Pongbot table tennis robot - review and comparison

[dingyibvs]

you misunderstanding, i never talked about 1st ball production of the amipong from my own experience. Honestly, so far i haven't trained 1st ball receive drills with the Omega. I am too busy training other bot drills for hours nonstop. There are plenty of yt vids showing/explaining how to setup the Omega (if not the Amicus) for 1st ball production, and they (incl. the Omega manual) explain that the "1 ring visible"-setting be the correct/optimal one. Sounds like that you've been operating your Amicus your whole life wrong, sh*t da happenz.

Oh you're travling rn? Oh i am on vacation myself rn. Always same same then. Whatever you throw at me, i'll throw back at ya.

Ah and ofc one could use the amipong on the higher-levelled settings (2 rings visible, 3 rings visible, etc) for 1st ball production. I was just repeating what the other authors told about the perfect setting. Regarding 1st ball production, i am not too keen AT THE MOMENT on testing it on my Omega, since i've experienced that the amipong's side spin production is different from a human's side spin (humans vs amipong use a different component pair out of the triplet <x-axis y-axis z-axis> for the ball's angular velocity vector), resulting in side-kicking bounces; one should expect that a Pongbot produces more human-like side spin, okay. But what about the 1st ball bouncing height?

So you're interested in 1st balls which land short (=are slow) and have heavy backspin? Sounds good to me. But lemme add the requirement that the height during the entire flight path should be LOW, reaching its max height right above the net line. I suggested 1cm clearance as an arbitrary measurement. So let's be clear, detailed, and informative: the net height is 15.25cm. Then my requirement be: the 'butt' of the ball can never surpass 16.25cm during its entire 1st ball flight path, until the ball is hit by the bat/racket/paddle of the human.

Here for clarity's sake a quick pic of the nomenclature used:

In other words, it's pointless to have a short slow heavy backspin serve coming at me, which bounces up to 17.25cm (measured from table level to ball's butt). High-bouncing balls are useless, even if they have much backspin and are short!

And to be clear (yet again), i never said or claimed that the amipong is or would be a great robot for 1st ball production. I do have higher hopes for this task with the Pongbot, at least with regard to human-like side spin. But first let's hear the full details and then witness how well your bot does that!

My own evaluation of the quality and capabilities of amipong's 1st ball production will be detailed in future, possibly in my amipong review thread. It will be an epic thread, discussing all details, possibilities and limitations.

Oh so you haven’t actually used your robot for services, but you’re gonna lecture me about how I’ve been using it wrong?

 

[Musaab]

The regular price is USD 15k. If you want to get one, please email us to discuss it in detail. @Musaab

let me put this in perspective for you, 18K euros could buy me years worth of practice time in specialized training centers in europe with guess what ? actual human-beings feeding me balls

 

[riemsesy]

let me put this in perspective for you, 18K euros could buy me years worth of practice time in specialized training centers in europe with guess what ? actual human-beings feeding me balls

Gross 🤮

 

[plunder]

Oh so you haven’t actually used your robot for services,

correct but even tho i'm on vacation i managed to quick-record this video-proof, demonstrating (the degree of) ball production & placement consistency, the lowness of the serve, the shortness of the serve, and the heavy backspin of the serve. And the non-degradation of the consistency, thx to the retrofit kit:

Now it's your turn i guess! 😛

 

[dingyibvs]

correct but even tho i'm on vacation i managed to quick-record this video-proof, demonstrating (the degree of) ball production & placement consistency, the lowness of the serve, the shortness of the serve, and the heavy backspin of the serve. And the non-degradation of the consistency, thx to the retrofit kit:

Now it's your turn i guess! 😛

Tell me, which ring is that?

 

[plunder]

Tell me, which ring is that?

That's the "2 rings visible"-setting, i.e. the Powerpong standard setting (unlike the Amicus standard setting which would be 3 rings visible, as described before)

 

[dingyibvs]

That's the "2 rings visible"-setting, i.e. the Powerpong standard setting (unlike the Amicus standard setting which would be 3 rings visible, as described before)

Right, and 2 rings visible is HIGHER than the Pongbot's height, as I previously measured. Wasn't your issue with the Pongbot's height? Wasn't it your assertion that it's too high to make low services? So why are you using a height setting that's even higher than the Pongbot's?

