Hey everyone, I recently picked up the Arducam UC-A06 Rev C (the 6-camera MIPI CSI-2 interface card kit) for my Jetson AGX Orin Developer Kit and I’ve hit a pretty frustrating physical interference problem.
The MIPI interface card sits right on top of the NVME SSD slot on the carrier board. The board-to-board connector Arducam used only gives about 5mm of separation between the cards. The underside of the Arducam card has these IC chips that sit below the PCB level and some brass mounting points — they’re making direct surface contact with the top of my installed NVME SSD. I’m seeing mechanical stress on both boards and what feels like a developing heat spot right where they touch.
Has anyone else run into this? Wondering if there’s a known workaround before I start rigging something up myself.
Quick update — I did try flipping the interface card 180° along its long axis to see if that would help. Problem is, that orientation puts it over the WIFI NVME slot instead. Those aerials are right there and honestly not sure what kind of interference or shielding issues that might cause. The design seems to intentionally route over the storage NVME slot, so I’m guessing there’s a reason they didn’t go the other way.
I spent some time looking for a male-to-female ribbon or flex cable that could go between the card and the camera interface socket to give me some mounting flexibility, but no luck finding anything compatible so far. Anyone know of a cable solution for this connector type?
Thanks for the detailed report. Can you confirm which exact B2B connector is on the Jetson carrier board side? If you have the Samtec part number that would be really helpful — we can look at what mating height options exist and whether a different stacking connector could solve this without needing to redesign the whole board.
Yeah, I pulled the datasheets. The connector on the AGX Orin Developer Kit carrier board is a Samtec QSH-060-01-H-D-A-K-TR.
NVIDIA’s design guidelines for mating connectors with this are the QTH-060 series. The last two digits in the part number control the stacking height:
QTH-060-01 → 5.00 mm spacing (this is what Arducam appears to be using)
QTH-060-02 → 8.00 mm spacing
QTH-060-03 → 11.00 mm spacing
QTH-060-04 → 16.00 mm spacing
So right now with the -01 variant we’re barely clearing 5mm. An NVME SSD with a heatsink label is already pushing 2-3mm thick, and the ICs and brass standoffs on the underside of the Arducam board eat into what little gap remains. That’s why there’s contact.
The QTH-060-03 (11mm) would probably be the bare minimum that gives enough clearance and at least some airflow. The QTH-060-04 (16mm) would be ideal — that’d give proper breathing room for an SSD with a heatsink.
That’s excellent analysis, thank you for pulling those part numbers together. I’ve passed all of this along to our engineering team.
Just to confirm what you’re seeing — when you mention brass mounts making contact, are those the standoffs on the Arducam card itself or something on the SSD side? Either way, the 5mm QTH-060-01 clearly isn’t giving enough clearance for a populated NVME slot underneath.
We’re actively evaluating the connector swap to a higher stacking variant. The QTH-060-03 and QTH-060-04 are both on the table. I’ll update this thread as soon as engineering comes back with a decision. In the meantime, I wouldn’t recommend running the setup with the boards in contact — that pressure plus the heat concentration could damage either the SSD or the camera interface card over time.
The brass mounts I’m talking about are on the underside of the Arducam card — looks like mounting/standoff points built into the board assembly. Combined with a couple of IC packages that protrude below the PCB plane, it’s multiple contact points against the SSD surface.
And yeah, I’ve pulled the SSD for now just to be safe. Not worth risking either component. I’ll keep an eye on this thread for the engineering decision — the 16mm option would honestly make this a non-issue entirely. Appreciate you taking it seriously and not just brushing it off as a one-off.
Glad you pulled the SSD — definitely the right call. The engineering team has confirmed they’re looking at this specifically now and are considering moving to a higher B2B socket on a future revision. No firm decision yet on whether it’ll be the 11mm (QTH-060-03) or 16mm (QTH-060-04) variant, but your analysis and the Samtec part breakdown has been really helpful in framing the options for them.
I’ll post back here once there’s something concrete. If anyone else with an AGX Orin + UC-A06 Rev C setup is seeing similar issues, please chime in — the more data points we have the better.