LED Bunker Light Teardown

I recently picked up this LED bunker light for $33 and I thought it'd make an interesting tear down.

AT5700 bunker light

When you remove the diffuser from the front the first thing you see is the PCB that holds the 32 LEDs.  Ideally you're meant to be able to loosen the screws and turn the PCB slightly and then lift it out, the only problem with that is that there isn't enough room to get your fingers down the side of the circuit board.  As it took me a couple of minutes to remove the PCB I'd suggest that in their next design iteration they route some indentations into the edge to allow you to hold onto the board.

Surface mount LEDs

The light is just a low power bunker light for non task illumination and is rated for outdoor and indoor use.

Box specifications

The LED driver is mounted to the back of the PCB and seems to be of the constant current type.  It's hard to tell but I think the voltage range of 12-25 volts indicates that it will regulate the output to 350mA and vary the voltage accordingly.  When measured the LEDs were drawing 350mA at 12V for a power of 4.2W.  The box indicates a power draw of 6 watts so the difference is likely to be what gets wasted in the driver.

LED driver specifications

The light was chosen because of its small size, almost too small in fact.  As we'll see later, the 70mm thickness doesn't leave much space for cabling.

Dimensions

The connector that supplies power to the board has a cover over its soldered terminals on the front side of the panel.  I assume that this is to prevent shorting occurring and not a safety thing as the voltages on the board are low and still accessible on the sides of the surface mount LEDs.  It's a good idea but I wish they'd put the same amount of thought into the installation process as it's quite hard.  Ripping a surface mount component off a board isn't that hard to do.

Insulator over connector solder points

The main problem I have is that the driver is mounted to the back of the LED PCB.  During installation you have to first connect the cabling to the terminal block.  This means that you need to leave about 8 inches of extra wire so you can get a screw driver in there before putting the board into the base and if you can't shove that extra cable into a wall or wherever the light is going, it has to be curled up inside the fitting and that's hard as there is only about 30mm of space behind the board.  I think a better solution would have been to mount the driver to the base of the fixture so that the driver can be wired in first and then connect the LED PCB via the header connector to the driver.  The connector on the PCB could be also moved to the front side to allow connection after mounting the LED board.  There is plenty of room, you just bring a small cable up the gap beside the PCB and connect it on the front side.

Rear of the PCB as it sits upside down in the base

Curious about how the LEDs were arranged I mapped out the layout of the board.  In the image below each different section of copper track has its own colour to clearly identify them.

Electrical layout of PCB

This shows that the 32 LEDs are arranged in 4 serially connected groups of 8 parallel lights.  I would usually have an issue with this because there isn't any form of load balancing, but I think the lights are under-driven enough that it isn't a concern.  I'll take a minute to explain that better.  If the LED driver is generating 12 volts with a constant current of 350 mA, you would assume that each component is getting 43.75mA at 3V, but due to manufacturing tolerances it's unlikely that the 350mA will get evenly split between the 8 LEDs in each group.  You then have a situation where one will be drawing more current than the others and if there isn't enough of a safety factor in the design it will eventually fail.  When this happens, the 350 mA will now be divided between the 7 remaining lights and another will fail this will continue until an entire group fails.  The reason that I'm not too concerned about this is that I think the packages are 2835 LEDs and they can usually handle more than the (0.04375 x 3) 131.25 mW per device.  The forward voltage of a white LED is usually considered to be above 3V.  This is what makes me think that to get the reliability they wanted, all they had to do was add extra LEDs to split the current and under drive each chip.  You would probably find that if there were only 7 LEDs in each group the higher current would cause the forward voltage to increase and instead of regulating at 12V, the driver may go to 13V.

LED schematic

It's not too bad for $33 but I think the design could be improved with only some minor adjustments.

Using a Solder Paste Stencil for a Prototype

If you've seen any of my previous post you'll know I'm in the process of making a PCB that contains a configurable LED Grid.  Why?  Something to do I guess.  Anyway I finally got around to assembling the board I had made by BreadboardKiller.  I've assembled small surface mount PCBs in the past and have always manually applied the solder paste by hand with a syringe.  It's hard to get right, you might not get paste in the right place, you can use to much or too little, but when it's a only a few parts it's not too hard to fix.  This board has 96 surface mount components on it, which means that doing it by hand was going to be a close to impossible.

I decided to get a laser cut Kapton stencil from OSH Stencils.  At 40 dollars it looks expensive for a piece of plastic, but that includes a one off cost for set of board holders the exchange rate wasn't kind either.  These are the black acrylic L shaped pieces in the image below.  The stencil was only 20 dollars US which was well worth it.

