You need an IEC cable to plug in a power station, so you go rummaging in your junk drawer and pull one out …
They’re all the same, right? It even says 10A on the plugs! You plug it in, turn it on and the unit begins charging. All good. Three minutes later, the smell of melting plastic starts to fill the air and the cable has turned into jello. Quickly, you reach for the power socket and switch it off. Thankfully, aside from some plastic fumes and a now very hot and gooey cable, nothing had succumbed.
What happened? Turns out, I’ve just encountered a counterfeit mains power cable. The scary thing is, it wasn’t the only one in my collection – I had eight and it was about time they were all retired to the garbage bin so this could never happen again!
Torturing Cables
While, at first, I was a little shocked at how close I’d been to starting (yet another) fire in my bedroom, my thoughts then drifted to trying to find out just how bad these counterfeit cables were. Under very unassuming light loads, they never seemed to produce any signs of distress. I decided to build myself a test rig so I could compare a selected set of cables with different cross-sectional areas to see how they all fared.
Building Test Apparatus
Universally, mains IEC C13 cords are rated for 10A of current, which is coincidentally how much current a general power outlet in Australia can supply. But to perform an experiment with the mains is a bit dangerous, so instead, I decided to do it using a more controlled source of current.
In reality, there is a subtle difference to current-carrying behaviour for AC as opposed to DC due to skin effect, which means that DC is a little easier to carry. So, I enlisted my Aim-TTi QPX750SP power supply to give me 24V at 10A, current limited. That way, at worst, my apparatus might have 240W being dissipated over it – much safer than using the mains where peak fault currents can be orders of magnitude higher than the regular current.
To test the cables, I decided I would run 10A through both conductors – through the active to the far end, looping back on the neutral back to the source.
To do this, I managed to scrounge up a quality panel-mount IEC C14 10A socket from an old piece of retired equipment (another junk box find). I trimmed off the old solder and wires and soldered a thick 2.5mm² short across the two pins – pointed downward to let the jack fit through the panel.
I designed a simple box in Tinkercad (although 5mm too short on one side) to accommodate an IEC socket (although I designed it for a screw-in one as well, as I didn’t know what I had available), a wall-plate for a single GPO (which was an old spare kept from when it was replaced with a double) and two sets of banana connectors (one for sourcing the 10A current, the other for measuring the voltage). I used more 2.5mm² wire for the current carrying leads and speaker wire for the measurement leads. It was tighter than I would have liked due to the socket being placed a little too close to the side – I was trying to avoid the socket and mechanisms that some other backplates have near the centre. But it all did fit …
With some overzealous use of the label maker to ensure nobody plugs this in by accident, I now have a test rig setup for cable torture. Add some outdoor space (for the acrid fumes) and protective terracotta tiles and the venue is sorted. To observe, I’ve got a camera and my Thermal Master P3 watching.
One side to source (in yellow), the other side to measure (in green) and both accepting safety shielded banana plugs. As much as I like these sockets, they’re probably more designed for crimp terminals rather than soldering, as they have a habit of melting and changing geometry while being soldered. I didn’t want to use crimp terminals, because the ones I have are … also somewhat melty at high currents. They don’t have enough metal.
The IEC C14 socket with the warning of it being a dead short. I’m glad the mounting apertures for the panel mount and screwed panel mount varieties are basically the same with the difference of two screw holes. I wasn’t going to go out of my way to buy anything for this experiment. Just in case anyone else discovers this box – I’ve added the warning on the other side for it not to be connected to the mains. That should do it, right?
[Label falls off in a couple of years]
I guess it’s now a breaker finder … whoops.
For compensation purposes, I also used a mains plug which I shorted with a similar length of 2.5mm² cable. This means the only resistances unaccounted for would be the IEC socket and plug assembly. While this means I can’t get the exact wire resistance – I’ll at least be somewhat closer than if I didn’t have this to “null” out the remaining resistances of the switch, mains socket and mains plugs. But this may not be quite 100% accurate due to the different contact materials and geometries and variations from mounting to mounting. I’m also not keeping the temperature of the copper consistent – it will vary in resistivity as it heats up or cools down.
