I had to wait 5 weeks or so to get the parts and the PCB fabricated and then it just arrived today so was really excited to solder it all together as it’s the final piece for the microscope (just doing it as manual so no Sangaboard or motors).
Unfortunately I had real trouble with the flush surface mounted LED, it was nigh impossible to solder to the board cleanly and when I did manage to solder it, it appears to be non-functional.
When plugged into 5V via a Raspberry PI I get no light. I traced the voltage with a multi-meter and it seem to be stopping at the LED. I think I may have destroyed the LED while attempting to solder it to the board.
It’s definitely possible that the LED is dead, but are you sure it is the right way around? You can use the 5V line and a series resistor (or a power supply with an appropriate current limit) and try to power the LED directly while it’s still on board (in both polarities to check if that is the issue).
The easiest way to solder these components is probably with solder paste, you can use hot air (just don’t blow the components away) or even a clothes iron to reflow it. Place a small amount on the pads (easier if you mix the paste with a decent amount of flux first) with a needle or a toothpick, put the components on, and apply heat until the paste melts. Don’t worry if the paste bridges the pads, unless there’s a massive excess it will separate by surface tension/solder mask interaction when melted.
The footprint for the LED is very tight. It can be hand soldered but its tricky. For the next rev of the LED board, we might want to increase the foot print ever so slightly to leave space to sneak in a soldering iron tip.
That wont have a negative effect on reflow soldering and it will make hand soldering much easier.
What part of the world are you in Mike? Maybe I can send you a completed one. I made a couple to use with the 5V, but it seems I will be going to the constant current driver so I will have some 5V spares.
The 5mm LED is easy to deal with by hand and uses minimum components that builders are very likely to have anyway. It performs well in the microscope.
I think we leave the PCB for manufacturing services. There are a relatively small number of people who have the techniques for SMD soldering using a hotplate or hot air soldering station who are their own manufacturing service.
Hi @mmca, oh wow so you were able to successfully solder the LED on? You must have a steady hand, nicely done. I’m based on Brisbane, Australia, how about you?
Okay @WilliamW, thanks for the context. I think it would be useful to update the documentation surrounding the electronic components to set expectations on this. It’s not apparent until you get the parts and try to solder them that it poses such a challenge and that the LED workaround might be the preferred option unless you can find a vendor to produce the entire board for you.
I’m going to have another attempt with soldering the surface mounted LED once I get the new parts. I bought 20 LEDs this time, so plenty of room for failure, and if that fails completely, I’ll likely go ahead with the LED workaround.
Anything beats what I’m currently doing which is shining my iPhone light down the hole!
Im in Los Angeles, USA so wouldn’t make much sense to ship out one PCB that far.
Have you ever been to https://hsbne.org/ ?
Maybe you can send them an email and see if you could use their hotplate or hot air for a couple minutes.
I’m sure they would be interested in seeing the OFM as well.
Hi @mmca no I’ve never heard of that place! Thanks for sharing.
The replacement LEDs came last night and with some patience I was able to repurpose on of the circuit boards I had and managed to solder it on!
Here are some of my first pictures one of red blood cells and another of a pine leaf!
Thanks for giving me the confidence to continue knowing you had also managed to get the LED on. It’s definitely tricky but with enough patience it is possible, even with a thick soldering iron and solder.
What are you using for the diffusion material? I still haven’t found one I really like.
A buddy is going to give me one of his broken LCD monitors, I am hoping that the diffusion stuff in there will do the trick.
@Mike those are nice sharp pictures. There is quite a colour cast which does not look as though the camera calibration has worked well. That is most usually when there is not quite enough light. You can run the camera lens shading / white balance at any time using in the settings in the web app. There should be no slide in place for that calibration.
You should get an almost completely even white/grey across the field of view. The yellow graded to pink at the edges of the background is what the low cost optics tend to look like before calibration. If there is not enough light you sometimes get a mottled pink/green background.
@mmca the diffuser material given in https://build.openflexure.org/openflexure-microscope/v7.0.0-beta1/parts/materials/white_polypropylene_sheet.html is what we have tested. As it says in that page ‘any white plastic sheet should do provided it’s thin enough to be translucent but not transparent’ . I have done some brief tests with printed white PLA either from support or brim material from 3D prints, or a specific printed disc, 2 layers thick. They seem to work, but I have not done any formal tests. It will primarily effect the evenness of the illumination across the field of view, it might also affect the resolution for very high resolution imaging (more than 40x objective lenses).
Awesome thanks @WilliamW, I wasn’t using the OpenFlexure software for those first shots, just the built in libcamera command so will flash an SD card and give it a go to get better results with respect to colour and lighting!
Thanks for the link on illumination that’s perfect!
