LongRunner's Mini-review Collection

[LongRunner]

Appliances

• Onix ON-PDF15 USB-rechargeable desk fan: While I don't generally buy from this name, I was willing to make an exception this time since it was on clearance for about AU$15 while the original price looked sufficient (was it around AU$40?) to make a reasonable product; the motor is brushless (although the sleeve bearing could have been better-oiled, but problem solved with an application of my Zoom Spout oiler) and it moves air well enough for its size (vaguely Vornado-esque in design although it has 5 blades and the front grille slats aren't angled for optimal flow-focusing, but at least they swirl the right way and seem to work adequately). There's a questionable electrolytic capacitor inside, but then we never really expected the good stuff in much of any home appliances anyway. Charger is not included, so we don't have to worry if it's safe or not.
• TaoTronics TT-DL22 LED desk lamp: The lamp itself (with 48 LEDs, 30 of 5000K and 18 of 3000K; can also combine both sets to approximate 4000K) is functional enough, but the USB charge output interferes with digital TV when loaded. Not yet sure of the exact deficiency, but it would have been perfectly easy to ground the metal base for a bottom shield (and I can't imagine a suitable top shield adding much more cost either).
  The 3 electrolytic capacitors inside are again your usual mediocrity (Kehong brand, "LOWESR" but no actual series stated).

Audio cable

• Jaycar/Electus "Concord Multimedia" WA7502 (3.5mm to stereo RCA, 2m total of which L and R are separate for 25cm):
  These cords are a con , with no shielding. (It does otherwise work, but clearly not worth the AU$12.95)

Power-boards/strips (formally EPODs) (Australia/NZ)

• Crest PWA04984 and PWA04985: The most-decent (not that the bar's been set high ) modern power-boards I've seen so far (surprisingly available at Coles), with reasonably sturdy casings and socket contacts of ordinary design but clean brass; the PWA04984 has 4 sockets, while PWA04985 has 6 sockets and a master switch (which is actually double-pole although not stated as such); otherwise they're much the same, and on both models the left-most socket has a wider spacing (55mm) than the rest (45mm). Both contain a single 14mm MOV (claimed to absorb 175J suggesting a nominal 470V or thereabouts) from active to neutral, protected by an adjacent thermal fuse plus F3.15A HRC fuse (although with the usual arrangement where the load remains on); the indicators are green InGaN LEDs driven half-wave through 3MΩ resistors, so heating there is negligible (and they are in opposite polarities to cancel out the DC component, however small it is, for as long as the MOV remains enabled). The cord (standard H05VV-F3G1.0 with a side-entry plug, manufactured by Conghao) measures accurate to the stated 1.2m, and at 20A the circuit breaker trips in about 15 seconds.
  Viewing their (admittedly bloated) site, related models include the PWA04980 (4 individually-switched sockets) and PWA04986 (6 individually-switched sockets); and for those who fancy USB charging (up to 2.4A rated), PWA04982 (4 sockets + 2 USB ports), PWA04987 (6 individually-switched sockets + 2 USB ports) and PWA06021 (4 sockets + 4 USB) all with presumably the same MOV (though I can't guarantee without checking individually); there's also a slimmer 6 socket + 2 USB model (PWA06205) if you aren't bothered about surge suppression.
  Even more interesting is their "Connect" series which have an IEC C14 inlet instead of a wired-in cord (they also sell 1m, 3m and 5m C13 cordsets in case you don't already have 150 of them ); these include the PWC4100 (4-socket master switched), PWC4400 (4-socket individually switched), PWC4402 (the same + 2 USB) and PWC6100 (6-socket master switched), none of which claim any surge suppression (although all have the same wide-spaced socket on the left). Will add those to this thread if I get one (or more)…
  The two "plain" power-boards they sell (CCPB4 and CCPB6) aren't so interesting (looking like an ordinary rebranded Chinese design), but at least they're priced accordingly and hopefully-compliant (the only recall under the brand was back in 2002).

So, it looks like non-electricians aren't out of options there.

[LongRunner]

Appliances

• DēLonghi HSX3324FTS: A solid convector/fan heater like I've been quietly desiring for some years (my Mum used to have a Goldair Turbo-Convector model 8050 in her office, from ≈1997 to about the mid-2000s; while those had the fatal flaw of plastic top grilles, I did like their neat layout and thicker steel front and back panels compared to the modern norm). Even the thermal cut-out (Klixon TH11CA181) is exactly the same as in the old Goldairs, although without the warning buzzer (advertised as "Sensortronic" back then). It also has a (proper) tip-over switch (unlike the cheaper models) and timer, and mica sheets (for enhanced electrical insulation?) inside the casing besides the heating elements. The PBT (+ 30% glass fiber) blower housing is shielded from direct heat exposure by the metal plate mounted to its front (I can actually keep my hand on the case area directly behind the blower), and the heating elements themselves are in very good form (and hardly buzz at all when on).
  One small design niggle is that one of the motor wires (which connect through a pair of bus-bars along the element assembly, being on the opposite end from the controls) is a bit short and therefore taut; still, I'll give it this:
  

  

