Power issues by USB

Hi there,

Having problems powering the board over usb - measuring capable output at 1.1A on my USB 3.1 ports and 0.6A on USB 2.0 on AMD Picasso platform under Windows 10 - the same platofrm I’m interfacing with CP2102, SAMD51, STM32F405 etc without issue.

The only way I can get Giantboard stable is by also connecting a Lipo (3.7v) to the board and then boot loops disappear and the board is stable enough to enter a serial console.

The board only pulls about 140mA from the USB port (measured), but not stable on USB itself - only when Lipo is connected - its like something is wrong with the Vregs if its only powering correctly with Lipo attached.

Presently the boards not holding in good stead to be usable for my competition entry.

Something definitely seems off. If possible, can you take a photo of the top section of the board, near the USB connector? There’s not a whole lot going on between the connector and the PMIC, so hopefully something simple. The fact it works while on battery is promising.

Hiya,

Here goes :slight_smile:

I am having exactly the same issue. Without a battery connected it gets stuck in a reboot loop. For completeness I tried another USB lead and a different brand of micro SD card. It is only when a battery is connected that it will boot. I just presumed that was a limitation on the design.

alistair, I had a board that was doing the reset loop you described but I never tried hooking a battery up to it (see Reset loop at Linux boot). It boots fine with the battery. Looks like I can confirm the same issue also.

Everything looks normal, it’s really odd. The trace coming from the 5V USB goes straight into the PMIC, so there’s nothing really going on between there. Out of curiosity are you trying to boot with the Wifi feather wing attached? If so a battery is necessary during high cpu load like booting as the wifi and giant board together come very close to the 450mA limit on the PMIC with current spikes coming from the wifi initializing. When powering from battery this limit is lifted and will provide the full power of the regulators.

Even just the bare board is boot looping by USB power.

Don’t seem to be able to initiate charging to occur either by editing the power management line out.

It definitely sounds like a defective part somewhere on the board. Without having a board to test, I can’t offer much advice. I’ve not run into this problem with any prototype or production boards that I have. It might be best to try and see about getting a replacement.

I’ve seen a few problems like this that were attributed to QFN packages not having the pad open in either the solder screen or mask. So of course, impatience just got the best of me and I lifted the PMIC off. Looked OK. In conclusion, my board is now probably not available for troubleshooting this issue anymore. :crazy_face:

Really wish I would have tried touching up the joints around the perimeter first. Sorry.

1 Like

I might smother the PMIC in TOPNIK in the morning and wipe the joints up a little.

1 Like

If you reach a conclusion on this then please let me know. It might or might not be related to my USB devices not showing up issue.

My board boots up from USB power with no battery. In fact I have not yet tried a battery. Current taken during boot-up varies from 40mA to 80mA give or take.
So anything much more than that is likely to be a hardware problem.
The LEDs have a story to tell during boot-up. It takes about 1 minute 30 seconds from power to getting the USB port login prompt.

On my board generally the 130 to 140mA over USB range , so yes assuming a problem. And don’t think a solder wipe will fix this. Shelving this board waiting response from Electromaker as this is a competition board.

I keep making parallels to a Pi Zero as a USB gadget to this board and coming to very quick and stark conclusions. If the QC was there I wouldn’t be doing this.

Yeah, I think I get what you are saying and the QC thing bothers me too, if I’m being honest. However, I understand the nature of this project and am glad it happened. Then again, I was excited more for the minimalistic easily-sourceable system-in-package design, than for being a Pi Zero replacement.

I have some more clues. While trying to rebuild the kernel to diagnose my USB problems the USB connection to the computer kept going down and then returning. I an communicating with the board over serial so I can see it is not crashing, but the PC keeps identifying a new drive that needs checking (presumably because the connection was lost while mounted) and the USB network connection (that I am sharing the Internet to thought the PC) keeps going down.

This started when I started using the network heavily to pull down files but is continuing afterwards. I unplugged the USB for 5 minutes and plugged it back in and it is still restarting the connection (just like it was being unplugged and re-plugged) every few minutes.

I am going to leave it for an hour or two and see if it continues to do the same. Then if it is stable again run it for a bit without using the network (over USB) to see what that does.

Could this be as basic as a duff power regulator? Can I inject some regulated power on other pins to test this?

1 Like

Great,

i’m also looking for a work aound, i’m using the featherwing, all works fine with the LiPo,
but for my project the LiPo doesn’t fit.

Could i put a beefier 5V into the JST connector of the battery?
I don’t think hooking up a more powerful 5V to the Vbus will help anything since it is still passing the PMIC, right?

The featherwing can’t be powered 5V, right?

cheers

Wim

I’m contemplating trying an Adafruit 500mA Micro USB Lipo charger I have if I have spare JST connectors to make a double ended connector to go from Giantboard to charger.

Will take a look at this when I get time - might be a solution to moving regulated juice into the JST connecor.

Wouldn’t the battery circuit just accept the 5V source?
Or does it identify that as faulty / cutoff.
I don’t want to risk it haha.

I could maybe do a voltage divider from 5V to 4V… but don’t have a lot of space for loads o stuff

Its this little gizmo I have :slight_smile:

That looks promising,
How about putting a capacitor on the 3.3V line to shave those peaks?