3.6V minimum VIN requirement for teensy 4.1 with USB communication might be too low

Trip

Member
Hello everyone,

I am posting this for anyone who might have the same problem regarding USB communication with the teensy 4.1 when using external power on the VIN pin.

For a project that involves a teensy 4.1 on a custom PCB I used the minimum voltage requirement of 3.6V for the VIN pin to conserve as much power as possible.
This gave me inconsistent serial detection and programming on my computer. After raising the VIN voltage to 3.9V by replacing a resistor on the PCB these problems dissapeared completely.
Serial detection and programming did work with 3.6V on VIN, but it could take up to a minute for it to start working. Once it was working however it was perfectly stable afterwards which was the weirdest part.
I was worried it might be a signal integrity issue at first, as the bottom d- and d+ pads are connected to a PCB underneath and routed to a USB mini port.
But it was the actual voltage on the VIN pin of 3.6V that was too low for consistent detection of USB. I dont know if this is reproducable on a different setup but if you come accross issues like very slow serial detection or slow programming it might be worth it to try this.
 
If I read the datasheet correctly, the maximum dropout voltage can be 475mV:

Capture.PNG

So, it's wise to supply at least 3.775V to the Vin pin.

Regards,
Paul
 
If I read the datasheet correctly, the maximum dropout voltage can be 475mV:

View attachment 24991

So, it's wise to supply at least 3.775V to the Vin pin.

Regards,
Paul

Usually that kind of dropout only happens when the regulator is very close to its max rated current in this case 1A. You can see that in the graphs of the datasheet as well. I measured the current with a usb voltage/current measurement stick and it only consumed about 50mA. So I dont think the dropout was the problem in this case. Although it could become a problem of course if you draw a lot of current.
The 3.3V rail was also perfectly stable at 3.6V and 3.9V. If a program was loaded it also worked without issues even at 3.6V.
The only thing which did not work properly was the USB link at 3.6V. Which as I mentioned before only started working after like about a minute although the exact time was random. Changing it to 3.9V fixed this problem it starts immediatly now when applying power.
 
Usually that kind of dropout only happens when the regulator is very close to its max rated current in this case 1A.
Agree.

As far as I see on the schematic, the only other pins of the processor that could be influenced by the Vin, are pins USB1_VBUS & USB2_VBUS [called USB_OTG1_VBUS & USB_OTG2_VBUS in the datasheet].
These pins are spec-ed as:

Capture.PNG

Minimum 4.4V? I'm confused...perhaps PaulStoffregen can shine some light on this?

Regards,
Paul
 
Agree.

As far as I see on the schematic, the only other pins of the processor that could be influenced by the Vin, are pins USB1_VBUS & USB2_VBUS [called USB_OTG1_VBUS & USB_OTG2_VBUS in the datasheet].
These pins are spec-ed as:

View attachment 24993

Minimum 4.4V? I'm confused...perhaps PaulStoffregen can shine some light on this?

Regards,
Paul

I believe according to the usb specification the power supply needs to operate in that range for client devices.
https://www.beyondlogic.org/usbnutshell/usb2.shtml (VBUS section)

But maybe the IMXRT1062 can still work with an even lower voltage I have not looked into that.
The pinout of the teensy mentions that VIN needs to be between 3.6V-5.5V. So I designed my PCB to supply 3.6V on the VIN pin so it draws a bit less power.
But for the USB link to function properly it needs a higher voltage atleast in my case. I am just leaving this information here so other people dont make the same mistake and spend a few days trying to figure out why their USB link is buggy in their system.
I already solved the issue I just dont know why a higher voltage fixed it. If someone wants to figure out why it does that, they are free to do so but its not nessecary for me atleast.
 
Perhaps the lower input voltage results in a slower V_3.3 rise time and the 1.5K pull-up resistor used to pull up the USB DP signal isn't pulling up quickly enough and the PC is lacking the signal needed to start enumeration.

You could test this hypothesis by powering the Teensy and waiting for a few seconds to plug in the USB.
 
Perhaps the lower input voltage results in a slower V_3.3 rise time and the 1.5K pull-up resistor used to pull up the USB DP signal isn't pulling up quickly enough and the PC is lacking the signal needed to start enumeration.

You could test this hypothesis by powering the Teensy and waiting for a few seconds to plug in the USB.

That was what I considered - but didn't want to search for Paul's startup order post that was relinked recently ... seems the startup sequence on the pjrc.com T_4.x bootloader chip page has sequencing - though not sure it links to that telling post.

Shortcut to test would be to put 3V on vBat as that helps startup ordering based on the post not linked.
 
Perhaps the lower input voltage results in a slower V_3.3 rise time and the 1.5K pull-up resistor used to pull up the USB DP signal isn't pulling up quickly enough and the PC is lacking the signal needed to start enumeration.

You could test this hypothesis by powering the Teensy and waiting for a few seconds to plug in the USB.

This seems to be a more likely cause of the problem. Right now I am working on another part. I will test this out and report it here if I have the time. The system also includes a soft start on the power supply maybe that has an effect as well. I might have a board that i could use as a guinee pig.
 
Here are details Paul noted on the startup and power response ... with the final line note about vBat powering as a way to get a resolution if this were the problem

pjrc.com/threads/64468-Teensy40-only-3-3V-power-supply
I will try that as well although the chip itself booted up just fine. It only was the USB serial link which displayed strange behaviour. The test program that I ran was blink with some text on the serial line. The led was blinking immediatly at startup but the serial only started like a minute later.
 
Back
Top