[sdiy] USB power (slightly off-topic)

[brianw]

The USB Specification is for 0.5 A maximum.

Unless you have everything following the newer specifications, on both Host and Device sides, you can't reliably go beyond 0.5 A per cable. This SSDs should have a non-USB power input to handle their 1.7 A needs.

Apple might have been the first to go beyond 0.5 A. Their chargers stick to the 0.5 A limit unless they detect the special circuits in Apple devices that guarantee extra current won't damage things.

The USB Spec eventually settled on ways to go beyond 0.5 A, but if you're designing hardware then you should become familiar with these new rules and verify that the SSDs are playing my the rules. You should be able to get the USB Device Descriptors to verify the maximum current that they advertise.

While Powered USB Hubs are "supposed" to shut down beyond 0.5 A, it's really kinda expensive to design circuits that will reliably shut down at precisely 0.5 A - not less, not more. So, what happens is that lots of computers, laptops, and hubs all seem to be capable of going beyond 0.5 A when they're really not supposed to.

My point is that you can't just strong-arm this. Although the USB market has a lot of cheap products that play outside the rules, you're still risking damage to anything and everything if you step outside the guidelines.

By the way, USB.org has all of these specifications available for free, and even some special articles on subjects like this where there have already been lots of problems in the face of customer desire to get beyond the limits.

Brian Willoughby


On Nov 8, 2022, at 11:28 AM, Rutger Vlek via Synth-diy <synth-diy at synth-diy.org> wrote:
> Dear list,
> 
> I hope you don't mind me tapping into our collective list wisdom for something slightly non-synth related. The case is as follows:
> 
> I'm building a NAS at home from a Raspberry Pi 4B and a couple of SSDs. The SSDs are connected to the Pi via a SATA-to-USB converter. They work, but stop working when used intensively. Not to my surprise, because the power requirement of the drive is 1.7A at 5V each. The Pi USB ports can only deliver 1.2A at most (and shared between 2 ports). So I need a more beefy supply. My initial idea was to buy a couple of these USB Y-cables that have two USB-A connectors on one side, allowing the user to provide the paralleled power of two USB ports. However, even on a single port, the Pi already reaches it's limits. So I thought, why not plug the second USB-A plug of the Y-cable into a USB-charge, so that it can provide all the current needed by the drives. However, I have a bad feeling about this...
> 
> I don't have a clue about the internals of a typical USB Y-cable, and have not been able to find any datasheets on them. In the worst case, power from the Pi's USB port would just be connected directly to the power output of the USB-charger, and that does not sound like a good idea. They might not be at the exact same voltage, and I since both of them are SMPS's, I don't know if they'll somehow start interfering with each other. If only there would be some series diodes internally present in the Y-cable, but I have no clue if there are... Does anyone know (or can measure)?
> 
> Plan B was to buy a powered USB hub, but I struggle to find one that can deliver the required 1.7A at 5V on multiple of its ports at the same time.
> 
> Plan C, that just came to mind, is that I might be able to find a beefy USB PSU with multiple ports, that I can use to feed the Pi from, as well as the drives via Y-cable. The Y-cable would then be combining the power from Pi's USB output with the direct output of the same PSU. Would that work?
> 
> Plan D would be to make my own power distribution board, where USB connectors pass data between the Pi and the disks (but no power), and power is injected from a DC barrel plug (and some filtering)
> 
> What do you think would be the best approach?
> 
> Rutger