Pongbot Nova S Pro owners' review and discussion thread

[smee0815]

I added slight modifications to the webclient:

• simple drills with no variations (one ball definition per line), no pause (set to 0) and not more than 9 balls will be handled by the bot itself (so in these simple drills no more stops showing two blue lights, no small hiccup between stopping after the first 9 balls before the client manages to send the next 9, same as the behaviour of the vendor app by Pongbot just with up to 9 balls)
• show grey buttons, if they are deactivated
• add button to remove the pause time, if set (manually changing to 0 was a hassle imho)
• the balls drills can be written with commas or spaces as separator, might be easier to read, see the example definition below
• there can be comments per line: starts with the semikolon character, ignores everything remaining in this line

500   4000  -50   6   0  1
1000, 4000,  20, 10, 30, 1 | 5000,  700,  40, -10, 20, 1 ; either fh or bh drive
4000   700  100   2  20  1

@plunder

• id01 What is the rationale here? I'm not sure I understand what you are trying to achieve? If you want to experiment with a ball definition, I suggest you duplicate the line, add ; as a comment character in front so you have a backup and change the values in the copied line, see the example below
• id02 added semikolon (used in my preferred programming language as a starting character for a comment extendend up to the end of a line
• id03 Not sure yet what is the best way forward for curating drills, ease of use, switching quickly, sharing drills etc. I would like to keep the client as small as possible for now.
  

500   4000  -50   6   0  1
1000, 4000,  20, 10, 30, 1 | 999,  700,  40, -10, 20, 1 ; either fh or bh drive
;;;; 1000, 4000,  20, 10, 30, 1 | 999,  700,  40, -10, 20, 1 ; either fh or bh drive
4000   700  100   2  20  1

 

[plunder]

synchronizing the balls from both bots might be an issue: the only mechanism available would by synchronizing the time of sending the drill,

just a heads up, i've ordered a second Nova unit (which i name "robi3" .. my third robi in the house) for 199€ (hopefully with no surcharge uncosts 🤑) to learn about operating two novas simultaneously. btw now with the webclient and comprehending my parameter list, it has become much more convenient/natural/fun to place the nova away from the table and create an annotated drill (with ";" comments) for such distant setups which are super easy to calibrate with the 6-clicks function.

@smee0815 maybe adding a visit counter and a date/version number ("webclient v1.1, last updated: 2025-11-01") wouldn't hurt. The Chrome bluetooth pairing seems more reliable to connect now, thx! Yeah, the commenting feature supersedes id01, and i am finishing the annotation of the parameter list this weekend.

EDIT: finished (last updated: 2025-11-03🫠), enjoy!
I would love to make a similar spreadsheet for the away-position: maybe the Nova should be positioned with the eject hole being flush with the edge of the table, instead of the proposed (and never really exploited) "529mm behind'n central" position? 🧐

 

[Faye]

@smee0815 @plunder Just wanted to give a big shout-out to both of you! After your contributions to the webclient project I finally decided to buy a Nova (instead of something much more expensive that I had initially in mind). It should arrive this week. Can't wait to get started.

Even though I had no possibility to check it out personally yet, so I am not quite sure if it makes sense: would it be possible to hardcode the parameter list of @plunder into the webclient? That way, users could compile new drills on the fly maybe?

 

[Mike_TT]

Hello ,
Fantastic hack @smee0815 ! You opened doors I did even imagine.
and eternal gratitude for @plunder to ease the set up job, very valuable !

@plunder : I confirm that your rubber mods are changing all the parameters, but it is a good starting point. If you sell them, what is your selling Price and how to pay ?

@smee0815 : Especially during the set up phase, It would be great to know what were the 2 or 3 last shots of the nova, in order to ease the changes. ( It works good on android and Win/chrome ).

Many Thanks Misters

@Mike_TT

 

[plunder]

@Faye congrats to your very first post on TTD, welcome! (((btw the pongbot developers already hardcoded their own conservative parameter lists into the original Android app, including an interactive visual representation of the ball flight trajectory, very cool; and as @Mike_TT confirmed, your Nova unit will operate best with a slightly different parameter list, adapted to your setup: your exact pos8, your balls, your rubber band thickness, your room temperature, etc. So my published parameter list works perfectly with my very setup, but will not work as perfectly with your very setup)))

@Mike_TT congrats to your very first post on TTD too, welcome!
Yes, i would sell some of my stash, and i'd guess i'd take 20€ thru paypal, then subtract my cost of shipping and then prepare (for shipment) a fair amount of rubber produce for the monies left over. But before doing so, i'd instruct you to order some from the manufacturer direct first, for a couple of weeks test drive at your own responsibility (risk of damaging the Nova internal plastic structure!), since i will NOT go thru the hassle of selling and shipping, if the prospective recipient hasn't proven before that he's been already actually using the rubber bands and simply needs a refill. As proof, the shopper could share a photo with comments ...

