I have read through some threads about possible issues using Watchdog timers to trigger a reset on Borons. The threads are quite dated, so I thought I would stick my neck out and ask again in hope things have changed since then
I am looking to implement a STMicroelectronics M41T65 to reset a Boron should the need arise. It seems better practise would be to get the Watchdog to rather power cycle the device as apposed to reset. I will also use the RTC on this device to wake up the Boron every 24-48h. Is there any merit to this still, or has all ‘issues’ been resolved?
There is a Particle application note, AN023 Watchdog Timers that shows a few options for watchdog timers. There is an example of using an AB1805 to handle both reset and power-down reset options.
I personally implemented the watchdog rickkas7 put together with the AB1805 per the application note. It works very well and integrated well within the Particle platform and Boron using the library discussed there. It covers multiple scenarios and so far it has been working perfectly for me. I can’t thank rickkas7 enough for putting this together as it would have taken me months to come up with something like this on my own and I now have a solid functioning hardware watchdog, deep 30 second power down reset, able to leverage the HIBERNATE for lowest power consumption sleep as well as just an external RTC.
@rickkas7 Dear Rick, how would one actually obtain an assembled AB1805-SC board according to the schematics you provided? What manufacturer/service do you use that will not only make the PCBs but also assemble the very advanced BGA-package components that clearly exceed what’s possible for most hobbyists who manually solder their custom PCB components? How did you obtain the photographed board in the tech note? Thanks.
I have seen some examples of people doing it but applying heat to the back of the board manually, however I would not attend this.
I would guess you will either need a hotplate or, what I will be using, a small reflow oven. This watchdog does seem great but overkill for my application so I opted for a NXP part. I am going with the QFN option due to very restricted board size but if I am not mistaken, it is also available is SOD package.
Lastly, you should be able to ask your PCB manufacturer to assemble some components if you’d prefer PCBWay offers this service, but be careful, it is quite a slow process. Hence me buying a small reflow oven.
If it was me, use a PCBA shop. Well worth it in my opinion. Although I don’t have a reflow oven nor much experience with soldering SMD components. I’ve had good results with PCBWay and other similar places in the past. It took me a little learning how how to go fro. An Eagle Cad file, to Gerber files and a BOM to properly order from but was worth the time to learn. Most of them have good tutorials or YouTube videos of the process. I’m still impressed how economical a fully assembled custom PCBA for only a few units can be. Your talking maybe $5-$15 a board depending on what else you have on it. Most of my orders were around 3 weeks from order placed to delivery. So not overnight but still reasonable. Once received I can just plug in the boron and away you go. At least for me, I’ll put up with a little patience in having someone else do the PCBA.
I agree. Having them assemble the boards for you is well worth it, especially for larger volumes. I normally assemble (reflow) myself only for prototypes. I would hate to have a PCB-House populate 15-20 boards - with sometimes expensive components - on a PCB I haven’t tested yet. Of course this depends on how sure you are you will get the design 100% the first time around.
Personally I prefer to test first. Sometime things do boom
I may need to go the reflow oven route on the next board I do from scratch for faster prototyping. I think I had 3-4 revisions of a PCBA before I got it right. My biggest thing is it seemed bare PCB still took 2 weeks so it was only another week for PCBA or are you fabricating your own PCBs too or maybe found a place local?
Haha, I wish it was. Working on one now, going crazy. Have been over this a million times, but quite a complex board for someone with my somewhat limited experience
This us my experience as well. Then we have another week shipping if we have not customs delay. My previous fabrication run at PCBWay took quite a long time. We have no one local in South Africa that I know of. In US there are some option, but from wat I could find out via the All-knowing Google is that they are way ore expensive (around $200 - $300 for ten bare boards).
These are great to start with. Just be careful on which one you buy. There are tons of reviews, I must have watched them all, haha. If you are looking at the cheapies, I would recommend that you go with QR-2520D or larger one. Stay clear from the T962 series. If budget is not a problem, there are proper ones available from US or EUR but will et you back around $3000.
There is one more piece of equipment that can speed up prototyping drastically. It has been on my my Christmas list for two years, but Santa is conveniently skipping over it It is the Voltera desktop PCB printer. Even though they might claim can you produce End User products, I am not too sure. BUT I can most certainly see the merit in being able to design, print, test and iterate within couple of hours. Then once I know the design is working, I would still send to PCB house for full assembly of all components accept MC's of course.
how would one actually obtain an assembled AB1805-SC board according to the schematics you provided?
I ordered the bare boards and solder stencils from jlcpcb.com and assembled them by hand as I only made a few boards. I use an inexpensive T962 reflow oven, which I haven't had any trouble with. However I only typically reflow a single small board at a time so the unevenness of the heat isn't as much of an issue.
If you need any quantity a PCBA service is the way to go, but being able to assemble your own boards is handy for prototyping.
My apologies, let me rephrase my comments in the reflow ovens;
There are a lot of case studies on the cheap reflow ovens available online with data supporting the findings. The key difference (IMHO) between the T962 and QS-2520 models is the heating process. The T962 and T962A models rely on IR heating only whereas the the QS-2520 models support a combination of IR and convection which allows for a more evenly distribution of heat. This slows the oven to better follow the preset reflow curve.
In addition to this the QS-2520 model also uses thermocouple and then of course there are the concerns about paper masking tape used for insulation on the T962 series and then the fact that the device is not grounded properly. There are couple of video’s online on how to ‘fix’ these so called problems, so I suppose it’s not the end of the world. It is also advised that you manually wire in a second temp sensor on the control board to compensate for cold junction. The T962 assumes room temperature is 25C.
Many of the T962 ovens have been reported to have a 35C to 40C variance in reported temperature and actual temperature. I have not tested this, but the videos clearly show this. I asked Phui about these concerns and the admitted to using masking tape as apposed to something like Kapton tape but assured me it is not a concern. They opted not to comment on the Grounding issue and denied the temperature variance. They claim a 3% variance is maximum their factories have reported.
Buying in South African Rand, the T962 price tag is much more attractive, but I had the option to choose, so I chose one that does not pose these concerns
PS: these are not facts established by myself, merely data acquired over the course of couple of weeks researching these!!
.
It is the 