#104 Reverse engineering an old linear power supply

Back panel connectors Antronics made by TPW

Recently I came into possession of two working ups devices from 1993. both of them had old capacitors and old 12v7a lead acid batteries inside the devices.

First thing I did was clean the cases and the PCB boards. Once that was done I replaced the old capacitors and the 12v7a batteries, then I tested both devices. both work fine but the design is old and a bit dangerous.

Secondary side
Primary side

So I decided to reverse engineer the circuit in order to better understand the design and to see if I could make any improvements to a design I would like to make.

While reversing the PCB I noticed that the mains earth and the GND of the circuit were connected together. I also notices sone discoloration from what looks like heat between the regulator and the transformer. Also the 330 ohm resistor for the led appeared to be discoloured from what also looks like excessive heat.

Mirrored for reversing
The original schematic I reversed
Schematic after I implemented suggestions

With these issues in mind I also noticed that the heatsink for the LM317T was very small and close to the transformer and the mains 1A fuse was placed after the choke and varistors instead of before them.

In conclusion I decided to choose between a different regulator at a fixed voltage or a chain of 4 LM317Ts providing around 6A of peak current, Since I do not need to adjust my voltage like the original circuit I should be able to get 13.75v by using a fixed 1k and 10k resistor. I also wanted better heat dissipation and Amps so I will definitely install good heatsinks with thermal compound. Depending on the size of the enclosure I get for the project I may add a fan.

The project files and components list etc. can be found on my Github here.


#103 Dangerous BK-357 USB charger teardown

Listing on Takealot
Claimed output…

While looking for a new multi port USB charger I came across the Model BK-357 sold on Takealot by OQ Trading. This charger had many positive reviews with a 4.1 star rating at the time of writing this article and for the low price of R149.00 I had to give this device a try.

Caps could potentially be repurposed from e-waste…

Once I received my charger I noticed That the fast charge USB port was working flawlessly but the 3 normal charging USB ports seemed to have current divided between them.

So I decided to open the charger to investigate further. I noticed the bottom part was glued into place and could be pried open carefully with a small screwdriver and spudger.

Primary and secondary sides with a thin line in-between…

Once opened the PCB was in good condition and contains 2 small switching transformers and a nice fusible resistor that also acts as an inrush limiter but that’s where to positives end. The interference capacitor was skimped on also the electrolytic capacitors are all different colours and brands It’s possible that they have been taken off old junk and re-purposed which is okey but they may have a diminished quality which is almost as alarming as the gap between the primary and secondary sides of the transformers. The biggest gap is around 4.8mm which is not to bad but right in the middle the gap closes all the way to 1.3mm!!!

Largest gap is around 4.8mm and the smallest is a whopping 1.3mm!!

This is very dangerous since there is 1.3mm of PCB space between you and mains voltage!!!

I will be posting my findings as a review on Takealot.

When purchasing multiple socketed USB chargers go for the larger more expensive ones.. As you can see in this case the tiny transformers just can’t output enough current on the cheapies.

The article from Europe’s Safety Gate Alerts can be found here. They identified the problems and measured the charger all the way back in December 2021 in Ireland and here in SA we are happily selling these.

I have also included an article from a Russian review back in March 2021. where The charger was also analysed and determined to be a potential hazard the article can be found here.

1x fast charge 3x normal charge with current divided by 3


#98 Commonly available MOSFETS in South Africa

When looking for affordable and commonly available MOSFETs in SA I came across a few candidates.
Now there are mostly big tradeoffs with N-channel logic level MOSFETs however, I find that most of the time I use them instead of relays for slow switching applications.
Most of my projects have not been using PWM so I have not had any issues using the logic MOSFETs . And I find that I use the N-channel MOSFETs almost all the time.

“real fet” characteristics
“fake fet” characteristics

Finding MOSFETs with decent specs for a decent price was quite tricky but I found a sweet spot with the IRL520N now there’s always the chance of getting fake chips and I might have fallen victim to this but the “fake chip” had specs that were quite close to the “real chip” in a comparison I did between two chips. The fake one also had a slightly larger and glossier form factor.

“real fet” left and the “fake fet” right

These are my top 3 N-channel MOSFETs which are common

  • IRL520N
  • IRLZ44N
  • IRFZ44N

I would like to add some honorable mentions as well. These N-channel MOSFETs are either not common stock with the online stores I use or are way to expensive for a hobbyist however sometimes they are mandatory requirements for specific projects.

  • IRF3710
  • IRF540N
  • IRF3205


#92 Repairing a D5 EVO PCB board

D5 EVO top enclosure
Big burn inside enclosure with semi melted screw.
Back of IR beam PCB board compared with new PCB
Burnt IR bean PCB positive in diode

After a huge thunderstorm I noticed that the IR beams on my old electric gate were not working so I decided to take a look at the IR beams connected my electric gates poles.

After opening the IR beams enclosure I was greeted with a burnt PCB.

Upon closer inspection I was able to determine that after a lightning strike the bolt flowed through the pole then through the screw inside the enclosure and then from the bolt to the IR PCB board,
The board fried and then the bolt transferred from the IR beam PCB to the D5 EVO positive output PCB terminal and blew up a SMD power mosfet above the 12v relay.

Full top of PCB board
Close up of burnt mosfet.

Strangely enough the D5evo PCB could still open and close the electric gate it seems that the only thing affected by the lightning strike was the output power terminal and the destroyed mosfet (55L104 N-Channel)

I had to purchase a set of two new IR beams for R950.00 but I decided to see if I could repair the D5EVO PCB myself. I decided to use an IRL520N N-Channel mosfet since this was commonly available at the time. Unfortunately I could only find the TO-220 package so I had to bend the mosfet a bit but it worked out in the end.

IRL520N N-Channel mosfet
Datasheet comparison
A few value differences