  Bronze-200px.png (15.48 KiB) Viewed 196582 times

  It's relatively expensive at AU$169 RRP, but I think the build quality is worth it (anyhow, in its lifetime it'll use way more in electricity). Power is 1000W low + 1400W medium (2400W on high), though blower operation is restricted to 2400W mode by the shared rotary switch.
  Germany gets a 230V 2000W version (HSX3320FTS, 89€) and the USA/Canada a 120V 1500W version (HSX3315FTS, US$120/CA$130).
  I suspect the HCM2030 is cheaper than DēLonghi really want to go, but that they decide they might as well at least make it competently (which they do).
  UPDATE: The Australian 50¢ coin (32.5mm corner-to-corner, larger than any Euro coins) can reach from the 1000W element to the rivets holding the back mica barrier in; so if you have chaotic children around, I'd suggest internally covering the rivets (and lower holding tabs) with neutral-cure silicone.
• Russell Hobbs RHT12BRU: A 2-slice (as 2 short slots, which lose less of the heat than 1 long slot) toaster with a nice solid brushed stainless-steel (albeit ferritic) casing, and wickedly powerful (1670W at 240V ) heating elements (also NiCr rather than FeCrAl) for fast toasting (although somewhat uneven, due to some parts expanding away from the mica and therefore glowing hotter; but nearly all conventional toasters suffer this to some extent).
  It also includes an additional safety measure using a relay (which, yes, is double-pole) so it can switch off even if the lever gets stuck down.
  It replaces a cheapo Heller TAH2S (from 2005). Recommended retail price is AU$60 although discounted to AU$48 at RetraVision today.

[LongRunner]

Amplifier

• Technics SU-Z780: Found it among some "rubbish" while on a walk late in May, once home it was clear someone already "had at" it in more than one place: The original custom fan (using a DC brush motor, as this was around 1989 so brushless fans were presumably relatively expensive), with its motor knackered, had its wires cut out and controlling transistors removed from the main PCB. The output relay had its contacts bridged by solder blobs (not because the relay was bad, but because the power module had been replaced with one lacking the control pins for the relay); I salvaged it for future use and jumpered it out more-tidily with copper wire (pairs of 0.5mm strands from a piece of 1.5mm² TPS cable as used on Australian lighting circuits; would have gone with tinned if I had a suitable spool). The other needed repair was to resolder the speaker terminals; and the potentiometers will need contact-cleaning before I'm done with it, but apart from that it seems to work fine (testing into load resistors home-made from sections of a salvaged fan-heater element, with another fan temporarily bodged-on).
  The replacement fan will be a Nidec D08K-24PH (I have several) salvaged from a Kyocera laser printer; the next step is to make the voltage regulator (from the internal unregulated +42V or thereabouts, a bit more than the 7824 is good for but an 8.2V 1W or 5W Zener diode should drop the input enough for it) and perhaps a new control circuit. I'm also cutting a larger hole for the 80mm fan, but this is hard (and noisy) work so I probably won't do it again if I ever repair another of these…

Audio cable

• DSE CA0106 (3.5mm to stereo RCA, 1.5m of which about 13cm is separable at the RCA end): Another cheapo (apart from having gold-plated contacts) without proper shielding, although at least I got it for only AU$3 at a Good Sammy. Also, the strain-relief sections on the RCA plugs are too large in diameter, and hence too stiff to be effective on the relatively-thin cable.
  UPDATE 2022-05-12: The left RCA pin broke off, since it's just a sub-paper-thin metal layer around a plastic core

Travel adapter

• Go Travel 095 (Schuko to AS/NZS 3112): So close to decent… This one is discontinued with its replacement being model 544, anyway I found it in another op-shop also for AU$3. The casing is PA66 according to the card and contacts are quite sturdy brass (0.6mm on power contacts, 0.8mm on earth clips; it indeed passes the rated 10A fine) albeit with aluminium (?) rivets, but the Schuko earth clips are a bit short and only make contact at about the same time the power pins do (and were overall quite loose as originally formed). It seems that the Australian standard for these includes no requirement for the foreign sides to comply with their respective standards. Original price was indicated as AU$12 (about fair for the materials used).
  Also, the company is a bit weird in selling both half-decent designs and death-dapters.
  UPDATE 2022-08-20: Now that I have a better model, I've decided to recycle the brass contacts and fire-test the casing. It did take some perseverance with the glow-wire (which probably also gets hotter than officially required) but I managed to ignite it, and it continued burning for over 11 minutes!

[LongRunner]

UPDATE: The newer Trent & Steele TS12 is the same sort of design, but bigger (10 cup) and a bit better-made.

I recently had an accident with the TS5: The lid got vibrated off the benchtop (by water-hammering when trying to turn on the cold tap) and landed on its knob; the glass held together, but the knob plastic broke. While I could have bodged it back together (by finding a longer screw with the same thread), I also noticed some rust on the inside and Mum had a voucher for Hardly Normal (the plan was previously to buy the new toaster with it, but I ended up getting that from RetraVision due to it being near the dentist we use), so we went there on Saturday (2021-09-18) and looked for something hopefully more-robust.
(We still haven't gotten around to installing a new cook-top…)

But before we send the TS5 off to e-waste facilities, I decided to finally address this:

The set temperature for the thermal fuse of 192°C does seem very high, given that its only distant thermal connections to the heater are via a thin steel mounting bracket, and the wire from it to one heater terminal. Would be interesting to stick a thermocouple in there and see what temperature it actually reaches in normal operation…

…which peaked at (on the outside of its insulation sleeve anyway) 107.2°C during a final run with 250g raw rice + 600mL water. (Not sure if different quantities would change this, but probably not by much.) This could increase if the thermal contact from heater to pot is poor, though.
SEFUSE documents say the body temperature can operate (for 10,000 hours?) up to 20°C below the “nominal” cut-off temperature (that in the part number, as opposed to the guaranteed cut-off temperature for agency approval), or 140°C whichever is lower; you can draw your own conclusions from this.

