[AllanG]

I had an interesting experience today that may interest people. Having finished building my PLUS rx board a while ago but not wanting to use it yet on my operational station, I've paired it with a spare tx board and have been using it for testing purposes.

Tx is in the garden in an unsealed enclosure but protected from rain, rx is in the house and both are no more that 15m apart and have been running only during the daytime. Over the last week I found that all worked well for the first hour or two after starting up, "tx efficiency" was about 85% or higher and I could watch the packet count updating at approximately 1 packet every 5-7 seconds or so. After that first period though, the rate would randomly drop to very low levels and sometimes I would lose data altogether for a minute or two, then it would usually recover partially or even completely for a time before repeating the behaviour. A few days ago I tried replacing the rx module on the PLUS rx but it made no difference. Today though after running properly for a time, the PLUS rx stopped receiving any data at all and didn't restart.

So I replaced the wifi transmitter on the outside tx unit. The system immediately restarted properly - and as I write has been running uninterrupted for about 9 hours. The "tx efficiency" is currently 87% and has remained very stable. I looked at the TX module I'd removed carefully with a magnifying glass. On the solder side of the pcb I noticed some white, powdery deposits around a couple of the solder joints - I also saw what I suspect might be a "dry joint", characterised by a dull grey instead of shiny metallic appearance, which is circled in red on this photo....



I am guessing the white powdery substance is caused by dampness in the air reacting with solder flux that hadn't been properly removed after manufacture and is probably causing instability in the module.

I bought a few of these modules on ebay but I can't remember which vendor. This is what the "other side" looks like...



The solder flux residue can be clearly seen around the connector pins, it doesn't seem as if there was any attempt to clean it off....

These modules are very cheap but unfortunately, in some cases at least and this is one of them, the quality is correspondingly poor!


Allan.

[Wizza]

Hi Allan, Interesting scenario, and will be interesting to see how it goes over the next week or so. Hey, would you consider testing something else..??? Looks like you are using a header socket on the board to plug the TX module into..?? What if it was the header socket causing the issue..??? If all is good over the next week or so, what about inserting the "suspect" 433MHz module again, and see if the problem repeats itself. The reason I say this, is "could" the header socket that the TX module was plugged into be part of the issue, and "seating" the new module when it was installed may have addressed the socket/header issue.... Don't know, and just a thought, and would be interesting to hear what you find, if you had time...

Mark R

[werk_ag]

Have anybody ever hear about dendrites? Google about it, it will worth.






It's a "disease" that attacks PCB,s mostly caused by the use of some types of flux. The most the PCB's are exposed to high levels of moisture in the air, the quicker the problem arrives.

Almost all RX and TX modules I saw, have signs of use of inorganic flux. They should always be cleaned with distillate water, not alcohol or acetone which only dilutes the flux and spread it in a tiny film along all the board. When you use alcohol or similar flux dissolvers, always wash the assembled boards with distillate water, and fully dry them before use.

[JT118]

At first glance I thought you were on a anti-smoking campaign!

[werk_ag]

This "disease" kills even faster than the cigarettes.

[AllanG]

Mark, that's a very good point about the header socket and more on it in a moment. A bit of background first...

The operational station tx unit is installed in a sealed, weatherproof enclosure and has been working perfectly for 14 months now. The "spare" tx I'm talking about was used for a couple of months last summer and early autumn while I was trying (unsuccessfully, thanks to awful English summer weather!!) to calibrate my own solar sensor. I didn't then have a spare rx, so enabled debugging on the Nano and used a connected netbook computer and the serial monitor. Over the winter and until last week the tx was stored in a dry but unheated garden shed.

This last week, using the new PLUS rx, I've been able to set up a complete second "test station" using the same spare tx, but now with a wifi module installed, one of a few that I ordered a year or more ago for the original project. I didn't check it (with a magnifying glass) before inserting it on the tx board so it is *possible* it was a little defective to start with. The tx board is in a box but open to the air on the bench at the end of the garden. It's winter here of course and although the setup is put in the shed at the end of the day's experiments, the daytime air is very damp at times.

Now, onto the header sockets...   I'm sure there's a potential for connectors in general (on tx boards) to cause difficulties, certainly in damp conditions. Yesterday I connected up my solar sensor again and this was the result...



Between around 13:15 and just after 14:30 the UV index continually dropped from around 0.2/0.4 to zero and then recovered. I thought that a bit strange, so shortly after 14:30 I went outside and re-made the connections *on* and *to* the solar pcb. As you can see, the readings subsequently appear normal.

Admittedly this particular tx board has not been very well protected from the elements and it seems likely that connectors *can* be a source of problems in that case. I'll certainly do as you suggest and, after leaving the replacement wifi module in place for the next week or so, try the original again and see what happens and post the results here.

Either way, I think there are two learning points here. 1 - Sockets of any kind have advantages *and* disavantages. 2 - The quality of some wifi modules is certainly marginal at best....

Best regards,

Allan.

[uncle_bob]

Have anybody ever hear about dendrites?

Apparently it's a major consideration with PCB's on spacecraft.

[AllanG]

Have anybody ever hear about dendrites? Google about it, it will worth.






It's a "disease" that attacks PCB,s mostly caused by the use of some types of flux. The most the PCB's are exposed to high levels of moisture in the air, the quicker the problem arrives.

Almost all RX and TX modules I saw, have signs of use of inorganic flux. They should always be cleaned with distillate water, not alcohol or acetone which only dilutes the flux and spread it in a tiny film along all the board. When you use alcohol or similar flux dissolvers, always wash the assembled boards with distillate water, and fully dry them before use.

Wow that's amazing, Werk_AG!! Unfortunately the best lens I have for my camera is a 60mm Macro that I used to take the photo of the wifi module. I'm sure if I had a more powerful lens I would see something like that first one of the two pads.

Yes, the tx board that the wifi module is mounted on is certainly open to moist air, I've described the circumstances in my reply above to Mark (Wizza). Of course in normal operation in a weatherproof and sealed enclosure, this kind of situation is unlikely to arise but it's good to know the potential problems that can arise. Personally, in future I'm going to clean and check carefully the wifi modules *before* I use them...

Allan.

[AllanG]

Have anybody ever hear about dendrites?

Apparently it's a major consideration with PCB's on spacecraft.

I guess that's the *ultimate* hostile environment, even worse than an English winter...

Allan.

[werk_ag]

Regarding the use of sockets for the RF modules (both RX and TX) personally I only use them in test units, and never on production units. I do it not only because of the possible problems with the connections on the socket, but mainly because the path of the RF signal increases and change. Ideally the path between the RF output (or input) of the module and the antenna should be as short as possible.