The CC3000 - does it work yet?

Hi there,

I’ve had my Spark Core sitting in a box here for a few months now. I’ve not started using it yet because I know the CC3000 module has been plagued with serious firmware issue(s). I’ve been waiting for Texas Instruments to resolve these before I start using Spark.

So I was really interested to read about the recent Spark “Deep Update” message. However, my hopes that the CC3000 issues had been fixed seems to have been dashed.

This recent (June) thread in the CC3000 forum is calling for a grievance letter to be sent to CC3000 senior management about the failure to resolve the issues.
http://e2e.ti.com/support/wireless_connectivity/f/851/t/345205.aspx

What is Spark’s position on this? Would they agree that the latest CC3000 firmware is still irrevocably broken? Or does the Spark team have some extra knowledge that suggests that it is possible to make the CC3000 operate reliably?

Best wishes,
Damon.

as far as i can see static ip still doesn’t work, adafruit tried. wep works now though

The main issue with the CC3000 referenced in the forum thread, is that the firmware locks up in busy networking environment (who doesn’t live in a busy network environment these days!) and requires a hardware reset to get back on a network. This will manifest as repeated connect/disconnects to the cloud.

Plus during a hardware reset the EEPROM is written to – so in a busy environment the device will be bricked with a dead EEPROM in about a year.

1 Like

The current state of the CC3000 after the deep update is, in my opinion, pretty stable, but YMMV.

Formerly in our office space (which is a co-working space with lots of people probably all downloading on torrent, i.e. a very busy network), a Core couldn’t hold a connection for more than a minute. With the CC3000 deep update patch, it’s pretty stable.

Despite the TI thread, I would encourage you to run the deep update and see what your experience is to see if it will work for you. I would say that for most people on most networks it works just fine now.

There are two issues that mean that the CC3000 will likely not see any further firmware updates:

  1. The processor in the CC3000 is basically out of memory. Even the current stability fix sounds like it was basically just a workaround, because the actual issue (which relates to the ARP cache) couldn’t be fixed entirely because of limited memory.

  2. The CC3200 is now getting distributed, meaning that the CC3000 is now old news.

What does that mean for the Spark Core? It’s useful in its current state for many use cases, but not for everything; the CC3000 is an inexpensive module that’s easy to get ahold of in small quantities, but it’s not as feature complete as other processors. We’re evaluating the CC3200 for our next-gen hardware, as well as other Wi-Fi modules from other semiconductor companies, which will lead to a wide range of improvements, because the big semiconductor companies are currently investing a lot in their embedded Wi-Fi solutions.

2 Likes

Hi Zack,

Thanks for the detailed reply. So you think its this fix from the 1.13 SP that has stablised things?
“Fixing the automatic ARP updates in the network stack which would make the long running TCP operations to hang.”

That’s the CC3000 fw side of things, what do you think about the stability of the host drivers? Lots of people on the TI forums suggesting lots of different workarounds there as well…

Ok, I’ll give it a go! :smile:

Damon.