I've also salvaged the C14 inlet, which is probably the most-useful part for any future projects (since it wasn't soldered, this is easier than trying to get them out of old PSUs; and for proper contact, you also have to avoid the non-compliant inlets commonly found in gutless wonders).

Anyway, its replacement is a Sunbeam RC5600, which is larger (7 cups) and accordingly slightly higher-power (240V 500W).
It has the same sort of lacquered brushed-steel finish (a cheap look-alike to stainless) with black plastic parts.
But the best feature is probably the “jar style” lid (which is hinged on the right-hand side and latched at the left; only an aluminium insert needs to be removed for washing along with the pot), which also solves the "spurting" problem mentioned before (no more need to coat the inside of the pot with olive oil! ).

The base isn't as easy to open (since it's mostly clipped-in, with just a single divided-slot screw for alignment), but I managed eventually and it definitely looks better‑built overall than the TS5 (no dent in the rice pot, which also fits closer in the unit; crimp connectors are actually fitted properly to the wires; etc.); then again, for double the price (give or take) you wouldn't want anything less…

Circuit-wise, the only substantive difference is that the keep-warm function is done by a thermostat (in parallel with the cook switch), instead of the low-power auxiliary heater in the TS5; comparative advantages and disadvantages would be:
+ regulated temperature
− but ripples with the on/off cycle
− could stick on (even if unlikely with ≈2.1A through contacts rated for 10A) and overheat, making the thermal fuse doubly crucial for fire safety
I believe the cook switch also has a slightly lower switch-off temperature (is it 140°C at the pot base, versus 150°C in the TS5?); I did find the rice to get noticeably crisp at the bottom in the TS5, but hardly at all here.

That aside, it uses the exact same thermal fuse and very much the same arrangement for the indicators (120k carbon-film resistors and about the same-size neon bulbs), although this time their PCB is epoxy-fiberglass rather than phenolic-paper. The wires are all white here, though.

The silicone sealant at the element ends could have been better-applied (my Breville BKE320 kettle, for example, has a much better silicone job).

While I'm reusing my existing shortened C13 cord, the one included is 0.9m long (this time of ordinary H05VV-F3G0.75, or 4V-75 sheathed under Australian standards), feels flexible enough; and loaded with a 2400W heater (this time the HSX3324FTS), passes after some 4 hours. Observing also a few generic fan heaters (the Target Essentials TEFH211B and an older Kmart Homemaker Everyday of unknown model number) which have 90°C cords without an apparent design reason, it seems that some Chinese designers may just opt for them “automatically” without bothering to individually check the temperatures in-situ…
Also, this time the C14 inlet is around the back, so it doesn't stop you from placing things to the left of the unit.

If you remember this statement from my TS5 post:

I might instead use the included cord with my PC (saving the longer cords for when they're needed)…

That I did, and since even 0.9m is more than long enough there, I've now swapped it again for the RC5600's included cord

The manual has this usual statement from Sunbeam…

Do not use your appliance with an extension cord unless this cord has been checked and tested by a qualified technician or service person.

…but observing that counterfeit IEC cords seem much more common than counterfeit extension cords, it would probably be more pertinent to warn about them.
That said, the current draw here is low enough that even the worst counterfeits I've seen would still survive (not that I'd use them at any load, of course).

I'm compiling cooking data like before, and will upload when I've gathered a reasonable amount.

Update 2023-04-22
The condensation collector is getting loose and falling off easily. Sunbeam should have at least used detents to hold it on, not relied on friction alone.

[LongRunner]

Appliances

• Kambrook KHP120: After damaging the non-stick coating on a KEF125 electric frying pan (now replaced in Kambrook's range by the KEF135), I changed tactics and got a portable hot-plate instead; I decided against an induction type due to the amount of non-ferrous cookware here, leaving a glass-ceramic top over bare resistance elements as the next-best alternative (granted, for pro-chefs the choice is induction or gas, but I'm not that fussy…).
  Anyway, for AU$120 you get two 1200W glass-ceramic tops (to fit the Australian 2400W maximum) with an outer diameter of 18cm overall (137mm for the element proper); they operate independently, each having its own neon indicator, thermostat and also thermal fuse (Microtemp G4A, Tf=216°C, mounted to the casing in the general vicinity of the thermostat and heating element). Kambrook also provide a single version (KHP110) for AU$70. Absolute temperature accuracy isn't as good, but it has much less hysteresis than the KEF125.
  The overall build is fairly solid, but with some alignment issues on the right-hand side (the neon bulb was a bit high up, and the thermostat position collar is rotated a bit counterclockwise from where it should be), and the screen-printed markings aren't durable at all. I'm also not sure how well the polycarbonate (?) indicator windows will fare under frying oil exposure (glass may be better for them, if more expensive).
  Curiously, inside the unit each hot-plate casing is actually earthed twice: One earth wire goes from the input to their rear mounting brackets, the other goes from input to the left-side bracket on the right-hand plate and right-side bracket of the left-hand plate. I'm not sure what this achieves when the cord has only one earth wire (which is much more likely to get broken) anyway, and the other metalwork is only indirectly earthed; but maybe there's some obscure requirement in the standards somewhere…

