Continually blowing UART RX on teensy 3.5

Status
Not open for further replies.

jasper

Active member
we have been using teensy 3.2 for a few years controlling a bunch of peripherals, including an Xbee on Serial port 2. No problems, more than 20 devices built and hundreds of hours of operation.

Lately we have switched to Teensy 3.5 and after about an hour or so of use RX2 suddenly stops working on the Teensy. Everything else works, including the TX2.
Any clues why the same circuit would work on the Teensy 3.2 and not on the teensy 3.5 ?

We now have about 5 teensys with blown RX2 pins, whats the warranty on Teensys?:p
 
Interesting! I'm wondering what the difference in the chips could be, assuming the same, identical, circuitry connected.
I'm wondering if there is an ESD (static electricity type of thing) happening? Maybe the '3.5 is a little more susceptible?
Another option is a differential in Vcc between the Teensy and the device it is talking to so that a "High" on the serial line from the sending device tries to pull the Rx pin higher than Vcc. Monitoring with a CRO would give an idea of that.
One thing then we do here is always have some small resistance in the Rx and Tx lines to external devices, baud rate depending. Commonly and even for 115,200 and short leads we use 100 ohms (may be a far too high if you have line lengths more than a few inches).
We've shipped many thousands of devices that have a CPU IRQ pin exposed to the "real world" and use a low capacitance transorb (transient voltage suppressor), then a small resistor, and use a schottky diode from IRQ to Vcc to try to catch and limit any overshoot as well. This has been highly successful for us.
No idea if any of that helps.
 
thanks yes we will follow those suggestions. appreciate the feedback. its a very short track from the teensy to the Xbee .
 
Have you tried using pin 9 (RX2) on any of these damaged boards as a regular I/O with digitalRead or digitalWrite, or as PWM with analogWrite?

Warranty service is tricky with DIY electronics like Teensy. Fundamentally, warranty replacement is about defects in materials & workmanship. If we made a mistake in manufacturing your Teensy (which wasn't caught by our testing, which does verify connectivity to every pin), or if someone component turns out to be defective (again, not caught by our testing here), then we replace it. Warranty also covers damage during shipping, which is usually easy to spot (torn packaging, crushed cardboard, etc). But warranty isn't about compatibility. It isn't about damage that occurs, which isn't the result of a manufacturing defect. Warranty doesn't cover failures caused by other parts.

Not every case is clear. DIY electronics is a huge challenge, because of pretty much limitless & complex ways the products are combined with other stuff. It's not like a consumer product like a blender or washing machine that's generally only used in specific ways. Robin & I try to be reasonable. If you're certain these blown boards are due to some manufacturing error on our part, please email Robin directly. Reference this thread.

My gut feeling from only the info here is there's very likely some circuit problem. What that could be, and why it would impact Teensy 3.5 but not Teensy 3.2 is something I'm having a hard time imagining. If you're willing to show photos & detailed info about your design, maybe something will become apparent?
 
exceeding the voltage can damage the pin, diagrams/pictures and schematics of how it's wired can show where the issue lies
 
I'd been a while since I've used any XBee products, but as I recall they all have 3.3V signals.

Hard to imagine how this could be damaging the hardware. Then again, this is still a case of so many unknowns, even just a confirmation that the damaged boards have their pin unable to do normal I/O when reprogrammed to use pinMode, digitalRead, digitalWrite rather than Serial2. One possibility is some other type of problem appears to be damaged hardware.
 
I was totally unaware of XBee till I looked them up a little while ago. If yours is 3V3 and powered off the Teensy Reg then I can't imagine any Vcc difference would exist to cause a problem. If you have a separate regulator then that would be a consideration, something to check closely.
I noticed there is a 5V version of the XBee. If you're using that one, and the Teensy 3.2 and 3.5 are 5V tolerant then there is the possibility the 3.5 is less tolerant - and of course that is a wild guess. If it is a 5V Xbee then I'd suggest a level transaltor - resistive divider or an active one etc.
Paul is right - there are too many unknowns. I doubt it is a systemic problem with the '3.5 - something else is happening.
I wonder; crazy idea, the XBee is a transmitter when it is sending data - is there any possibility that RF is getting into the RX pin and causing a problem? I seriously doubt it, but, well, in the absence of other ideas etc. Ur. Thinking more, I would have thought RF would cause latch up and the processor would recover on a power down them up cycle, there wouldn't be enough energy in it to damage silicon. Maybe not such a clever idea. I presume you've done that, cycled the power (sorry that is a stupid question).
 
Status
Not open for further replies.
Back
Top