#apertus IRC Channel Logs

I've been using backblaze's normal backup product for 3 years and am very happy with it

I've also built one of their storage pods too, I like that they open-sourced their design

hmm, i will need 60 days to upload my data

then I'd probably look at tarsnap, those are the only two i'd trust with my data

good morning

so when I'm ready to talk to the sensor, should I run cmv_snap3 ?

that should return with a single register value, which is the sensor temperature

yep, send it to /dev/xdevcfg

yes, please do so :)

i.e. activate the check_pin20.bit in the FPGA

then run ./check_pin20

it will show you all shorts/connections between lines

without the sensor, but with the socket

pull-up JX1_LVDS_2_P

pull-up JX2_SE_1

that doesn't look too bad :)

are any of those problems?

now try to bridge an LVDS pair on the socket

hrm, let me see what I have to do that with

there is another way you can test the connections if you prefer to use a scope/analyzer

which row is unused in the socket again? the bottom one that's short two pins?

oh neat

hrm, I'm trying to short them in the socket with a 30GA wire, but am seeing nothing new on check_pin20

pings 7 and 8 from left on top row

try a different one or try a single one to GND (via 200R)

don't have an easy way to do it to GND via a resistor

i.e. you want to put two tiny cables in the socket and then lock it

and put that into the ZIF socket

N_01_P/N aka JX1_LVDS_13_N/P

http://vserver.13thfloor.at/Stuff/AXIOM/BETA/Sensor_Check.pdf

yes, that's what the diagram shows (blue is GND)

well, that's not really good :/

maybe, but they are all labeled/marked

really? strange

i.e. that is keyed, same goes for the power board

so we can assume, you have those right

the dummy interface has arrows on it which point upwards

hrm

yep

also there shouldn't be a problem, as both the sensor and dummy interfaces are rotation symmectric

so while you won't get the correct connections, it shouldn't put GND on a power rail

the design is such, that the sensor can be rotated 180 degree without problems

oh wait, it was probing as 3v0 with the lock closed

I was probing with a tiny DMM probe that fits in the holes

yeah that looks to be what happened

as I said, I usually stick in a uBGA socket (strip) and lock it in place to probe

i don't have such a small size available here

okay then, just make sure that there is no short circuit when you close the zif socket

well it it shorted 3v0 to GND we'd know I think

it LVDS pairs are at 2.5v, but the chip says mav 2.4v

for the test the zynq uses CMOS25 config

which will have the expected swing

I've verified the voltages, I guess I might give the sensor a try

be back in 5

cmv_init.sh prints out 0

./power_on.sh

/cmv_init.sh

that means that the SPI communication with the sensor is not working properly

check the light pink pins on the top-right corner

yep, index pin A1 is on the bottom left

aha

so once I get to the sensor board the PIC and FPGAs are already working

it's always fun being the 2nd person to bring up a system

let's see what scripts/code you have (sec) I don't think we have the latest tools for that on the image

remove the SFE (just to make sure)

then activate the icsp.bit (md5: f671a9a98f7bbbc38636aab43954e979)

booting back up now

feels about right

# openssl md5 icsp.bit MD5(icsp.bit)= 86005abfb43edbc4383e8450968c5d2b

no

icsp.py, dump, prog, off

http://vserver.13thfloor.at/Stuff/AXIOM/BETA/icsp_sel.py

give me a few minutes to get the hardware you have together here

do you have the 3xPMOD assembled as well?

okay, it would be helpfull to get at least the LEDs soldered on

anyway, my setup should now be similar to yours

when you load the icsp.bit the blue led on the microzed goes on

./icsp_sel.py A

(or at least the bus to the PIC)

got he blue light

I get all --'s except for 70: 70 in colum 0

now run ./icsp_picid.py

okay, sec

anything else I'll need? I have to shutdown and swap the sdcard to put new files on right now

I don't have ethernet nearby here

laptop yes

1 sec

macbook, no ethernet

yeah

there is, I just don't have one here

aha a folder

# python icsp_picid.py

b'L'

b'^'

well, that's not good

i2cdetect showing nothing now

yeah, that is kind of expected

do you have a gpio.py script?

http://vserver.13thfloor.at/Stuff/AXIOM/BETA/ICSP/gpio.py

run it :)

okay, that looks fine, let's update your kernel and devicetree, maybe the image still has the broken serial driver

Linux beta 4.0.0-xilinx-00054-g3a450582-dirty #3 SMP PREEMPT Sun Oct 25 07:08:41 UTC 2015 armv7l GNU/Linux

Linux version 4.0.0-xilinx-00054-g3a450582-dirty (root@beta) (gcc version 5.2.0 (GCC) ) #3 SMP PREEMPT Sun Oct 25 07:08:41 UTC 2015

simply copy it to /boot

and reboot the Beta

icsp again?

much better

similar but different result for B

better but still not perfect

which should be ['0x2000', '0x305b']

so, for whatever reason, you PIC is not responding

power enable and power good was fine on the gpio.py reading

which one where you looking at?

they are both 1 right now

they will get reset during the picid

try to specify A and B to the icsp_picid.py

and do they show the correct PIC ids?

no

means the pic is not properly responding

onse sec

https://kelley.ca/temp/Main_Board_v0.33_r1.1-Bottom.jpg

still bad IDs on this one

but that shouldn't be a problem IMHO

yep

can you attach a proble to *_SCL and *_SDA ?

well, DMM is tricky to see, pulses are short with the default I2C

but we can do a test, sec

what you could do is to write a small VHDL code for the zynq to toggle the SDA/SCL at a known low frequency (<1Hz) which you can measure with the DMM

yeah I can hook up a logic analyzer tomorrow

sorry, I've never used PICs before, strictly ARM and MIPS

will show a number of sequences via clock and dat

is ICSP I2C based?

ah ok

you can hook up the JTAG port of the RFE lattice

and program it directly

ah

but the unprogrammed PIC should result in sane defaults for the other FPGA pins

ok

the PIC firmware (and source) is on your beta image

in /opt/PIC/PIC16/PIC16F1718

oh, nice

ok cool

5am here, going to grab 4 hours of sleep

thanks! and thanks again for the help debugging