TXT Controller and I2C speed

Alles rund um COMPUTING inkl. Microcontroller & Co.
Everything about COMPUTING incl. Microcontrollers
Forumsregeln
Bitte beachte die Forumsregeln!

Re: TXT Controller and I2C speed

Beitragvon PHabermehl » 09 Aug 2018, 11:25

Hi Frans,
This is excellent information, thank you very much!

Peter
PHabermehl
 
Beiträge: 870
Registriert: 20 Dez 2014, 23:59
Wohnort: Bad Hersfeld
Alter: 45

Re: TXT Controller and I2C speed

Beitragvon hamlet » 09 Aug 2018, 17:59

Hello Frans,
Many thanks for your photos and the type info. I'm sure this will help others, who are also puzzled by the unexpected gear ratio. ft could have just enclosed to their building sets a small sheet informing about this change, mhm.
I'm still a bit puzzled by the large prime in your measured 76 1/6=457/6 ratio. I guess there exists a three or four stage gear combination 3 * a/b * c/d * e/f * g/h resulting in a ratio very close to 76.1666. But who cares as your measured ratio works out nicely.
Kind Regards,
Helmut
hamlet
 
Beiträge: 279
Registriert: 12 Jan 2011, 22:41

Re: TXT Controller and I2C speed

Beitragvon fotoopa » 09 Aug 2018, 18:42

PHabermehl hat geschrieben:Hi Frans,
This is excellent information, thank you very much!
Peter
Thanks Peter.
hamlet hat geschrieben:I'm still a bit puzzled by the large prime in your measured 76 1/6=457/6 ratio. I guess there exists a three or four stage gear combination 3 * a/b * c/d * e/f * g/h resulting in a ratio very close to 76.1666. But who cares as your measured ratio works out nicely.
Helmut
With my latest tests I am almost sure that the right ratio is closer to 76,166666. With this value, my own hardware quadrature encoder no longer gives an a deviation even after 30467 pulses on the FT motor 135484. FT would indeed give some more information, you still need it to determine your distances. I think Fischertechnik scores very poorly on giving elementary information.
Still, I think the TXT Controller is a great device. Especially the Robo Pro software and the TXT Controller allows you to create an HMI interface via your PC. If you use a UHD screen, a lot of information details will fit on it. The transfer of parameters over an I2C connection then allows me to solve all the problems that are too slow on the TXT controller itself.
It's high time now I made my new project with the colour pearls. This is in the first place a mechanical problem, how to guide the pearls to a point for the color measurement. The color selection process must then be completed smoothly.

Another small update:
Today I also opened an XM motor. Of course there is no encoder on it, but the PCB is just the same. In both cases, the PCB "layout" is provided with but not present with a real quadrature encoder, i.e. with 2 hall detectors. For the encoder connection, 3 or 4 pins are provided, to be fitted according to motor model. There is no magnetic disk on the XM motor. Only the 075F fuse is present. There was no mark of the type of motor used.

Frans.
fotoopa
 
Beiträge: 81
Registriert: 05 Okt 2017, 11:44
Wohnort: Belgie
Alter: 75

Re: TXT Controller and I2C speed

Beitragvon fotoopa » 10 Aug 2018, 11:47

One last measurement to accurately measure the encoder pulses of the FT135484 motor. I have brought all the encoder signals to my FPGA hardware and run the motor 10000 revolutions. My own hardware quadrature encoder with 40 pulses / rotation indicated 400,000, the FT measured pulses via an input to the FPGA indicate 761716 pulses.
This gives me a very accurate measurement of 76,1716 pulses/revolution. The ft motor was running at full speed without load and took a little more than 30 minutes to measure.

Bild
Higher resolution: https://www.flickr.com/photos/fotoopa_hs/42152572070

Frans.
fotoopa
 
Beiträge: 81
Registriert: 05 Okt 2017, 11:44
Wohnort: Belgie
Alter: 75

Re: TXT Controller and I2C speed

Beitragvon fotoopa » 16 Aug 2018, 19:57

Here's the start of my new project to sort 10mm color pearls. This is the test setup for measuring the color via an I2C sensor module TCS34725. I still have to make the feeder for the pearls. My first tests for this were not good. But I'm now working on a chain version and it will probably work.
Here is a picture of the poor feeder module:

Bild
HD version: https://www.flickr.com/photos/fotoopa_hs/43147651895

Bild
HD version: https://www.flickr.com/photos/fotoopa_hs/43147652395

Now some pictures of the color test setup:
Bild
HD version: https://www.flickr.com/photos/fotoopa_hs/43168635685

Bild
HD version: https://www.flickr.com/photos/fotoopa_hs/43356493894

The color recognition works well. If I can now supply the colour pearls correctly, I'll have to succeed. The user has to choose a color and the number and the right pearls have to fall into a tray, the others have to go back to the feeder. There is still a lot of work to be done, but I think it will work!

Frans.
fotoopa
 
Beiträge: 81
Registriert: 05 Okt 2017, 11:44
Wohnort: Belgie
Alter: 75

Vorherige

Zurück zu Robo Pro / Computing / Software

Wer ist online?

Mitglieder in diesem Forum: 0 Mitglieder und 3 Gäste