Teensy 4 dead

Status
Not open for further replies.

KG5NII

Member
Is there a schematic, component layout ETC for the Teensy 4? I've a problem with mine. Was running one of the ILI9341_T3 graphics example sketches overnight which was working fine, but when I checked on it about 18 hours or so later, it was dead. The 3.3V pin is reading about 0.8V or so. Measuring the resistance from the 3.3V pin to ground shows the fault: 14.2 Ω. A schematic would be helpful, but I'm afraid it's toast. Ran good for 3 days. Live fast, die fast I guess. I was really getting attached to the little guy....

Measured voltages with only USB plugged into Teensy:​

PIN Voltage
VBat2.3v
3.3V0.80v
On/Off2.3v
Vin5.0v

Curious about where VBat and On/Off pins are getting those voltages from.

Cheers,
Ron, KG5NII
 
Schematic is now up on main website: https://www.pjrc.com/teensy/schematic.html

My guess is you already tried it, but make sure nothing but the USB is connected to it and then double check voltages, including what is going into VUSB... Also if you soldered pins in and the like, look for any shorts...
 
Oh there is a short on the entire 3.3V rail. Resettable fuse too hot to touch resulting in only 2.5V to the 2305 mosfet which in turn feeds that 2.5V to the 3.3V LDO. LDO is in current limit due to the 14 ohm short on the 3.3V rail. After examining the schematic its easy to see where the 2.5V on VBat and On/Off pins are coming from; the 2305 mosfet. Probably the RT1062 has the short. Could be the flash memory or the Kinetis chips or even a tantalum but I doubt it. Either way it's toast. Kaput. Finis.

These devices are not Arduino which are relatively speaking almost indestructible. The processor on the T4 is extremely sensitive and very unforgiving, thus the Teensy has to be handled with extreme diligence with every precaution taken while handling. Not a criticism, but a testament to the nature of the beast when dealing with devices such as this. Thought I was careful being it was working and stopped on it's own accord. Will have to order another one, but will wait until all the 'I gotta have it now!' hysteria is passed, whenever that may be.

Thanks for caring enough to answer my query. Appreciate that and you as well.

Cheers,
Ron KG5NII
 
Without implying anything about your setup which I know nothing of, I notice the Adafruit version of their ILI9341-based touchscreen https://learn.adafruit.com/adafruit-2-8-tft-touch-shield-v2/downloads https://cdn-learn.adafruit.com/asse...fruit_products_cap_touch28_tft.png?1438860497 was designed for the original Arduino boards, meaning pullup to +5V on SCL, SDA lines for example. There may be other non-Adafruit boards of similar design which also use 5V interfacing levels.
Some of the Teensy 3.x had 5V-tolerant pins, but Teensy 4 is not at all 5V tolerant, so that is one potential difference from what people may have been used to.
 
The Teensy 4 is as good and robust as all the other Teensy 3.x's/LC as abused here. Eight full months of beta units and ham handed handling caused no trouble with any of them - all still working.

Not sure if anything was soldered there? But soldering the side pins is more clear of PCB parts than some.
 
sounds like perhaps you fed 5 volts into the teensy and that caused a issue. the teensy 4.0's are not 5v tolerant on any pin but the VIN pin (where the 5 volts from the usb plug goes to)
 
firehopper, nope. Was being powered by the USB. But do agree with what you said otherwise.

Cheers,
Ron KG5NII
 
Last edited:
defragster, am positive about no solder shorts. Have examined it closely under my microscope. Yes, I get your point about it's robustness ETC. That's the whole point of me trying to identify why mine failed. What is clear is that something happened after about 15 hours into the run. The PC connected to the Teensy during the test is my Linux workstation. Examining its syslog files shows quiet USB activity for approx 15 hours. Then all of the sudden there are tons of USB connects, disconnects for about 60 seconds consistent with the same messages one gets when plugging / unplugging the USB from Teensy. Then total quiet after the last disconnect. Was running one of the ILI9341_T3 example sketches at the time.

Cheers,
Ron KG5NII
 
JBeale, that is an interesting possibility. Purchased this display from pjrc. That having been said though, I seem to remember there being a 3.3v reg on the display but will need to double to check. It's been quite a while since I've looked at it. I will make some measurements though. I would think if that was the problem it would have fried the Teensy immediately though. Instead it ran for 15 hours driving the display before dying.

