↓ Skip to main content

Modding a Cheap Uv Lamp for Fun and no Profit

·1141 words
Tools Reverse Engineering Upgrade
Author
Werti
WTF am I even doing?
Table of Contents

Checking out the lamp
#

I have a cheap UV lamp that I want to use to cure UV-resin for PCB repairs.

Here should be a image of a 'Cheap NoName UV Lamp'
The UV lamp features 6 LEDs of unknown output

While the lamp was very cheap, it does indeed output UV light, as the resin cures. The lamp has a large issue however. It gets hot! This would not be a huge problem, if it would not turn off roughly \(20\,s\) after pressing the power button.

Here should be a image of a 'Cheap NoName UV Lamp turned on'
The UV output reacts differently with two types of paper underneath the lamp

The reason why I believe this shutdown feature was incorporated is the high heat generation and the costs that would be associated with a reasonable thermal design. The PCB is of the aluminium backed type, but that is not enough to carry away a lot of heat, as I will see in the thermal camera image.

Here should be a image of a 'Cheap NoName UV Lamp PCB'
The PCB of the lamp does not feature a lot of things and the modifications i have in mind should be simple. The back side is aluminium, so it will cooperate nicely with an additional heat sink.

Getting temperature under control
#

Lets measure how warm this device gets after one operational cycle.

Here should be a thermal image of a 'Cheap NoName UV Lamp'
Thermal situation after one power cycle. The hottest part is an LED with \(55.3^\circ C\), which is quite toasty for just \(20\,s\) operation.

I looked into my box of heat sinks and found one that has an area larger than that of a cpu dye from a low power PC that will fit the PCB of the lamp perfectly. It also has an active cooler so for very long runs, this might help.

For testing I ‘mounted’ the PCB to the cooler with a simple clamp

Here should be an image of a 'UV Lamp PCB clamped to a heat sink'
Heat sink and PCB clamped together without thermal compound for simple test if this is even remotely promising.

Well the results speak by themselves! The LEDs are not even registered as the hottest part anymore. The hottest point is now only \(32.3^\circ C\) and likely a reflection of my hand on the metal of the USB port. This is HUGE! The temperature delta is more than \(20.0^\circ C\)!

Here should be a thermal image of a 'Cheap NoName UV Lamp PCB clamped to a heat sink'
Thermal situation after one power cycle again. The hottest part is not an LED anymore!

Understanding and modding the circuit
#

Ok, now that I can say the temperature problem is ‘solved’, lets look at the electronics

Here should be a schematic of a 'Cheap NoName UV Lamp PCB'
Reverse engineered PCB. The six LEDs are hooked up in parallel and are fed 5V from USB via two parallel \(1\,\Omega\) resistors. SW1, IC1 and Q1 are part of the timer. IC1 is not identifiable as it has no markings.

Luckily, the schematic is very simple and we can easily modify for continuous operation by just removing the parts responsible for the time switching and bridging the pads of the N-channel MOSfet. Doing that will result in the device being on as soon as power is applied to the USB port. The following schematic will be the result.

Here should be a schematic of a 'Cheap NoName UV Lamp PCB'
Modified schematics without SW1, IC1 and Q1.

All SMD parts but the LEDs and the two \(1\,\Omega\) resistors need to go. I realize, that there is not a lot left after this mod and I could have build this from scratch as well. Keep in mind that this thing is already produced, so i might also put it to use instead of being the reason for more e-waste. Aside from that, the circuit board is well-suited for modification thanks to its aluminum backing and simple electrical design.

Here should be an image of me using a hot air to remove parts from the PCB
Removing everything but the LEDs and the two \(1\,\Omega\) resistors.

After the modification, I soldered a wire to LED+ and V- each to power the whole thing directly with \(5\,V\).

Here should be an image of the PCB with soldered on wires
Adding two wires for supplying power.

The image shows the bridged pads from the transistor. Operating the board in this state results in a power consumption of approximately \(10\,W\).

4.99V1.85A9.23WINPUT: 67.85VV-SET: 5.00VI-SET: 2.58AOVP:   62.00VOCP:   24.20A

Finishing with a housing
#

To get things together, I threw up a quick design in FreeCAD. The design should feature:

  • Clamping pressure on the LED PCB to the heat sink
  • Letting the light through (kinda obvious)
  • Place for a switch (i found a two level switch in my arsenal)
  • Place for a DC-DC buck converter
  • Place for a USB C PD receptacle
  • A tripod thread mount
  • Should be FDM printable
Here should be a rendering of a housing
The Housing was quickly drawn up in FreeCAD. Part of it is a hollow box with a screwable lid to house switch, buck converter and USB port. The hollow parts in the back have the purpose of fitting around the two remaining resistors, the capacitor and the pads, where the wiring terminates on the PCB.
Here should be a rendering of a housing
On the front I tried to make the openings for the light as open towards the outside, while having material on the other side to press down the PCB onto the heat sink via the 3D print.

Ok The print settings were less than optimal. I have left the slicer in the draft setting as i wanted a quick result. Especially when printing the hefty chamfers around the LEDs towards the bed this was not a good idea.

Here should be a rendering of a housing
Oof. This is rough and looks almost comically bad compared to the renders…but hey it does work fine, so this is good for now.

As this will be a one-off project, the electrical work inside is hand wired and fixed in place with hot glue. The input is a module with a USB C port and an IC that is requesting \(12\,V\) via USB PD. After the input, there is a two way rocker switch with a diode between the output terminals. This way, the light is on in stage I and II but the fan only operates in stage II. Up to this point the Switch is only switching the \(12\,V\) coming in via USB PD. Because that would fry the LEDs, I have put in a mini360 DC-DC Converter, set to \(5\,V\). It is the same DC-DC converter as I used in the RICOH Theta SC revival.

Here should be a image of a 'The inside wiring of the lamp'
Does not look perfect, but this will work for a tool, that I wont need every day.

I have added a \(\frac{1}{4}"\) screw terminal to the bottom so i can hang the lamp over a pcb with UV curable resin with a magic arm.

Here should be a image of a 'The finished lamp'
Yay it works!

And now the only thing left is a test on the resin i was doing this whole thing for.

Here should be a image of some 'cured resin on a random PCB'
The resin cured fine on a random PCB in about five minutes.

So the quick mod works for the few PCB repairs i need to do from time to time. If i will ever modify this thing further, i will definitely reprint the housing with more appropriate settings. Keep in mind, that the five minutes needed for curing would have required 25 presses on the power button and me being present for that every time.