... after which i'll be willing to ship a refill, sorry for the inconvenience hopefully fair enough😌

Point being, most "users" of the Nova woht need the rubber band mod (not now or anytime soon i suppose) or doht feel urgent enough to disassemble the wheels (btw, imho it is IMPOSSIBLE to clean the Nova wheels without removing them from the interior). Only hardcore Nova users with no fear😱 of doctoring hands-on🧰🛠️ would be eligible. ((good news, my current renewed pair of rubber bands are installed completely without glue and THIS TIME haven't fallen off or shifted sidewards .. which means that it does take some real practice/technique/experience/tricks to install the rubber band perfectly evenly/regularly/properly and that it cannot be done by a bumbler who is all thumbs👻 — he'd be dissatisfied with the mod/transaction/results/consequencesifany and likewise i would regret having shared with him: if you ask the rubber manufacturer, who only sells at bulk, for a quote, you'll understand how adversely the stuff is to come by))

 

[Faye]

Okay, my first test with the Nova was a bit disappointing.
The first disappointment was that the separately available carry bag has no space for the ball container. It's actually ridiculous to design a dedicated bag without enough space for all the robot's parts—and then charge a relatively high price for it.
But back to the Nova:

I haven't done any serious training with it yet, I just wanted to familiarize myself with the basic functions. So before I venture into the web client right at the start, I wanted to familiarize myself with the basic functions of the device by trying out the drills on the official app.
Is it just my impression, or are the pre-programmed drills all boring or useless for serious/systematic training?

The few community drills I looked at in the short time I had available weren't really any more exciting or useful either. But maybe that was because I was approaching the whole thing without a plan.

I also noticed that after this short session, there was already significant buildup of residue on the wheels, even though I only did spin-intensive drills very briefly towards the end of the session. So I may need to consider the rubber ring modification.

Then I'm not sure if there's something wrong with my device, but I had to set the “Shorter & Longer” parameter to the maximum value of 10 during the basic calibration of the device so that the ball lands in the middle of my half of the table:

I was also confused by the “Backhand & Forehand Flip” drill from CHICO TT. I tried all the settings when calibrating this drill, but only about every 10th ball made it over the net... Is there something wrong with my device if it seems to work in the CHICO TT video?

Or perhaps I should rephrase my question: is it fundamentally wrong to use the official drills, community drills, and calibration function of the app?

 

[Faye]

Ok, I gave up on the premade drills and decided to copy plunder’s drills from page 1 of this thread instead. Might need some time, but I will learn some things along the way.

Some observations that I made below. Would be grateful if someone could confirm.

- After creating a custom drill, it is not possible to change its angle anymore. So if I want to mirror a custom drill from position 7 to position 9, I cannot duplicate it. Instead I must create a completely new drill from scratch. Is this true?

- It is not possible to order the list of custom drills. They are only shown in the order of creation. So annoying. Am I overlooking this functionality maybe?

- I initially created 1-2 Tomahawk serve drills with 21° angle. Visualization of the flight curve was fine. Then I made some of the Tomahawk serve drills with parameters from his Excel sheet to match the 9° angle. @plunder I thought you chose in-app angle 9° instead of 15° (IRL) for visualization purposes in the app. But none of my 9° angle drills shows any visualizations. I am pretty sure that I copied your values exactly. Do you actually see visualizations on your 9° Tomahawk drills?

https://www.tabletennisdaily.com/forum/topics/pongbot-nova-s-pro-owners-review-and-discussion-thread.36322/post-502672

 

[plunder]

Re calibration. I use "0 0" (not "-1 10") in my pre-programmed drills calibration aka Programmed Drill Calibration. Result being, my modded wheels shoot the 2 "Combo Test"-balls far out yipeeh 🔭 🤩. Playing with the "Shorter & Longer"-setting teaches us what's happening: the head changes its frikkin Ball Height (i.e. the auswurfwinkel)! I audaciously interpret this farce as 'if you want lower/harder balls at all pre-programmed drills, then make the 2 test balls land shorter here'. I completely ignore this settings page. "0 0" and far out test balls are fine with me because i am happy with the premade drills and the custom drills. Also note that my pos8 is defined as the rear rubber feet being flush with the table's edge.

