Learn how to recover from a mistaken chip order and improve SMD soldering techniques using a mini hotplate and vacuum cleaner stencil holder. Watch the mesmerizing results under the microscope!
[0:00] I messed up big time. There’s no way this 24
[0:03] pin chip is going to work on a pad that only has 20 pins.
[0:07] How did I get this so wrong?
[0:09] I’ve been working on my jumbo seven segment display. One of the issues with version one
[0:14] is that the main PCB is really big which means it costs quite a lot to manufacture. I’ve been
[0:19] working on a modular version of this with PCBs that can be daisy-chained together.
[0:23] This would let me have a lot of digits on my display and be very cost-effective.
[0:28] The boards are based around the TPIC6595. This is a really cool device, it’s a shift register
[0:34] with open-drain outputs that can sink a large amount of current.
[0:38] This makes them ideal for switching a large number of high power LEDs.
[0:43] So, what actually went wrong?
[0:45] There is nothing wrong with the schematic,
[0:47] I’m using a TPIC6595 and this is the correct symbol.
[0:51] We can also check on the datasheet and we have all the pins correctly labelled.
[0:56] Well, it turns out these chips are actually quite hard to get hold of in a surface mount package.
[1:01] On Mouser, there’s no stock until November.
[1:04] And on RS it doesn’t even show up in SMD format.
[1:08] There is a variant of the TPIC6595
[1:12] a TPIC6A95 that has the same functionality
[1:16] and seems to be more widely available. This chip actually has a higher current rating
[1:21] in hindsight, I should have probably used this version in the first place
[1:24] as the higher current rating means my LEDs could be much brighter.
[1:29] Somehow during ordering, I managed to confuse this version of the chip with the TPIC6595.
[1:35] Unfortunately for me, it has a completely different number of pins and pinout from the TPIC6595.
[1:42] There is a 20 pin version according to the datasheet, but it still wouldn’t have worked as the pins are different.
[1:47] In some ways, I’m quite glad that I ordered the 24 pin version as I would have happily
[1:53] soldered up the 20 pin version and then spent ages trying to troubleshoot why it didn’t work.
[1:58] Somewhat annoyingly I didn’t notice my mistake until I’d already added
[2:02] solder paste and placed a bunch of components.
[2:05] At this point, I must apologise to anyone with OCD for my haphazard
[2:09] resistor orientation. Normally I would be more careful, but I got a bit overexcited
[2:14] about trying out my MHP30 mini hotplate - I promise to be more careful next time.
[2:19] It was only when I tried to put the chip on the board that I realised something was very wrong.
[2:24] I’ve been hunting around trying to find a supply of SMD versions of the TPIC6595
[2:30] and I’ve not had much luck - I guess this is all part of the global chip shortage,
[2:34] so since I already have them in my hands, I’m switching my design over to the TPIC6A595.
[2:43] As always, the PCBs came from PCBWay and they did a great job.
[2:47] Fortunately, I ordered some other boards that don’t have any problems so I’ll be
[2:50] kicking off some new project videos soon - so don’t forget to subscribe and hit the bell.
[2:56] Fixing things is pretty simple.
[2:58] There’s no symbol in KiCad for this chip so I ended up duplicating the
[3:01] existing symbol for the TPI6595 and updating it to match the datasheet for the TPIC6A595.
[3:09] There are 4 more power ground pins to fit in so we just need to make the symbol a bit larger.
[3:14] In hindsight, I could have just put all the
[3:16] pins on top of each other and kept the symbol the same size.
[3:20] Updating the schematic is pretty painless, we just need to shuffle
[3:23] things around to make a bit more space. Everything else stays exactly the same.
[3:28] Laying out the PCB is also quite straightforward I had a reasonable amount of space to play with
[3:34] so the longer chips fit without any problem and we just need to route the traces again.
[3:39] Following the advice of one of my fellow YouTubers this time I’ve printed out the PCB at 1:1 scale
[3:45] and I’ve confirmed that the chip actually fits on the pad.
[3:49] It’s an annoying mistake to make, I had two different boards manufactured using this chip and
[3:53] made the fatal mistake of doing a copy and paste, so the error propagated across both of them.
[3:59] But, since I have some boards that I can’t really use I thought I do some experimenting and learn something new.
[4:05] I’ve seen a few of the cooler kids on YouTube using a vacuum cleaner to
[4:09] hold their solder stencil over the PCB and I really wanted to give this a go.
[4:13] I’ve created my own version,
[4:15] the vacuum cleaner attaches to this adapter and the PCB drops into this slot.
[4:20] We just need to align the stencil over the top of the PCB
[4:24] and make sure all the pads are aligned with the stencil holes.
[4:27] I can then use my foot to switch on the vacuum.
[4:30] This holds the stencil really firmly against the PCB so there is no danger of any gaps.
[4:36] It’s really well stuck and you can’t move it at all.
[4:40] We can now just spread the solder paste over the stencil.
[4:43] I’m not entirely sure I’m using the right tool for the job here, and
[4:46] Any suggestions in the comments.
[4:48] I need a bit of practice. At the moment it feels like my technique is a bit poor.
[4:53] But I think it will improve over time
[4:55] With that done we just turn off the vacuum and we can lift up the stencil mask.
[5:00] Doing a quick check under the microscope it doesn’t look too bad.
[5:04] I didn’t have things perfectly lined up and I think I smudged things a bit when
[5:07] I lifted off the stencil. I’ve seen a more sophisticated version that has springs to
[5:11] lift the stencil up when the vacuum is released. Maybe I’ll try it out.
[5:16] Adding the components is a bit fiddly - especially when you’re hands are not very
[5:19] steady - maybe a DIY pick and place machine should be added to the list of projects.
[5:24] Or maybe I need to make sure I’ve had breakfast before trying to do fiddly work.
[5:28] Let’s try out the MHP30 mini hotplate. This is a lot smaller than I was expecting,
[5:33] but the board does balance nicely on it.
[5:36] I’m going to heat the board up to 220 degrees and see what happens.
[5:40] It works really well - I love watching the components move
[5:43] into position - it’s quite mesmerising.
[5:46] The MHP30 is pretty small, but you can shuffle the PCB around to get all the components soldered.
[5:52] One word of caution - the PCB does get hot. So don’t use your fingers to move it around.
[5:58] Looking at the results under the microscope we’ve got some good results.
[6:02] I can see a few solder balls along the sides of the resistors,
[6:05] but these come up very easily when we clean the board up with some IPA.
[6:09] Looking at the side view of some of the resistors I think this looks great.
[6:14] It’s certainly not as good a reflow oven, but for my little projects,
[6:17] I think it will do the job. I’ll be using this for a few projects
[6:20] and I think as I gain more experience my SMD assembly will improve dramatically.
[6:25] It’s definitely better than my attempts at my
[6:27] hand SMD soldering which you can watch in this video coming up now.
[6:31] So, quick update. I managed to score some of the SMD TPIC6595 chips on eBay - I’ve
[6:38] no idea if they will actually work, but I’ll give it a go. I’m going to have a
[6:42] lot of these modular boards to play with - I’m sure I’ll think of some good projects.