It’s been an age since I last had the time to write up a random post. Normally, if I’m feeling well, I’d be happy to churn one out every fortnight or thereabouts, but lately, I just didn’t quite have the energy or motivation to do much at all. Of course, the clutter eventually needs to be “cleaned” … and so a random post is born.
Concession Opal Card
One of my biggest achievements in the past fortnight was to obtain a concession Opal card. Just before I left for Hong Kong, the Opal network had been (finally) completed on 1st December 2014, ahead of time (or many years too late, as I reckon). As part of the rollout, they had phased in the introduction of different classes of cards – those for tertiary concessions came last.
At the present moment, the scheme is only open to select participating institutions on a data sharing basis. Information about a student, and their eligibility, will be shared from the institution with Transport for New South Wales. UNSW was one of the first institutions to jump on board, despite protests about privacy. In reality, the NSW Government wasn’t going to back down, and students might not have a paper ticket choice after the end of this year so it is inevitable that we’d have to pick convenience over privacy.
UNSW used every avenue to alert students of the Concession Opal’s availability, including their Facebook feeds. The application process formally opened 2nd February, and involves students logging into myUNSW online, reading through a long document and granting permission for their information to be shared through the online form.
After granting permission, if a user revisits the page, the end text is changed to allow students to withdraw their permission, however this makes them immediately ineligible for the concession Opal card, and appears to de-activate any linked concession Opal cards.
The process takes two days between granting permission, before returning to Opal.com.au to continue the process.
To sign up, users simply need to click the Concession card under “Order an Opal card”. A pop-up appears to inform users of the participating institutions and dates – this has now since been updated since I applied:
Users need to click “Continue ordering an Opal card” and follow the instructions, filling in all personal details and ensuring that the details entered match the details shared with TfNSW. There is a step where your details are checked with the database, and if they are not matched, you will not be able to order a Concession Opal! You will also need to add some value to the card during the sign-up process.
Once the order is done, you will need to wait several days for your card to come in the post.
When you get the card, it will be on a form letter with some pertinent card details printed on the form. There are some identifying details on the card itself too. The card will need to be activated by logging into your Opal account and entering the four digits after the lock symbol on your card.
Aside from the card, you also get a pamphlet which explains the Opal ticketing system, how to use the card and the associated benefits. Thankfully, for the Concession holders, the Daily Travel Cap is reduced to $7.50 and the Weekly Travel Cap is reduced to $30 to be fair. $2.50 Sundays also applies.
The back of the pamphlet also has a feedback form, with a $15 Opal credit enticement should you be chosen for participation in the survey and an advertisement for their Opal branded accessories (smartphone covers, wallets, etc).
It’s as simple as that, and now I finally have an Opal card. What took so long, huh?
The nice guys from element14 had given me a little “Christmas Bonus” in the form of an Arduino Uno, Arduino Yun and an Infineon RGB LED Shield. They arrived just after I got back from Hong Kong/China, but I didn’t have much of a chance to play with them or review them until just recently, when I saw them sitting on my shelf and felt sorry that they were still in their plastic bags.
As a result, I went on a little blogging marathon, producing three parts so far:
- Internet of Holiday Lights Off-RoadTest – Part 1: Unboxing
- Internet of Holiday Lights Off-RoadTest – Part 2: Hardware Preparation & First Light
- Internet of Holiday Lights Off-RoadTest – Part 3: Current Waveform, Optimizing LEDs & I2C Commands
It’s expected that there will be several more parts to come, as I explore the Arduino Yun and try to make it all nice and “internet connected”, but at least for now, the basics work.
I don’t know what it was, but this past month seems to have been a bit of a storage failure filled month.
USB 3.0 External SSD
I normally use a USB 3.0 connected SSD as my main storage drive for my university work, but it seemed that it struggled to stay connected. Assuming the contact was poor, I unplugged it and replugged it and it was completely dead.
Further examination shows that the cause of failure was nothing related to the SSD, and everything related to the cheap enclosure I had used. You may have guessed that the connector had simply broken its solder joints, but you’d be wrong.
In fact, the connector itself was made of such brittle plastic, probably dimensionally incorrect to boot, that it cracked and failed to support the pins carrying power to the drive. As a result, the case is pretty much bin material (no point replacing a USB 3.0 connector, surface mount and all).
Above are two macro photos of the connector tongue from two different angles. For contrast, here is a connector from another product that seems perfectly happy.
LaCie Porsche Design 16Gb USB 3.0 Key
I had problems with this key before, showing up as an Innostor device with no storage and once corrupting the root directory entirely. After a few formats, it seemed to behave itself well enough for a while that I continued to use it for non-critical tasks.
Struggling to find a spare USB key to do a Windows 7 install, I grabbed this key and made an install from the Windows 7 USB Download Tool, which all went well right until I was installing it when it complained that there was a failure unpacking the files. I immediately suspected the machine had a hardware failure … but alas, that wasn’t the problem.
The USB key was at fault. I tested it after removing all files and formatting with SDFormatter.
I doubt it was SDFormatter’s fault, but I tried formatting it using native Windows. Alas, no dice.
Of interest was that the failure involved a fixed number of sectors at the same locations. This is unexpected, as any flash failure is expected to take out cells, which would be moved around as the wear levelling and flash-translation algorithms remap the logical sectors around the pool of workable physical sectors.
The locations of the errors were not at the rear, nor were they at the front of the logical space. In fact, they were scattered throughout the first ~5Gb of space. No errors occurred after that. I knew what to do to preserve the key, because otherwise it would be bin material – leave the first 5Gb unpartitioned.