With my non-calibration and modded wheels, the ChicoTT “Backhand & Forehand Flip” drill works perfectly for me: both balls land. The cerves are short, low, and have virtually no underspin at the racket. I've seen these balls before: obviously his chosen cerve could be found in my parameter list (by editing the drop point).

@Faye, all your 3 observations are correct. Re the 15° IRL angle vs +9°/-9° set angle situation, i ignore the visualizations, because no, i doht see all of them either. (((I had chosen (and keep using) the 15° IRL angle, so that the Nova's head schwenkbereich could comfortably cover the player's far right and far left at the same time. For the ball production it doesn't matter if you had used +9°/-9° or +13°/-13° or +18°/-18° in your "setting", and i experimented that +9°/-9° would give the best visualization results from pos9/pos7 but apparently some visualizations wouldn't show because of the extreme ball parameters. When i do one Tomahawk cerve-receive drill, i doht look at the visualizations, i simply start the drill, hoping that the Nova (with its natural minimal streuung) would land all 4 balls of that drill; all 4 cerves are long cerves, because the Nova (and all other 2-wheels robots) is incapable of producing short Tomahawks with high spin rate. And if one of the 4 balls keeps hitting the table edge or is too short/easy, i'd try to tweak/finetune its ball settings. The table was for original wheels; with my modded wheels very few ball settings had to be adjusted anyway (and i didn't share the updated settings because they are optimal/unique to my very Nova system).

In my next post (it'll be ongoing work for weex!) i'll try to rebuild all modified Programmed Drills (Basic, Complex, Combined) for use with the webclient. Tailored to my very Nova system and using safe/secure, i.e. conservative parameter values (no net balls, no edge balls).

 

[plunder]

In this post we're rebuilding the Programmed Drills from pos8 (set angle = 0°), folder Basic which contains 6 drills. I'll be composing them at my PC thinking 🤔 with the help of the parameter list, then testing and tweaking🧐 them IRL. Let's post the codes here overtly (easier edit/copy/paste/run) than on a live Google Docs spreadsheet:

Programmed Drills — Basic (modified)

• b01_Push(b).txt (status: tested, good!)
  

  ;
  500,3500,-40,-6,0,1 ; s40b (some arbitrary shortish cerve)
  500,3450,35,-7,0,1 | 500,3350,40,-7,0,1 | 500,3250,45,-7,0,1 ; u35a u40a u45a attack it!
  500,3850,25,-6,0,1 | 500,3650,30,-6,0,1 ; u25a u30a
  500,4500,15,-5,0,1 | 500,4000,20,-5,0,1 ; u15a u20a try to loop it!
  ; this semicolon facilitates finger copy pasting on your portable Android device

comment: b01 is a 3-ball drill with max underspin, the pushes become righter, lower, longer, spinnier, thus harder to bh-flick. but as you can tell, all balls have very little underspin (even if realistic). the Nova simply caht produce high-quality pushes, not even from pos8!

• b02_Push(f).txt (status: tested, good!)
  

  ;
  500,3500,-40,6,0,1 ; s40b (some arbitrary shortish cerve)
  500,3450,35,7,0,1 | 500,3350,40,7,0,1 | 500,3250,45,7,0,1 ; u35a u40a u45a attack it!
  500,3850,25,6,0,1 | 500,3650,30,6,0,1 ; u25a u30a
  500,4500,15,5,0,1 | 500,4000,20,5,0,1 ; u15a u20a try to loop it!
  ;

comment: same as b01, just mirrored to the fh. since the balls have very little underspin, you should be able to fh-flick most of them.

• b03_Drive(b).txt (status: tested, good!)
  

