Build a flipper "table" but got issues with sensors

[mderie]

Hi,

I've started to build a flipper "table" and I want to use my Mindstorms so that the flipper can play alone !-)

In short I've build a ball launcher (around a touch sensor) and two flippers (left & right) around sensors (I've two kits : EV3 & NXT)

Here's the problem I got with both Light / color sensors or with the sonic sensors : I've got erratic value readings (out of range, nonsense)...

My question here is : do you have any suggestion on how I could use the sensors in a way that it can react quickly on the ball passing at proximity ? And be quiet when nothing happen !

I've included a picture to show my purpose. I guess that you all know what a flipper is :) The ball launcher is not included but in my head it is working !

I'm really stuck with the sensors they are too sensible... Thanks in advance ! I've started with the moon plates. I've tried multiple ball sizes, but still ball detection is loose :(

I've done a very simple program in Sratch and with EV3 classroom tool and still fighting with the lack of accuracy of the sensors : they are too tiny spot oriented

Regards

Edited January 20, 2021 by mderie
Added the missing image :(

[Mr Jos]

Hey, I don't see any of the pictures? Have you posted the correct link for them?

[mderie]

Unable to upload any image :(

It weight 101258 bytes...

Edited January 20, 2021 by mderie

[Mr Jos]

Try using www.Flickr.com, then upload your photo's, and copy the link to them here.

[mderie]

Thx for the tips Mr Jos ! Here's the link :

The main issue is that the readings from the sonic or the light sensors are erratic, useless.

They report uncoherent values. I need something like the ball is just below me or not !

Regards 

 

 

[Toastie]

Hi @mderie,

what about more contrast? A white ball on a gray background may not be easily recognizable by the sensor. Also, is your erratic reading from static situations, or from a ball swiftly moving? What is the temporal resolution of your code and what is the required resolution (speed of ball) to distinguish from "it is there" or "not there" and then more importantly "almost there", as the flippers, when operated with finesse, need to know all that.

Best
Thorsten

[Mr Jos]

You might want to try putting the sensors on an angle, now they point 90° (Downwards). Try mounting them from behind the tip of flipper watching to the top of the board, and point down only around 30° not 90°. This way the measured distance will be larger, less error readings will occur. I would like to make a test-setup to show/test but I'm 3months away from home (pile of Lego), only got my 6DoF with me. If needed I can make a sketch on paper and upload it. Got time enough sitting out 15days quarantaine not allowed to leave the hotel for 1minute.

[mderie]

@Toastie : good point, I've switched from the sonic sensors to the light ones without thinking about the contrast

@Mr Jos : yep this is really a good idea, I'm on it... I'll try post the image of the next configuration ASAP. Maybe back using sonic sensors in cross fire configuration !

For both & the eventuals other readers :p Thx, I'm always nicely surprised about people helping each others on almost all possible topics.

BTW, I'm still using EV3 Education / EV3 Education Scratch... Let say that it it is though for a programmer guy like me !

Do you all use MicroPython ? I've just flash an SD card with EV3DEV (Linux distro)... Not yet tried :(

[mderie]

So as promised, here's the new design. It provides better readings ! I guess that I should involve a cube ball to be perfect :)

Anyway, its good enough, thx for the feedback... I'll try now to finalise the complete board and the ball launcher located at the bottom when the ball evades the flipper

 

 

For the programming part, I'm still using the "ugly" editor... And a very simple ball detection algo, maybe this can be improved as well :)

 

[mderie]

Hi, here's my final project... stopping as is since I'm becoming a bit mad with the sensor misreadings ! Thanks for watching / helping