There are demonstrations on how to use the stencils online but I'll show my setup.  The board holders are taped down and the stencil is aligned and taped down on on side to act as a hinge.  This allows a board to be put in the holder, have the stencil flipped over it, the paste applied, the board removed, and the process repeated.  Application of the paste is easy.  Squirt some out of the syringe where it needs to go, and use the paste spreader (basically a credit card) to swipe the paste across the stencil.

PCB Assembly Set-Up

The stencils are easy to use and make sure you apply the right amount to each pad.  For this project I chose to use lead free solder as I assumed that I'd get it everywhere when using the stencils.  I was right.  I wasn't sure how to clean it but mild soapy water did the job.

Solder Paste Stencil

Part of Stencil for LEDs

Part of Stencil for Resistors

Stencil label

Once again I used my toaster oven to solder the boards.  It has no automatic controls, I stand there and watch the board and time the steps by counting aloud to myself.  The temperature is also set manually by turning the dial.  The process is described in a previous blog post.  It helps to put a little bit of solder paste on a fiducial mark so that I can see the moment the solder liquefies through the oven window.

PCB in Oven After Reflow

It's important to shield the board from direct IR radiation from the heating elements, that's why there are trays above and below the board.

PCB in Oven after Reflow

After the surface mount parts were soldered the through hole parts were added by hand.

Assembled Board with LEDs

When I designed the board I screwed up and made the holes in the footprint for the terminal blocks too small.  This means they had to be enlarged by hand, and because of the way the tracks were laid out, the positive terminal block has to go on the other side of the board.  No biggie.

Assembled Board

Assembled Board

Surface Mount Resistors

The parts seem to have been soldered nicely.  There isn't an excess of solder or too little, and there's that nice little fillet you expect to see as well.

0.25 Watt 1206 Resistor

The LEDs are harder to judge as the pads are under the board, but they all work.

Osram GW JCLMS1.EC-GUHQ-5L7N-1 LED

Just to prove they all work.

LED Grid

To show the reconfigurable nature of the board I disconnected the jumpers that power the middle section of the LED grid.

LED Grid with Sections Turned Off

Over all I'm very happy with the results.  Obviously boards produced this way aren't going to be of the same standard as professionally made ones, but these are rather sturdy and fine for prototypes.  I'd gladly use OSHstencils again just for the time that it saves me.

Homemade Surface Mount Vacuum Pick-Up Tool

Lately I've been experimenting with surface mount components, trying to better understand how to design for them, how to solder them, and how to place them.  I'm getting better at designing for them and understanding the different soldering methods, but I still have issues placing parts on PCBs.  Using tweezers is okay, but I wanted to see if I got better results using a vacuum pick-up tool.  After all, industrial pick and place machines use a vacuum to lift components, why can't I.  So with that in mind, I decided to build my own, hopefully learning the idiosyncrasies of the process during the build.

Home made vacuum pick-up tool

My design was reasonably basic with only a few requirements.  The vacuum pump needed to be controlled by foot, freeing up my hands to place parts.  Ideally the hand-piece should be about the same size as a pen to make it ergonomic to use.  Finally the hand-piece should be fitted with a luer-lock connector to allow tips to be easily changed.  The rest of the design wasn't too important and was dictated mainly by what parts I could get my hands on.  After an initial failed attempt I'm happy with the results, both aesthetically and functionally.

Hand-piece with power supply, foot switch, and vacuum pump

A quick demonstration of the system's capabilities

The Build

A couple items necessary for the build were a foot switch and a vacuum pump.  The easiest way to get these was to order them through Little Bird Electronics.

The foot switch was sourced from Seeed Studios and would easily do the job.  It's rated for 10 A at 250 V and has NC/NO contacts.  The colour coded wires on my switch were wrong, so it pays to double check the wiring with a multi-meter first.

Foot Switch - Image from seeed studio

The vacuum pump was sourced from SparkFun Electronics and is capable of a vacuum of 16 inches of mercury which is about half an atmosphere.  Ideally if there aren't any leaks the flow rate shouldn't matter as there will be almost no air flowing when parts are attached to the tool.  The pump requires a 12 V, 1 A power source which is supplied by a switch mode wall-wart that was surplus from another project.

Half an atmosphere is more than enough to lift surface mount components.  In a perfect demonstration of the Square cube law, as the parts being lifted shrink, their mass shrinks by the cube of the scaling factor, while the area available for lifting shrinks by only the square of the scaling factor. For example, if the dimensions of a part are halved, it now weighs an eighth as much and has a quarter of the area to lift it.  This means that ideally you would only need half the vacuum pressure to lift it.  This means the vacuum pressure to lift components decreases as the parts get smaller.

Vacuum Pump - Image from sparkfun.com used under a CC BY-NC-SA 3.0 license

A simple cable harness was built to connect the foot switch, power supply, and vacuum pump.

