's an exception: the wide long-line underspin balls, i.e. the "Loop(B)" from pos9 (or likewise the "Loop(F)" from pos7), would sometimes fly too long ("out!"); for a particular difficulty level (⭐, ⭐⭐, or ⭐⭐⭐), feel free to Calibrate the starred drill, setting the flight path Shorter By 5 (out of 10) increments. Not all three difficulty levels of a potentially affected Programmed Drill would need this calibration adjustment; the calibration is saved separately for each difficulty level of each Programmed Drill, so better use it only when really needed. Note that the calibration setting remains saved, even when you switch the robot position to pos7 or pos8 or etc.

 

[craptiger]

Apologies if I've missed this in the thread somewhere (I'm going to read it properly at a later stage, promise!) but has anyone any tips on ball collection and the Nova S Pro?
I've just "upgraded" from a cheaper machine that recycled balls and am really struggling with having to keep filling it up. The room I play in is not that big so I have to crawl under the table to retrieve them and it's a massive pain in the backside. I'm thinking of making some kind of net that will collect them and return them down a pipe into a box by my feet but wondered if anyone has done anything similar before.
I particularly want my net to be a good couple of feet high all the way to the middle of the table so any that bounce backwards stay on the table too. It must be easy to take apart (I play in my living room) and re-assemble as well.
I will use these cheap clamps to attach my net https://cpc.farnell.com/pulse/apc01-b/panel-clamps-edge-mount-black-pr/dp/ST04451 and probably copper pipe to slide it over. I may use an old king size duvet cover for the "net" as I can then stitch tubes to slide over the pipes. Unless anyone has better suggestions of course?

 

[plunder]

@craptiger I talked about my standalone rolling catch net in this thread, and it's a funtastique companion for the Nova. I roll it to the player's side when i program'n test a Custom Drill, or i roll it to the robot's side when i do a full training session with the Nova. I wouldn't recommend anything else/anything other than such a type of catch net, for use with the Nova. The catch net is big and "cannot" be disassembled.
I placed a hula-hoop under the table; if a balls stays under the table, the hula-hoop serves great as a collecting (or pushing) tool.

If space is a real limitation, then yes you could try a fixed catch net, as many users do. ( i can only facepalm when i see my club mates doing ball bucket training without any kind of catch net ) Yesterday it took me maany hours to redo the 12 Tomahawk Serves from scratch (see the edited post#5), optimizing the flight path of the 48 balls AND having to clean the wheels 3 times omfg during the process. Without a catch net on the player's side, the creation of the 12 drills would have 😵☠️me.

If you appreciate the Nova so far, maybe you'd be better off getting an Omni at Black Friday Sale? Was only 1000$ shipped or so.

🙃

 

[plunder]

In this post we're rebuilding the Programmed Drills from pos8 (set angle = 0°), starting with the folder Basic which contains 6 drills.

Programmed Drills — Basic (modified)

Here's a live link which allows me to edit the values on the fly:

Programmed Drills â Basic (modified)

docs.google.com

And here's a screenshot of it from 2025-01-22:

Comment:

• From pos8, there are 6 Basic drills, 12 Combined drills, 18 Complex drills, while from pos2 the numbers are 6, 2, 0, respectively. Each Programmed Drill comes with 3 difficulty levels. Doing de maths, (6+12+18+6+2+0)×3 = 132, yet the ads state that the Nova comes with 264 pre-programmed drills wth. Please doht tell me that hitting the "Random"-button counts for doubling my 132 number!
• I was able to confirm that the full range of vertical head motion (called "Ball Height") is indeed 50.0°. Since there are 150 increments between "Ball Height = -50" and "Ball Height = +100", this equals 3 increments-per-degree, exactly. For example, if you want the Nova to throw out the balls by +1.0° steeper, you must increase the Ball Height by 3 increments.
• "Ball Height = 0" means that the Nova head is perfectly horizontal, throwing out balls at 0°math (measured against the horizontal). "Ball Height = -50" means that the Nova head is fully pointed down, in practice it's angled at (quite exactly) -17°math. "Ball Height = +100" means that it's pointed fully upwards, angled at (nearly) +33°math.
• It was easy to measure that, (6+12+18)×3 = , all 108 pre-programmed drills from pos8 are angled at "Ball Height = +50", i.e. the balls are shot out at +16.5°math, which makes them very easy, too easy to hit, because they fly so high (AND with little speed AND with little spin). And that's the reason/motivation for this post: By re-creating the 6 Basic drills as Custom Drills, we have the chance to lower the Ball Height.
• For each Basic drill, let's have the Nova throw out 3 variations of the ball with slightly different ball settings. Use "Repetition = 1" for each of the 3 balls for an equal probability of 33.3%, and activate the "Random"-button before starting the drill. This makes the modified Basic drill a bit more interesting and challenging. As for the Frequency we choose an identical value for a steady regular hitting rhythm.
• The 12 Combined drills and the 18 Complex drills are basically just combinations of the 6 Basic drills. With the above spreadsheet, you have all the building blocks to re-create those multiball drills. For each ball, feel free to adjust the Frequency number and the Repetition number to your liking. Obviously, when combining the modified Basic drills, you must agree on a common Ball Height first, e.g. "Ball Height = +40" would work for both Push(b) and Loop(f) from the above spreadsheet but also(!) for Loop(b), which i prefer to practice separately/independently/byitself at "Ball Height = +30".
• The max underspin amount in Push(b) and Push(f) is "Spin = 5" because of the low "Speed = 1.5" which is a short push. Apparently, from pos8, the Nova is not capable of producing a short push with any high amount of underspin. Imho that's disappointing and a negative .👎
  If the Nova isn't capable of that, then how about the Omni? 🤔Doht expect a detailed honest answer from any Omni buyer! 🤥

