Poll: Are you interested in roadtesting I2C Real-Time Clock with Integrated MEMS Oscillator

It seems quite an incredible device, just 3x4mm (there's a crazy 2.1x2.3mm variant too), and still easy-ish to hand solder (0.65mm spacing).

I think it would be very handy for IoT products, and especially outdoor units at temperature extremes, where crystals don't operate.

It's annoying having to wait for NTP when you're powering up (if it's even possible - it may be on a Bluetooth or similar network) and want the minimum on time, so an RTC can make a lot of sense. Plus, it's near-essential to have time capability otherwise TLS won't work (needed for secure comms).

Recently I've been needing higher precision clock to measure the Allan deviation of a clock. I'll probably settle for a GPSDO, but I still need to do some research on which one to get...

Arduino connector is something like "industry standard". It has some benefits like support for stacking shields. You can use this board as lowest shield in your stack, disconnect some jumpers and since that moment you can reuse all pins (including i2c by design) in upper layers. Board itsef do not consume any pin from stack unles you want to use any of advanced signals.

Shiedl is designed as evaluation board. This means that it is not designed for production use (but you can use it in production if you want) and MAXIM expects that you will use it before designing your own board with this chip to check that chip supports everything that you need in your design. Arduino connector makes completly sense because you can do your experimets with chip on any arduino or any board with compatible connectors (for example you can evaluate it on previously roadtested Cypress PSoC 62S2 Wi-Fi & BT5.0 Pioneer Dev Kit, also on previously roadtested Bluetooth 5 Radio SoC EVM, and also for example on board mentioned here Seeking Feedback: Arduino Compatible Shields to Include in Upcoming MicroZed Roadtest) and these boards are very common and popular. You can evqaluate it without need for use wires on any platform that has this connector and there are thousands of these boards on the market. PMOD is also industry standard but less frequently used in comparison with Arduino connector. It is used mostly on FPGAs boards and less freqently on MCU boards. There are many other similar incompatible standard like mikro bus, groove, RaspberryPi connector and many others. If you want to use it on other platform you can use wires (you need at least three - SDA, SCL and GND) to connect signals between arduino connector and another evaluation board without Arduino connectors.

PC app is not mandatory. It is used for evaluating features of chip if you do not want to write any code. Evaluation using PC app requires (WIndows ) and bundle PICO board to be placed on board with correct frimware flashed (board which come with this shield probably come with correct firmware flashed from factory). This is one of reason why PICO board (and connector for this board) is present. You can use bundled PICO board for evaluation instead of external arduino. It is normal ARM Crotex MCU.

I think MAXIM provided lot of ways how to evaluate this chip and their strategy is good.

This is one of those dev boards I just don't understand. The clock chip looks interesting and it's cheap. I'll probably try one and I'll either add it to a prototype design where it doesn't much matter if it works or not or stick it on an adapter board and hand wire it to an ST Nucleo.

I really don't want to mess with a PC app and an interface board using  a micro that I'm not interested in - if I didn't think that Maxim could make a chip that pretty mcuh works as advertised I wouldn't consider using it at all.

If I want to spend the time actually checking out the hardware then I'm interested in the nuances of getting it to work in a real design.

For my money a dev board in PMOD or similar form would be far more use.

MK

To be honest, you don't need GPS PPS necessarily, assuming you can run it for a few weeks. NTP disciplining a system clock over the internet regularly gets into the tens of milliseconds range of accuracy. Sure, you're not going to find the exact offset, but that is very dependent on the RTC's temperature as well, so testing at extremes is going to make things a bit more evident. Testing at regular room temperatures, if stable, could be hard to discern errors assuming it is well trimmed from the factory. Operational voltage can also have an influence.

Being able to resolve the effects of the internal temperature compensation, checking the behaviour in terms of power consumption from battery, trickle-charge capabilities could be interesting.

- Gough

Yeah, I went and took a look.

I did the maxim temp sensor roadtest, it claimed to support win7(which I have) but in reality didn't. Until I get windows, I'll step away from maxim RoadTests.

Sadly, GUI software is not compatible with Linux and MacOS because downloaded file https://www.maximintegrated.com/en/design/software-description.html/swpart=SFW0012990A (you must register/login to download it ) is named with .msi extension which is Windows Installer... On the same page is also mentioned that only supported operating system is Win10 (but after analysis described later I think it works even with older Windows). Application is written in .NET using Windows Forms (because it looks beautiful, I first thought it is created in WPF but after short reversing using ILSpy I have seen that it realy uses older Windows Forms and no WPF). I tried run it using Mono on Linux (I have to extract .NET executable from installer by installing it on Windows and copying installed files). It successfully started:

After a while app crashed with stacktrace pointing to X11 keyboard. I am not sure if it was caused by bug in Mono or some Mono incompatible visual tweek done by Maxim to make WIndows Forms app looks so beautifull.

But because I do not own the device, I cannot verify that USB communication (the most important thing) also works. Application use hid.dll to some (maybe not all!) communication with device. After quick browsing over internet I think that hid.dll does not work and communication with USB device will be unusable from Linux even application successfully started.

I2C protocol used by RTC chip is easy to understand and Maxim provide Arduino library (downloadable at https://www.maximintegrated.com/en/design/software-description.html/swpart=SFW0012750A ) so I think WIndows is not necessary to evaluate and roadtest this board. You can also try write your roadtest application around building opensource alternative to this app yourself....

I wish these GUI systems were linux or macOS compatible.

GPS is good point. Also thank you for navigating to accurancy over Vcc and temperature range. If someone else is interested in it:

I should not do such calculations in my head Yes, its about half a minute over two months, I think somewhere I multiplied it by two again... (as a back of the envelope calculation - 1ppm means 1 second deviation in 11 days).

But the 30 seconds are actually the worst case figure (at 1.6V supply and 90°C) - at 3.3V you will see _much_ less than that. If you don't want to run a long-time test, you probably need to check it against a GPS-disciplined generator (or at least a GPS 1PPS output).