Hi there,
And Welcome hereā¦
So your analysis of the
Seeed Studio 114991193 USB to CAN Analyzer is spot on. The behavior youāre seeingāintermittent data loss and malformed frames on older firmware (0264) versus stability on newer firmware (8134) strongly points to a buffer management or timing bug in the earlier versionās firmware.
Upgrading firmware on these specific low-cost Seeed adapters is difficult:
- No Official Tool: Seeed Studio typically does not provide a public user-facing firmware update utility for the adapter
Most users encountering firmware-level bugs on these units are forced to replace them with the newer versions (like your 8134 units) or switch to higher-end hardware if long-term stability at high bus loads is required.
If you cannot replace the faulty units immediately, you can try these workarounds:
- Reduce VCP Baud Rate:
While counter-intuitive, setting the VCP to a lower standard speed (like 115200 or 921600) sometimes forces the host driver to use different polling intervals, which can occasionally stabilize āfragileā firmware. - Increase Inter-frame Gap: If you have control over the CAN source, adding even a few microseconds of delay between the 10 burst messages may prevent the internal buffer from overflowing.
- Use the Latest Software: Ensure you are using the latest version of the USB-CAN Analyzer tool from GitHub, which includes some protocol optimizations.
HTH
GL
PJ ![]()
PS always buy driectly from Seeed if you want to make sure you dont get an old device.. What da?
Thanks for the informative and speedy reply.
I appreciate your information on workarounds but my 10 message test scenario was just a first contact trial. The ultimate CAN bus bandwidth requirements of my project will be significantly higher also with multiple CAN sources.
Should I find the 8134 version holds up under a more representative test I take on board cgwaltneyās comment below about buying direct. Iām in the UK so not quite so easy, but I will see if they offer international shipping.
Did you check out this unit>
Thanks for the suggestion.
A nice simple device like the USB CAN analyzer is perhaps the ideal solution to my needs, connecting a PC to a CAN bus. However if it canāt work reliably I may well be looking to a solution using the board you suggest along with one of the XIAO boards.
Hi there,
So , That is a SOLID recommend with the correct ESP32 or MCU with the TWAI controller Peripheral built in
I find this info to be pertinentā¦
Nearly all modern ESP32 microcontrollers support the CAN bus protocol through an integrated controller called the Two-Wire Automotive Interface (TWAI). While they have the controller built-in, you will still need an external transceiver (e.g., SN65HVD230) to convert the signals for the physical CAN bus.
the seeed board.
ESP32 Variants with Native CAN Support
ESP32 (Original Series)
: Includes a single TWAI controller supporting Classical CAN 2.0B at speeds up to 1 Mbps.- ESP32-S Series (S2, S3):
- ESP32-S2
: Features one native TWAI controller. - ESP32-S3
: Features one native TWAI controller. Often used in boards like the
ESP32-CAN-X2
to support dual-bus applications (one native, one via SPI).
- ESP32-S2
- ESP32-C Series (C3
, C5, C6):
- ESP32-C3
: Supports one TWAI controller. - ESP32-C5
/
ESP32-C6
: These newer variants include two integrated TWAI controllers, allowing for dual-bus communication without external SPI controllers.
- ESP32-P4
: The high-performance, non-wireless variant features three integrated TWAI controllers From my early testing the C5 is packing the big punch now so , S3 before that.
HTH
GL
PJ ![]()
You also take note the 2Mbit speed is over the standard spec on the bus side , but CAN-FD may be more to the app, Iām testing for robotics and thatās the new standard. ![]()
Seeed should be looking at a Dual Bus device with the C6 and send the competition packing ![]()