Let me repeat myself, the Pongbot's height is not an issue because settings lower than its height is not very useful. Your video proves what I said.

 

[plunder]

Then let's see your video of slow short LOW serves with heavy backspin!

Here an omni vid from 2021-Nov, i.e. over 3.0 years ago:

 

[Jslick89]

Then let's see your video of slow short LOW serves with heavy backspin!

Anyway, you said "The service mechanism simply doesn't allow for the speed/landing spot combo which creates a short backspin service".

That's completely bull 🐂

I made a quick video. Probably could have goofed around with the settings a little more if I wanted to but you can see the Omni S pro can easily serve low, slow, with heavy backspin. I didn’t count the misses completely but I think the Omni had a a slightly higher miss percentage but that’s honestly negligible when you consider I’m using a few different types of balls of all different ages and my table is not a competition table. Even if you say the power pong performed better, can you say it performed $1000+ better? Definitely not.

Also, This sort of serve isn’t actually very realistic anyways because good table tennis players will have two bounce serves. Not serves that bounce a million times on the table or back into the net.

In fact, I never had a problem with the power pong 5000 doing pure backspin shots. Where I had the biggest issues was trying to replicate side spin shots..like a short tomohawk serve off the wide forehand side of the table. Or a short pendulum serve off the backhand side of the table. The power pong cannot do it like the Pongbot can.

 

[dingyibvs]

Then let's see your video of slow short LOW serves with heavy backspin!

Anyway, you said "The service mechanism simply doesn't allow for the speed/landing spot combo which creates a short backspin service".

That's completely bull 🐂

Did you not read the whole paragraph? This is what I said at the beginning of the paragraph: "You can't use the low ring position on the Amicus/Powerpong to make short heavy backspin services."

I was clearly responding to your assertion that the Robopong's height is too high, and that you're being dishonest when you insinuated that you can't make a low service with its height. Thanks for showing a video that you can, in fact, make a low service with a height that's even higher. 😂 Though greater thanks to @Jslick89 for showing that the Pongbot can do the same.

Look, I want to repeat that I don't think the Amicus/Powerpong is bad, but it's got limitations. @Jslick89 pointed out the issue with sidespin services, but IMO that's its biggest problem with services. Its biggest problem is that in order to get relatively low services, backspin and topspin services need to be served from different heights. Backspin services need to be with 2 rings visible, while topspin services is best when only 1 ring is visible or else the service will be too high. This means that you can't run drills where you have alternating backspins and topspins.

let me put this in perspective for you, 18K euros could buy me years worth of practice time in specialized training centers in europe with guess what ? actual human-beings feeding me balls

LOL, I don't think those are meant for personal use. They're more for big training centers with a lot of beginner students. If you compare the cost to hiring a full time multiball server working 8 hours a day then you can see how the economics might work. TBH though even for that application IDK if it has enough advantages over a regular robot to justify the much, much higher cost.

 

[dingyibvs]

I made a quick video. Probably could have goofed around with the settings a little more if I wanted to but you can see the Omni S pro can easily serve low, slow, with heavy backspin. I didn’t count the misses completely but I think the Omni had a a slightly higher miss percentage but that’s honestly negligible when you consider I’m using a few different types of balls of all different ages and my table is not a competition table. Even if you say the power pong performed better, can you say it performed $1000+ better? Definitely not.

Also, This sort of serve isn’t actually very realistic anyways because good table tennis players will have two bounce serves. Not serves that bounce a million times on the table or back into the net.

In fact, I never had a problem with the power pong 5000 doing pure backspin shots. Where I had the biggest issues was trying to replicate side spin shots..like a short tomohawk serve off the wide forehand side of the table. Or a short pendulum serve off the backhand side of the table. The power pong cannot do it like the Pongbot can.

You actually can do those types of services with the with the PP/Amicus, what you need to do is to physically rotate the head to the side. The problem is that you then can't do a drill where it sometimes serve to the left, sometimes to the right.

 

[plunder]

off the wide forehand side of the table. Or a short pendulum serve off the backhand side of the table. The power pong cannot do it like the Pongbot can.

"off the wide forehand side" "off the backhand side"?? The halo could do it i guess and the nova maybe, but not the omni. The halo can be located at the backhand side, so can the amipong, as suggested in the omega manual and demoed in yt's. That's why the amipong central unit has 4 big rubber feet!

