Parallella Community

I'm wondering if I could get away with supplying single-cell Li-ion voltage (~ 3.6V) to Parallella instead of 5V. From a cursory review of the schematics, it looks like

1) switchers will run slightly more efficiently,
2) 3.3V LDO dropout is < 100 mV @ 300 mA, so that looks ok,
3) HDMI won't work,
4) USB host may not work--looks like it depends who is supplying VBUS?

Given those caveats, anyone see other problems with this?

Mark

Interesting idea and also curious as to whether this would be feasible. Could be nice for certain embedded/headless applications.

Cheers,

Andrew

AFAICT it should work. That's what I was planning on doing myself

I can actually give it a shot on the rev0 board.

Hmmm... looks like the I2C pins on the expansion connector may not work, as they run through the 3.3V <-> 5V level translator. Why is that anyway? Pull-ups to 5V are going to disqualify a lot of potential I2C peripherals. Oh well, I probably don't need I2C for my project.

The Intersil I2C interface needs 5.0V signaling based on datasheet and Intersil AE feedback. Will now check again whether this is hard requirement for robust operation.

Does anyone else have inputs regarding this issue. Were folks assuming that the I2C interface on the PEC_GPIO connector was 3.3V? Immediate feedback appreciated!

Andreas

I definitely found weird that the I2C was 5v (it's on PEC_POWER btw, not PEC_GPIO . Actually having any 5V is a bit weird on such an advanced board.

But it's not a big deal, level conversion of I2C is pretty straightforward ...

And where did you see the ISL requires 5V ? AFAICT it requires whatever is the maximum voltage is fed to the inputs, so if you feed it 3.7V, it should be happy with pullups to 3.7V.

Cheers,

Sylvain

Aha, in my haste I hadn't even noticed the 5V I2C running to the Intersil PMIC. Hmmm. The TPD12S016 datasheet does not specify whether we will get any I2C signals on the "5V" side when VCC_5V is out of spec (4.5 to 5.5 volts). Worst case, it just goes dead and blocks communication with the PMIC, and we would need some hand rework (feasible thanks to the pullup resistors?), which would mod the board to battery supply but disallow 5V input.

As for the expansion conntector... I did assume any I2C available on the connectors would be 3.3V or 1.8V, or really anything but 5V. But, it's not that big a deal.

The ISL9305 datasheet says, "When the PMIC and host run at different voltages, the I2C pullups should go to the higher rail." Since the input is specified down to 2.3V for the switchers, I would interpret this to mean that I2C pullup to 3.7V is fine when that is the input voltage to the switchers/LDOs. If there is any problem it's likely to be at the HDMI level-converter chip, which could cease to function when VCC_5V = 3.7V. Hopefully not.

please have in mind, that only a fully charged barrery provides 3.7V. So voltage will drop while feeding the board.
IF your voltage limit is too strict, the battery would be too early declared "unusuable and empty" while still holding some charges inside, wasting capacity.

Actually a fully charged one will be more like 4.2V.
If you have a decent sized lipo ( like > 2000mAh ), you'd be taking current at maximum 1C, and if you look at the voltage curve, with a 3.6V cutoff (to have the 3.3v rail), you'd be able to use > 90% of the battery capacity, which isn't bad.

Cheers,

Sylvain