How to connect the external trigger input

[Kai Golibrzuch]

We have an DAVIS346 mono, which is supposingly able to time stamp an external TTL trigger via the synch in input. We built a cable since the support told us we only need to connect the Trigger IN (going onto a BNC with 5V TTL 50Ohm). However, connecting this cable frequently causes the camera/ software or even the PC to crash. 

We then rebuilt the cable to also connect the 5V input and GND, which seems to work better. However, it is still very fuzzy to get the proper triggering to work.

[Tobi Delbruck (UZH-ETH)]

Did you check all the options for pulse high/low and timing? And read
the synchronization guide?  I'm not sure where is the latest version of
the doc.

[Kai Golibrzuch]

The Trigger seems to work reasonably well when all inputs are connected. The support first told us that connecting the Trigger In is sufficient (which would be reasonable and according to manual for 5V TTL). Well, it is definitely not. Trigger worked, we went for lunch, restarted the software (not touching anything) and did not work anymore. It also seems to be very sensitive to noise which the cable picked up. We are pulsing a lot of voltages up and down in the lab.

[Tobi Delbruck (UZH-ETH)]

You need ground too, not just IN. Your cable has both?

We did a test using long (>10m) BNC cables but not over very long
runtime. We also successfully synchronized 4 Davis346 cameras over a big
area to collect the DHP19 dataset over long recording sessions over
monthso of operation.

Here is older version of sync guide
https://docs.google.com/document/d/1U7dQqRliF5-vg2YOSl2V2UB2W2TPEUDdSR5gHPANwqw/edit?usp=sharing;
it might be helpful for cabling.

[Luca Longinotti]

Hi Kai, support saying that the signal alone is enough was a mistake,
I'm sorry for that. SIGNAL_IN, VDD_IN and GND_IN are needed for
triggers to work, as per:
https://inivation.freshdesk.com/support/solutions/articles/36000267742-synchronizing-with-external-cameras

The signal input is pretty simple: there's an electrical isolation and
voltage translation component (which is why VDD and GND must be
connected, as they provide the reference and power for the input stage
of this component), and the output of that, referenced to internal GND
and 3.3V of the camera board, is sent to an FPGA IO to detect changes
and timestamp whenever a rising or falling edge is detected. This does
not do any signal filtering or debouncing, so any fluctuations would
indeed be picked up too.
No part of that directly supports TTL logic, it's all CMOS components,
which makes the proper functioning of the translation component with
all three connections vital.
The translation/isolation component is an Analog Devices ADUM1200
(model ADUM1200WSRZ-RL7).

Could you elaborate on what kinds of crashes you saw? This part not
functioning properly should not be able to crash the software or your
PC, at most the camera would be disconnected and the software just
stops with a notification of the camera device not being there anymore.

--
Luca Longinotti (llongi)

Software Engineer
iniVation AG - https://inivation.com/
Zurich, Switzerland
Office: +41 44 500 32 14

Support - https://inivation.com/support/

[Tobi Delbruck (UZH-ETH)]

Yes, this is very important point. Since the first implementations done
in the Sensors group, the sync has been improved by inivation to adapt
to the supply voltage by connecting all 3 pins as Luca points out. These
aspects are not covered in the older sync gdoc that I shared.

Tobi