#timvideos

mithro: so now, I have a logic analyzer. But I don't know what I'm doing with it

and can't find the UART on the TOFE pinout in the opsis docs

tumbleweed: You want to attach the logic analyzer to the I2C bus and reset line between FX2 and the FPGA

tumbleweed: one problem you have is that the I2C CLK isn't on a header anywhere

I'll need some help finding it

UMAC1 is probably the best IC to use

tumbleweed: Look at the schematic do at https://github.com/timvideos/HDMI2USB-numato-op... and page 7 you can find the UMAC1 wiring

Title: HDMI2USB-numato-opsis-hardware/Numato-Opsis-v3-prod-3-g83e7bd5.pdf at master · timvideos/HDMI2USB-numato-opsis-hardware · GitHub (at github.com)

U23 is nice and big, so I'm giving it a shot

tumbleweed: yeah

SDA signal is also avaliable on P18

The reset signal is available on K3 == JFX2-RST

yeah

trying to figure out how external clock is supposed to work

tumbleweed: hrm?

how do I get it to use the I2C bus' clock?

tumbleweed: You just want to capture the clock + data signals?

presumably

tumbleweed: So you need to attach your logic analyzer to pins 5 and 6 of UMAC1

yeah, I gave up on external clock, and just captured at a higher clock rate

I have some analysis which is basically a screen full of bus errors

tumbleweed: oh, yeah you definitely want to oversample it

tumbleweed: you also want to trigger on the JFX2-RST

aha, let's try that

JFX2-RST seems to stay high a lot

tumbleweed: That is expected, the FX2 is only running when it is high IIRC

you pull it low to reset the FX2

that's annoying for triggering

tumbleweed: you want to trigger on it going high

tumbleweed: the SDA and SCL should be idle for the rest of the time

tumbleweed: but just after a reset it will read a bunch of stuff out of the EEPROM first, and then from the FPGA second

tumbleweed: The data will only occur after a reset

ssk1328: ping?

mithro: no, I mean, the clock line is just staying high now

tumbleweed: the clock line will only toggle when it is sending data

tumbleweed: Do you know how I2C works?

well it's serial, so yes, that's what I'd expect

but that's not what's happening

which is why I'm confused

I don't know the I2C protocol

but I do know that these logic lines probably shouldn't be staying high

aha, because I'm being a numpty

well that was obvious

but you know...

tumbleweed: When idle, they will remain high

tumbleweed: https://learn.sparkfun.com/tutorials/i2c

Title: I2C - learn.sparkfun.com (at learn.sparkfun.com)

no, there seems to be a bug in this software

if you use too big a capture buffer, it's all corrupted

I noticed this earlier

mithro: Pong

mithro: did you see the doc I sent here

ssk1328: yes

ssk1328: I don't quite understand the first section

mithro: that's a problem I am unable to solve

ssk1328: You can have multiple sinks in a component

ssk1328: Why doesn't the adder block just have two sinks?

I tried that

mithro: That gave a error

Maybe not using the correct way

ssk1328: What was the error?

mithro: https://corelli.tumbleweed.org.za/stefanor/i2c-...

mithro: Let me reproduce, but it was something about Endpoint can't have two definition of sink

mithro: also https://corelli.tumbleweed.org.za/stefanor/capt... (FX2-RST seems to be having some fun)

tumbleweed: Okay, lets start with FX2-RST - the correct behaviour is that on boot it goes high, then at some point later it will get pulled low for a period and then go high again

it is rather wobbly

it's line 2 in my capture

tumbleweed: That is the FPGA resetting the FX2 to get it to read the new firmware

I dunno how to read a sla file?

ssk1328: Can you give me edit access so I can play with your block diagram?

mithro: with SUMP

mithro: at 1.8ms, it seems to settle down

mithro: Sharing setting changed

tumbleweed: https://github.com/timvideos/HDMI2USB-misoc-fir...

Title: HDMI2USB-misoc-firmware/fx2.c at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)

when I say wobbly I mean it's toggling faster than the I2C clock, at times

tumbleweed: https://github.com/timvideos/HDMI2USB-misoc-fir...

Title: HDMI2USB-misoc-firmware/opsis.py at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)

tumbleweed: It shouldn't be toggling....

well yeah, that's why I mention it

mithro: This is how am I doing it right now, and surprisingly no errors turned up

Title: Ubuntu Pastebin (at paste.ubuntu.com)

mithro: I have started the make gateware for this, lets how this works out

mithro: Made a typo in above code

mithro: I am gonna pastebin the error I got here

Title: Ubuntu Pastebin (at paste.ubuntu.com)

mithro: Also the correct code which cause this error http://paste.ubuntu.com/21109705/

Title: Ubuntu Pastebin (at paste.ubuntu.com)

ssk1328: That isn't an error?

mithro: The second pastebin is the terminal output of error I get

mithro: I'm guessing it could just be a very slow transition from low to high

but there are still some spikes to low

nafc

ssk1328: Ask _florent_ or in #m-labs

mithro: Okay

tumbleweed: It's hard for me to tell if the problem is actually your FX2 line or your equipment?

ssk1328: I updated your document with my thinking about the pipeline

mithro: I can't tell, either

mithro: and it's not necessarily a problem. How did the rest of that I2C dump look?

mithro: I saw the changes