Random Connections of the Meshcore World

I've never claimed to be an electronics person and, though I always try to pick things up as I go along, I'm still, and always will be, a learner.

That, however, doesn't mean I can't ask questions that get to the heart of the matter. And this is what I've been up to this week. And it seems a lot of people don't have the answers, seemingly because they've been connecting things up without much thought as to maximum voltages and potential longer-term damage to their Mesh boards.

First, let's remember that there are two main types of Meshcore (or Meshtastic, if you prefer that) board: Heltek, now at v4, and the more expensive but often preferred RAK boards. With both Heltek and RAK boards now being optimised for low current draw, one can conclude there is little to choose between them. The only board I've had software, rather than hardware issues, is in fact the RAK board (a database locked error, requiring a re-flashing).

The solar charging of these two boards are not the same. The Heltek v4 has integrated MPPT charging, whereas the RAK boards do not.

Many folk use the external MPPT charging board shown below. It costs around £10 and, in my case, is rated for 12V solar panels, with 5V output. 

MPPT charge regulator. Fine for RAK boards, no good for Heltek.

My real-world test of this board shows it is entirely successful with the RAK boards. You connect any solar panel, of any rating between 6 and 24V, to the 'SOLAR IN' socket. The output goes to the input of the RAK board. Your battery connects to the remaining board socket. Battery charging begins when the MPPT regulator sees about 3.7V. It will only ever provide around 4.2V in full sunshine, which is exactly what we need. The RAK board has a 'never exceed' input limit of 5.5V. 

My test so far this spring shows that the battery is charged up to around 80-81% and only discharges by around 1% overnight. In winter, the charge and discharge will of course suffer more, but we'll cross that bridge once winter arrives; we have summer to enjoy first!

When it comes to the Heltek V4, MPPT is already on-board and you can't therefore connect an external regulator in series, because the battery will see little or even no charging. The trouble with the Heltek is that you need to provide the correct input voltage (not more than around 6V), without going through an MPPT regulator first. That's usually achieved with a simple (and very small) buck converter. As far as my tests show, this seems to be successful, though I haven't tested it for more than a few days. 

One alternative for Heltek is to use the much more complex Waveshare regulator. This is a bit more expensive, physically larger and uses around 16mA continuously, simply by virtue of being connected up. Whether that is too much current draw depends on where you are and how active your board might be.

Waveshare solar power manager module. Nice item, though the screw connectors are never my favourite, and it consumes 16mA at all times when connected, which may or may not be an issue, depending on your location and board activity levels.

 

Another option is to buy a 6V-rated panel which has a buck converter and USB connector glued to the back. This works very well for me on a field-deployed Heltek V3 (not V4, though it should be fine for that, too). The only issue is that these boards tend to be of the flexible type, which can be a problem in field deployments, where a solid surround is usually needed. Also, the USB connection needs either to be modified to a permanent, wired solution, or else protected with sealant of some form, so that the whole thing doesn't rust to bits in the outside environment (which will happen quickly otherwise).

Be aware that directly connecting a supposedly 6V panel, especially of the 'CCTV', smaller types that are widely sold, can result in failure of your boards, regardless of type, because they will often exceed 7V output. Those types of panel, whilst fine in very sunny locations with little cloudy weather, are absolutely useless otherwise. Even on a bright day with cirrus obscuring the sun a little, the output falls to very low levels.

  

 

  

LoRa Coax Losses.

Recently, I installed a test solar repeater at a remote location. This was entirely to assess the reliability of such a system and whether anything might need improving in time.

My test solar repeater, with its short RG174 coax (now changed).

 

As a test, I simply used a 1m length of RG174 coax as part of a pre-made connector to hook up the McGill Microwave Systems 4.5dBi colinear vertical to the Heltek V4 board.

On showing a picture for interest on Facebook, the inevitable 'oh, you shouldn't do that!' comments began to roll in. 

First, I 'shouldn't put an antenna in somebody else's field at random, without permission'.

Er, I've been given a personal easement to use this field and structures 35 years ago.

Then, I shouldn't use 'that very thin wire for a long run; it will defeat the gain of the antenna'.

Er, the cable is 1m long. 

The pole the antenna is fixed to is simply to hold it clear of a roof and little else. Great height isn't feasible and neither is a larger antenna at this very exposed, remote location. The antenna-coax combination was an entirely reasonable system, given what I had available and that it was a short run and that a 1/4 wave antenna using the same RG174 cable assembly had already been shown in earlier tests to give perfectly adequate performance from the same location.

Another factor, significant for me, was that thin coax needs only a narrow hole to be cut through brick walls - in my case, by hand drilling. Wider cable needs a larger hole - very difficult to drill by hand, especially above head height, and otherwise excessive and unsightly for some on a house wall. 

All the same, the question as to whether RG174, which is widely used, almost as standard for LoRa applications, is really so lossy as to require replacement in most cases.

Undoubtedly, RG174 has extremely high losses at or around 868MHz; Farnell's datasheet gives 105dB per 100 metres, or 1.05dB per metre. So any runs much greater than my own 1m length would, indeed, present losses that most people would deem unacceptable, except perhaps where a directional antenna was in use. Even then, a run of more than 2m would really need better coax.

LMR400 coax.

 

The usual alternative is LMR400. In ham circles, we've been pre-conditioned to expect high prices for long runs at HF or VHF, made worse by US-centric discussions online, which usually have very much longer runs to antennas than those in the UK.

In fact, a pre-assembled LMR400 coax of (standard) 3m length for LoRa comes in at just  £19, delivered (mainland UK). This is the kind of length many would choose for an antenna situated on a house wall to be fed from indoors, for example.

RG174, thin coax.

 

The benefit of LMR400 is the low loss: at 868MHz, it's around 12.8dB per 100 metres, or 0.128dB per metre. With a 3m length, the loss is 0.38dB, or already 0.66dB less than just 1m of RG174.

So, is the use of LMR400 and its relatively high price, worth using?  For runs of 2m or less, for most people, no. Beyond that kind of run, LMR400, or even the cheaper and slightly higher loss (~0.193dB/m) X-400, rapidly leaves RG174 in the dust, keeping losses within tolerable limits until one reaches around 15-20m runs or so. Above this length, a better antenna and/or even more expensive, lower loss coax will come into consideration.