HDMI reverse power USB issue
The_Praiodanish 2017-12-11 18:29:33 1553 Views1 Replies Hi Folks,
as I'm new to this forum I'd like to introduce myself, I'm an electronics engineer designing products with and around x86-based hardware, so it was just a question of time to come across the Lattepanda
In one design I've encountered a rather strange issue and I'd like to know whether anyone had the same problems.
I designed a "UPS style" addon board which contains a LiPo backup battery and the required DC/DC converters to generate the +5V power for the Lattepanda. When the whole device is in the "off" state, I disconnect the internal +5V rail by an electronic switch from the power input of the Lattepanda. The backup power board also contains an USB2.0 hub (USB2517), several PCM2900E codecs and a FTDI FT232.
Weird things start to happen when I connect an HDMI monitor to the Lattepanda and switch on my device afterwards: The +5V on the HDMI line seem to screw up the internal USB hub on the Lattepanda board, if I connect a monitor first and apply power to the LP board afterwards, it is very likely that devices attaced to the Lattepanda's onboard USB2 hub are no longer recognized (Arduino FTDI, the two external USB2 ports). If I boot up without a monitor connected, everything is fine.
Doing some measurements it appears the 5V on the HDMI line pulls the USB's VBUS rail to roughly 1,6V, is this enough to "reverse power" the internal USB hub and bring it into an undefined state while the rest of the circuit still has no power? If I cut the +5V line in the HDMI cable everything is fine, too!
USB3 is not affected.
Any ideas?
Thanks in advance,
Thomas
as I'm new to this forum I'd like to introduce myself, I'm an electronics engineer designing products with and around x86-based hardware, so it was just a question of time to come across the Lattepanda
In one design I've encountered a rather strange issue and I'd like to know whether anyone had the same problems.
I designed a "UPS style" addon board which contains a LiPo backup battery and the required DC/DC converters to generate the +5V power for the Lattepanda. When the whole device is in the "off" state, I disconnect the internal +5V rail by an electronic switch from the power input of the Lattepanda. The backup power board also contains an USB2.0 hub (USB2517), several PCM2900E codecs and a FTDI FT232.
Weird things start to happen when I connect an HDMI monitor to the Lattepanda and switch on my device afterwards: The +5V on the HDMI line seem to screw up the internal USB hub on the Lattepanda board, if I connect a monitor first and apply power to the LP board afterwards, it is very likely that devices attaced to the Lattepanda's onboard USB2 hub are no longer recognized (Arduino FTDI, the two external USB2 ports). If I boot up without a monitor connected, everything is fine.
Doing some measurements it appears the 5V on the HDMI line pulls the USB's VBUS rail to roughly 1,6V, is this enough to "reverse power" the internal USB hub and bring it into an undefined state while the rest of the circuit still has no power? If I cut the +5V line in the HDMI cable everything is fine, too!
USB3 is not affected.
Any ideas?
Thanks in advance,
Thomas