Unreliable Operation: FTP, Telnet, Power, ...

[mhepp]

Hi,

I got my five WiPys last week and tried them out today. Overall I must say that I am kind of disappointed.

First of all, soldering anything onto the pads for the headers etc. is a bit difficult, because of the type of the finish. I had to apply SMD flux before soldering; otherwise, the soldering joints would be pretty unreliable.

My concrete problems are as follows:

1. Power supply. The spec says the board needs 14 mA in operation and less than 1 mA in sleep mode. I tried to power it from a 1.5 to 5 V step-up converter from a single 1.5 V cell, which works fine on all of my other uP projects, e.g. with PICAXEs or Pyboards. The "heart beat" LED only flashes at a high frequency and no Wifi will be found on my Macbook. When I use a lab DC power supply with 5 Volts or a Breadboard power regulator with 5 Volts, the WiPy seems to work fine; at least the "heart beat" LED flashes slowly as it should.

Adding a 10 - 470 uF capacitor in parallel to the step-up converter slightly improves the picture, as the LED flashes correctly then, but no Wifi network will be found.

With an oscilloscope, one can see that the 3.3 V pin on the WiPy suffers from regular drops by about 1 V, likely due to heavy power consumption by the Wifi component when sending data.

I might be able to fix that by adding external power circuitry, but I think it is kind of disappointing that the WiPy is so sensitive as far as power supply is concerned.

2. FTP and Telnet bugs

The Telnet and FTP connections are very brittle. I manage to set up a Telnet connection. But when I try FTP, this typically

- kills the Telnet connection and
- freezes after 1-2 transfers

I have not found a mode that allows reliable operations. I would expect to be able to have the FTP and the Telnet connections work fine in parallel so that I can upload and test new versions of software with ease.

Does anybody know how to fix these problems?

[kfricke]

You might take a look at some more specific specs of the SoC used on the WiPy... the Texas Instruments CC3200

There they do tell:

Code: Select all

- Advanced Low-Power Modes
  Hibernate: 4 µA
  Low-Power Deep Sleep (LPDS): 250 µA
  RX Traffic (MCU Active): 59 mA @ 54 OFDM
  TX Traffic (MCU Active): 229 mA @ 54 OFDM, Maximum Power
  Idle Connected (MCU in LPDS): 825 µA @ DTIM = 1

[danicampora]

First of all, soldering anything onto the pads for the headers etc. is a bit difficult, because of the type of the finish. I had to apply SMD flux before soldering; otherwise, the soldering joints would be pretty unreliable.

Like I already replied before, we hand soldered the headers on more than 2000 WiPy's, and we had no needs for flux, just good quality multicore wire.

I already replied to your other post regarding the power supply. It does consume 14mA (on average) when in STA mode, but it boots in AP mode, and anyway, even in STA mode WiFi inherently has peaks of current consumption during Tx, that's the way it is.

Regarding FTP and telnet, it is possible to use them both at the same time, I do it constantly, and other users have been doing as well. Did you check the FTP instructions and tips here:

http://micropython.org/resources/docs/e ... ftp-access

I recommend Filezilla, it's free, multi-platform and works like a charm, what else could you ask? For telnet Putty on windows it's also the right choice. If you are on Linux, then just use the stock telnet and ftp clients, both work flawlessly and can be used simultaneously.

I am not saying that there aren't things to improve on the telnet and FTP servers in the WiPy, but certainly not what you claim.

Cheers,
Daniel

[danicampora]

Regarding disconnection, the server(s) have a timeout, but it's configurable, check here (you need the latest software)

http://micropython.org/resources/docs/e ... ass-server

How to update:

http://micropython.org/resources/docs/e ... er-the-air

[blmorris]

@mhepp - More information on your power supply would be helpful if you can provide it. Specifically part numbers and / or a link to a full spec for the step-up converter and the 1.5V cell that you are using. (Is the cell a standard AA or AAA alkaline, or is it something else?)

What kind of solder are you using? The Pb-free versions have a higher melting point and tend to oxidize quickly while soldering; I usually use a Pb-free solder with a water-based core, because I can spray the finished board with hot water and have it clean up perfectly. It is finicky to use, however, and I routinely pre-treat any surface that isn't a gold finish with a water-based flux pen when using this solder. It is more work, but the results are nice and clean.

I don't have a WiPy yet, and without the specs on your power supply it is hard to be sure, but I suspect that all of your operational problems could be due to the power supply not being able to handle the peak current requirements.

-Bryan

[mhepp]

As for the soldering: I use high-quality multi-core 60/40 solder and have been doing so for almost 30 years. Anyway, a bit of flux does the trick, but this was unexpected as e.g. the Pyboard gold-plated pads solder much better. I understand that some people argue that Leadfree-HAL is cheaper and more suited for joints that can receive mechanical forces.

As for the FTP access: I finally found the instructions on the constraints of the FTP interface. Maybe printing them on the box would help.

As for the power supply: A note on the specs as for the peak currents would also be nice.

And please: I understand that running the kickstarter campaign was a bigger thing than you had expected, but there is no need for a slightly hostile tone in your replies, particularly after quite some delays and excuses.

I like the basic approach and it might turn out to be useful. But in comparison to the PyBoard, there are quite some edges.

[danicampora]

And please: I understand that running the kickstarter campaign was a bigger thing than you had expected, but there is no need for a slightly hostile tone in your replies, particularly after quite some delays and excuses.

I am sorry, I didn't mean to be hostile, my sincere apologies, but please try not to double post.

[danicampora]

As for the soldering: I use high-quality multi-core 60/40 solder and have been doing so for almost 30 years. Anyway, a bit of flux does the trick, but this was unexpected as e.g. the Pyboard gold-plated pads solder much better.

I agree, gold plated finish is much better for soldering, and that's what we ordered with the assembly factory, but they messed it up and the boards ended up with HASL finish. We could have rejected it, but decided not do so in order to avoid more delays. The expansion boards came out right, with ENIG finish.

[mhepp]

Hi, thanks for your fast reply and explanation!

I think what will help future users to avoid my problems is adding a one-page instruction sheet to the box that

- discusses power supply options briefly (my assumption is that a 3.7 V LiPo battery is the simplest for mobile solutions, and a 5 Volt breadboard supply unit (like http://www.amazon.com/Breadboard-Power- ... B00BXWV2F6) or any stabilized 500mA+ V source for breadboard experiments.

- point to the limitations of the FTP protocol support (passive mode, no SSL, one connection at a time).

That would have saved a bit of time

[deshipu]

There does seem to be some tricky problems with the FTP server. While it does work with the default "ftp" Linux client, and with FileZilla, the WiPy will freeze each time I try to log in using, say lftp. I haven't looked into this in detail yet, but my guess is that some ftp clients try to use some optional advanced commands, and that trips the board's code into a crash, instead of a graceful error.