So how is it working? I'm asking because apparently they've chosen to completely ignore my cries for support so I can't find out first hand
Anyway, I'm just curious, I already scratched my head enough and found a way to do things without their precious board.
A replacement board is due to arrive today. I'm told this one was tested before being shipped and is working. The most recent batch appears to have problems (as we have both seen). I'm surprised to hear of your experience with support though - things have taken a little while to sort out for me but I have always got prompt replies to email and hopefully i will have a working board waiting for me when I get home this evening.
the board works. I still see some timeout errors running dxl_monitor from the rpi3 but re-running a couple of times seems to do the job. now having problems with the ps3 controller demo (for the hros1) but will find another thread for that
This one sort of surprised me. I would have thought they would routinely test all of the boards especially the Pro, which is not a cheap board...Originally Posted by jjennens
Glad this one is working for you!
As for PS3... Good luck! This has always been somewhat temperamental! Which board are you using to communicate with it? Edison or RPI? I have not touched my Edison's for a long while now. And now that Intel has discontinued them, probably wont again.
If RPI, what image are you using. Have not looked lately to see if there still is an RPI image that Trossen originally did and if they updated it to more recent versions of Raspian? My guess not.
Side note: over the weekend I did setup a new RPI zero W and for the fun of it, I was able to use the current Raspian bluetoothctl to be able to pair/trust/connect up to an PS4 controller.... I bought this controller earlier as I had better luck pairing to it than I did to PS3 controller.
Trossen does not have a factory in-house. They, like most other people who sell electronics, contract the manufacture of their electronics boards to electronics manufacturers (colloquially called "board stuffers.")I would have thought they would routinely test all of the boards especially the Pro
From the board stuffers, you will typically actually get a fully assembled and sealed box -- board + static bag + collateral/inserts + box + sealing.
If they were to test each board, they'd have to open each box, break the seals, open the anti-static bags, and then test the boards.
But, that's something they already paid the board stuffers/factory to do! Part of the manufacturing plan is a testing plan.
From what I know of that business, it's likely the factory somehow failed to run the proper quality controls, and Trossen was left with a bad batch. While they may sample a board or two out of a batch (opening them, meaning they can't then be sold as "new") that could easily miss a defect rate of 10, or 20, or even 50%.
My two boards didn't come in any sealed box, just packed in a sandwich of styrofoam. Which wouldn't have been a problem if they worked
I'm running on the RPI3.. I installed a hros1 specific rpi3 build with a December 2016 date on it that I think I found referenced in the forum somewhere. I managed to pair the controller and run the jstest - no problem there. When it comes to the "demo" program though, the robot assumes the squatting position but the controller does nothing. For some reason decided to sudo apt update and upgrade and that caused the dxlmonitor to stop recognizing any of the servos and randomly not seeing the controller on id 200, sometimes seeing it on 201 or 202. Reformatted the card and re-installed the hros1 raspian. All the timeouts seem to have disappeared and all servos recognized again. Unfortunately no difference on the Bluetooth ps3 demo. Tracing via printf has lead me to this fn in
HROS1-Framework/Framework/src/controller/PS3Controller.cpp
void run_server(int csk, int isk) { struct pollfd p[2]; short events; int err, ncsk, nisk; rbt_state = 0; serv_restart: p[0].fd = csk; p[0].events = POLLIN | POLLERR | POLLHUP; p[1].fd = isk; p[1].events = POLLIN | POLLERR | POLLHUP; if (debug_enable == 1) printf("Starting PS3 l2cap server.\n"); while ( 1 ) { p[0].revents = 0; p[1].revents = 0; err = poll(p, 2, 100); if (err <= 0)
poll always seems to return 0
Haven't figured out why yet though...continue;
It has been a long long time since I looked at any of that code... Did you try running it with sudo... It looks like it is trying to talk to HID... So maybe a permissions issue?
Yes, ran it with sudo. I did notice a couple of odd things. It defines USE_JOY_MIXER which spins up a thread to execute a fn that is almost totally commented out and does nothing useful as far as I can see. Meant to try commenting out the #define but not obvious that will fix the problem. It does get through the StartPS3Server fn ok but
if (bacmp(&bdaddr, &bany))
printf("Bluetooth HID daemon (%s)\n", addr);
else
printf("Bluetooth HID daemon\n");
that bacmp always returns 0. Haven't figured out whether that's expected or not yet.
just found this (from you a couple of years ago) which looks relevant
http://forums.trossenrobotics.com/sh...ontroller-code
Yes as mentioned I was having issues with the PS3 code base. Like having it create all of the extra threads and the like...
I know I had a branch of the code that I was playing with that removed most all of this... https://github.com/KurtE/HROS1-Frame...ick-controller
But I don't remember the state of it, nor if I was able to get some of these changes into the main code base...
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
Bookmarks