With the main components sorted, all that was left was the hand piece.  After thinking it over for a while I decided to use brass.  It should be strong, but easy enough to work with and join.  It did require me to buy a MAPP blow torch to silver solder all the parts together, but you can never have too many tools.  I also had a fun couple of days figuring out how to solder, only burnt myself once too.

I didn't have access to a lathe or a suitable size of brass tubing, so I had to work with what I could find.  I came across the injection moulded brass body piece shown below on ebay.  I have no idea what it does or what it's for, but it was something I could work with.  I started by cutting the pipe to length by removing the thread on the end.

Brass body

Brass body without thread

The barbed fitting on the end of the pipe was too large for the 1/4 inch ID tubing I had to fit the vacuum pump.  This meant that another fitting needed to be added to the pipe to connect to the smaller tubing.  The thread was cut off a barbed fitting similar to the one below and the barb soldered to the body of the hand-piece.

Barbed fiting

The barbed fitting soldered to the main body

In the close up of the soldered joint below you can see that it's a bit messy.  Silver solder apparently works best when joining two closely fitting metal pieces, it isn't ideal for filling.  I think If I had another go I could do a better job, but after a clean it'll look fine.

The soldered joint

The next step was quite nerve racking.  Up until now the parts used were cheap and easily replaceable, the next part wasn't.  The tiny luer lock connector to be fitted to the tip of the hand-piece cost forty dollars (you can get them cheaper overseas, everything costs more in Australia), and because I'd melted some test pieces when practising soldering I was a little hesitant.

Luer lock connector to hold vacuum tips

Before soldering, the edges of the the two parts were bevelled to make a stronger joint.  I'm not sure if it was required, but it seemed like a good idea.  The nickel plating on the fitting was also removed where it was going to be soldered.

Bevelled fittings before soldering

I bit the bullet and soldered the luer connector to the main body and everything turned out perfectly.  The solder was sucked into the joint and ran all the way around to make a beautiful joint.  After a little filing you can see the result below.  The grip on the luer connector was also sanded off to make the tool more comfortable to use.

Although the hand piece will accept any luer fitting, it's preferable to use a metal one.  If a plastic fitting was used, the air flow through the tip could cause a static charge to build up on the plastic and discharge into the part being lifted damaging it.  If a metal fitting is used, static build-up can be controlled.

Now for some tests.

Metal Tip fitted to the hand-piece

Vacuum tip lifting an MSOP-8 package

I know I said metal tips were preferred, but I have a set of plastic ones that I'm using to understand what size tip is needed to lift what package.

21 gauge blunt tip bent needle fitted

22 gauge vacuum tip lifting an 0603 resistor (1.6 x 0.8 mm)

The rubber suction cup in the images below was from a cheap pickup tool I bought onilne, so I serious doubt it's an ESD dissipative rubber, but they can be bought and when used with a metal tip would create an ESD safe tool.

14 gauge vacuum tip with a rubber suction cup attached

Suction cup lifting a PCB

To securely connect the tubing to the hand-piece an ear clamp was used.  I'm not sure what the tool was that I used to crimp the lugs.  From what I understand, it might be a fencing tool.  Edit:  I've now learnt they are called pincers.

Ear clamp with the tool used to crimp the lugs

I was happy with how things turned out, but the hand-piece needed something more.  It was just crying out to be polished.  It was rubbed down with 800 grit wet or dry sandpaper.  The curved parts were done by cutting strips of sandpaper and wrapping them around the hand-piece.  To finish it off it was rubbed down with steel wool and then rubbed with Brasso.  There are still a couple of fine scratches that I could get out with a finer grade of sandpaper, but I'm satisfied with the end result.

After a quick polish

After a quick polish

Overall things worked out well,  but I still have to tweak the design a little.  When the foot switch is released the vacuum isn't released straight away, which means the parts don't come off the vacuum tip for sometimes 40 seconds.  Ultimately it means I did too good of a job making sure there were no leaks.  I just assumed that the vacuum would release through the pump, but it seals tight when power is removed.  Reducing the length of the tubing would also help as it acts as a large vacuum reservoir.

I enjoyed this project. I made something I'm proud of, got to practice silver soldering, and made a tool that's usable and just what I want.

Cheap Magnifiers For Inspecting Surface Mount Soldering

Over the last couple of months I've been trying to learn more about the theoretical and practical aspects of surface mount soldering, and one of the biggest hurdles I've come across is inspecting the resulting solder joints due to the small size of the parts.  Ideally I'd like some sort of microscope but at the moment I don't have the room for one, and on top of that I'd like to keep things relatively cheap.  With that in mind I've gathered some small, cheap, tools that make the job a little bit easier.

