Print

[uminded]

I got an abandoned classic Pleo with broken vertebrae with the intention of only needing to 3d print a few new ones and have a working pet. On the upside the skin is in perfect shape.

Turns out there was no battery and when I use my lab power supply and a 10k resistor to the 3rd terminal Pleo boots, moves his neck and goes "huh?" in a slow loop. I am going to order a 1.25mm header so I can get the serial debug messages. As I have the skin off is there any voltage test points I should probe and make sure the voltage regulators are working?

[pnhicks]

Hi unminded,


The serial port route is tricky!  Have you considered using dino-mite to connect to pleo
I am on vacation at the moment and cant give you the test point info off the top of my head.
Sorry about that!  There is a post by Xi Ping (I think) that may have a picture with some voltage on the input board.
Cheers
Peter

[uminded]

Can't use the USB port to connect do to the power cycling and "huh?". The port shows up in device manager then disappears right away.

I was hoping the serial port would dump more info as it must boot enough to make the "huh?" sound.

[pnhicks]

Hi,
Did you try booting from a an sd card with a script like statsRb (works with Ugobe too)?
 Cheers, Peter

[pnhicks]

Look at post from Xu Peng about Pleo not reading battery voltage. It has those test points you wanted.

[uminded]

Well the serial port revealed some more info. Apparently the battery is fine and it is in a boot loop.

I tried the 1.0 downgrade and it said version mismatch and Pleo barked like a dog. I tried the 1.0.2 and it gave some stats on the head camera then did an Adler32 calc before rebooting.

Boot log is here: https://pastebin.com/TCwmracv

I found the POP downgrade and it worked! at least it fully boots before restarting. I get this every time:

*************
pleoPOPPopState(): Current State: IDLE
*************
INFO: anim: motion_play(8207[bubble_low_idle_v2]) returned 0

[pnhicks]

Hi,


I would now try to install an "normal" os.  The POP (Point of Purchase) is not a full OS.  It was used as a demo in stores or given to users who had bricked their ugobes.  You can install OS via usb or sd card.  The usb versions are on the forum as well.


There is a large (3 page) discussion) about the upgrade/downgrade process as well.


Cheers!


Peter

[uminded]

I was still not able to do the V1.0.1 upgrade from POP but had an idea. I took the POP script and replaced the img files with the ones from v1.0 and they installed. I then was able to install v1.0.1 normally with the images written to all three controllers.

I now get a full boot but I still am stuck in a boot loop. The warning messages make me think I have a PLL chip or something not running at the correct frequency or something. I also don't get the missing files messages as the upgrade worked.

Code: [Select]

WARN: high: timing: sys_application TOOK TOO LONG: 605109us!
WARN: high: timing: sys_application TOOK TOO LONG: 698766us!
> WARN: high: timing: sys_application TOOK TOO LONG: 97262us!
> WARN: high: timing: sys_application TOOK TOO LONG: 146663us!
WARN: high: timing: sys_application TOOK TOO LONG: 26988us!
WARN: vm: No main function found. Will stop trying.
WARN: drive: no function with the name of emotion_eval: File or function is not found
WARN: high: timing: sys_drive TOOK TOO LONG: 27860us!
WARN: high: timing: sys_drive TOOK TOO LONG: 49976us!
WARN: high: timing: sys_drive TOOK TOO LONG: 41761us!
WARN: high: timing: sys_drive TOOK TOO LONG: 40314us!
WARN: high: timing: sys_drive TOOK TOO LONG: 523260us!
WARN: high: timing: sys_animation TOOK TOO LONG: 47672us!

This is the last set of messages before it reboots

Code: [Select]

NFO: vm: Loading fatigue_birthing(4012)
** bvr init: fatigue_birthing
** bvr main: fatigue_birthing
INFO: anim: Opened 0x304F[fat_birth_pre_pre] with 1 entries (1 properties each) at 8897.
INFO: anim: setting command to 12367 at 8898
INFO: plog: plogged id 12367, ts 8 @ 0x8006
INFO: anim: AnimSystem: find_entry returns entry 1 with weight 79
INFO: anim: Entry found in command fat_birth_pre_pre
INFO: anim: playing new Primary motion: 8260
INFO: low: motion: ts 8901: Load motion 8260
INFO: low: motion: Opened motion file birth_pre_pre
INFO: anim: playing motion birth_pre_pre[8260] as Primary at 8903
INFO: anim: full-body primary. stopping and disabling secondary and tertiary
INFO: low: motion: Set playback speed to 100% of normal.
INFO: plog: plogged id 8260, ts 8 @ 0x800A
WARN: high: timing: sys_animation TOOK TOO LONG: 47672us!
INFO: <0xff>

