USB hub SNAFUs

Hi,

I was connecting another 4 port USB hub to a machine tonight "identical" to one already in service, there. Once in place (and operated in the same fashion as its sibling predecessor), it failed to operate as expected! (i.e., PC wouldn't enumerate any of the drives attached).

Suspecting the hub might be bad, I tested it elsewhere. Replaced cables. Added external power source. etc. Works like a champ on another machine (with different devices dangling off of it).

[remember, "old" hub still works fine on the machine of interest]

Tried a *second* hub (which is actually the *third* of this type being discussed) with similar absence of joy.

After enumerating the behaviors of the three different hubs (one "good", two "bad" -- yet all *functional*), I took a peek inside. Of course, different chipset in the good vs. bad ones.

Aside from an intellectual curiosity, this is not a real issue for me *personally*. I'll just pick up another couple of hubs.

But, I had been debating what to use for external interconnects on a high volume project. USB (host) *seemed* the natural choice. It's always "just worked" (for me) in the past!

In light of this little experience, I started thinking about how I would respond to a customer making the claim that *my* device "wasn't working" -- when he attached (this particular) hub/device. Yet, *swearing* to me that these things worked properly "elsewhere".

I've been giving serious thought to using ZigBee/BT for external device connections as it offers several other advantages over USB (no wire tangles, longer "virtual" wires, no cables to keep track of, no connectors poking through the case, no galvanic paths back *into* the device, no concerns about supplying "enough" power to the external devices, etc.). Of course, the ubiquity of USB and the huge cost advantage have (previously) been weighing against the wireless approach. :<

So, for folks with USB-capable (host/slave) devices, what have your experiences been:

- problems as a host NOT recognizing some bit of COTS kit dangling off your port (that the user thinks you *should* be able to recognize -- e.g., "generic hub")

- problems as a slave NOT being recognized *through* some particular combination of fabric to the host And, how do you address these problems when interacting with the customer? How happy is he/she with your remedy ("buy a different hub", "buy a different disk drive/printer/smokeshifter", "plug in the cable BEFORE applying power", "plug in the cable AFTER applying power", etc.)

(I.e., the wireless solution may win because it is LESS ubiquitous... less likely to be "expected" to talk to "anything with the correct connector onboard")

Thx!

Reply to
Don Y
Loading thread data ...

Sounds like they didn't get the firmware right on the new chip type. I'd say that even though it works "elsewhere", it MUST have a spec conformance failure to not function as the spec demands.

What brand is it? I want to be sure to avoid them. Especially if you let them know about it, and they fail to fix the problem.

Swap out the old one for it and see if it works "in place of" the original. If it doesn't, you have a legal case, but depending on the maker, that may not matter. A bigger maker will want to make it right, at the product level at least. A chinese shit shop wont.

There should be a web site somewhere where we list bad designs and their makers.

I hate going to Fry's and having to buy things I have no example of the working of.

Reply to
Archimedes' Lever

I've had some USB memory sticks that certain hubs didn't want to enumerate. I never did figure out why, although given how heavily cost-minimized USB memory sticks are, I wouldn't be surprised to find it was something in the memory stick and not the hub.

I've also been told that some memory sticks don't present themselves as just a simple USB device -- they act as a hub themselves with a memory stick and various other peripherals connected (e.g., a DRM device). You often see people with these sorts of memory sticks having problems when they try to playback MP3 files on them through a car stereo -- there are apparently some car stereos out there that, while they are a USB host, just don't have the software in place to be able to see "through" a hub. (...which I can kinda understand, given that supporting a hub does take a fair amount of software, and we're often talking just little

8-bit CPUs in some car stereos).

Bluetooth isn't really any panacea either -- there are a LOT of compatibility problems out there between various BT adapters and specific phones, due to the software of one or the other not being quite right.

---Joel

Reply to
Joel Koltner

Are these cases where *you* implemented the host and (later) encountered sticks/hubs that "you" couldn't enumerate? Or, cases where you plugged a stick/hub into some other host (e.g., PC) and encountered problems? [The remedies are obviously different :> ]

Yes -- though (many) hubs are almost as cost-minimized!