Your username sounds familiar to me but I can't quite place it.

Cheers,
Ron, KG5NII
 
defragster, am positive about no solder shorts. Have examined it closely under my microscope. Yes, I get your point about it's robustness ETC. That's the whole point of me trying to identify why mine failed. What is clear is that something happened after about 15 hours into the run. The PC connected to the Teensy during the test is my Linux workstation. Examining its syslog files shows quiet USB activity for approx 15 hours. Then all of the sudden there are tons of USB connects, disconnects for about 60 seconds consistent with the same messages one gets when plugging / unplugging the USB from Teensy. Then total quiet after the last disconnect. Was running one of the ILI9341_T3 example sketches at the time.

Cheers,
Ron KG5NII

That is odd - I've left ili-9341 and ili-9488 and st7735 and one and two st7789 displays running on one or more Beta T4's for days or longer since January and no issues - each one still working - the first 1052 is tucked away working - but three used and abused beta units in a heap on my desk still good. I only soldered one production T4 with long pins and that is the one where the long pins allowed running the twin ST7789's and that came some weeks back from first production batch #1 units so it has had long run time already.

Of course only that final unit is a production unit - and only a sample of one - but the biggest known/apparent change was the silkscreen - after one component exchange and the SD pad layout.

Other posts were wanting to be sure that the display in use didn't somehow have 5V pullups on the SPI lines or other signal levels that might have let 5V touch a T4 pin.
 
Other posts were wanting to be sure that the display in use didn't somehow have 5V pullups on the SPI lines or other signal levels that might have let 5V touch a T4 pin.

Yep, that is on my to-do list. But as I have indicated elsewhere, one would think if that were the case, the Teensy would have been toasted almost immediately. Still for the sake of completeness I will be doing some measurements on the display when time permits.

Cheers,
Ron, KG5NII
 
The Adafruit display used 10k pullups to +5V. If you connected that display to a chip clamping inputs to +3.3 V + some diode drop, you might only have around 0.1 mA flowing into the pin. I'm not sure if you can get away with doing that for days on a T4. Some devices might be ok with it(?) I'm not willing to try it myself :) Of course it might be something else completely.
 
Results of display measurement and possible conclusions

The Adafruit display used 10k pullups to +5V. If you connected that display to a chip clamping inputs to +3.3 V + some diode drop, you might only have around 0.1 mA flowing into the pin. I'm not sure if you can get away with doing that for days on a T4. Some devices might be ok with it(?) I'm not willing to try it myself :) Of course it might be something else completely.


Just wanted to let you and the others who contributed to this thread know that after checking the display, It is definitely a 3.3v display. It has an onboard 3.3V LDO whose input is taken from the VCC pin. Interestingly there is a jumper that one can short which will bypass the LDO regulator completely, presumably if one were to provide their own 3.3V to the display. The only line which appears to have a pullup is T_IRQ which is a 10k to the display's 3.3v rail. No pullups on any of the other pins.

As far as the Teensy 4 failure goes, the only other possibility I can fathom would be the close proximity to RF from my HF rig. It's a homebrew rig only running 100W output power, however, I was making some tweaks to the it's RF PA during the timeframe I was running the Teensy test. BUT, the Teensy seemed to be running fine during and after that. I suppose it is possible that RF could have been the culprit, I was at one point outputting some really nasty spurious harmonics even though that was into a dummy load, it is the only conceivable explanation. The rig and the PC connected to the Teensy has a USB connection, so the Teensy ground was not isolated from the rig. I am leaning very heavily towards thinking a RF ground loop situation caused the issue. I've worked with dozens and dozens of microcontrollers in the past and have seen some pretty bizarre behavior due to RF ground loops, but this would be the first outright controller failure I have experienced due to one. Yes, ground loop sounds like a definite possibility to me.

Proper analog and digital grounding is not trivial. In fact it can be a nightmare in some situations.

Cheers,
Ron, KG5NII
 
As far as the Teensy 4 failure goes, the only other possibility I can fathom would be the close proximity to RF from my HF rig.

It would be interesting to know if HF traveling over some connecting lines, due to some non-linear circuitry in the input of Teensy, could generate very high rectified voltage.
 
Status
Not open for further replies.
Back
Top