Boot info is

Code: [Select]

Pleo boot loader
7361 - Nov 15 2007 21:34:31
Starting image 8128 - Mar 19 2008 06:24:52<0xfa><break>


Code: [Select]

PleoPM version $Revision: 8139 $
INFO: high: setting monitor command to: stat power

                  Battery Voltage = 75, Min 75, Max 76 (tenths of a volt)
              Battery Temperature = 0, Min 100, Max 0 (degrees C)
                  Battery Current = 264, Min 257, Max 335 (milliamperes)
             Battery sensor value = 70 (% capacity remaining)
      Battery sensor direct value = 70, Min 0, Max 70
              Battery Voltage A/D = 667, Min 661, Max 669
          Battery Temperature A/D = 875, Min 874, Max 877
              Battery Current A/D = 37, Min 36, Max 47
            Thermistor Resistance = 129938, Min 129785, Max 130291 (ohms)
      Battery Internal Resistance = 0, Min 100000, Max 0 (milliohms)
        Battery Capacity Consumed = 0 (mAh)
                       Powered up = true
                  RTK Adjustments = 0
                           RTK IK = 0
                           RTK VB = 0
Power monitor reset early warning = 0 (/LLO)
> WARN: high: timing: sys_application TOOK TOO LONG: 97262us!


[pnhicks]

Hi Un,


I was going to try and sound like I understood any of your post..


Sorry, I'm like a dog looking at fire at this point, its very interesting, but I've got no clue really.


If I was forced to make an opinion, humph, I'm still of a mind that your OS is corrupted and that an install of 1.0.2 via usb "LifeOS 1.0.2 update TWEAKED (via Windows USB)" available on the forum is your best bet.


I'm not sure that the stats power is telling you the whole story about the condition of your battery.  Did you run the update with this battery?  I might even try to run the update with the power supply IMHO.


I have about a 25 percent success rate so take that with for what its worth!


Cheers!


Peter

[pnhicks]

Hi Un,


Here is a schematic from Xu Peng of the input power board on the bottom of the Ugobe Pleo (the board with the on-off switch).


If you were still looking at power issues.





Cheers!


Peter

[pnhicks]

Hi Um,


I was re-reading your initial post and I may really have a good clue!  If this Pleo had damage to the neck vertebrae, it is quite possible that the data lines going to the head cpu got damaged (pinched chaffed or cut).


They are quite small (like 26AWI) and suffer from crosstalk and EMI problems due to their length and the EMI noise from the pleo servos.  If one or more of them is compromised, it may corrupt the data being sent to and from the head cpu during upgrade and or boot.


The wire harness has a choke mounted around it to help prevent noise but I have seen some chokes cracked due to rough handling of the Pleo.


Just another of my wild ideas!


Cheers!


Peter

[uminded]

This collection of messages is concerning:

Code: [Select]

WARN: vm: No main function found. Will stop trying.
WARN: drive: no function with the name of emotion_eval: File or function is not found
INFO: high: application: No App loaded.
INFO: vm: No startup script to load.
INFO: vm: Function main not found.
WARN: vm: No main function found. Will stop trying.

I think maybe dino is rebooting because their is no app and the watchdog resets it. I have him powered off a lab supply and I know the current and voltage are stable.

I am using the SD card images and a 2GB SD card. Maybe its still to big for a 2008 embedded controller? I just formatted the SD as FAT and pasted the files onto the drive. Anything else I am suppose to do?

Until I solve the boot loop issue I can not access him via the USB port.

[pnhicks]

Hi Un,


Is it possible that the POP script did not work/finish correctly?  Would it be worth it to try to run it again?


Did you get a chance to look at the multicolored neck wiring harness?


Cheers,


Peter

[pnhicks]

Hi Un,


I have used a 2Gb SD card without problems on my Ugobe. I'm not sure if I used it on version 1.0


Cheers!


Peter