  ;
  4300,500,-40,-6,0,1 ; c40c (some arbitrary long topspin cerve)
  7000,2000,10,-7,0,2 | 6800,2000,15,-7,0,2 | 7200,500,20,-7,0,2 ; t10c t15c t20c lands at 85%
  6500,1000,25,-6,0,2 | 6300,500,30,-6,0,2 | 6000,500,35,-6,0,2 ; t25c t30c t35c @90
  5700,500,40,-5,0,2 | 5500,500,45,-5,0,2 | 4900,1000,50,-5,0,2 ; t40c t45c t50c @95
  ;

comment: b03 is a 3-ball drill with ~max topspin, the robot's drives become righter, higher, slower, less spinny, a bit longer, thus easier for active handling. Use this as robi2pon session template, e.g. for editing ball repetition and deactivating lines. You could add balls with similar trajectory but less topspin, i.e. copy balls, reduce their UW, and increase their LW. btw the on-the-fly editing of drills for temporary validity comes very natural, normal, automatic, and is part of a robipon session. Editing means that you're temporarily shifting your focus to another small aspect within the template.

• b04_Drive(f).txt (status: tested, good!)
  

  ;
  4300,500,-40,6,0,1 ; c40c (some arbitrary long topspin cerve)
  7000,2000,10,7,0,1 | 6800,2000,15,7,0,1 | 7200,500,20,7,0,1 ; t10c t15c t20c @85
  6500,1000,25,6,0,1 | 6300,500,30,6,0,1 | 6000,500,35,6,0,1 ; t25c t30c t35c @90
  5700,500,40,5,0,1 | 5500,500,45,5,0,1 | 4900,1000,50,5,0,1 ; t40c t45c t50c @95
  ;

comment: same as b03, just mirrored to the fh. Not a challenging drill because all 3 balls can be smashed easily and safely, prolly because the topspin bites into your rubber. So add balls with less topspin.

• b05_Loop(f).txt (status: tested, good!)
  

  ;
  500,3875,-45,6,0,1 ; s45b (some arbitrary shortish cerve)
  500,4450,45,7,0,1 | 500,4350,50,7,0,1 | 500,4300,55,7,0,1 ; u45c u50c u55c
  500,4800,30,6,0,1 | 500,4700,35,6,0,1 | 500,4550,40,6,0,1 ; u30c u35c u40c
  500,5650,15,5,0,1 | 500,5200,20,5,0,1 | 500,5050,25,5,0,1 ; u15c u20c u25c, nearly #unrealistic
  ;

comment: same as b06, just mirrored to the fh.

• b06_Loop(b).txt (status: tested, good!)
  

  ;
  500,3875,-45,-6,0,1 ; s45b (some arbitrary shortish cerve)
  500,4450,45,-7,0,1 | 500,4350,50,-7,0,1 | 500,4300,55,-7,0,1 ; u45c u50c u55c
  500,4800,30,-6,0,1 | 500,4700,35,-6,0,1 | 500,4550,40,-6,0,1 ; u30c u35c u40c
  500,5650,15,-5,0,1 | 500,5200,20,-5,0,1 | 500,5050,25,-5,0,1 ; u15c u20c u25c, nearly #unrealistic
  ;

comment: long underspin balls which you're supposed to open up (with a loop, nospin loop, inverse loop, drive, inverse drive, inverse punch, or similar) instead of pushing or lifting. the balls become righter, faster, lower, spinnier, thus harder to loop. the last ball is a tough cookie and easiest looped off the bounce. but as you can tell, all balls doht have much underspin (even if realistic).

 

[Mike_TT]

Hi, I would like to share an idea with you, observing the inside mechanism of the nova S on the snapshots of @plunder , there is room to implement a rack and pinion system to turn the head of the robot with a small electrica motor.
This will allow to switch from a chop, to flip, to topspin quickly
In this way, the nova could become a halo...
The biggest constraint remains the ability to send information , from a single line, with the bluetooth connection of @smee0815, to the motor.
What are your thoughts about this Id?

@Mike_TT

 

[smee0815]

Hi! I didn't have time to continue working on the webclient, but it seems, that having predefined "interesting" balls and drills is quite helpful. As an inspiration for easier experiments, I translated (hopefully) all builtin drills of the app into the format used by the webclient. Each line has the name, the number of stars and the drill text that can be pasted into the webclient (untested so far, sorry).

Especially the random drills need to be adjusted by removing the line breaks and replacing them with the | character, so the webclient actually randomizes the balls.
The table with the drills can be found in this table

 

[plunder]

translated (hopefully) all builtin drills of the app

are you saying that the values in your table are original untouched values extracted (intercepted\ausgelesen) from the app?

Thanks for the massive work, it is helpful to see the original data.

