Connection problem (red light)


#1

We have updated the micro SD card and now we can see the cockpit on the internet browser (earlier we couldn’t see it) but it is still not working; on the bottom left of the cockpit it is saying there is no connection (red light).

Do you know if this is a known issue?

The only part that is not yet complete is the ESD connection to the board, can this affect the functioning of the connection?

Thanks a lot


#2

have you programmed the arduino on the cape?


#3

mmm... i'm completely missing this part,

I only bought the KIT, connected everything as per the video tutorial and turn it on.

the only modification has been inserting a new .img on the sd card.

The ESDs are not yet connected because as soon as I connect them and turn them on, the ethernet connection is dying (but i hope this is a secon problem)


#4

http://wiki.openrov.com/index.php/Update_Software

There are some that have problems with updating the cape software:

http://openrov.com/forum/topics/trouble-loading-code-into-cape-bd

regarding the ethernet dying when powering the esc's; this sounds like a short circuit somewhere. try connecting/powering one esc at a time.

we have made a wiring schematics thats based upon the video tutorials: https://github.com/OpenROV/openrov-electronics/blob/master/Openrov-wiring-rev1.pdf


#5

Via email from Eric...

I think the reason your system is shutting down when you turn the ESCs on is because you have the "Servo PWR" jumper in place (based on the photo you attached). That jumper powers the 5vdc pins off the board if only servos and no ESCs are plugged in, but since the ESCs provide their own 5v power to the bus, that jumper should be removed when you're using ESCs. Hopefully that fixes the power problem your describing, but if not, let me know and we'll try other ways of solving the problem from there.
On the servo and ESC outputs for the Cape, the signal pin (usually a white or yellow wire) goes to the trace with the square solder pad. 5vdc (the red wire) goes to the center pin, and ground is the pin on the end that does not have a square solder pad. This connection scheme is written on the board, but it's a bit hard to see when the headers are in place. Another way to describe how to plug things in is that the signal pin (yellow or white wire) should be closest to the letter of the port (A, B, C or D).
The ESCs and servo should be attached "Port" (left) to A, "Vertical" to B, "Starbord" (right) to C, and "Tilt" to D. If Port and Starbord are backward (it's possible the code may be written that way but I'm not sure) just switch which ESCs attached to A and C. Also, I should mention that we've been having trouble getting the "tilt" servo to actually tilt the E-Chassis. It seems that there is just to much friction against the tube when everything is mounted into place. When I get back to the United States (I'm still on my way back home from Antarctica) I'll try to figure out some ways to reduce that friction, but in the mean time, tilt might not work very well. Let us know if you come up with a good solution there!
The connection problem is a little more challenging. I'm still experimenting to see exactly what factors cause connections to be good or bad, but so far I've noticed the following:
1. Because of cross-talk, the connection quality is significantly worse when the tether is tightly coiled. If you've been testing with the tether coiled as you got it, try unraveling it across the floor and see if that makes things better. When you recoil the tether, try to make the bundle looser, and you should find that your connection improves.
2. Check that all the connections between the Ethernet adapters are good. I've often found that numerous strands of wire often break where they are screwed onto the terminals of the Ethernet adapters which increases the attenuation of the line. Make sure that there are no breaks in these strands, that they are screwed down tightly, and that all other connections linking the tether to the Cape are good. Also, keep the set of conductors twisted as much as possible. The twist in the brown and white-shielded conductors keeps interference from creating voltage biases which cause data errors, so try your best to keep the two close to each other and twisted.
3. Make sure you're using a modern web browser. I'm not sure exactly what the effects are of using older browsers are, but Cockpit was designed to work with newer versions of Chrome, Firefox, etc. Also, Ethernet cards in different computers may vary, so if you're still having problems, try changing computers just to check if that makes a difference.
Okay, I hope that's at least a little bit helpful. If these problems persist, or if you encounter other issues, let me know and I'll try my best to help guide you through it.

#6

Eric says:

When I get back to the United States (I'm still on my way back home from Antarctica) I'll try to figure out some ways to reduce that friction, but in the mean time, tilt might not work very well. Let us know if you come up with a good solution there!

I found some measures how friction can be reduced significantly - my electronics board now rotates VERY smoothly:

  1. make sure that the zip tie holding the beaglebone/cape is fastened with enough strength so it doesn't touch the acrylic tube
  2. same for the zip ties holding the camera
  3. to avoid that the (black) diodes on the LED boards touch (and scratch!) the acrylic tube, cover them with some tape. LED boards should be glued very tightly to the camera carrier (I used a thin layer of plastic glue instead of hot glue or epoxy)
  4. make sure the ethernet transmitter does not touch the tube (fix it with zip tie or velcro strip)
  5. cut out some millimeters of the (transparent) plastic RJ45 connector that plugs the ethernet transmitter to the beaglebone at the side it collides with the tilting servo. Be careful not to harm the wires in it. This gives you some extra millimeters the beaglebone/cape can be moved away from the acrylic tube towards the center.
  6. a VERY convenient way to reduce the amount of cables in the tube (and thus removing friction caused by it) is to remove the original USB cable of the camera and use the short white cable coming with the kit instead. See my suggestion here.

-Stefan



#7

I forgot to mention: Make sure that the camera plate sits tightly on the circular/semicircular plates on the side. I cut off the wedge-shaped ends of a cable tie and secured the camera plate with it so it doesn't impose pressure to the acrylic tube.

-Stefan


#8

This is awesome! Thank you thank you thank you!

E