What I do like about the electric frying-pan principle is that the embedded heating element gives tight thermal coupling (although the heat is concentrated somewhat over the element itself), and perhaps better efficiency than with any cooktop (though of course sacrificing flexibility). In normal operation they're also about as safe as an induction cooker, although a thermal fuse is conspicuously absent from those I've examined (perhaps because thermal fuses with a high-enough holding temperature weren't available when the standards were last updated, though Microtemp now boasts Th=220°C for Tf=257°C).

If a thermal fuse is to be added into a frying-pan probe, then I've got the idea of also connecting a “REPLACE PROBE” indicator across it, to inform the user that the failure is not in the pan itself (although this would only be much good if the probe is made available separately for replacement).

[LongRunner]

Lamp
You may have noticed that I've not had a very good track record with them. Let's see if this one fares any better…

• Emporium “Beckett” table lamp (model CWTLE1412 according to box, or C1312TLE27 on the label): Rather old-fashioned looking with a large glass surrounding the bulb, and what looks like some sort of cement base. Somewhat confusingly, the box shows two models (the other one, Zana, with a B22 socket instead of E27) and doesn't make very clear which one is in there (only stated on the bottom above the barcode).
  Electrically there's not a great deal to see; but I can say it at least passes double-insulation, and the plug pins are the correct length. The wires inside the lampholder are sleeved with heatshrink (normally good to 125°C, which is probably enough for the stated maximum 25W incandescent). I also checked the air temperature around an 8W CFL (the stated maximum for an electronic lamp) and that's fine too (maybe 12.5 above the room).
  But it still has only a single-pole cord-line switch, so is unsafe in the event of upstream miswiring (or a broken neutral with other loads connected); and following the logic of this official recall, that would be reason enough to recall this one too. I'll have to think about whether or not to count this in my own recall thread (and to be consistent, I'd have to apply the same reasoning to the toggle switch in the T-869; not to mention many other desk lamps)…
  (After all, toasters are required to have double-pole switches; so why shouldn't these?)
  At least it can be replaced with the HPM D5M (and these cheap factory-fitted cord-line switches seem to break easily anyway).
  UPDATE: Having since broken the surrounding glass and lacking a reasonable use for it, I opened up the switch to discover poorly terminated wires which are also copper‑clad aluminium (or some other non-ferrous white metal). Under load, it indeed gets much warmer than normal for 0.75mm².
  Also strangely, I counted 27/0.18 strands (0.75mm² can be 24/0.2 if Harmonised, or 30/0.18 per Japanese standards; this matches nothing I know of).
  So it earns a recall in the end; and given the circumstances, perhaps it's time to move on to native LED lamps (even if some have other problems).

Large appliance
I'd check more of them out too, given (a lot) more space; it's not knowledge which is limiting me.

• Ariston N11HPLWAU: Since the timer motor burned out in Mum's old(-ish) Fisher & Paykel DE45F56A (not exactly surprising given the hair-thin wire used to make the coil), we went to Harvey Norman to get a new clothes dryer; this is actually a condensing heat-pump model (instead of the cruder vented, resistance-heated type), so besides saving energy it also collects the moisture rather than releasing it into the room (and requiring an exhaust fan running). It has a 5-year warranty (especially valuable on inverter-driven appliances, although I'm aware that some makers cheat by providing a long warranty only for the motor itself and not the driving electronics ), and is made in Poland with the cord (1.44m of H05VV-F3G1.5, although limited to 10A by the Australian plug anyway and drawing up to 850W steady-state) from PATELEC (plug model 161) rather than the usual Chinesium.
  The refrigerant is stated as 400g total, R1234ze 58% / R134a 42% (a.k.a. R450A for the blend); nice to stay with something non-flammable, even if we don't get much choice in fridges and freezers there…
  Operation has turned out to be a bit weird though: It has a habit of getting stuck with the time display flashing (I think that's supposed to be paused) and then being unable to continue. I've not been able to work out why from the manual, so make of this what you will

Things worth noting about the F&P one: It only has one heating element, the warm cycle is (or rather, was) done by cycling it on and off periodically in the timer. This obviously ain't good for the switch contacts (which look quite worn) nor the element itself, and also poses a bit of a trap: Depending where the timer fails, it may leave the heat on continuously (thereby going hot). It's also pretty lame when most portable heaters have 2 elements and 2 or 3 power settings . The thermal cut-outs are only on one side of the drum, which when wall-mounted (as was the case here) ends up being the underside; I would therefore prefer that it had both “top” and “bottom” sets thereof.

See also, my impressions of the Vornado 6303DC.

