Usb 1.1 -> 2.0 Passthrough Device?

[mindlord]

The 2 port hub listed above has a power regulation feature that allows full power to be delivered to any one of the two plugs as needed. It's quite a nice piece of gagetry.

 

[OrR]

Allow me to demonstrate my intentions [st]through interpretive dance[/st] with a picture.

No more bulky than the usual USB cable. Apart from the second USB lead hanging off, which could easily be removed, but I won't be doing that.

This would be an "unpowerd USB hub". IF designed correctly (no guarantee and quite a few aren't) then it will limit its outputs to 100 mA (I think, need to check the spec, but I am sure it is substantially less than the 500 mA of a normal port on a powered hub).

So with this, you may be no better off than with the USB OTG port.

Actually, the way it should work is that the hub asks for 5 power units (=500 mA) from upstream (the Pandora) and once he has got confirmation he gets them, he can then supply four of them to the attached devices (because it is assumed that he needs one himself). The distribution depends on how many devices are attached and what they need so if only one device is attached, it can get up to the whole four units which would mean 400 mA. At least that's my understanding so far, and I realize that many devices probably implement it incorrectly, too.

 

[Chip]

How much power can the Pandora actually supply on its USB port?
The protection circuit limits it to 750mA peak / 500mA continuous.

 

[Squidge]

Actually, the way it should work is that the hub asks for 5 power units (=500 mA) from upstream (the
Where did you get these "power units" from? The USB specifications states that you specify in your configuration descriptor your maximum power usage in 2mA chunks upto 500mA. So a hub would state it needed 50mA by specifying 25 in its MaxPower field for example. As devices are plugged into the hub, they state themselves how much power they require and some hubs add up those fields for internal use. A copy of the configuration descriptor is then sent to the host so it can then see which devices are using what power (along with other essentials, such as the number of interfaces).

If a device (including a hub) states a value which the host can't provide, one of two things happen: The host asks the device to disable itself, or the host shuts off power to the port. If there's more than one device on a port (via a hub) then the host can signal the hub to turn off one of its ports (or relay a message to disable a device).

Its done this way because different host interfaces implement power in different ways. Some offer 500mA per port maximum (the USB spec), and others implement the same power globally (ie, 500mA shared over X ports, as they can't supply more than that globally).

 

[Eniko]

Every time the whole USB thing comes up I get confused all over again. I wish someone who knows their stuff would write a USB Consumer's Guide or something.

 

[OrR]

Actually, the way it should work is that the hub asks for 5 power units (=500 mA) from upstream (the
Where did you get these "power units" from? The USB specifications states that you specify in your configuration descriptor your maximum power usage in 2mA chunks upto 500mA. So a hub would state it needed 50mA by specifying 25 in its MaxPower field for example. As devices are plugged into the hub, they state themselves how much power they require and some hubs add up those fields for internal use. A copy of the configuration descriptor is then sent to the host so it can then see which devices are using what power (along with other essentials, such as the number of interfaces).

If a device (including a hub) states a value which the host can't provide, one of two things happen: The host asks the device to disable itself, or the host shuts off power to the port. If there's more than one device on a port (via a hub) then the host can signal the hub to turn off one of its ports (or relay a message to disable a device).

Its done this way because different host interfaces implement power in different ways. Some offer 500mA per port maximum (the USB spec), and others implement the same power globally (ie, 500mA shared over X ports, as they can't supply more than that globally).

The USB (2.0) spec calls 100 mA a unit load in chapter 7.2.1. According to that chapter, a device can only be either 1 unit load or 5 unit loads so what I said above is wrong, specifying any power draw other than 100 mA or 500 mA seems impossible... But what you say about more detailed power drawing arrangements doesn't seem to fit with this, either?

 

[gruso]

...and here's the real thing. (sorry for the blur)

The seller deserves a plug:
CODE
http://myworld.ebay.com.au/youlovethisdeal/

 

[arrrgh]

That looks pretty cool. But is it a full-fledged hub with USB 1.1 to 2.0, or just a splitter? Guess we'll just have to wait for the Pandora to find out
edit: nevermind looked on the product page and apparently:
QUOTE

• # USB 2.0 (compatible with earlier versions) transfers data at speeds up to 40 Mbps

 

[gruso]

Although I was already 100% confident that it's the right hub for the job, I think I can now confirm it. My funny little computer seems to have USB2.0-only ports on the rear. My keyboard only works when plugged into the front. This hub allows the keyboard to plug into the rear.

 

[Chip]

Do you have the means to test if it has any sort of power limitations? You would think that if there is only one device plugged in, whatever power the host can supply (minus a negligible amount for the hub chip) should be available to that device.

I'm just wondering if they have some sort of regulation circuitry that only sends half the power to each port even if there's nothing plugged in to one of them. It seems highly unlikely they would have designed it like that, but if you can test it, we will know for sure.

If you have a USB-powered 2.5" HDD to try, that will definitely take the full 500mA (and probably a bit more).

