U4 going short on teensy 4.1

C

csreades

Active member
I am trying to trouble shoot an issue with U4 going short on a few teensy boards I have. The challenge is mostly that of the 30 boards I am using I have only had U4 go short on one setup. I want to think it is something to do with upstream of the USB cable. If this is the case what would I be looking for? Issues with the hub, cable etc?

I am also struggling a bit to work out what version of U4 I have, and what version to replace it with, is there a trick to identifying U4?

I have had a look at my powered hub that I am using, it is https://www.startech.com/en-gb/usb-hubs/st53004u1c I got a USB cable and attached my oscilloscope to the 5V and GND wires and a 100k resistor in parallel. I wanted to see if the hub was behaving. When inserting/deserting the USB cable there are some pretty large voltage spikes. This seems to be the case across a few of these hubs so I am not sure this isn't "standard".
 

Attachments

  • MicrosoftTeams-image (28).jpg
    MicrosoftTeams-image (28).jpg
    186.1 KB · Views: 48
Last edited:
It’s the 3.3 V regulator on the teensy 4.1. I am not sure which specific part I have on mine, I believe they have been changed twice?
 
Yes, U4 has changed twice. Teensy 4.1 made in 2020 and 2021 had TLV75733P. Due to chip shortages, we changed to NCV8186AMN330TAG in early 2022. The NCV8186 is easy to identify (using a magnifier), because it has 8 pins rather than 6. In early-mid 2023, we switched to TLV75533P. Visually it looks almost identical to TLV75733P (and I suspect internally it is the same silicon with the overcurrent limit set lower). You can tell the difference between TLV75733P and TLV75533P by looking at the Schottky diodes. The early boards with TLV75733P had two BAT54C dual diode packages (3 pins each). All later boards starting with the change to NCV8186 have a single BAS40-05V quad diode part with 6 pins.
 
Last edited:
Mine is an 8 pin IC with "GDT". I thought it would be the NCV8186 but the datasheet says that is "GAX"? Or am I reading that wrong?

Edit: After spending more than 5s looking and scrolling down I can see GDT is the 3v3 non-discharging variant.

Now I know all this I am still looking for possible failure mechanisms?
 

Attachments

  • MicrosoftTeams-image (30)_proc.jpg
    MicrosoftTeams-image (30)_proc.jpg
    126.2 KB · Views: 50
Last edited:
Are you using the Teensy 3.3V regulator output to power a lot of peripherals?

There has been some evidence that hanging large amounts of capacitance on this output can sometimes cause reverse current flow through the regulator during power down. If that is a possibility, more info on that type of failure can be found at this link
 
csreades said:
Mine is an 8 pin IC with "GDT". I thought it would be the NCV8186 but the datasheet says that is "GAX"? Or am I reading that wrong?
Click to expand...

This is the info I see in the datasheet on page 9.

1701127436035.png


And this on page 10.

1701127515344.png


If this doesn't clear everything up, can you give a link or screenshot to the documentation you're seeing about "GAX"? From the info I can see, GAX would mean a 1.75 volt regulator. I have no idea how to interpret the date code digit.
 
Thanks for the information previously. I replaced the part and it has since gone twice more. Any thoughts on what could be potentially causing this? My first idea was to use something like (https://uk.rs-online.com/web/p/voltage-regulators/1005939) to create the 3v3 rail. The idea that a bigger part might be more robust to what ever is causing issues. However I have looked at the power on sequence and not sure that is compatible.

This teensy is on a board that I have 10-20 actively running, only one of them specifically is having this issue so I am not sure it should be the capacitance as previously stated?
 
Are external things connected to the 3.3V power? Or it is only powering the parts on Teensy 4.1 (no wires connected to any of the 3.3V pins on the edge of the PCB).
 
It powers:

DAC (MCP47F) ~ 380uA
Couple of LEDs ~ 3mA
Half of a buffer (TXS0104EPWR) ~ 10uA
RS485 RX (MAX3096CSE+) 2.4mA

So lets say 15mA if we are super conservative? A decouple cap for each IC (I used 2x 10uF and 1x 1uF).
 
Could noise on the USB line cause the fault on the teensy? I am getting USB port saying over voltage and turning the port off at the same time as this happens.
 
You must log in or register to reply here.
Back
Top