Rand: SSD Dying (or not?), Ankle CT, iMac Temp Sensor Hack, Fake Batteries, More Opal, etc.

It’s been a while since the last random post, so here comes another, which is jam packed with carefully curated observations and events that have occurred since the last posting.

Gough’s “almost” heart-attack

Before anyone gets too scared, no, this one isn’t a literal heart attack, but more of a figurative one. I was working away on my main desktop, just launching Photoshop as usual, but the computer gets slow and starts stalling. Eventually, nothing opens … so I force the machine to shut-down with a long push of the power button. Clunk. It turns off.

Pushing the power button again, I expected a regular boot but instead, I was greeted by this:


Crap. Crap crap crap! Something’s wrong … really wrong. Is this real life? Or shall I just try again? I go and disconnect everything peripheral wise, and enter the BIOS and turn off the non-chipset RAID ports just in case. No dice. But the message was different.


Lets try safe mode? No dice there either. Was this the end of my OCZ Vertex 3, which had reached 14TiB written? Lets try booting Linux from a USB key – it always saves my bacon. Maybe we can still make an image or copy off the files and reinstall, not that I really would look forward to that.


Alas, the drive was not reporting itself as ready with a read error and ATA Bus Error. At least that was a hint – looks like it’s having trouble talking to the board. The drive’s SMART data looked fine, but curiously the OCZ Vertex 3 doesn’t report Interface CRC Errors as a SMART parameter, otherwise it would have jumped out at me.

Maybe that’s a sign that the motherboard’s on its last legs (a second hand Gigabyte 890FXA-UD7 running a 1090T BE at 3.90Ghz, overvolted)? I didn’t touch anything in the computer so why would this happen now?

As it turns out, it wasn’t anything too bad. I pulled the drive, chucked it in a USB 3.0 case, which imaged it fine – so the drive was okay. I plugged it back in, using the same cable, and it booted just fine. So, maybe a bad connection due to connector oxidation? Who knows – but alas, even with less pins, it seems SATA can have the same cable issues that plagued IDE more often.

It did turn into a good thing, as I moved to also upgrade the machine – I turned one of the eSATA ports into an internal port to add one more SSD to the system by looping an eSATA cable from the outside to the inside, and using an eSATA to SATA cable. I can probably do it for the other port as well to bring my drive count up to 10.

That being said, I did discover that the chipset is capable of working with port multipliers, but they’re just too expensive. I mean, a 2 to 1 port multiplier would be nice, 6Gbit/s on the uplink, two 6Gbit/s on the down with up to 3Gbit/s split evenly would suit almost all hard drives with no performance loss – but nobody makes them. Instead they’re mostly pricey 6Gbit/s 5-port models, or 3Gbit/s uplink 2-port models which will bottleneck the faster hard drives.

I also shuffled my PCI/PCIe cards around, so now I’m running a GTX580 as primary GPU, with a GTX560ti running as a secondary GPU. This means I can run the four monitors that I normally have, all on hardware graphics, and retire my Matrox Dualhead2go and Displaylink USB solutions! The improvement in graphics performance and lack of window-positioning limitations is very welcome.

Dual-Energy CT Scan of the Ankle

I suppose it’s a little personal, but readers would probably know of the ankle issues I’ve been experiencing recently. A specialist referred me to get a dual-energy CT scan of the ankle, looking for gout crystals, and referred me to Superscan’s facility in Fairfield Heights. We had booked in the procedure ahead of time, and it was completed very quickly, pretty much first time, using their Siemens Somatom Force unit. The scan itself was fairly quick, and the unit made only a mild whirring noise – kind of like a washing machine with no clothes in it.

The best part of the scan? I got to take home a DICOM CD-ROM of the scans, with a selection of results printed on paper (rather than film). I really like DICOM files – they’re easy to deal with, don’t involve analogue media like film, and they can be opened directly in Adobe Photoshop.

The disc was written by a piece of equipment made by Sorna, and the DICOM viewer included, called “lightbox”, actually works – which, is a surprise in my experience because the included viewing software often screws up. Unfortunately, the software seems to have written the files in a semi-random interleaved order meaning that importing them into Photoshop gives you a non-sequential layer order. Oh well, lets do some manual layer rearrangement – it’s less painful than trying to acquire from film.

A total of 7 sequences, with a total number of 570 slices were exported. The DICOM disc contains 470.6Mb of data, which I’ve turned into a range of animations. Unfortunately, due to the size, you must click on the image to open the animation in a separate tab/window.