The first inspection magnifier I used was an old 10x jewellers loupe that belonged to my grandfather (shown below on the left).  It allows me to see a lot more detail but it's slightly cloudy and makes some finer detail hard to see.  I decided to buy a selection of new ones from Deal Extreme.  Although buying cheap tools online can be a bit hit and miss, something like this that's been around for a long time, and is a simple design is usually a safe bet.

For under 8 dollars (postage included) I was able to pick up the two jewellers loupes on the right in the image below.  The single lens model is a 30x magnifier and the dual lens loupe has 10x and 20x lenses.  They're a lot clearer than the original loupe I had, and having a range of different magnifications gives me some flexibility depending on what I'm looking at.  They even come in their own little case.

Jewellers Loupes

Jewellers loupes are handy but sometimes you need a little bit more magnification, for these situations I bought something called a pocket microscope.

Pocket Microscope

The pocket microscope has a magnification factor of x100 and a built in LED to illuminate the subject when the microscope is opened, and for 9 dollars I was quite impressed.  The depth of focus is quite small, but the focussing dial allows the focus to be changed easily.  The magnification is quite impressive, for example, the detail on a human hair is just visible.  As an aside, I think this would be an ideal inexpensive gift for any kids that show an interest in science.  Who wouldn't want a portable pocket sized field microscope?

There was however a manufacturing error in the microscope.  The wire that attaches to the battery terminal sticks out and doesn't have any insulation over it.

Battery Compartment with Bare Wire

When the battery was inserted, the wire cut through the protective coating on the battery and shorted the positive case to the negative terminal.  The first indication of this was the battery getting hot.  Pushing back the tab that the wire is soldered to fixed the problem as this lowered the wire at the same time.  Although it should be fine, I don't feel comfortable leaving the batteries in permanently and make sure that I take them out after each use.

Damaged AA Battery

I'd love to be able to show some photos of what I can see, though the eyepiece but am unable to.  I recommend buying some magnifying tools like this.  By no means are they a replacement for a good stereo microscope but they're inexpensive and come in handy in numerous situations.

Soldermask Expansion on DFN Footprints

Recently I've spent some time redesigning the PCB I had made to test out the MCP9808 temperature sensor from Microchip.  I'm satisfied I can solder the small footprint DFN packages without too much hassle.  This then allows me to remove the large footprint components and arrange the board the way I want, so that it satisfies some mechanical constraints.  I basically want to water proof the board and provide strain relief for the cables.

In the process of redesigning the board, I tidied up the footprints and thought it would be a good opportunity to talk about solder mask expansion.  Solder mask on a PCB is there for a couple of reasons, it provides the underlying traces some protection from the environment, but what I'm mainly interested in is its ability to prevent bridges between pins during soldering.  In an ideal world you'd design the solder mask so that it only exposes parts of the board that you want to solder, but due to inaccuracies in manufacturing you need to make the holes in the solder mask slightly larger than needed to account for things like misalignment and shrinkage.  This means that even if there are alignment issues, the whole of the pad is still exposed.  The amount that the openings are increased is called the solder mask expansion.

Typically you'd want an expansion of 5 mil to be safe, but in certain instances you have to have a smaller amount.  For example the DFN component below has no expansion.  The red line indicates the space between pads, 0.2 mm, and the green line indicates the pad width 0.3 mm.  The black area is solder mask, and the light green areas are pads.

DFN Footprint - No solder-mask expansion

In the image below, after a 5mil (blue line) solder mask expansion is applied, all of the solder mask between the pads disappears, which is a valid way of doing things called a gang solder mask, but it's not what I'm after.  I'm experimenting and trying to figure out the pros and cons of different methods.  The dull green area is bare board, the black area is solder mask, and the light green areas are pads.

DFN Footprint - 5 mil solder-mask expansion

To make sure that there's solder mask between the pads, the expansion needs to be decreased.   When doing this, another requirement needs to be met, the minimum-solder mask web.  If the slivers of solder-mask between pads are too thin they'll lift off and cause problems during manufacturing.  A safe minimum value for this parameter would be 5 mil.  This means that the solder-mask expansion in the diagram below has to be smaller than about 1.4 mil (blue line) to allow the black section of soldermask in between the pads to be wider than 5 mil.

DFN Footprint - 1.4 mil solder-mask expansion

Using a solder mask expansion of 1.4 mil is less than optimal.  Ideally an expansion of 5 mil would be better.  Ultimately it all depends on the capabilities of the board manufacturer.  If there is good alignment of the solder mask with the copper layer it doesn't really matter.  I'm considering doing two runs of the board, one with an expansion of 5 mil and one with an expansion of of 1.4 mil, the boards don't cost a lot, so it's a cheap way to learn and get some experience.