I think I might add a loop of wire to the Earth terminal and add a “flag” label to the loop to indicate it as a dead short and not to plug into a mains socket … lest it also be a breaker finder. But on the upside, with this plug, I could also do some testing of power boards and extension leads.
In all, these are probably boxes that nobody else needs, but me at this very specific moment …
Cable Collection
There must be cables to compare – so here they are:
The master of melt, the counterfeit DHT Tech 0.75mm² cable. This particular cable came from a new-old-stock external 5.25″ USB to SATA enclosure branded Macron. It’s interesting, as shorter length pre-moulded cables have been allowed to go down to this cross-sectional area for a rating of 10A, when formerly, 10A was almost universally associated with 1.0mm². Nevertheless, this rating is, likely, very exaggerated.
But you wouldn’t know it just by looking at it – the IEC plug has 10A on it and the mains plug has 10-16A in one place and 10A in another. You can’t blame me for not knowing!
This next example is a Sun Fai proper 0.75mm² cable.
Stepping up a little is a cable from Hong Shan Chuan for the US market that I cut the head off and replaced with a local plug. I realise that’s not ideal as the colour coding is different and the insulation technically is only rated for 125V (although, in reality, it doesn’t seem to be an issue). Because it uses AWG, it’s cross-sectional area is 0.824mm² which is a little more than the last, but not quite 1.0mm².
Then, I also have a Baoheng which is the traditional 1.0mm² in cross-sectional area.
Finally, a very odd-ball cable that is rare to find – a Hong Shan Chuan cable in 1.5mm². That now means the cable is more of the thickness needed for a 15A cable, but with 10A plugs on both ends. Lower losses are definitely welcome, but it’s not something you’ll easily find.
Results
I decided to take the voltage of the shorting plug and subtract that from the voltages registered for each cable after a minute of operating. I realise this is going to result in the copper being at different temperatures for each cable (thinner cables lead to higher dissipation leading to higher temperatures which causes increased resistance). From there, I also computed the measured resistance (which will include the IEC plug/socket resistance) and the power loss per meter of cable. From there, I used the annealed copper resistivity to back-calculate the expected cross-sectional area and expressed it as a percentage of the actual rating on the cable.
To be clear, because of the methodology, this test does not measure the resistance of the cable alone. The resistance figure also includes a contribution from the IEC plug/socket as I didn’t want to just destroy the cable and test the wire itself. Therefore, the lower calculated cross-sectional area is likely a result of further additional resistances that are not accounted for, such as the wire-to-pin resistance as well.
But it would seem that my genuine cables are registering about 55-65% of the calculated cross-sectional area. The fake cable registered just 3.19%. Its losses were 10x greater than the next closest lead. With its resistance-per-meter figure multiplied by 0.6 (to compensate), it was performing like an 0.23mm² (AWG 30-31) cable. Perhaps it actually is a 0.5mm² cable just heated up high enough (over 220 degrees C) that it increased its resistance accordingly – based on the temperature coefficient of copper, a 200 degrees C increase would increase the resistance by a further 78.6% or close to double.
If you want to see the result – there it is in the video above and the photos below. After the smoke started to slow and it seemed that a short was going to be inevitable, I turned off the power remotely.
Melted insulation turned glossy, while inside, insulation had burned into a crumbly, bubbly char.
The printing on the cable is distorted, while internal wire colour insulation can be seen melting together with the outer sheath.
Watching the video, it’s clear the cable drooped as it heated up – but the plugs remained resilient. This suggests the plugs themselves may indeed be 10A rated.
Once cooled, it was no longer “jello” but more of a stiff, brittle, crispy mess. Where the cabled had crossed, the heat concentrated melting the two insulation together. While proper cables under load don’t usually get this hot, if you keep leads coiled or too many leads under load in a confined closed space, they could overheat and fail similarly – so uncoil your extension leads when in use!