 

[arrrgh]

Gruso, you seem pretty knowledgeable about the Pandora's USB, will this:
CODE
http://cgi.ebay.com.au/USB-A-Female-to-Mini-USB-5-Pin-Male-Adapter-Converter_W0QQitemZ270366510391QQihZ017QQcategoryZ163924QQtcZphotoQQcmdZViewItem

work with the Pandora's OTG port? I'm pretty sure that it will, just wanted confirmation.

 

[gruso]

I can't know for sure, but I don't think it will. OTG adapters are very specific things, which apparently tie one pin down to earth so the device knows to switch into host mode. If the adapter doesn't have OTG in the name, it probably won't have this function.

Here's what you're looking for:
CODE

http://shop.ebay.com/?_from=R40&_trksid=p3907.m38.l1313&_nkw=usb+otg+host+cable&_sacat=See-All-Categories
http://www.amazon.co.uk/s/ref=nb_ss_ce?url=search-alias%3Delectronics&field-keywords=OTG+Cable+Adapter&x=3&y=23



Chip: I'll look into it.

 

[marshal]

I just got my Griffin USB hub today (delivered this morning). I just couldn't hold off ordering one of those

It small and seem to be constructed sturdily. Also the black finish should look great with the pandora too.
The only negative point is that is says "Macbook air and other notebook" at the back... they should have leave the Macbook air out and put Pandora instead

Other than that I connected it to my computer and it get recognised as a usb hub. I don't have a 2.5' HDD or anything that use that much power so I don't see how I could test it further really.

 

[arrrgh]

Found this:
CODE
http://cgi.ebay.com/USB-A-Female-to-Mini-B-OTG-male-adaptor-Host-Mode_W0QQitemZ350118250585QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item350118250585&_trksid=p3286.m20.l1116
with the search you suggested - looks really cool, no cables or anything.

 

[gruso]

^ Looks the goods. He even mentions the pin shorting in the description.

I've given the hub a little run. I don't have a HDD on hand (I have a 2.5" enclosure with nothing in it, unfortunately) but I've got a digicam, Huawei E220, and wifi stick. I tried combinations of these on the hub and everything seemed to run smoothly. I wasn't doing anything too intensive, just things like searching for a network while opening photos. It seemed very normal and hubby.

I'll see if I can dig up a 2.5" HDD this week and give it a good hammering.

 

[pelrun]

Arrggh, the first adapter you posted will definitely work - if you look closely at the mini-usb plug you'll see that it's slightly differently shaped to a mini-b (and has a white plastic insert); it's a mini-A, which will have the id pin grounded.

All OTG cables should have one end set as master like this; it's purely an initial suggestion when connecting two OTG devices together, as they can then negotiate to switch host-client roles.

However, I am a little wary of these small adapters (having built my own in the past) as they give plenty of leverage to rip the socket from the device if you aren't careful. So I'll stick with those that have at least a short cable.

 

[Squidge]

The USB (2.0) spec calls 100 mA a unit load in chapter 7.2.1.

7.2.1 says the following "The power source and sink requirements of different device classes can be simplified with the introduction of the concept of a unit load. A unit load is defined to be 100 mA."

Note the world "simplified".

Table 9.10 in the USB 2.0 specification specifies a bMaxPower field "Maximum power consumption of the USB device from the bus in this specific configuration when the device is fully operational. expressed in 2 mA units (i.e., 50 = 100 mA).".

The USB 3.0 specification no longer specifies the simplified unit, and also states that the standard supports upto 900mA (in 8mA units) for SuperSpeed devices.

 

[OrR]

Ok, thanks for the explanation... This stuff is so bloody confusing...

Arrggh, the first adapter you posted will definitely work - if you look closely at the mini-usb plug you'll see that it's slightly differently shaped to a mini-b (and has a white plastic insert); it's a mini-A, which will have the id pin grounded.

Yeah, the picture seems to be mini-A but the decription says mini-B... On the other hand the second one with OTG in the title has mini-B in picture and descriptrion. IMO, only a mini-A plug makes sense but as both adapters do not seem to be allowed in the (now deprecated) standardization of mini-connectors, I guess it doesn't matter much... In the end, both should fit mechanically, and it comes down to the how the ID-pin is wired. For host mode, the ID-pin (4) should be connected to ground (5) with a resistance of less than 10 Ohm (shorted). For device mode, it should not be connected to ground (or connected with a resistance greater than 100 kOhm).

 

[pelrun]

Actually, I'm not surprised that the sellers say mini-B by mistake; they neither make the adapters nor use them, and so they're not likely to bother understanding the difference between the two types (even if it is critical). Or they expect people to make the same mistake and they want to show up in the search regardless.

You won't find professionally manufactured mini-B adapters with the ID-pin mod; they *will* all be mini-A. If you get an actual mini-B, it's the wrong cable entirely. I take that back, some of those cables sure do seem to be using hacked mini-B's. Probably for the same reason I came up with the mini-B hack in the first place - you can't find the bloody plugs!