BTW ... give me some weeks and months before i share my final verdict/conclusion about the whole pongbot/nova/situation (pro's con's good bad ugly recommend or not, stars rating, personal feeling/thoughts, etc). You may have noticed that, until now, i have tried not to use judging words/phrases like "that's fantastic, that's bad, that's cr*p, i like, i love, i hate, etc". So far i've been trying to be objective, descriptive, neutral, balanced, honest, fair, whenever possible. However, when time comes (= the final verdict/conclusion), i will not hold back in expressing my opinions, feelings, and evaluation and judgements. One of the next posts will be, for example, about the build quality. Even in that post, i'll try to not judge (commend or disapprove). Only in my final verdict/conclusion would i judge, regarding all topics in this discussion thread.

 

[plunder]

In this post we're reviewing the aspect of Frequency, i.e. the timing of balls within a drill.

Frequency

Comment:

• The "Freq." number is an inverse measure of the time you have left to prepare for the ball which you are editing. For example, if you're editing the ball settings for ball#3 of a 5-ball-drill (5 different balls: ball#1 ball#2 ball#3 ball#4 ball#5, each with "Repetitions = 1"), then setting "Freq. = 20%" means that, right after the Nova has shot ball#2, you have 1.43s of free time🥹 before the robot shoots ball#3. It's basically a pause (or "prepause") before the ball gets shot out by the robot. The bigger the number, the less time you have.
• Trying to remember what "Freq." means as 'the bigger the number, the faster the ball gets repeated' or 'the faster the ball gets shot out in terms of timing' does not work so well in my mind because you still would be confused about when the pause takes place: after the ball gets shot out, or before it gets shot out, or or or (postpause? prepause?).
• With a simple stopwatch, by measuring the pause between two succesive shots (say ballA and ballB), you'll know the "Freq."-setting of ballB. This comes in handy, if you were to recreate the multiball drills from the Programmed Drills.
• I did that for all the 2 starred ⭐⭐pre-programmed multiball drills (i.e. Combined, and Complex), and you'll notice that many of them are set to "Freq. = 10.0%" vs "Freq. = 60.0%", which often feels natural and correct, but sometimes feels off, so i corrected them.
• If you set "Repetitions = 4" for ball#3, then all 4 identical balls (ball#3 ball#3 ball#3 ball#3) come with the same prepause. After the fourth ball#3 has been shot out, you'll experience the prepause of ball#4.
• Calling this setting "Freq." AND measuring it in percentage is counter-intuitive imho. But you'll learn it 😛
• As one can see, the +10% increments for the Frequency are linear whereas the Prepauses are not. Maybe 🤔 it would be more practical or feel more natural to have a linear Prepause scale between 0.67s and 2.00s instead? I believe so. Oh well, we'll learn to live with this nonlinear time scale 🙄
• It's a pity that one cannot set a short break/pause after a run through a sequence. The "Freq." of ball#1 is going to be that pause! And that pause can never be longer than 2.00s. — robi1 has several ways to incorporate longer pauses between the repetition of sequences.
• What would be a natural frequency? Here are two reference data points: The Adriana Diaz counter drive rally had a frequency of 75bpm (= (14-1)hits/10.30s × 60s/min). And when i train at my near-table positioned returnboard rb1, the frequency is pretty exactly 60bpm, which feels nice/natural but is a bit on the slow side because the rb1 is like a steady block, not an active blocker or counter puncher, the ball comes back at harmless speed and one has plenty of time to execute the next topspin. So for the Nova setting one should aim at a frequency of 50% or 60% in rallying drills.
• It is helpful to know how to determine the BPM's yourself from youtube videos, gpt knows how to.
• For a balleimer single ball drill, a prepause of 1.67s appears common/natural; so use "Freq. = 10%" as reference or base line.