Sequence 1 – Axial Bone


Sequence 2 – Coronal Bone


Sequence 3 – Sagittal Bone


Sequence 4 – Axial Soft


Sequence 5 – Coronal Soft


Sequence 6 – Sagittal Soft


Sequence 7 – 3D


So yes, that’s my left foot, and interestingly it seems to be fine bone and joint wise with no evidence of crystals although, again, they did discover something akin to an effusion somewhere. Back to the specialists, it is!

iMac Late 2009 Drive-Upgrading and Temperature Sensor Issues

Just before the new year, I helped a family friend upgrade from a 1Tb hard drive that was slowly grinding to a halt to a new Intel 730-series 480Gb SSD. During the upgrade, which we “bodged” by looking at some guides online and by trying our own things, we discovered a few things which were noteworthy.

Note 1: Removing the front glass

Removing the front cover glass is actually pretty easy on the iMac. The method developed by the family friend involves using laminated paper sheets inserted in-between the aluminium body and the cover glass on both the left and right sides. Once one sheet has been inserted on both sides, a second sheet is tucked in on both sides, which moves the glass far enough from the magnets that it is simply pried off with hand force.

There is no need for suction cups and no risk of accidentally pulling too hard or unevenly.

Note 2: Removing the LCD screen

Removing the LCD assembly is not too difficult provided you have a T10 torx screwdriver that’s long enough. Only eight screws (if I recall correctly), four along the left and right edge of the LCD frame, need to be removed. The LCD fits snugly and requires a little persuasion to slide out slightly. Then, you must be careful to disconnect the sync, LVDS, backlight and LCD temperature sensor wires before lifting the whole screen out of the chassis (which will then have a tendency to tilt backward on its own).

There is a tough part when it comes to reassembly, and that is trying to get the screws back in with the influence of the magnets surrounding the frame of the iMac – but with enough patience and technique, it is possible. A magnetic pick tool is very handy to fetch any errant screws.

Note 3: Replacing the Hard Drive with an SSD

The internal drive is a 3.5″ drive, and so it is best to use a 2.5″ to 3.5″ adapter caddy. The original drive will have a temperature sensor cable to the base board.

If you remove this cable and replace the drive with something else, as I found out the hard way, it will result in the hard drive fan in the iMac ramping up to full speed after a few minutes of being on. This results in a (potentially annoying) whining noise and can shorten the life of the fan.

Unfortunately, since the temperature monitoring is very special to the iMac, there are several options which some have found. The software based fixes don’t seem to be universally applicable, especially with the new MacOSX. One involves the use of a specially crafted cable, which emulates the temperature sensing for any drive, but is difficult to get.

In theory, it seems based on measurement that you can just replace the temperature sensor with a regular 1N4004 diode, as long as you orient it the right way. If you want to hedge your bets, you could probably put two back-to-back and then install that.

But there is an easier way, especially if you’re willing to sacrifice the original drive. I discovered that the temperature sensing on the original hard drive functioned without drive power applied – so it was a simple task to remove the original hard drive PCB, wrap it in paper, slip it behind the SSD and add some sticky tape – with the temperature sensor cable installed as originally done.

Problem solved, and no extra parts required!

Gigabyte WindForce Frustration

With this “issue” of random posting coming about as I was rotating hardware around in my machines to optimize utilization, I was frustrated to find my two Gigabyte graphic cards had developed bad fan bearings. One was a Gigabyte GTX580SOC, the other was a GTX560tiOC. It’s just one reason I absolutely hate fans on graphic cards, even though they are necessary for the highest performance cards.

Both cards were seeing temperatures over 88 degrees C, a steady decline in cooling performance even though the heatsinks were decently clean. Upon disassembling, I found that the assembly people put a stupidly large amount of thermal paste on the heatsink and that was probably part of the cause.

The other cause was the failing bearings which resulted in a clattery idle speed and a noisy “whiney rattle” at full speed with a reduction in top fan speed.

Unfortunately, many others suffered bearing failures relatively early on, and so spare parts aren’t as easily available. What you need is a 75mm fan, with 40mm triangular mounting. I managed to find one source of replacements with a four-wire PWM connection.

The replacement procedure required disassembling the whole heatsink to remove the fans, and cutting the original fan connector shrouds to plug in the new fans (due to their extra pin/pins). A bit of work with a multimeter was needed to find the correct plug orientation, and creativity was needed to route the long wires from the new fans in an unobtrusive manner.

After replacement of the fans on one card, the load temperature fell to just 73 degrees C. Much better, and much more quiet too. I’m still awaiting replacement fans for the GTX580SOC which needs three fans.

More Fake Batteries

While the more seasoned people will say “what’s new?”, it’s always a frustration to receive something fake when you’ve paid for something decent. This time, it was a battery for an old Icom IC-40S handheld CB transceiver – this one was ordered as a “replacement” battery of 1600mAh capacity, but was so poorly built that it fell apart as soon as it got here. The contacts were also angled and very thin, the plastic was too thin and creaky. They used a very similar label to the original Icom label, and claimed to be 1650mAh and “Made in Japan”. Yeah right.