A slight bend and the copper sticks out of the charred insulation. As far as I can tell, it looks copper coloured and is quite flexible. It doesn’t seem to be CCA (copper-clad aluminium) as some other fake leads have been found to have.
Wondering what the burn mark was on the face-plate? I took the picture of the box after the test – indeed, this cable had left its mark.
I decided to cut a nice “core sample” using a sharp knife and compare it with a different 0.75mm² lead from an old appliance. The difference in the copper area is very astonishing.
Zooming into the maximum magnification and having the two side by side, I used Photoshop to measure the area, cutting out some of the voids in the right space where the strands did not seem as tightly packed together. Based on the counted pixels, the estimated area seemed to be a half – so ~0.375mm² (AWG 27-28) for the counterfeit cable.
One thing to note is that the counterfeit cable is about 1mm thinner compared to a proper 0.75mm² lead. It’s also a bit lighter and feels a bit more flexible owing to less copper being inside. But it would seem that, in an effort to reduce costs, they couldn’t even spare enough cheaper plastic to make it the same diameter.
Conclusion
It’s a fair assumption that interchangeable leads can be used interchangeably especially when the ratings on the leads are able to meet the ratings of the source and appliance’s demands. But that assumption is definitely challenged by counterfeit cables. It would seem that what is printed on the cable simply cannot be trusted. If it had been used in its original low-power application, it might have never caused any major trouble, except for increased power losses. Thankfully, I was around when the cable started turning into jello and became something I could smell. A swift move to cut the power before conductors finally touched prevented what might have been a brief “bang” and the conductors fusing open, but not before filling my room with acrid PVC smoke. The actual thickness of the copper in the counterfeit wasn’t conclusively determined – the microscope suggested about 0.375mm² (AWG27) while the compensated resistance test suggested 0.23mm² (AWG 31) but without accounting for the wire temperature, which may have indicated up to 0.41mm² (AWG 26), so there is definitely some overlap in results by different methods.
Of course, I’m not the first to be burnt (literally) by this sort of thing – digging around the internet, it seems that one other reference to DHT Tech cables cropped up in 2013 on the Hardware Insights forum, posted by a fellow Australian. While my samples definitely came from new-old-stock external drive enclosures, it would seem that such products might continue to haunt people into the future especially as these are the sorts of cables we would usually save and toss into the junk drawer “just in case”.
This also brings me to an observation – an increasing number of products that have IEC leads bundled but with the strict instruction to “only use supplied cable”. In some cases, this is necessary for regulatory compliance – especially for cables with internal shielding or with a ferrite suppression bead in-line, as that usually indicates a potential issue with EMC in the case of using a plain mains lead. But other times, it may simply be to avoid potential brand-risk in case someone uses a counterfeit cable and it burns up. After all, I’ve not seen (proper) IEC leads not be rated the full 10A, but perhaps it’s also the difference between 10A “peak” loading from heavier motorised or cooking appliances which have varying demands which might be just fine on 0.75mm² or 10A continuous loads (which heavy power stations, electric cars may be doing) which are better off with 1.0mm² cross-sectional area cables – after all, in the past, it would seem that 1.0mm² was the standard until copper became expensive.
But if a mains cable seems a tiny bit thinner than expected, or a tiny bit more flexible, or a tiny bit lighter … it could be a counterfeit.
Bonus: Site Update
Due to unavoidable VPS infrastructure maintenance, older images may be missing from blog posts for a ~20 minute period starting on 10th May 2026 13:50 UTC (or 11:50pm Sydney Time).
A little over a week later, the main site will be expected to be down for a short period (~20 minutes) on 19th May 2026 11:25 UTC (or 9:25pm Sydney Time).
It is hoped everything will be back up and running soon after maintenance completes, assuming no disasters occur. However, if the site goes down for a while longer, rest assured – I’ll be doing whatever I can to bring it back up.
Your patience is much appreciated.