SPWF04SA - Discontinued soon, says my FAE! Why in god's green earth did ST push this part to the market when it has so many problems ?!

[Mario Ghecea]

Posted on March 06, 2018 at 23:16

I was recommending this to my company for productizing in IOT after working on it for seven months! 

Disappointment is beyond reproach ST...What a great way to lose your credibility as a solution integrator! 

[Tesla DeLorean]

Posted on March 07, 2018 at 00:16

Is there something better on the roadmap?

One might take a lesson from the Chinese with respect to product life cycles, ie design quickly, ship all available parts to market, move on to the next generation. If you've ever bought USB WiFi dongles you'll see considerable churn as the vendors move from one chipset to the next. If you want continuity you have to buy all the parts you need in a single purchase. Design cycles in the US which drag on for multiple years before deployment don't fit well into fast moving markets, nor can they sustain parts which ramp slowly to meaningful volumes.

One of the issues with the SPWF04SA and the prior iteration where they we're NOT from the Micro Processor group.

[Mario Ghecea]

Posted on March 07, 2018 at 00:21

I spent my own money setting up a solution for my company...What is the alternative here Clive ? LoraWan, SigFox, forget about IOT ? lol 

[Mario Ghecea]

Posted on March 07, 2018 at 00:31

Forget about IOT when ST is implementing it, is more like it!

:(

[Tesla DeLorean]

Posted on March 07, 2018 at 00:38

Depends on what the draw of WiFi is... presumably the consumer-off-the-shelf hardware that gets you to the internet. But in the corporate market *nobody* wants third party IoT stuff on the existing network (see Target and Chillers), so rental/contract type equipment tends more to cellular.

SigFox and NB target VERY low data volumes. Most of the demos show dumb things like real-time gyros, when the usable application is watching grass grow, or reading a meter twice a day.

LoRa has a vast array of module choices, and handful of chips if you want to roll your own. The software stack is sufficiently open that you aren't beholden to a specific vendor or implementation. Comcast is rolling out LoRa into the neighbourhoods.

For IoT you'd want to develop something that is agnostic to the physical implementation, so you can apply that quickly to the module-de-jour. If you like Micro Python, that can be ported to any number of STM32 Cortex-M4 parts. Build a main board, have the radio on a plug in form factor, that way you're not tied to a specific carrier or radio technology.