Why did I laugh? Because the battery fell apart immediately on unpacking and revealed this:


It’s clearly 1300mAh at the most because of the labelling inside. I’d place my bets that it’s even less. It’s also wired with very thin wires, which is no good for a 5W transceiver that can be pulling 1.6A in high power from the cells.


The battery is also a poor fit, with large gaps between the battery casing and the body of the radio. The belt clip also cracked on the first application of pressure. Absolute rubbish. They had the hide to charge me AU$25 for it – but they refunded me and cancelled the transaction as soon as I gave the pictorial evidence.

While I was at it, I also had several Canon NB-2LH compatible batteries from my old DSLR that I had sitting on my shelf because of short run-time or outright failure to charge which I thought I’d take apart too. They’re worth a laugh, so I decided to lump it in with this post as well.


The left cell was bought from eBay in 2011, whereas the right cell was bought from a stand-alone retailer in 2008. I had a pair of the Powersmart cells, but the first one failed in one year – this one only failed in 2011. The Powersmart claims to be 750mAh, which is a bit high, but the other one is just a plain joke. If you can fit 1800mAh into that package, you’re doing the impossible!

Lets start with the Powersmart cell:

20150307-1457-3867 20150307-1457-3868

Inside, there’s a 14430 cell marked ESBC. No information can be had on the cells, but they’re likely 600mAh in a series configuration for 600mAh total. Under the shrink-wrap, the cells have some printing but it’s not really helpful in finding out the cell’s origin.


Being the older battery, the PCB itself isn’t actually too bad.


There is a permanent fuse protection for short circuit, a MOSFET and likely a controller for cell protection and balancing. There’s even some ceramic capacitors scattered about. That’s pretty good for safety and functionality although it probably doesn’t match the genuine.

20150307-1503-3871The unbranded cell on the left looks like this – it uses XSL branded 14430’s dated 29th May 2011. The best lead I have on it has some relation with Ultrafire, which doesn’t make these cells quality at all.

Another lead seems to imply they are SZSXSL‘s cells, but their site is hardly navigable. I’d have to go with 600mAh rather than 1800mAh which they claim.


There wasn’t any writing under the shrink wrap from unwrapping a cell, but one thing that was noticed was the very dirty spot-welds. The PCB is a later design, which seems similar but omits the fuse protection which is not as good for safety. It looks to be a cost-focused design.


So, I guess the mAh ratings on eBay and other places are pretty much as useful as a road-sign to a blind person. To pay more for a battery with a higher mAh rating is a risky proposition, because you might not be getting any more in reality – but then, buying one with a lower mAh rating will probably have even less capacity than you expected. It’s a bit of a risk either way.

Later replacement batteries are probably prone to cost-reduction and corner cutting in the protection and safety circuits of the cells as well, which is disappointing, as this is something that cannot be checked without taking it apart.

An Adult Opal Card Arrives

As it turns out, my brother (who is not eligible for concession) needed an Opal card and ordered one online. After it got in, I got a chance to scan the documents, so now I know how the concession Opal package differs from the Adult card.

The order process is pretty similar, without the need to do any verification unlike the concession case. The card comes in an DL envelope, with a cover letter that’s pretty similar.


The card itself is black on the front, as most local Sydneysiders would know, and the rear is pretty much the same.

Adult-Opal-Card-Front Adult-Opal-Card-Rear

Opal-Card-Starter-PackHowever, the included information packet is much thicker and is called a starter pack. Inside this pack are the guide to using Opal, which seems to be periodically refreshed (this one dated Jan 2015), as well as a few other miscellaneous flyers.



Interestingly, you get the set-up auto top-up brochure even if you have done so already. What a waste.

Set-Up-Auto-Top-Up Ready-Tap-Go

Other Transport Observations

I’ve begun using my Opal card, and so far, it seems to operate satisfactorily. The readers are sometimes a bit slow to recognize a card, and other people seem to have a lot of problems because they’re not holding their card close enough for the whole duration of the transaction. Anyway, I’ve managed with it, but they’ve already managed to incorrectly bill my fare.


This transaction on 18th Feburary was to catch the 418 from Sydenham station to the MSY near Kingsford on the way into uni. To be fair, the bus was running horrendously late due to a building fire, and the driver was flustered. When it arrived, I tapped on, before finding the driver furiously bashing the console to try and get it to co-operate. As it turns out, it was not tracking GPS and so stops had to be manually selected. The wrong stop was selected when me and two others boarded and they were incorrectly charged.


The straight line distance between stops was 1.32km, almost one whole section. In reality, it was probably a whole section due to the loop the 418 does on the approach to Sydenham station.