In this particular case, I know the devices (USB drives) are recognized by the PC. *Other* hubs allow the devices to be enumerated properly. The "new" hub is recognized (as a "Generic USB hub"). *Other* devices enumerate properly in the hub. But, these *particular* devices won't enumerate behind this particular hub.

When I get a chance, I will systematically try different devices and hubs to see if this is a common problem (I've not encountered it before). Then, for those that don't seem right, I'll try a couple of other OS's to see if it's something that can be handled with a fixup table in the OS (or if it is a genuine "defect" in the hub/device). The "next" hub that I pulled out of the hub box worked fine (as had the one used prior to the "new"/bad one). I'll have to peek inside any that "don't work" and those that "do work" and see if there are any common aspects (perhaps a particular controller with "issues"?)

Yes. The same holds true of "N-in-1" memory card readers, etc. But, in each prior case, the physical device was at least

*acknowledged* in the enumeration. These devices look like they aren't even *seen* (by the hub *or* the PC).

I'll use other OS's with more verbose reporting to get a better look at what is actually happening when the devices are plugged, etc.

Understood. A "justifiable" design decision -- though perhaps not obvious to the end user (who just sees a USB connector and wonders why he can't do what he does *elsewhere* that connector presents!).

One crude analogy that probably users would easily relate to: adding a cube-tap on the end of an extension cord. I.e., to them, the hub just looks like a "connector multiplier" -- much the same as a cube tap represents an "outlet multiplier" for the extension cord.

They've got a bedside lamp plugged into the extension cord and it's worked fine. They plug in a clock radio and *it* works fine, as well. They plug in the cube tap (so they can use the clock radio and lamp at the same time) and the lamp no longer works -- but, the clock radio *does*! Suspecting a bad outlet in the cube tap, they swap the connections of lamp and radio (expecting the *radio* to now fail) and the radio *still* works (using the cube tap outlet that had failed to light the lamp) AND the lamp still doesn't work (using the outlet that had previously powered the radio)!

If someone were to describe this behavior to you (the manufacturer of the radio, clock or extension cord), you would be tempted to "conclude" that the user was "confused" :> Surely that can't be an accurate description of the situation!! (?)

Agreed. The "advantage" BT (and ZB even moreso) has is that it

*isn't* as ubiquitous. People don't expect to plug "anything" into (e.g.) their phone (via BT). There are just far fewer BT "offerings" to entice/corrupt the user.

OTOH, people see a USB connector and they figure they can plug anything into it "just like on their PC" -- begrudgingly acknowledging that they might need to find a *driver* for the device... but, the system will *see* the device and tell them about this missing driver (whereas this appears to be a case of the system/PC not even *seeing* the devices behind the hub).

Reply to
Don Y

I had a PNY Attache that did not work in a Intel board, worked fine in any other hub. But an older model of the Attache worked fine in every hub I tried.

Cheers

Reply to
Martin Riddle

I'd suggest D-Link or Belkin. Not that these people are brilliant, but they tend to have something that resembles customer service. Generally D-link makes good stuff.

One painful idea would be to get a small industrial SBC that has sufficient USB connectors and roll your own hub. I've seen single board computers with 14 usb ports.

Reply to
miso

That's why Belkin had a class action lawsuit against them for one of their wireless routers.

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

I'm not concerned with solving "my" hub problem -- I simply dug another out of a box and replaced the "problem" hubs.

What concerns me as a developer is how a potential customer might find himself in a similar situation "blaming" my product for a problem that is "beyond my control" -- a bad hub design or "external device".

I will examine the hub, other hubs and other devices to see to where this "fault" can be attributed. Likewise, the host OS (hard to imagine MS wouldn't have stumbled across a problem like this and installed a fixup workaround *if* the problem can be handled "in software").

