Rug Warrior Pi: Build Blog (a PiDA bot)

[alanmcdonley]

I installed the new Pololu 2Amp supply set at 5.3v which results in a very steady 5.2v at the Pi bus and USB ports.

So time to crank the cores and see how hard this robot can think!

====================== THROTTLING (and voltage) TEST ==========
launched four fibanacci.py & processes
Shut down when frequency throttling started for temperature managment (80 deg C)

Start of test:
********* RWPi STATUS *****
2016-07-22 00:36:31.886675
battery.volts(): 7.3
battery.hoursOfLifeRemaining(): 7 h 32 m
Processor Temp: 42.4'C
Clock Frequency: 600 MHz
throttled=0x0
currentsensor.current_sense(): 337 mA
irDistance.inInches: 10.8
usDistance.inInches: 53.0
bumpers: ALL




Last status before throttling:

********* RWPi STATUS *****
2016-07-22 00:40:55.165051
battery.volts(): 7.2
battery.hoursOfLifeRemaining(): 6 h 40 m
Processor Temp: 80.6'C
Clock Frequency: 1.20 GHz
throttled=0x0
currentsensor.current_sense(): 802 mA
irDistance.inInches: 10.7
usDistance.inInches: 56.1
bumpers: ALL

Throttling to limit temperature began:

********* RWPi STATUS *****
2016-07-22 00:40:58.223945
battery.volts(): 5.4
battery.hoursOfLifeRemaining(): 0 h -15 m
Processor Temp: 80.6'C
Clock Frequency: 1.20 GHz
throttled=0x20000
currentsensor.current_sense(): 788 mA
irDistance.inInches: 10.6
usDistance.inInches: 56.0
bumpers: ALL


Last Throttled status before stopped test:

********* RWPi STATUS *****
2016-07-22 00:41:47.498234
battery.volts(): 7.2
battery.hoursOfLifeRemaining(): 7 h 7 m
Processor Temp: 82.7'C
Clock Frequency: 948 MHz
throttled=0x20002
currentsensor.current_sense(): 694 mA
irDistance.inInches: 10.5
usDistance.inInches: 55.8
bumpers: ALL

One minute after test stopped temp dropped to 58 degC:

********* RWPi STATUS *****
2016-07-22 00:42:45.689743
battery.volts(): 6.0
battery.hoursOfLifeRemaining(): 0 h -9 m
Processor Temp: 59.6'C
Clock Frequency: 600 MHz
throttled=0x20000
currentsensor.current_sense(): 354 mA
irDistance.inInches: 10.7
usDistance.inInches: 55.9
bumpers: ALL

(I have to figure out why the Pi thinks all the bumpers are activated - they are active Low with pull ups configured. - something weird going on there.)

Note that this six minute test ate 7 hours of "idle" juice. Pretty high price to think so hard.

[alanmcdonley]

I must have fried some of the mcp23S17. Ordered some new and some 1 LSB MCP3208 chips also-supposedly pin compat to the PiDA 3008. I'll have to figure out the i/f change and modify my v/bit constant. It will be interesting to see if the readings are more stable.

[mikronauts]

Sorry to hear that some magic blue some escaped... but at least the chips are socketed

You are correct, you can use a 12 bit MCP3208 on PDA, but you will need to modify the analogRead() and analogReadRaw() functions slightly for the 12 bit (instead of 10 bit) result.

Please note, analogRead() is still supposed to return a 10 bit result, use analogReadRaw() for 12 bit results for compatibility.

I must have fried some of the mcp23S17. Ordered some new and some 1 LSB MCP3208 chips also-supposedly pin compat to the PiDA 3008. I'll have to figure out the i/f change and modify my v/bit constant. It will be interesting to see if the readings are more stable.

[alanmcdonley]

Posted on the 1yr anniversary of my Rug Warrior Pi project plan: Robot Resurrection: Another Dumb Robot

After replacing the 23S17 Digital IO expander, upgrading the MCP3008 to 12bit MCP3208, adding a current limiting 1k resistor to the active low bumpers circuit, and going back to the 5v fixed power-supply, my Rug Warrior Pi is again "whole".

I updated my battery voltage sensing / battery life estimator code, the infrared ranging sensor, and the current sensor for 12bit voltage readings, and confirmed all subsystems (especially skirt bump directions and encoder interrupts) are functioning correctly.

The system boots and runs with no "panic attacks" (under-voltage events) when running full-time speech reco, motors, tilt-pan servos, with ultrasound and IR distance sensors all going ( and also a test with two cores at 100% running fibonacci, three wireless ssh sessions logging data, and the sensor threads going).

This week I am doing an online course so not much progress will happen, but it is good to reflect on the last year's progress and know that my "Pogo" is healthy again.

[mikronauts]

Looking good!

Posted on the 1yr anniversary of my Rug Warrior Pi project plan: Robot Resurrection: Another Dumb Robot

After replacing the 23S17 Digital IO expander, upgrading the MCP3008 to 12bit MCP3208, adding a current limiting 1k resistor to the active low bumpers circuit, and going back to the 5v fixed power-supply, my Rug Warrior Pi is again "whole".

I updated my battery voltage sensing / battery life estimator code, the infrared ranging sensor, and the current sensor for 12bit voltage readings, and confirmed all subsystems (especially skirt bump directions and encoder interrupts) are functioning correctly.

The system boots and runs with no "panic attacks" (under-voltage events) when running full-time speech reco, motors, tilt-pan servos, with ultrasound and IR distance sensors all going ( and also a test with two cores at 100% running fibonacci, three wireless ssh sessions logging data, and the sensor threads going).

This week I am doing an online course so not much progress will happen, but it is good to reflect on the last year's progress and know that my "Pogo" is healthy again.

[alanmcdonley]

I'm about to start another cycle of working with my robot.

I still miss having a common platform to be able to benefit from other people's brains and efforts. I don't see a robotics card as good as the Pi Droid Alpha. The most active, common platform is probably the GoPiGo, which seems too limited for my dreams.

Besides, I think I have all the hardware done now. It is time to start creating behaviors and an integration architecture.

Alan

[alanmcdonley]

I'm having great fun starting back at zero with my bot API. I decided rather than do everything myself, I would build a GoPiGo compatible API so I could run programs written for that bot with *little or no* changes.

With the pan servo and the ultrasonic sensor implemented I was able to run the us_servo_scan.py program with my bot located 59cm from a corner:

Code: Select all

Map:
--------------------------------------------------------------------------------------------------------
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                                                                       |
|                                                 o o o o                                               |
|                                                 o o o o                                               |
|                                               o o     o o                                             |
|                                             o o         o o             o                             |
|                                           o o     x       o o           o o                           |
--------------------------------------------------------------------------------------------------------
dist_l[i_center_reading]: 59 ang_l= 90



[alanmcdonley]

Just ran
sudo apt-get update
sudo apt-get -y dist-upgrade
sudo rpi-update
and rebooted so now kernel 4.9.65 (was 4.9.41)

and now SPI and the motor PWM do not appear to be working:
tilt, and pan servos work
ultrasonic works
ADC does not
motors do not
encoders do not (interrupts off DIO)