[LongRunner]

Since it's been too crowded at home to really take on more physical products, I'll temporarily change course for now (especially being on Linux):

Media players

• Audacious: Not as featureful as foobar2000, but it works alright overall. The provided graphic equalizer is only octave-band with 1dB steps (whereas you can get an add-in equalizer with third-octave – or more correctly tenth-decade – bands, and 0.1dB steps, for foobar2000), but at least it doesn't have the nasty time-domain artefacts of the built-in foobar2000 equalizer. The blur scope visualization is wonky though (can only show about half the recordable amplitude range within the window, with positive amplitudes going down and negative going up).
• VLC: Bringing it up after noticing that the playback speed (and hence sound pitch) drifts from time to time; it turns out this problem has been known for donkey's years. I now use it for video files only, but even then it's quite annoying to hear with absolute pitch.

Image viewers

• XnView MP: Functional as always, with one substantial technical shortcoming: The scaling isn't gamma-compensated. (Some versions of IrfanView do compensate according to this page, that seems about it as far as viewers go.)
• nomacs: FOSS and it basically works, but doesn't seem to support color management so isn't much good for non-sRGB images (or depending on your monitor, anything). Nor does it gamma-compensate during scaling.

Image optimizers

• E-Mage: An effective tool (including for JPEGs), but also quite a brutal one in that it strips off basically all non-“critical” metadata (including quite important stuff such as ICC profiles) without providing settings to change this.
• PngOptimizer: Doesn't deal with JPEGs (predictably), but provides options to keep a few tags (although ICC profiles aren't among them) so is otherwise somewhat preferable to E-Mage.

Maybe someday I'll dabble with the source code, but for now I can't bring myself to do more sedentary tasks than is already the case…

While I'm at it, here's something I can review without even owning it:

• Elliott Sound Products Project 99: Rod Elliott typically knows what he's doing, but this seems to be a bit of an exception; it's a high-pass filter (tuned to 18Hz as designed) to suppress infrasonic noise (from vinyl records etc.), nothing wrong with the basic idea. Despite what the project page claims, however, the implementation is quite far from a true Butterworth filter; here's its frequency response simulated against some other filters:
  

  

  Infrasound filter simulation.png (65.59 KiB) Viewed 194506 times

  P99 is shown on the blue (amplitude in dB) and red (phase) traces; a true 6th-order Butterworth filter is shown on the magenta and green traces. The cyan and olive traces represent a 6th-order Linkwitz–Riley filter (equivalent to a cascaded pair of 3rd-order Butterworths, but only really meant for loudspeaker crossovers), which still performs considerably better than P99. Finally, the grey and black traces show a 4th-order Butterworth filter (suggested on the page to be “marginal”), which still manages a sharper initial roll-off than P99 (only losing ground below 5Hz, by which point it's already 46dB down…)
  Now, the true 6th-order Butterworth filter I simulated has three Sallen–Key stages, so can't actually be built on the P99 PCB. But the 4th-order Butterworth sure can (by jumpering out C1 and C4, and omitting R1 and R4), and works pretty much as well as P99 (or arguably better); while I used E96 resistances for the other filters as simulated, you can use E24 values alright (with the given 150nF capacitors, R2 = 56kΩ, R3 = 62kΩ, R5 = 22kΩ and R6 = 150kΩ).

The simulator, incidentally, is Qucs 0.0.20; not flawless, but so far seems workable enough for me. Users of recent Linux versions (Ubuntu 20.04 and similar) will need to get a compatible version of Qt4 here.

[LongRunner]

Computer mice

• Crest CZBWLM: A cheap wireless mouse, basically functional but seems to have a small movement dead-zone. You can use it in a pinch but I won't recommend it; that said, it does at least use a microswitch for the middle button (rather than a less durable tactile switch).
• Logitech MX Master 3: The site RTINGS has lately been at the forefront for reviews of headphones and the like, but are their recommendations for computer peripherals much good? While they don't examine the internals, I've opened mine and the left/right button switches are the 10M cycle Omrons; however, the vertical scroll wheel uses a flat tactile switch (which generally don't have such a high cycle rating; I can find only a few references to 10M cycle, with 1M more typical), as do the side buttons in a smaller form. (The plastic flexure mounting of the scroll wheel assembly will also have some limit.) So depending on your usage, you may wear those out before the main buttons… It's not as large as ideal for a giant like me, but closer to the mark than many mice.
  If Logitech would at least state the cycle life of the middle and side buttons, then I might have given it a Bronze award. Another oddity is that while the USB contacts on the dongle are gold-plated, those on the charging cable (USB A to C but not wired for SuperSpeed) have none discernible.
  There's also a warning that it's "not intended for use by children under 14 years old"
  Logitech are too lazy to provide any configuration software for Linux; you can get some options back with LogiOps, but it seems to have a side-effect of randomly enabling high-resolution scrolling (which is practically impossible to control properly in detent-scrolling mode), so I gave up on that.
  EDIT: It also seems prone to missing a line when scrolling up at a certain rate, so maybe the electromagnetic scroll wheel isn't all it's cracked up to be
  UPDATE: The rubber grip deteriorates with skin contact, so that was a bad choice

[LongRunner]

While on a walk yesterday, I stopped and browsed at Coles before moving onto a local hardware store (for a few screws and nuts).
(I wasn't even hungry really, but did feel a bit thirsty.) Among the other (boring) heaters, I found this little guy for AU$29:

Overview.JPG (910.63 KiB) Viewed 193750 times

Being that many of the other (also cheap) heaters there are this same brand, I suspected it may well be yet another zombie-brand (it seems to me that brands which still provide decent products, like DēLonghi, don't have to bang on about their age in this fashion); and this was quickly confirmed by a web search once home. Still, I already knew what 500W of heat is good for (the 600W setting on my repaired Omega Altise OMC15E1 is sufficient in my bedroom most of the time, often even on the high side of what's necessary), so it didn't seem entirely fair not to have covered any model of that power.
I suppose it's also a cute little thing (although this could perhaps be said of even the 2400W fan heaters, compared to the room size they can warm)…

Clamp.JPG (63.37 KiB) Viewed 193750 times

The cord is 1.55m of H05VV-F2X0.75 (which is fine at the power rating) made by Ningbo Chaoyu Electric, but upon shaking the heater, I heard a rattle; opening it up, I found the stripped-off piece of insulation from the neutral wire, which to add more concern, has significantly uneven thickness (actually sitting right on the edge of the permissible tolerance minimum of 0.44mm at the thinnest point for 0.6mm nominal). At least the pins are correctly dimensioned. The cord clamp is badly‑dimensioned and the plastic severely stressed, though it still holds for now. At this power level even a C8 inlet could be used if desired.
Internal wires are all silicone-insulated 0.5mm².

Loose screw.JPG (127.93 KiB) Viewed 193750 times

The push-button switch feels decently clicky, but that only applies to the plunger movement; no snap-action applies to the actual contacts.
Also, one of the mounting screws is 1.6mm loose as it came.

Apart.JPG (756.11 KiB) Viewed 193750 times

Frame screws.JPG (486.5 KiB) Viewed 193750 times

The PTC heating elements are actually in two columns (separately connectable if desired), and even at this power level, it's a pity that the opportunity was forfeited (given the small rooms it would suit, although the box sells it as a "Personal" heater as seems to be the convention for such models ).
One of the assembly's 4 mounting screws is shorter than the rest (about 9.5mm vs. 12mm), with no visible justification in the casing design.

The heater frame is marked as PA66, which is not so good for the working temperatures of these unless specially stabilized (I've already seen the fate of such plastic as the top grilles on a Goldair 8050 Turbo-Convector and this operates in the same temperature realm, although may be somewhat mitigated by the limited life of the fan anyway). It is GF-reinforced going by a slice test, but IIRC so was the plastic on the Turbo-Convector.
Both the front and back housing are ABS, although at least with a steel front mesh (as on the other PTC heaters I've seen); still, I've seen in an Everdure HCC1800TB that the front shroud melted slightly as a result, and the same will in all likelihood happen here.

Both the bimetal cut-out and thermal fuse are mounted on a mica piece at the top of the heater frame. The bimetal is set at 80°C and the thermal fuse at just 98°C; this is because it's arranged such that the fan airflow keeps it cool in normal operation, while if the fan fails, the heat will rise up from the heat exchanger.

Fan from behind.JPG (443.03 KiB) Viewed 193750 times

Fan label.jpg (327.95 KiB) Viewed 193750 times

The fan has a custom frame, but is otherwise just a standard 80mm 12VDC sleeve bearing type. Ironically, the frames of these are usually GF-reinforced PBT, which (as briefly noted regarding the DēLonghi HSX3324FTS) at least according to figures I could find, is significantly more heat-resistant (up to 200°C for PBT+GF and 150°C for PBT alone, vs. maybe 140°C for typical PA66; but for long-term use, the temperature has to be lowered) .

Converter.JPG (333.77 KiB) Viewed 193750 times

Primary caps.JPG (182.42 KiB) Viewed 193750 times

The buck converter powering the fan uses "HTIANH" (?) input and Jwco output caps, and is driven by an LN8K15.
Its input is only half-wave rectified, although at least with two diodes in series (type M7) for redundancy.
At least it has a fuse, which is more than I can say for the Dyson AM05's control board (no wonder Dyson had to recall it)

Air temperature gets up to 126K above intake in the middle of the front, which is well above anything sane for the ABS shroud .
It survived being covered for 10 minutes, but barely (the thermal cut-out took over 2 minutes to trip and the shroud indeed deformed a bit).

The given warning slip has these two out-of-place items:

Do not use this heater in small rooms when they are occupied by persons not capable of leaving the room on their own, unless constant supervision is provided.

Fine in the manual for a full-size (1500W to 2400W) heater, but doesn't make much sense for this thing

To reduce the risk of fire, keep textile, curtains, or any other flammable material a minimum distance of 1 m from the air outlet.

That's already inaccurate enough for ordinary fan heaters, but makes even less sense when there's a plastic frame directly around the heat exchanger (And the manual itself even says 1.5m; by that distance the airstream is undetectable.) Predictably, a warning to "not plug several devices into the same outlet" is also copy-pasted from the higher-power types; but at least it doesn't warn against extension cords specifically.

To conclude, here are the grades:
Build quality: D−. Not entirely disastrous, but they've spared most of the expenses they can.
Heat distribution: D−. Like most PTC fan heaters I've seen, the heat exchanger is as small as possible resulting in feeble circulation.
(Granted, it's somewhat less critical for the smaller rooms you would be heating with it.)
Fan noise: C−. Significantly quieter than many fan heaters (though this also applies to the HCC1800TB), but still no wonder.
Warranty: D−. (1 year residential, no term given for "commercial" use)
Overall, it's another close call.