I then found it was not possible to contest this fare via the mobile website, which was extremely annoying. I could do it when I get back to a desktop, or I could call them. I fancied neither option in the end, because I probably wouldn’t have saved any money when counting the remainder of the journey when I continued my trip as a transfer in front of MSY.

But this does illustrate an issue – if users don’t carefully check and contest their fares, then they can be mis-charged without knowing it. The Opal site doesn’t allow you to look up the stop from the site – you have to do your own homework to determine if you’ve been billed at the right stop. Even worse is that as a metadata retention scheme, they now think that I travelled from Petersham Road even when I didn’t. Lets just hope Opal transaction logs aren’t used as location evidence because it’s obviously not infallible.

That bus ride was also noteworthy for a second reason – I actually helped the driver troubleshoot a problem with the bus. Apparently, the bus stop bell system was ringing immediately upon door close, indicating a stuck button. When the bus stopped, I explained to the passengers what was going on, and pushed every button – this released the stuck button (none appeared to be physically stuck) and allowed the stop bells to return to normal operation. Because he was running so late, he couldn’t be bothered fixing it himself!

A while back, I did mention the intelligent vandalism of Sydney Trains livery to say “shiT service operated by Sydney Trains”. This time, I was quick with the camera, so I managed to capture it.


There has also been a replacement of many of the disabled signage near lifts to a new sky-blue colour which is more eye catching. I suppose this makes their facilities easier to find for those who need them.



Another academic year, means another o-week. I still remember the o-week from my undergraduate days, and lets just say, over the years the o-week festivities seem to be shrinking. This year wasn’t particularly exciting in terms of stands in attendance (mostly the same old companies), or in terms of giveaways. Most of the stands with decent giveaways saw long lines that made it pointless to even try.

Technology wise, in prior years, we have been graced by Microsoft/HP, Asus and Google to name a few. This year, we don’t see any of the big names, but we do get Huawei.


I found it rather interesting that Huawei would choose UNSW o-week as a way to promote their brand and increase their presence in Australia. Maybe it will work out for them? Who knows. I have a feeling that new students will still be buying the major brands *ahem* and probably already have their needs met technology wise. After all, high-school life is not as technologically barren as it was when I was there.

No Thanks Australia Post!

I ordered some nice computer bits from a company near Melbourne VIC to be shipped via eParcel to me in metro NSW. Normally the time frame is two business days, but Australia Post took a record five business days this time. Many thanks for giving me a heart attack. I was waiting and waiting for this one.



That’s it – the end of another random posting filled with random observations. Needless to say, life keeps throwing up challenges … some of them make for good random posts :). See you next time!

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in Uncategorized and tagged , , , , , , , , , , . Bookmark the permalink.

5 Responses to Rand: SSD Dying (or not?), Ankle CT, iMac Temp Sensor Hack, Fake Batteries, More Opal, etc.

  1. drew wollin says:

    Hi Gough. Be wary of early OCZ SSDs, especially the Vertex 3; they have a bad reputation for dying, usually totally, losing everything. OCZ is now owned by Toshiba and is a bit more reliable. Intel is still the gold standard for SSDs and not much more expensive. Regards Drew VK4ZXI

    • lui_gough says:

      Hah! Everyone I know has mentioned something similar as to the cause of the demise of OCZ (and subsequent rebirth), but as far as I can tell, SF2281-based SSDs have been the most solid SSDs in my collection despite their compression performance variability. Early BSOD isues were due to OCZ’s partnership with SandForce allowing these products to get to market with buggy firmware that’s all but ironed out.

      Of the two OCZ SSDs I’ve owned, both outside their 3-year warranty with over 20TiB written, neither of them are dead or show any reallocations at all. I really think the OCZ issue is more one of “early adopter’s syndrome”.

      – Gough

  2. journaldulapin says:


    Is this possible to have a photo from the PCB from the hard disk with the iMac ?

    And a screen capture from a soft who “see” the temperature ? It’s interessant !


    • lui_gough says:

      Unfortunately it’s a bit too late for that. We were wary to get it done as quickly as possible since it wasn’t my machine and I was visiting, and the annoyances of the fragile flexible-flat sync lead which likes to start flaking apart at the end really meant that we wanted to get it done quickly so as not to cause any damage.

      The temperature sensor is *somewhere* on the Seagate PCB on his drive. It looks like a regular Seagate PCB except for it dedicates two of the normally unused diagnostic pins for temperature. A small thin 2-wire lead is used to connect from the diagnostic port to the temperature board.

      I’m sure there are many widgets for Mac which can see the temperature – but one thing’s easily proved – the iMac fixed in this manner stops whining like a jet after the old hard drive PCB is connected.

      – Gough

  3. journaldulapin says:

    OK, thank you

Error: Comment is Missing!