Making moulds using CNC, with urethane foam, wax, silicone and epoxy

I’ve been doing a bit of work on making moulds for the injection moulder. I’ve been making the base moulds with urethane foam and machineable wax. The was is definitely better since the foam is too grainy. I’ve ordered some urethane boards that will probably be really good for this, but they haven’t arrived yet. If you’re interested in some pictures of the process, read on. I’ll be posting some videos later on.

Continue reading Making moulds using CNC, with urethane foam, wax, silicone and epoxy

T-962 Reflow oven

I’ve wanted a proper reflow oven for a while. The hot-plate works well, sometimes together with the hot air gun. But for some projects it would be better with an oven. Many people build their own, but while I’d love that project, getting hold of an oven here in Norway would end up being almost as expensive as this cheap alternative: T962 from Puhui. I did a lot of googling before buying one, and it has some bad reviews. In fact I’ve found it difficult to find positive reviews. But I bought it anyway, and so far it seems the reviews are wrong and it works fine. The reviews do indicate that the soldering profile is slow (the watch is not real-time), which seems correct, but the oven works well. I’ve only used it a few times, but except two tomb-stoned resistors (attached on one end and lifted up so it doesn’t attach on the other end), I’m happy about the results. I think the tombstone is probably just a matter of tuning the profile + stencil, and since I’ve only used the default profile No. 1, I can’t say I’ve tuned it to perfection. My first impression is at least that I can recommend this product and that the reviews are not balanced. It seems many of the reviews are people who bought it from ebay or aliexpress, but I bought it straight from the company website, which might be better. It arrived here 4 days after I ordered it, shipped with DHL.


Using a laser cutter when etching PCBs

No, I can’t use my laser cutter to actually remove the copper on the PCBs. Unfortunately. Well, I haven’t tried as I expect it won’t work. But I can use it to remove the photoresist, instead of using a transparency and UV box. I’m not sure it’s simpler, but I wanted to give it a go anyway.  Here is a picture of the board after I’ve ‘engraved’ it using the lasercutter.

And here is a picture after it’s been etched, the normal way, using FeCL. The FeCL solution was a bit old, so the quality of the etching is a bit uneven. It shows promise, though I must admit I’m increasingly using professionally produced boards. But for a quick prototype, this is quite simple and fast.

Manufacturing boards using surface mount and a hot plate

I thought I’d write a little note and post some pictures of how I’m assembling the LCD display boards. The PCBs are manufactured at Seeedstudio and are really nice quality. I’ve tried to use as many surface mount components as possible, but have some through-hole also (headers etc). The LCD itself is also surface-mount, but I solder it on manually, not using the hot plate.

I basically use solder paste in a syringe and manually apply solder paste on each surface mount point.  It would probably have been slightly faster using a surface mount stencil, but using a syringe is also quite quick and not as messy as a stencil.

I then place the surface mount components onto the paste. In my case these are just simple capacitors and resistors, so it’s not very complicated. I used 1206 size originally, but have now started using 0805 size capacitors.

The board is now ready to be fried. There are several ways of doing this. Some people use a reflow oven that surrounds the entire board with heat, but I find that if the surface mount components are only on one side, using a hot plate such as this is the easiest. Not sure what the pros and cons are of each, but Sparkfun have also recommended using hot plates. I use a thermistor to control the heat, but since my hot plate has a thermostat I just use the thermistor to guide me towards the best setting.

Here’s a picture of some of the boards on the hot plate. I can do several at the same time.

The whole process of applying paste, heating them on the hot plate and then manually soldering on the LCD means I use about 1 hour on 10 boards. Of this, just doing the surface mount takes maybe 10-15 minutes for 10 boards. It’s definitely faster than manual soldering.  I might get more efficient as I move along. Right now I’m also soldering on headers on all the boards as it makes it easier to test, but I won’t do that in the long run. I’m shipping 14 boards to the store tomorrow. I will keep a few to sell here on the site. Unfortunately, Arrow have told me the remaining batch of 222 LCDs is delayed until July, so if you want to test a board early, it might be good to get one of this first batch.

Here’s a zoomed in picture of one of the boards, before the solder paste has melted. The picture also shows my silk-screen error (P1.4 is really P2.0). Other than that and a slightly inefficient LED-connection, the first batch of boards seem to be pretty ok (the points that need to be connected to turn on the backlight are a bit far apart, but no big issue). The next version will also have the SMT components on the bottom, as it looks nicer.

And here’s a picture of the LCD connector soldered onto the board. I do this manually, but it’s pretty easy when using flux to ‘guide the solder’. You can barely see the liquid flux on the board around the connector.