If you want a 500W convection heater and you're in a 220–240V country, my money would be on the DēLonghi HCC500 (a fanless convector) as being the (probably) best-built; although I'm not sure if feet are provided for it (maybe only the wall-mounting brackets, as are more-or-less standard with convectors). While it's only sold in a few European countries, you shouldn't otherwise have much difficulty beyond importing it.
But if you can't get it or similar from another decent brand (or don't want to wall-mount it), and space isn't that tight, you may be better-off just using the low setting on a 1000–1500W heater; if the element terminals are separate from the thermal cut-out(s), you might be able to re-arrange them so that they can operate in series for an even-lower power setting (I may try this in the OMC15E1)…

[LongRunner]

UPDATE: If you don't have over-wide plug-packs and aren't bothered by the length of a normal board, the Kaifeng KF‑MSD‑4A (and KF‑MSD‑6A) are at least as well-made and cost about half as much as these; you can find BizLine-rebranded versions of them at Rexel.
However I'm not yet aware of a better-made cube, and these are reasonably priced for their more-complex construction, so you can still consider them.

I was at Bunnings for some unrelated (and mostly non-electrical) items, when I came across this:

Overview.JPG (154.63 KiB) Viewed 193017 times

If you remember the original Allocacoc PowerCubes, I didn't have the best outcome with their build quality. Could this actually be better-made?
This version (also available with red or blue socket faces) provides 2.1A USB charging and 4 mains sockets (3 usable when mounted), but as is the case with the PowerCube, you can also get a non-USB version with another mains socket (and black socket faces) for AU$10 less.
They're listed at AU$32 (USB)/AU$22 (non-USB), which is similar to (or even a bit cheaper than) the Allocacoc versions.

Jackson Industries are one of the prominent name-brands for power-related end-user products in Australia (and NZ), competing with Crest (which I mentioned before as providing the lesser-evil among conventional power-boards). My prior impression of this company was so-so, but perhaps this will improve on that…

It's a sphere truncated to flat faces at the sockets, rather than a cube proper; it's also a bit bigger than PowerCubes, at 71mm flat-to-flat (85mm spherical diameter).
Here, the circuit breaker reset button is accessed via a hole at the top of one socket; the mounting base is screw-on instead of twist-lock. This base supports adhesive (though not as easily removable) or screw mounting, but unlike the PowerCubes, the screws are actually included. The hole pitch here is 26mm.
The finish is a more-pleasant matte texture (versus the gloss on the Cubes), but whether the plastic will last as well before yellowing is another matter (it doesn't say what type, and feels most like ordinary ABS to me). The cord is shorter at 0.9m, but at least that's stated accurately (within 2cm); it feels flexible enough.

The casing screws here, while not hidden behind a nameplate, require a notched-blade screwdriver to engage; in my case I took an old cheap screwdriver and cut the notch myself. Depending on the width of the screwdriver, you might also need to grind its sides down to fit. The two above the socket (as referred to their normal orientation at wall outlets) are slightly longer (12mm) than the two below it (10mm), although any justification for this isn't apparent in the casing design.

Outlet block

Inside.JPG (251.03 KiB) Viewed 193017 times

Contacts.JPG (222.69 KiB) Viewed 193017 times

Peering through the packaging, I could see that the contacts are a different (and hopefully more robust) design from the majority of modern Australia/NZ (and many USA/Canada) power-boards: A sideways-wiper arrangement, like that used (in a smaller form) in many IEC 60320 C13/C15(A)/C17 and C19/C21 connectors.
(The prevalent form involves busbar sections stamped to both sides to form a "box" around the pin; contact area of this is decent enough, but resilience less so.)
While no shutters are present here, those in the PowerCube (at least for Australia/NZ) don't work so well anyway, so it's not a great disadvantage.

Neutral.JPG (132.49 KiB) Viewed 193017 times

Earth.JPG (125.16 KiB) Viewed 193017 times

A few of the wire strands were pushed aside where inserted through the neutral and earth tags, although the connections remain good enough to survive 15A without scorching heat (see below).

Clamp.JPG (202.25 KiB) Viewed 193017 times

One of the cord clamp's screws has pressed into the clamp, although it still holds onto the cord; at least nothing's outright broken here.
I added washers under the screw heads (as you can see in the above pictures) to salvage it.

No surge-suppressor is included (nor claimed), but I prefer this to having MOV(s) with no failsafe (especially in a plastic housing).

USB supply

USB-front.JPG (247.85 KiB) Viewed 193017 times

Glue.JPG (246.82 KiB) Viewed 193017 times

USB-back.JPG (294.33 KiB) Viewed 193017 times

Unsurprisingly, this is the weakest part; on top of questionable capacitors, this one also has the bad glue (the PowerCube used silicone here). (Neither piece I see is on or next to a major heat source, but still.) The single-insulated mains wires (UL AWM 1007, 24AWG) are also run somewhat close to the secondary components; some heatshrink there would have been appropriate. It does have a proper fuse (T1A 250V) rather than a fusible resistor, though.