 

[smee0815]

Doh!... Good catch @plunder Indeed, I had 5/5 as calibration set in the original Pongbot app. So all balls were lower and more to the forehand side.

I sat down and recorded all drills again with the default calibration 0/0. Forehand and backhand are now symetric (parameter drop-point). The new table with all drills (from position 8, followed by the near-net drills position 2) can be found in this table

 

[plunder]

I haven't tested your parameter values yet, but the sextuple for "pos 2 Push(F), 3 stars" is "upper wheel out of range: 115" ("115,1825,50,10,0,1")

@serve_and_volley Congrats to your very first post on the forum, welcome to TTD! 💐

@The_mfk_Dude Congrats to your very first post on the forum, welcome to TTD! 💐

 

[smee0815]

Well, here we hit the issue of how to keep the hardware itself safe: The wheel itself doesn't actually rotate with speeds <500 rpm, which means, each ball will contact a stationary wheel at the exact same position every time, leading to increased wear and tear at this part of the wheel. This might not be that big of an issue, if the other wheel rotates rather slowly, but for faster speeds the bot might tend to faster-than-expected wheel break-down.

So, either the webclient needs to get quite a bit smarter than now (which combinations to allow, more validations, etc.) or we stay on the conservative bounds hardcoded now (500<=rpm<=7200). For now, I would prefer to keep it as-is.

 

[smee0815]

Hi, I would like to share an idea with you, observing the inside mechanism of the nova S on the snapshots of @plunder , there is room to implement a rack and pinion system to turn the head of the robot with a small electrica motor.
This will allow to switch from a chop, to flip, to topspin quickly
In this way, the nova could become a halo...
The biggest constraint remains the ability to send information , from a single line, with the bluetooth connection of @smee0815, to the motor.
What are your thoughts about this Id?

@Mike_TT

I don't really understand your proposal. Also, I can do software, not hardware. If you modify the hardware, the firmware of the bot itself would need to be updated, such that the parts can be controlled. Since we don't have access to the proprietary firmware, we won't be able to change it

 

[serve_and_volley]

@plunder : great work you put in. I have yet to try your suggested parameters for the different drills.

@smee0815 : I tested your template right out of the box. Haven't installed the official app yet
It works like a charm. Thank you SO much.
I just have a few questions (maybe I can answer them partly by myself after more testing tomorrow):

- Is it possible to have a set time after each drill stops and restarts?
For example it stops after one minute and restarts after 10seconds?
- In your drill example you have two random balls in the second line, which is already super cool. Is it possible to have more than two. I mean, can I just add a third one by using the "|"?
- a save as button, or any other method to change the default file name, would be very nice

Because of you and plunder I bought the pongbot nova. I am so sick with my amicus prime. The software is a nightmare!!!

@Jslick89 have you tried the template in a omnis s pro? I ask, because the omni is also getting more affordable. But the template is an absolute game changer.

Has anyone tried it on an omni, or on a halo?
BTW: are there other big differences between the omni and the nova? I mean apart of the net and the automatic ball feeder?
I heard the omni can produce way better quality shots. Is that right? Has it also a two wheel design and how does it produce sidespin?

Thanks a lot guys. I am happy that I found this forum...

 

[floxe]

Then I'm not sure if there's something wrong with my device, but I had to set the “Shorter & Longer” parameter to the maximum value of 10 during the basic calibration of the device so that the ball lands in the middle of my half of the table:

When calibrating the ball should land 1m behind the net, not in the middle of the half, which would be 0,685m.

 

[serve_and_volley]

After another session, I am somewhat disappointed with the robot.
It can not create a topsin shot with a lot of spin. I tried the highest setting (7200rpm) and I could easily block it anyways. Even countertopsins where no problem.
I mean I am a not a good player (1420 TTR).
I am not sure if I should keep it, or send it back to Amazon.

The good thing of the session was, that know I know that it is possible, to have 3 (maybe even more?) random shots in a line.

 

[plunder]

It can not create a topsin shot with a lot of spin. I tried the highest setting (7200rpm) and I could easily block it anyways. Even countertopsins where no problem.

i had reviewed such robi2 limitations before but you're right it's better to experience it on your own IRL nm. anyway, the bot can shoot out hard fast balls but they're less spinny then, e.g. try t5c:

7200,4900,5,-6,0,199 ;99 less spinny, low superfast #unreal