OpenServo.com

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

Hey again,

Sorry for the influx of topics. I have a motor that when I plug it into the OpenServo Test application and run Scan Bus, it returns the OSIF and then an "other" device with code 0x7f. I am unable to flash it then and the console reads.

jharvey · Posted: Mon Jun 21, 2010 9:41 am Post subject:

Is this in the servo? Did the OSV3 complete a basic test before it was assembled? To me it sounds like the wire on the motor is shorting on the edge of the motor case or something similar. I'm guessing you installed the ceramic cap on the motor. Perhaps that cap is shorting on something. Another place a short can be a problem is on the power supply for the POT. Check that those aren't causing problems.

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

The company I work for ordered them preassembled so I didn't actually install the board. I did just open it up though and everything appeared to be in order. And I believe the board was installed correctly originally because it showed up on a bus scan.

If I remember correctly, the behavior started after I did a firmware update, but the version of firmware was correct from the CVS and the OpenServo Test App said it went through fine, so I'm a little puzzled.

jharvey · Posted: Mon Jun 21, 2010 9:54 am Post subject:

Ah I see it notes Spain in your address, I think I know who you are. Sounds like a firmware upgrade may have gone sour. Can you do a hard reprogram? Normally the flash is done via boot loader, and I seem to recall the device ID stuff is part of the boot loader section of the program. So if it got into a 7f condition after a firmware upgrade, it likely means the boot loader is buggered. Can you do a hard flash, instead of a boot loader flash?

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

I probably can...If I were to do that would I...

need a different connector harness? (I have the SPI and the normal 4 harnesses as detailed in the Wiki)

need to use AVRStudio or AVRDude?

Thanks,

Will

jharvey · Posted: Mon Jun 21, 2010 10:09 am Post subject:

It can be fully re-flashed via OSIF. However, you will need an extra wire(s) on the connector. The normal operation of the OSV3, only requires 4 wires. The hard reflash will require a wire on the bottom row of the OSV3 connector.

As for software, I forget if the gui software from robotfuzz can do it, or if you have to do it with a command line. I think you have a perl script that can program a 7f device in the first 3 seconds after it was booted. When that happens it will do a hard flash. Or at least that's my understanding of it.

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

Alright, I will give it a try and let you know how it goes. Thank you.

jharvey · Posted: Mon Jun 21, 2010 10:38 am Post subject:

I think there is a pin out note on the wiki, I'm not fully sure if or how that works, so if you figure it out, please update it.

Before I go, and in-case ginge can comment on a hard flash with the OSIF, these servos are using the latest non dev firmware with the OE compile options. I seem to recall there were a couple new I2C register created for the OE option, but the rest of it should be the same as a normal build. I believe the new registers will not allow it to work with the GUI, but it should work with a flash, ect. As long as the new registers aren't needed. I believe the rest of the registers are the same, os most thing should work via GUI, but not all.

I haven't done a full re-flash, but would like to. So I'm also interested in the details of how to do it. I have a couple boards I made with out connectors, and I would like to flash them before I put on the connector.

They should be in fairly good condition, and they have that ceramic cap across the drive silicone. The wires are unlikely to short on the motor, but with vibration it's possible, unlikely but possible. I ran these unloaded with a Perl scrip before sending them out. They would bounce around full rotation like in a video I posted a while back.

For the power supply, perhaps if you can put a scope on it and make sure it's not a bulk cap issue. Look for short term spikes. My bench supply has a highly capacitive front end, so it can typically deliver short spikes, other supplies might not. Even then mine is/was set to 5 amps, I seem to recall it would get wonky when set to around 1 amp, even though the average power was down in the ma range. Basically the inrush can cause brown out.

Any how, I need to disappear now, good luck, and I'll be keeping an eye out for your progress.

ginge · Posted: Tue Jun 22, 2010 6:50 pm Post subject:

are you running the linux app as sudo? thise damn libusb drivers require sudo access (without changing configuration options)

Also, when you see address 7f it means you are in bootloader, and you should be able to upload some new firmware.

If you download the OSIF windows drivers, there is a testapp.exe that can be used to reflash the openservo

testapp flash 0x7f file_to_flash.hex

also note that if you use the -dev tree from CVS, you will be on your own in terms of interfaces.

Cheers
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

jharvey · Posted: Tue Jun 22, 2010 7:47 pm Post subject:

About the sudo thing, I was talking with a fellow who thought that had something to do with my problems in windows. I didn't track it down.

Also I just added that to http://www.openservo.com/OSIF_programmer

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

Ginge: I have not tried sudoing the app, I will do that tomorrow when I go back to work. That would make a lot of sense. And for now, I am staying away from the -dev part of the tree. I read that they have different registers than what is listed, so I figure once I get them working with the normal branch, we can then move into developing our own software.

Jharvey: I will post back tomorrow with my results tomorrow morning.

Thanks all
Will

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

So I tried loading the testapp.exe and console kept returning "This program cannot be run in DOS mode." Is there a way I can apply those same parameters using Windows Explorer and right clicking properties in the file? I am sorry for the numerous questions, as my location says, I am in Spain, but I am an American. My computer is in Spanish and I haven't quite gotten used to the whole thing in Spanish.

ginge · Posted: Wed Jun 23, 2010 8:08 am Post subject:

Can you try this version, I just tested it on an XP machine, and it behaves as expected:

http://www.headfuzz.co.uk/files/robofuzz/testapp.zip

cheers
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

wpotter · Joined: 18 Jun 2010 Posts: 50 Location: Valencia, Spain

Your app worked! Thank you!

Note: I just used the 4 wire harness. I thought that I was supposed to use the SPI one for hard resets, but the SPI Harness failed and then I used the 4 wire harness and it worked fine.

Jared: When rebooted the motor with the faulty firmware, I used the OpenServo CVS repository V3 [non dev] firmware and it does not work with your python scripts, but works perfectly with OpenServo Test. If it wasn't before, It is obvious now that you two used different firmwares.

Thank you both,

- Will

jharvey · Posted: Wed Jun 23, 2010 8:44 pm Post subject:

The different firmwares are generated from the same source code, but with some different configs set when compiled. Sounds like you got this 7f servo to come out of it. Did you reflash with the OE firmware?