Posts: 1
Joined: Sun Aug 16, 2015 2:59 am

Hoping for AT Command set and firmware - a wish list

Sun Aug 16, 2015 3:57 am


I have spent some time evaluating the ESP8266 vs the EMW3165.

The ESP8266 has two fairly critical flaws; it does not have a sleep mode via the AT command set -and- just about every user of the part says it crashes fairly regularly and particularly with the AT command set. That (crashing) in itself is enough to make me take the part less than seriously. It is a 'toy' for those reasons and confined to applications where these attributes are non problematic.

I will readily admit that these two criteria may change in the future; a new AT release that solves both those issues may surface. For now (August 2015) they are real problems.

I have commercial need and those two problems in a person portable device are show-stoppers.

The ESP8266 however has a very good (comparatively) and reasonably active support forum. There are downloads of the SDK and documentation etc. Sure, it is a little difficult to find things like actual downloads at times, bit it is all there.

If I were going to spend time to focus on developing firmware for a WiFi device I think it would be for the ESP8266 for the above reason. Support.

I do not have the experience or time or inclination.

I can see the attraction of the ESP8266 for the Arduino users - and it is a great attraction. Until another product offers the same simplicity (AT command set) the ESP8266 will rein supreme and all the others will probably languish. The Arduino users are possibly ok with periodic crashes, can mitigate against it with power cycling techniques and maybe toggling RST low etc. Commercial users will possibly generate their own firmware for the device or bypass it or, like me, get one for play then determine where to go from there, probably not progress past the 'play' phase.

I like the EMW3165 far better and I wish it had the AT Command set and the reliability I've dwelled on (painfully probably) for the reason of its power consumption possibilities. However I have neither the skills nor the time to produce the firmware/software to use it. I need to focus on my application which is probably the same for 99,9% of the users out there.

I've looked through the posts on this forum and I can see some clever people here who clearly have both the skill and hopefully some of the time required to bring this part into a commercial success, assuming they are willing. I also see that the support from the manufacturer is somewhat less than the ESP8266 and perhaps this is as a result of time - it is a new part. I mean no disrespect, just that from my observations I see this.

The reason for writing all this is my hope than someone will go to the trouble of producing what I believe to be 'winning' firmware for the device and the manufacturer will actively support it.

You may accurately say I'm lazy and don't want to go to the trouble of generating my own firmware. That is true. It is also true that I could do it in the time available to me. I need to focus on my application - same as most others.

I would welcome feedback from developers/others who have used this device successfully.

Kind regards

I am a developer of person portable embedded systems with experience in mainly assembler for the MSP430 series processors as well as many older cpus, digital design and some minor Arduino experience,

Site Admin
Posts: 61
Joined: Tue Jul 14, 2015 7:29 am

Re: Hoping for AT Command set and firmware - a wish list

Sun Aug 16, 2015 7:45 am

Hi Richard,

First of all, welcome to the forum. The points you make regarding the ESP8266 are quite valid. I completely agree that it hasn't quite yet matured to a level ready for broad commercial development. I think it's a great chip, full of potential and has the potential to spring up a new generation of low-cost IoT devices, but it certainly needs some more time to smooth out some rough edges and bugs, but it really packs in a lot of potential for it's pricepoint. The active community of developers sharing code and information for this device is really impressive, and definitely a strong point for the platform, and it's honestly my greatest wish that such a strong community can be built up to support the EMW3165 as well.

Right now, the EMW3165 is really in it's infancy, however unlike the ESP8266, there is no custom silicon on this module. It's just an integration of two very tried and true devices, the STM32F4xx and the BCM43362. STMicro and Broadcom are well-established and field-proven, their parts are well documented, and the necessary development tools exist. The major hurdle is that it's a currently costly to develop on this platform, in both time and tools. I will agree with you that in order for the EMW3165 to really take off and be a success, this hurdle needs to be removed by development of some projects which allow the device to be used in rapid prototyping and programmed via UART/USB/OTA and without a heavy IDE / compiler. I am sure that this will come in time, as more and more people get their modules and start getting their hands dirty and sharing the fruits of their labors.

While I completely agree on the need for some firmware to turn this module into a great platform with low barriers, I don't particularly think that an AT command set is most useful. The AT command set implies that this module would be controlled by another MCU, and that's certainly going to add unnecessary cost and complexity to any product you'd want to develop. The STMicro on this module is an ARM M4, which is practically capable of running linux. There are very few tasks that it couldn't handle. It seems silly to think of an 8-bit MCU sending AT commands over UART to an M4. I mean, I am sure there can be a few situations in which it seems reasonable to do things this way (ie, your only concern is ultra low power operation on a device which spends very little time actually connected to WiFi), but to make the most of the module and really reap the benefits of it's price point, one should really make the most of the on-board M4 by putting all the embedded logic on it.

Keep in mind that within this module, the STMicro is using Broadcom's WICED SDK to communicate with the BCM43362 Wifi radio, and the WICED SDK is essentially an AT command set implemented over SDIO with higher bandwidth than UART. You can then see that inserting another layer of AT commands over UART makes for a less-than-ideal system architecture. I believe that this is a case of "be careful what you wish for", I think the end result of an AT command set for this module would be disappointing in many ways, and not make this module any better suited for commercial products than it is in it's current state right now (GCC + WICED SDK development directly on the module, or using MXCHIP's MiCO with IAR compiler)

If you still think the best thing for your project is rapid development and you're willing to add a few dollars to your BOM in exchange for fast time to market, and you don't need all the bells and whistle's of the on-board STM M4, then I highly encourage you to check out the Particle P0 WiFi Module:

Their module is built on the same Broadcom Wifi Chip, and has an on-board STM M3 MCU, but they've got awesome firmware already making programming it an absolute breeze, their web back-end is incredible and to me, it seems like it's exactly what you're looking for right now. Their development tools are easy to use. The end result is that the module price is $10 instead of $8, however you'll save months in development time. Later on, when you want to get those two bucks back, hopefully the EMW3165 will have matured to the level needed. Or perhaps the particle module pricing will drop :-P

Anyways, just offering my two cents on this topic. By all means, I encourage development of firmware for the EMW3165 which will bring down the barriers, but at the same time, I am also a huge fan of making sure everyone picks the best parts for their projects.

Return to “Compiler & Tool Chain”

Who is online

Users browsing this forum: No registered users and 1 guest