As I said, I am not worried about my personal needs. I put a couple of 7 port hubs on the machine in addition to the four port hubs there already. It's main appeal to me is as a test bed on which I can get repeatable failures in the USB "field" subsystem. (I will have to check the problem hubs on other machines as well)

Reply to
Don Y

A few years ago we bought a $10,000 signal generator. It came with a 128MB USB stick. This, I assume, is because this stick actually worked while a random stick lying around might not have. Of course, the stick got 'borrowed' when someone needed one for something else and we now have to play pot luck like everyone else (or we would do if anyone actually used the USB connection).

One solution to both these problems would be to simply change the connector for one that isn't standard. No problems with users trying incompatible parts, and no standard parts 'wandering'. But as a user I'm not exactly thrilled with this idea...

Theo

Reply to
Theo Markettos

Or, just a "convenience issue". Sort of like giving gifts to young children complete *with* batteries (so they don't start nagging their parents immediately to go BUY those batteries!)

That's the appeal of the wireless solution.

Nowadays, it would take very little time for to publish a description of how to build an "adapter cable" to work around the proprietary connector. Then, you're spending customer service dollars troubleshooting external devices that were never intended to mate to yours (since a wise customer would carefully withhold the "little detail" that he now has a 4TB USB drive plugged in where the 128MB memory stick was intended to be placed -- since admitting this "hack" would undoubtedly immediately be met with The Party Line of "We don't support other devices. Use of such devices may void your warranty." yadda yadda yadda)

If you can obtain the "peripherals" that you can "reasonably" be expected to want to connect to the device in question (and, at "reasonable" prices), the biggest hassle I see is the risk of

*losing* that peripheral (your case). Similarly, damaging it.

Or, trying to support the device beyond it's published end-of-life (I keep an XT keyboard around for an old logic analyzer that is not fond of AT keyboards!)

The complexity of something like a wireless link sets the bar considerably higher for "unapproved attachments".

I'm just disappointed that USB isn't as universally *implemented* as its ubiquity would suggest...

Reply to
Don Y

Paraphrasing Jamie Zawinsky: "You have an USB problem. You are trying to solve it by using Bluetooth. Now you have two problems."

Reply to
Przemek Klosowski

Your (her) count would be off by a minimum of one as well. You forget the user. That's three problems in this instance.

Reply to
Klinger

The typical device these days is a micro SDHC. Typical, at least as it relates to modern design for thing like cameras and MP3 players, and Oscopes, etc.

They fit into a tiny slot, and are usually left in place. The only reason to remove it is to do a snail transfer to it, and then replace it again. One can also make a holder with an eyelet and a lanyard, keeping it attached to the product it was made for.

Reply to
Klinger

I had a couple of Belkin USB hubs that were flaky as hell, well, until they just went dead. Belkin is over-priced crap, as far as I'm concerned. Might just as well buy non-name Chinese junk.

Reply to
krw

You are. Belkin slaps their name on it. They made their name selling overpriced printer & serial cables with gold flashed connector housings. I have a couple barrels full of them to recycle.

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

But Beklin, Dlink, and probably Linksys is just Chinese ODM. Linksys probably does their own coding.

I'm not a big Belkin fan. I did like their insane lifetime replacement. However, I haven't had good luck with any of their high tech stuff. Even USB to 232 converters. Dlink on the other hand makes stuff that works, but support doesn't last very long.

Reply to
miso

Linksys is owned by Cisco, you stupid twit.

Reply to
WoolyBully

So English isn't your first language then.

Reply to
Pomegranate Bastard

That's not anything like so bad as trying to return something at Fry's.

I once bought a printer and a box of floppies at Fry's. Both were bad, and both had been previously sold and returned by someone else. Fry's just kept resealing the boxes and reselling the stuff. Even then, it was hell going through the return process.

--
John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   
 Click to see the full signature
Reply to
John Larkin

Indeed. Apparently this was quite institutionalized as well. Wikipedia says: "In 1997, Forbes reported on a series of issues about Fry's customer service and somewhat unorthodox business practices. Among the allegations was that the company had an internal policy, identified as "the double H" or "hoops and hurdles", to delay or prevent customers from obtaining refunds.[6]" --> From this article:

formatting link

Reply to
Joel Koltner

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.