The @Jslick89 clip exhibits the same degree of consistency ("~80%") in ball production, i.e. some balls (20%?) doht clear the net either, while the other balls have to be low over the net. In direct comparison of the two vids, the omega performance appears slightly more impressive, i agree.

And yeah for serve receive practice, i do believe you guys that pongbots are superior because of their human-like sidespin production (z-axis vs y-axis). I'll test serve receive practice on the amipong more extensively in near future. I am interested in tomahawk serves off the backhand side, landing fast wide long on my backhand side, with extreme sidespin.

 

[Jslick89]

"off the wide forehand side" "off the backhand side"?? The halo could do it i guess and the nova maybe, but not the omni. The halo can be located at the backhand side, so can the amipong, as suggested in the omega manual and demoed in yt's. That's why the amipong central unit has 4 big rubber feet!

The @Jslick89 clip exhibits the same degree of consistency ("~80%") in ball production, i.e. some balls (20%?) doht clear the net either, while the other balls have to be low over the net. In direct comparison of the two vids, the omega performance appears slightly more impressive, i agree.

And yeah for serve receive practice, i do believe you guys that pongbots are superior because of their human-like sidespin production (z-axis vs y-axis). I'll test serve receive practice on the amipong more extensively in near future. I am interested in tomahawk serves off the backhand side, landing fast wide long on my backhand side, with extreme sidespin.

I didn't mean starting from wide backhand/ wide forehand side and then serving. I meant a serve aimed at the opponent's wide forehand/backhand. With the "Amipong" if you program balls to aim wide with sidespin, the trajectory and spin is not realistic. Maybe your DIY tape job solves the problem, but i'm not sure how it would when the ball has to be deflected off the deflector plate to be aimed that way. As @dingyibvs suggested, i suppose you could position the Amni pong's head so that its aimd at a wide forehand/backhand, but that wouldn't be the best practice for a multiball drill.

With my OMNI S, i can program a short underspin/pendulum serve to my backhand, then a pure underspin shot to my wide forehand, then any number of top spin balls placed where ever I like on the table. It's a way more realistic practice on the robot. When I had the powerpong Omega, my first serve receive always had to be pure underspin because the sidespin did not work how i liked. In my match play, most of my opponents are having side spin on their serves. not pure underspin.

 

[plunder]

In my match play, most of my opponents are having side spin on their serves. not pure underspin.

I find practicing serve receive super important. So today i practiced just serve receive with the Powerpong (i.e. no multiball drills starting with a Serve) and found out its limitation, which can be worked around, fortunately. The amount of side spin isn't too scary. I doht mind separating serve receive training from multiball drills!

I've been wondering how it stacks up against the Nova's Serve quality.

 

[riemsesy]

Fix was VERY EASY in the end: just some tape (I used paper like tape you'd use for painting walls) around the metals responsible for feeding the balls in the base of the robot. See picture. Never had a jam for over a year now. This is really a design issue imho.
View attachment 33663

Thanks for this tip. It fixed our bot too.
I applied some duct tape around the springs and no more interruption of the machine feeding the balls.
💪🏻👍🏻

 

[dingyibvs]

I find practicing serve receive super important. So today i practiced just serve receive with the Powerpong (i.e. no multiball drills starting with a Serve) and found out its limitation, which can be worked around, fortunately. The amount of side spin isn't too scary. I doht mind separating serve receive training from multiball drills!

I've been wondering how it stacks up against the Nova's Serve quality.

For pure service return practice the Nova would probably be the best IMO. Before the Nova came out I actually specifically got a cheaper table top robot + projector stand for service practices. The location + height adjustments really allows for practices against all sorts of services.

I've since moved away from it though, because robots these days are just a bit limited. Once you get to higher levels you're gonna want to receive short services off the bounce, and the key to that is to judge the service's length very quickly so you can decide whether to step in or not. I don't think any robot currently can serve quality serves at varying lengths randomly right now.

 

[rova]

Thanks for this tip. It fixed our bot too.
I applied some duct tape around the springs and no more interruption of the machine feeding the balls.
💪🏻👍🏻

Glad it helps! Just plain crazy that we need to do this for a top of the line robot.
And, Butterfly support does not recognize this and just blaims the customer for their own design flaws. This is the not I received from support:
- "a gesture of goodwill"
- "due to disturbances in the environment"