EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: asgard on August 26, 2014, 08:44:40 pm
-
Setup:
I have been working on a little adapter device to bridge between an RPi and a lapdock originally used with a (now obsolete) Atrix 4G Evo cell phone. There is a male micro HDMI connector and a male USB connector in a saddle configuration. The adapter, called the SweetiePie looks like this (schematic, too).
(https://farm6.staticflickr.com/5582/14826220760_0ee63e8e32_m.jpg)
(https://farm4.staticflickr.com/3860/15012860535_c84989feac_m.jpg)
(https://farm6.staticflickr.com/5513/12656871875_f4b450db11.jpg)
That's some of the setup. Additionally, I have since learned that the Atrix will not turn on the 5V USB unless the HDMI cabling shorts the DDC/CEC/GROUND to the other grounds, which is the case inside the Pi's and the BeagleBone Black. So if any of these boards are connected to the HDMI output of the SweetiePie then the 5V is turned on in the USB. The way the Atrix seems to accomplish this is to direct about 10V across a 160K pullup resistor at DDC/CEC/ground, with a current limited to approx 64 uA. When that signal is shorted to ground it is supposed to indicate to the Atrix that the phone is cradled and therefore turn on the 5V and the video on the lapdock. I have connected a hacked cable directly from the Atrix to the Pi with success, but I wanted to reduce the spaghetti, so the SweetiePie fits precisely in the Atrix cradle. My problem is that for some reason, because of the P-channel MOSFET on the high side the output voltage is not at 5V solid, as I expected, but rises only to about 4.1V for 100ms, then does a slow (capacitor-like) fall back to ground for another 150ms, which then repeats the cycle. I am wondering if I have missed something. I have given some thought of putting a beefy electrolytic (or tantalum) capacitor from the high-side drain to ground, but that would only be a guess on my part right now. I always thought an N-channel would only look like a series resistor to the voltage going through it.
J.R. Stoner
-
1) you probably want to articulate your issues in a few sentences;
2) you probably want to make your schematic simpler and easier to read;
3) you probably want to figure out how n-ch or p-ch mosfets work.
Hope it helps.
-
Following up a bit...
I put a 1uF capacitor across pins 4 and 8 of the FET, which adds some AC decoupling from the P-channel gate, so that cleaned up the gate level a bit, but the sort-of sawtooth effect on the 5V output remains.
I have done some more measurements and have discovered that the 0.8V supply feeding into the HDMI is going into the rpi, and colliding with the 5.094V measured at the USB end. The obvious thing would be to put a diode in series with the USB output 5V supply to the rpi. The USB spec says the 0.7V diode drop puts the available power at the USB end at about 4.3V which would be too little to supply the pi, but can the 0.8V from the HDMI be useful to counteract the diode drop?