 

[plunder]

In this post we're rebuilding the Programmed Drills from pos8 (set angle = 0°), folder Combined which contains 12 drills.

Programmed Drills — Combined (modified)

Here's a live link which allows me to edit the values on the fly:

Programmed Drills — Combined (modified)

docs.google.com

And here's a screenshot of it from 2025-01-26:

Comment:

• Frequency, Repetitions, Spin Type, and Drop Point have been copied from the Programmed Drills, whereas Spin Amount, Ball Height, and Speed have been changed to produce more challenging, higher quality balls.
• I tried to already introduce variations in the same shots such that whenever you hit, for example, a "Drive(f)", it's not always exactly the identical shot for you because the ball comes flying with more topspin, lower, or faster. Feel free to edit these 3 settings anyway, if you like a particular drill and want a different ball quality, e.g. want to hit a forehand drive on a float ball.
• The original Ball Height setting was "Ball Height = +50.0" which i found hilarious because every "Drive"-ball —and even the "Loop"-balls!— could be smashed. Well i can.💪 With my forehand, and with my backhand, and I am not even a very skilled club level player! Now with the lowered Ball Height, the balls come indeed flying in low and sharp 👌, so there's no need to worry about changing the physical elevation of the Nova.
• Maybe my wheels are worn down yet or what, anyway, my Nova unit does have problems generating decent amount of underspin on any kind of slow Speed balls. And actually, the app doesn't allow you to set "Spin = 10" for a slow Speed underspin ball. I think i mentioned this problem in a previous post already hah. Only on fast balls, the Nova manages to impart decent amount of underspin; but then again, when IRL do you get fast balls in with high amounts of underspin? Not realistic scenario. — robi1 has a contrary problem, it produces too much underspin as soon as one ups the parameter for underspin (robi1's "Spin = -2" is immense and could be fixed thru firmware revision).
• The Repetitions number in red is personal, i sometimes like to get a repeated ball, so that i can try different shots on the same ball (or simply try again the same shot haha), maybe even for the purpose of getting some rest in. For example, on a "Push(b)" i would want to rest my arm by executing several push shots (at different bat angles) and then finally try to execute a backhand flick/flip.
• For a standard exercise like '2 backhands 2 forehands', you could simply run c03 and use "Repetitions = 2" for either ball. Imho it's a likable way of going about things.
• The fact that the Ball Height cannot be changed within a sequence doesn't turn out to be a party pooper , a human multiball trainer also doesn't change the height or angle much, does he? However the regular/frequent buildup on the wheels and their cleaning is starting to get annoying: if you want to avoid any buildup, simply use lower values for the Spin amount setting. It's logical physics mechanics chemistry that no buildup could form, when the Spin is low.

 

[plunder]

In this post we're rebuilding the Programmed Drills from pos8 (set angle = 0°), folder Complex which contains 18 drills.

Programmed Drills — Complex (modified)

Here's a live link which allows me to edit the values on the fly:

Programmed Drills — Complex (modified)

docs.google.com

And here's a screenshot of it from 2025-01-26:

Comment:

• The fact that the Ball Height cannot be changed within a sequence doesn't turn out to be a party pooper , but it starts to get limiting in your ball settings, if the three balls are to form a realistic scenario. One has to find optimal settings which still make sense. I did that for us 🥳. The above screenshot be not the end all be all. Let's train with these settings over weeks and months and learn which values need further tweaking for optimal training satisfaction.
• The drills in the Programmed Drills folder are not numbered for easy identification. As a result, the app developer dropped the ball no pun intended on a few drills, for example x06 originally says "Push(f)-Loop(f)-Drive(f)" which is the same title as for x10, huh? Running the x06 drill, we see a "Loop(f)-Loop(b)-Drive(f)" sequence, so i named the drill exactly that. Likewise i noticed off timings between shots, e.g. in x05, which is the Falkenberg drill, and a few others, so i corrected them.
• Whenever you see Frequency values of "10.0%" or "60.0%", then these are original values from the respective pre-programmed drills. When you see "20.0%" and "40.0%", then these two are mine because they feel more natural. Feel free to change these values when you're running a drill and feel the need for adjustment of timing. The fact that the 2-stars pre-programmed drills mostly work with the "10.0%" and "60.0%" setting lets me believe that the drills haven't been fully tested IRL by the Chinese coaches and pupils because optimized drills would have more variation in the Frequency values, wouldn't they? So it's our task to do the optimization. I am going to use the above live spreadsheets to finetune/tweak the values to my liking, to each his own.
• Drills with random balls are marked in magenta color (x07 x08 x13). You need to activate the "Random"-button before starting the drill. What's random then, is simply the order of the balls in the drill. Each ball is still well defined thru its given ball settings. Fortunately you're not getting served a permutation of the balls but, from what i can tell, a ball from the drill gets picked by chance thru its Laplace probability. You can increase the probability of a ball to be picked more often by increasing its Repetition number in the settings. For example, if in a drill with 5 distinct balls and Repetition values (1 1 4 1 1), you could possibly get served ball#3 10-times in a row. So for equal probability within the randomness, those values should be equal as in (1 1 1 1 1) or (7 7 7 7 7).
• Fast (or slow) balls which leave the robot with underspin can turn into balls with topspin because the bouncing on the table imparts a small amount of topspin due to the sliding friction coefficient of the table's surface. Hence it depends on the Speed, how much Spin (Under) amount you need to set for the ball to become an absolute zerospin ball by the time it hits your racket. In x01 i created such a float ball which you're supposed to hit with a drive shot. If you make the robot shoot out a "floater" (i.e. "Spin = 0.0") at high Speed, then the ball will have notable topspin by the time it reaches your bat. Using the Type/Speed/Spin triplet of "Under/3.0/1.0" shoots out an underspin ball whose underspin gets cancelled out by the table after the ball's first bounce on the table: that's a real floater in my opinion, a truly dead ball. Very interesting to see and learn!
• Let us know how long it took you to copy the above "Custom Drills". It takes ages, doesn't it? robi1 app has a sharing function for sharing drills and drill groups with friends, super easy and fast. With the Nova app, one cannot create a folder or a group of drills. Since there is no such a management possibility, my list of Custom Drills looks ridiculous 🤪:

 

[plunder]

In this post we're revisiting the important topic of wheel maintenance by opening a new perspective (and i should win the pulitzer for it).
Wheel maintenance Final(?) Solution

Apart from preliminary positive testing i haven't done full training sessions with this solution yet but from now on, all my Nova usage will be done with it until "further notice". The wheel's original surface is BASF hard rubber which is engineered to be wear'n tear-resistant and long-lasting (no aging, etc); most users will be fine with it for a couple of years to come, maybe 3-5 years. Hardcore users like myself who try to use the highest/max Spin setting whenever possible run rather quickly into the "problem" of wheel maintenance. So my idea/suggestion is to treat the surface as consumible/consumable by covering it with/adding a softish rubber band and see how that goes. My fear was that the balls could get cracked by the added pressure between the 2 thickened wheels (diameter increase from hard-50mm to squishy-52mm) or that something could get wrecked in the Nova head assembly. So how did it go?
Wonderfully! The Nova operates as if there weren't any rubber band on the wheels, maybe even better!

Comments:

• The rubber band's surface is not sticky/tacky but has about the same sliding friction coefficient as the (clean) BASF hard rubber surface.
• I have yet to see if there's buildup or a performance drop after hours of use at max spinrate.
• I have yet to learn how long 1 rubber band will hold up, i.e. how often i'd have to replace it with a new one.
• One can source these closed rubber rings easily from any local adult toys store 🍑🍆, i.e. legumes store
• I've got an inexpensive supply which should last me for many many hours, sessions, months, maybe a few years. Non-aging UV-resistant rubber bands are also available, they're a bit pricier.
• I am stoked, literally😭 crying, to have found this wonderful solution, and tbh i am confident that this is it, that the suggested solution is going to hold up over time. Only time will tell. Ofc i would update this post with more IRL observations/reports.
• For now i am keeping the nominal dimensions of the rubber band (advertised diameter×thickness×width) undisclosed since i am the only one employing this mod.
• If time tells that this is it, then all new Nova users should install the rubber bands too, AND the company should include them as spare parts in the accessory bag. Prolly the company had experimented with squishy bands as well and eventually decided to go for the BASF hard rubber solution because hard rubber requires less frequent exchange (replacement) than a squishy consumible. But tbh i'd rather exchange/replace a supercheap rubber band once in a while (say every 6 months) than clean the hard rubber surface every hour of Tomahawk serve receive practice.