Believe it or not, that configuration passed multiple tests … even after odd writes, it didn’t seem any wear levelling reallocations moved the bad areas into the partition. This is particularly interesting because it implies there isn’t an effective wear levelling algorithm on this controller. Strange.
HP Microserver VB250EAVER Hard Drive Failure
Who still runs 250Gb hard drives anymore, I hear you ask? Well, since my HP Microservers came with the VB250EAVER (Seagate OEM) hard drives, I had decided to make the most of them and use them. The first drive developed 11 bad sectors in 2 years, and got binned. The second drive (as I bought two Microservers) had developed 5 bad sectors in just under three years.
I couldn’t stand the impending threat of failure, so I decided to migrate the server drive over to a spare 2Tb “green” drive. The drive had to be 2Tb or less, as the server doesn’t understand UEFI/GPT.
Being a bit short on machines since they were all busy, I attempted a daring challenge – use the other HP Microserver, running Windows 7, to clone the drives across while the system was still hot as it was busy with some experiments. Despite the hard drive racks stating clearly that they are Non-Hot-Swap, I had confidence it could work.
The first step was to slot the drives in and click them in. Because of the design of the SATA power and data connectors, no damage is likely to occur to the drives. The drives spun up immediately but did not appear on the computer just yet. I had to fire up Device Manager and click on the Computer and Scan for Hardware Changes. Bingo!
Once the drives turned up, I got out WinHex and cloned the source drive to the destination drive, sector for sector. That went uneventfully.
Finally, I had to unmount the drives – heading into Device Manager, deleting the drives from the tree and removing them didn’t cause the system to freeze – so that worked fine. I re-inserted the clone, rescanned the system for hardware changes so that its partition table would be redetected. Then I headed to Disk Management, and since the drive was not the boot drive, it let me resize the partition without any fuss. I then removed the drive from Device Manager and unplugged it.
Whacking it back into the other Microserver, it booted as if nothing had changed. That’s what I call success – and I didn’t even have to shut down or reboot the other busy Microserver to do the rescue. It was a challenge but it definitely saved me from getting out my live Linux USB, a monitor and a keyboard (since my Microservers run headless, VNC controlled). Glad to know it’s possible!
On the way to a family dinner, I swung by my Aunt’s place because her internet wasn’t working. I had checked online, and her modem was dropping in and out every few minutes and eventually stopped connecting altogether. My Dad had prior tried to diagnose it, saw some modem diagnostic screens, and connected to the wireless network as well, which seems to imply the modem was okay. When asked whether the line crackled, it was reported that there wasn’t any crackling. That seemed strange. It had to be one or the other, preferably the modem (because we can fix that).
I arrived and checked out the line. It’s about as quiet as one can expect. I checked out the modem – a spare 2wire Telstra Bigpond Home Gateway I had been given ages ago. Its power LED was red, and it didn’t broadcast any Wi-Fi anymore. I gave it a reboot, and its internal relays clicked, but the boot-up never completed. Eventually the relays would click again.
I bought along my trusty Agilent U1241B, and disassembled the modem. I suspected the 12v switching PSU had failed (as the 2wire ones often do). I measured the voltage under the load of the operating modem, and for a 12v supply, it measured a stable 12.34v which didn’t waver even with the clicks of the relays. The boot-up was never completed and it sat there looping.
Being convinced it was the modem and not the power supply, I set it to one side and replaced it with a different D-Link spare I had lying about as a diagnostic modem. It connected and synced at 13Mbit/s once given the right authentication details, and configured correctly for their requirements.
Upon closer inspection, the 2wire modem had two bulging capacitors. If it were only for that, this wouldn’t be a very noteworthy post. The problem was that the two bulging capacitors were both solid polymer style capacitors – the ones which are often installed in modern motherboards, touted to be miles more reliable than the aluminium electrolytics which used to dominate those products. It took me by complete surprise, as I expected they would never fail – but I suppose collateral damage may have something to do with it. Most of the solid caps are rated for 10V or thereabouts for these sorts of equipment – I was measuring over 90v DC over one of the bulging caps, and slightly reverse polarized to nothing over the other. Maybe they are swollen because something else in the modem failed due to lightning, or age and caused the capacitors to be abused beyond their ratings.
Whatever it was, I felt it was a rather interesting result – definitely not one I had predicted.
Lexmark Toner RFID Chip
This is probably the most random part of this post, so it got saved for last. Just before the end of last year, I saw a few toner cartridges in the recycling bin, when a small PCB on them caught my eye. This was from a Lexmark workgroup copier, an X-series unit like the X864de.
It was a rather interesting piece because of the constant corporate desire to protect future revenue by investing in strange DRM schemes. Initially they were one-shot fuses, then there were some little contact-chips, and now we’ve gotten all the way to a contactless chip. I suppose the lack of contacts could make reverse engineering slightly more difficult and developing a compatible chip a more involved process. It must have been a pretty expensive thing to develop as well – but it could also have quite a lot of cryptographic smarts in it too.
The chip itself is a bare silicon “gobtopped” onto the PCB former which is also the antenna. A surface mount capacitor “holds up” the voltage enough to allow the chip to work. The PCB is dual-layer, but only for the through hole via that allows for the outside of the antenna to “connect” back into the chip.
The back of the PCB is marked FX 160K95122 K1 HFD1-T MDK 338V-0 W and has the Underwriters Laboratory logo on it. It doesn’t activate anything under a standard 13.56Mhz IEC style reader, so it probably operates in the 900Mhz “industrial” RFID band. Further details are not known.
Here’s a close-up of the die – no markings, just the scratchings from the “sawing” process during manufacture.
Alas, my mind is clear … sort of. Clearer than before. In the meantime, my Valentines Day special post didn’t attract any e-mails … which is … rather unfortunate.