eSATAp over adapter?

[sulu]

I couldn't find any info on that. Most posts mentioning "eSATAp" are from 2014, which I think is older than the concept of having an adapter.
So does the eSATA adapter forward the 5V lanes of the USB port?

 

[Letalis Sonus]

Given that the whole reason for this adapter solution was the actual lack of eSATAp ports available for purchase that were not hitting EoL, I'd guess that you'll only get a normal eSATA port that can't provide any power supply lines.

 

[sulu]

I thought the reason for not having a standard eSATA jack on board was that there were none available that would fit within the limited internal space constraints. An adapter wouldn't have that problem, so the variety of possible jacks would increase. Among these additional options there COULD have been some with eSATAp.
[doublepost=1478167710,1478167639][/doublepost]I thought the reason for not having a standard eSATA jack on board was that there were none available that would fit within the limited internal space constraints. An adapter wouldn't have that problem, so the variety of possible jacks would increase. Among these additional options there COULD have been some with eSATAp.

 

[Letalis Sonus]

There already was a fitting PCB prototype with an eSATAp port: https://pyra-handheld.com/boards/threads/a-bunch-of-pictures.74706/ https://pyra-handheld.com/boards/threads/populated-mainboards.74728/

The front of the port isn't shown, but the amount of contacts on the bare PCB and the width of that one USB port makes it rather obvious that it's more than just USB 3.0.

This thread contains the announcement that it's being replaced:

Sadly, the USB/SATA-Combo port is discontinued and not available anymore.

Here is a picture of the adapter, looks pretty much like a normal eSATA port.

 

[sulu]

The port on the Pyra's PCB for sure supports eSATAp. Essentially it's just forwarding the 5V lanes of the USB port to the SATA connector.
My question is, if the adapter [1] forwards these lanes too.

[1] https://www.pyra-handheld.com/wiki/index.php?title=Peripherals

 

[Letalis Sonus]

The adapter redirects the 5V to USB, obviously. The SATA connection only uses the additional twisted pair data lines from the USB 3.0 port, leaving the USB 2.0 part fully available, which is why the adapter exposes it as well. Just use an USB SATA power cable and you're good to go, eSATAp was rarely used anyway.

 

[sulu]

The whole point of eSATAp is not to need a 2nd cable.

 

[levi]

I'm not sure the 5V from USB will be able to provide enough current though. Normal internal SATA gives you some 4.5A at three different voltages, and I'm not sure the Pyra will be able to provide that (even without a battery, it's still only getting some 2.1A at 5V in from the power port). 2.5in drives may well run okay off less than 2A, but that limits your available storage quite dramatically.

 

[sulu]

I'm only interested in 2.5" HDDs. These run on 5V@0.5A. Theoretically that should just work from a single USB 2.0 port within specs.
But the current required for starting the HDD is usually a little higher (e.g. 0.6A) and there's also the controller of the external enclosure that also needs some power. So bus-powered 2.5" HDDs usually run slightly out of USB 2.0 specs. It mostly works anyway though.

I'd have confidence in ED and Nikolaus to design the Pyra in a way that it can handle this kind of overcurrent if they provide an eSATAp adapter.

 

[levi]

Letalus Sonus's picture suggests the adaptor will provide eSATA and USB2.0. You can get 5V a some milliamps through the USB2 connection, but it doesn't look like you're going to get eSATAp as such.

 

[lukey]

@EvilDragon *poke*

 

[Pocak]

When running on battery, the charger chip (if it's the bq24296) can provide 1.5A to VUSB. And if you connect a charger, I think that will directly power the other USB ports.

 

[Letalis Sonus]

When running on battery, the charger chip (if it's the bq24296) can provide 1.5A to VUSB.

It has to feed 3 USB ports, though. That's a worst-case load of 1.9A (2.3A if you already tread the SATA one as USB 3.0)

 

[levi]

The one you plug the eSATA adaptor into isn't USB3, despite the colour of the port. A future CPU board may make it USB3, but you'll probably lose eSATA then. Before then, the microUSB port is USB3.

 

[Letalis Sonus]

The one you plug the eSATA adaptor into isn't USB3, despite the colour of the port

But it was designed to be able to work as one with a different CPU board and therefore must still be able to deliver 900mA, the charger chip that provides it sits on the main PCB AFAIK.

 

[levi]

Yeah, but it doesn't talk USB3 power protocols, so you can't ask it for more than 500mA. Although maybe it won't bother with that stuff, and just give you 5V at whatever you want until it can't.

 

[Letalis Sonus]

Yeah, but it doesn't talk USB3 power protocols, so you can't ask it for more than 500mA.

It is able to, this is really just a limitation of the SoC. A future CPU board with more than one USB 3.0 host controller can use it as a fully featured USB 3.0 port - as long as the power supply is not handled by the CPU board, it *has* to be able to provide the full 900mA to do that.

 

[levi]

Yeah, sorry. I should have said that there I was talking about the current OMAP5 CPU board connected to the current main board.

I don't know exactly how 5V power is routed to the USB ports, but I would expect the traces on the main board to be able to cope with 900mA without burning out, so provided whatever powers it is able to put that out, it should be okay when we get a future CPU board with two USB3 interfaces.

 

[Pocak]

So, the schematics are out, time to dig out this topic!
It turns out the USB host ports are powered by a TPS2505, a 1.1A boost converter. The chip has the capability to limit current per port, but the chosen resistor value is such that there's essentially no limiting; either port can sink all the boost converter's output, which can be quite a bit more than 1.1A. Should be plenty for a 2.5" HDD, even with a USB trinket also plugged in.

BTW, there's no direct path from either charging port to the host ports, contrary to what I wrote above.

As for a hypothetical upgrade to USB3, there's clearly not enough juice, on paper. But it's likely to work anyway. I guess the OS could be set up to not give a new USB device permission to increase power use if that would overstep the TPS2505's limit, but I think there's no need to be so anal about it. Let the user figure out that plugging two power guzzling gadgets into a handheld is a bad idea, especially when the battery voltage is already low (which is when the boost converter is going to have trouble providing 1.4A). Ideally, the OS would pop up a message if the ports are overloaded (and not just hide it in dmesg). There's also the possibility that a new CPU board would come with a new main board with a stronger chip, and a note that the new CPU board mated with an old main board can't meet USB spec.