Fingers crossed🤞 that this is it. So far, things are looking good though👀, so i have reason to have high hopes and be confident. If the solution holds up acceptably, say rubber band exchange needed after every 6.0 months (no matter if used or unused, simply due to aging), then this would catapult the Nova, not the Omni, to cleared skys 🚀🌤️.

 

[Racket42]

I've joined TTD to let you know I think this thread is great. I plan to replicate your modified drills on my Nova S Pro when I have time. My wheels are white too - have cleaned them but still some residue. Spin is good though.

Do you use iPhone to connect? Mine takes two attempts every time before connecting, it's very annoying. The only solution I've found is to leave the app open all the time and it will connect the first time.

I'd love it if the devs could add a method for us to share custom drills in a way that allows us to easily download and upload them within the app, rather than have to type everything out manually. However, this would also require us to be able to manually change the robot position which (I think) is not currently possible. For example, if my position 7 angle was set to 13 degrees and yours was different, your drills would not be the same for me. I'm hopeful that the app is improved in the future.

Also, check out the ball bucket extension on Printables.com. You can add to the ball capacity if you have a 3D printer. I added one to mine and it now holds ~250 balls. However I think I will extend that to 400 by making a larger version.

 

[plunder]

Hi @dstt , welcome to the forum ( i am a newbie myself here and mostly joined to chat about my returnboards and robots ) and thanks for your feedback on the thread's content. I use Android phone or Android tablet to establish the bluetooth connection and yes it takes 2-3 attempts (and time!) until my personal smart device finds the Nova bluetooth signal; it is annoying and imho weird (because robi1 is instantly bluetooth connected) but no big deal as it happens only at startup of the training session. On a related sidenote and I will discuss it separately in another post (again) ... you may have realized by now that all your Custom Drills are not stored locally on your personal smart device but directly in the cloud, i.e. on the Shanghai Future Mind web server. The advantage is that you can use any other smartphone/tablet and as soon as you login to your app account, you'll have access to your Custom Drills. But you're right, this should not be a technical reason why we cannot share (e.g. robi1 uses downloadable files for sharing Drills and or Drill Groups) drills thru the current app version. On the Nova app, one cannot create a folder or group in the Custom Drills folder, e.g. "\Tomahawk" for all 12 Tomahawk serves, and then export/share it in form of a file for others to download/import. Clearly, the app is not as far/deeply developed as the robi1 app ("Win 3.1 vs Win95" development stage comparison ), and it took years for the robi1-devteam to deepen the app's development stage, it was a long path of evolution. We cannot expect notable changes in the Nova app anytime soon imho but i am hopeful too for the future.

It is thinkable that sharing be possible through adding publicly shared folders in your pongbot app account, since the Custom Drills are stored in the cloud anyway. Nowadays there shouldn't be any need to download/upload actual computer files (~potential compatibility issues). Let's do the sharing file-free hooray!🥳🙌

The "Ball Bucket Extender for Pongbot Nova S Pro" looks great. I would build one from cardboard or paperboard lol. Owners of a rolling catch net doht really need a larger container imho, so i am guessing that you pick up balls from the ground instead of using a rolling catch net?

White wheels, cleaning, 🫠 .. i plan to have my next extensive robi2pon training session on Fri, seeing how the rubber band mod holds up. Maybe the band breaks/snap or whatever (i doht expect so).

 

[Racket42]

Interesting that the Bluetooth issue happens on Android as well. Must not be an iOS problem as I thought. I hope they fix it in 1.0.6 which I can see is "available" in the app but not yet available in the App Store.

I do have a catch net, but fishing them out and refilling the robot every few minutes can become annoying. I guess making an extension out of cardboard would be pretty easy, but I have a 3D printer upstairs that doesn't get much action, so I'll give it some more work to do.

Good luck with the rubber bands, let us know how that goes!

 

[plunder]

Just to confirm that in the Android app it says the same:
"New Version Found V1.0.6 — Fix some issues for better user experience. — <Upgrade>"
with the <Upgrade>-button leading to Google Playstore and with no update download to be found (yet).

The +0.0.1 update is prolly for the app in general, not for the Nova app in particular. Let's see if we can notice any app change for us Nova users.