The output capacitors are 2×1000μF 10V 8mm Qunio TM series, rated for 576mA ripple each (although these are slightly shorter than specified); this isn't entirely adequate for a 2.1A flyback converter, especially with the hot rectifier diode (SVM1045V) right on the back of the PCB.
The primary capacitors are 2×10μF 400V 10mm Qunio GR and a 10μF 25V 4mm one is on Vcc for the FT839ND1.
The cramped design provides no room for these to vent, so who knows what can happen then…

There is also some flux residue around BD1 (ABS10), and a solder chunk stuck to the transformer tape.

Load testing
Now that I've modified an MPT-301 to provide current limiting (with a potentiometer adjusting the threshold), I can load the mains outlets with much more accurate current control than by powering heaters through it (and also without wasting electricity, and opening windows or cranking up the A/C, in summer ).

10A load through the active and neutral paths went fine, overloading to 20A tripped the breaker in 4 to 5 seconds. However, there's an additional test we can try:
Passing current neutral-to-earth (to bypass the breaker), so I can overload it in a controlled manner. I believe IEC 60320 (though I haven't seen it firsthand so don't quote me on that ) requires its connectors to survive a 25% overload (12.5A in this case), so I tried that; 2 hours later and it isn't even unduly warm.
Up to 15A for an hour; noticeably warm, but still holding together. (I'm using a rewireable Clipsal plug in the sockets here; molded plugs often get much hotter due to weak connections, especially after repeated strain.)

As for the USB supply, I have five 5.1Ω 5W resistors to hand, so have paralleled two of them with three 100Ω. Measured voltage at the end of the USB cable I used falls to about +4.7V (so about 2A actual), and the nameplate gets quite warm. It seems to work, but as made, I don't trust it long-term. The ripple is pretty bad too, over 150mV at room temperature (slightly lower once warmed up); a poly-mod gets it barely within the equivalent of ATX specification.

Conclusion
The outlet block is somewhat better-made than Allocacoc's, while the USB supply is of about the same quality (middling at best).
So I can recommend the non-USB version (a first for any modern power-board or equivalent!), but will have to leave this USB one with:

HM-200px.png (16.21 KiB) Viewed 193017 times

Still, I was able to recap it (with polys on the output, Rubycon AX on the input) and remove the glue; and also ran an earth wire to the USB output (given my standpoint that any SMPS powerful enough to require Y capacitors should really be Class I).

Next up are some fixed-installation electrical accessories from a small Australian company; will Clipsal be credibly threatened, or remain as royalty?

UPDATE: Non-USB model (PT5711)
I bought three of these online (from DSE since OfficeWorks, the only other reasonably-priced seller, was down to only one in stock).
The end socket, unlike in Allocacoc PowerCubes, is its own unit connected to the main block by wires:

End socket rear.JPG (162.52 KiB) Viewed 191855 times

End socket open.JPG (452.23 KiB) Viewed 191855 times

Versus Crest PWA04984

Contact comparison.JPG (411.52 KiB) Viewed 191855 times

The contacts here are somewhat similar to the conventional type, although made of phosphor-bronze and with curved sections for a bit more spring.
Many (though by no means all) of Jackson's regular power-boards have the same arrangement of curved sections in their active and neutral contacts, although in brass (you may be able to see it through the packaging if held up to light at the right angle).

Bad joint.JPG (96.55 KiB) Viewed 191855 times

End-on.JPG (78.02 KiB) Viewed 191855 times

Resoldered.JPG (93.03 KiB) Viewed 191855 times

Unfortunately, one unit had a failed solder joint on the earth connection (which is a bit cramped to reach in-situ) so I had to resolder it.

Circuit breakers tripped in 28, 32 and 40 seconds with a precise 20A. The cord clamps appear to be a slightly-whiter plastic and no screws have pressed into the hole. All casing screws are 10mm long, and the provided base mounting screws are also a bit different (with some added texture on the head).

Since these are still as good as ready-made options get in Australia now, my recommendation just holds. But if you want to build something really solid…

Top: (2 empty), headphone amplifier, camera battery charger, external HDD, internet gateway. Bottom: Anti‑static point (via Clipsal 463SAS) + Stylux T‑869 (with 4W 4000K LED), heating pad (Sunbeam EP5000), monitor (Dell U2713HM), PC, desk lamp (TaoTronics TT‑DL22), printer (Kyocera P2040dn)

Desk-back power outlets.JPG (411.58 KiB) Viewed 191855 times

…you can follow the lead of 1970s Kambrook: Buy enough twin outlets for the purpose, along with the same number of back boxes (with conduit ports where needed), some conduit for wire protection between the boxes, at least 3m of cable (I used 1.5mm² 7-strand TPS, but probably flexible with ferrules for any future rewiring), a 10A thermal circuit breaker (which I salvaged from a cheap UPS) and C14 inlet. If I had a 3D printer I'd design a frame enclosing the mounting nuts for the IEC inlet, but I don't so I've instead formed plastic sheets (salvaged from PSUs) to provide barriers between the wiring and screws/nuts; otherwise you could use nylon screws (which I did in an earlier iteration), or possibly make an earthed metal plate to put under the nuts (with star washers in-between).
If not mounting it to an existing structure, you'll also have to fashion your own baseboard.