Solar node for top of tall building using the Waveshare D solar power manager

This solar node uses the Waveshare D solar power manager because that has the advantage that the RAK node will shut off cleanly if the batteries go flat, and switch on again cleanly when the batteries have enough charge again. The design is intended to be placed on a very high building and basically left there un-touched for a very long time. Lightening protection is already present that the intended location.

Parts list and some comments about the build:

Node box (plastic, best to allow bluetooth through): 158 mm x90 mm x60 mm https://amzn.eu/d/00FG0PgX There is a rubber o-ring to seal the lid, and I have no yet tested how well this works in terms of waterproofness as the node as not been outside for long enough.

The parts of the node (batteries, RAK, Waveshare solar manager) are all fixed to an acrylic plate that is then just slotted into the box with a tight fit. There is not much play-room in this build.

Tools:

Step drill (Stufenbohrer) https://amzn.eu/d/0fqQX0UJ for drilling hole in node box for cable glands

Drill bits for metal: 1.5 mm, 2.5 mm & 3 mm for acrylic plates inside node box, 4 mm, 5 mm, 7 mm, 4.5 mm metal drill bits required for this build.

Filler in case holes need water-proofing: https://amzn.eu/d/0ioZV5fS I had to fill some gaps for one of the glands.

Hacksaw to cut off and bolts that were too long and stick out of back and to cut acrylic sheet: wide blade like this one https://amzn.eu/d/03X0dSQH

Fasteners:

M5 x 20 mm bolts nuts and washers

M5 x 10 mm pan head hex x 2 to fix antenna base to node plate

Various bolts and nuts and washers.

M5 x 60mm for constructing strain relief from part below and many M5 nuts. See image.

Strain relief: https://amzn.eu/d/0ftWa4Pv

Screws to fix battery holder to acrylic sheet: https://amzn.eu/d/01viGf88 This is done using pre-drilled holes of very small diameter, e.g. 1.5 mm, into which the slightly wider screws then screw nicely.

Solar panel: https://opencircuit.shop/product/solar-panel-6v-5w-glass

Solar panel adaptor plate to mount solar panel to angled mount: Baumarkt 19xmx5cm Lochplatte x2

Bolts for solar panel adaptor plates: M4 x 10 mm cross-head screws/bolts with nuts and washers

M5 washers: used a lot in this build.

Solar panel angled mount: https://amzn.eu/d/0dj29GHY I used M4 bolts and






an M4 and M5 washer to stop the bold head from slipping into the cut-outs of the mount.

Solar panel post mount: https://amzn.eu/d/04aH5to1 This has sharp edges, so be careful when building.

Antenna, which comes with a post mount: 868 MHz LoRa / Helium Rundstrahlantenne, Wetterfest, 5dBi

(ART7198) from https://www.wlan-shop24.de/

Plate to mount node box to antenna mount: Hagebaum (DIY store) Lochplatte 10 cm x 30 cm galvanised

Antenna cable N male to SMA name: https://amzn.eu/d/0dCOuSu8 heavy duty and for outdoors. It is a good idea to place some heat shrink tubing over this cable before fixing it to the antenna and feeding it into the node box. You can improve the waterproofing of the connector at the antenny by heat shrinking a tube at the end. This cable plugs into the following:

IPX to SMA antenna connector: https://amzn.eu/d/0aXeZmtL It is also worth have a bit of heat shrink around the connector. The connection must be good and it can be fiddly in the box. The gland needs to be located as far left as possible to that the SMA female connector can connect to the antenna’s sma male well. Remember that an antenna needs to be connected if the RAK is on!

PTFE plumber’s tape: https://amzn.eu/d/0ixILV2O to making the threaded join between antenna and antenna cable water-tight.

Antenna cable feed-through for node box: M20 cable gland e.g. https://amzn.eu/d/0hhkeOBR

Box solar panel feed-through: M16 cable gland , might be larger than needed. This depends upon if you want to use the jack plug on the solar panel to feed into the solar manager, or if you want to use the screw terminals, in which case the gland feed-through can be even smaller in diameter as only the cable itself needs to pass through the gland. The M16 allowed the jack to of the solar panel to pass through, just.

Node: RAK Wireless Wisblock + Lora = SKU: 110082 und SKU: 115000 https://store.rakwireless.com/

Solar power manager: https://www.waveshare.com/solar-power-manager-d.htm?sku=26363 SKU 26363 with option WITH battery holder (for 3 x 18650 batteries) that is SKU 26363

3 x 18650 li-ion batteries, I am using flat-topped (unprotected) ones, not sure if that is optimal in terms of the batteries sitting tight (with raised positive end might be better), but they seem to work fine. (balanced !! take care - 3 used in parallel - so not use e.g. one discharged one and two fully charged… they need to be inserted when they are each about the same voltage). It is always best to buy managed cells e.g. from here https://www.akkuteile.de/lithium-ionen-akkus/18650/keeppower/keeppower-18650-3100mah-3-7v-li-ion-akku-geschuetzt_12016_1020

, but make sure that the length of the cell is still ok for the battery holder. If the cell is too long, the holder will not receive them properly, if at all, and the fit in the node box is very very tight anyway. Update: the above 18650 protected batteries with raised postive poll fit better (tighter) than flat top ones. I have tested if the battery protection circuits interfer with the Solar Power manager D’s switching on and off of the RAK, and at least when the batteries are flat they do recharge and the node does activate. This is a fiddly issue though. The protection circuit kicks in at 4.2 V and I hope this does not happen before the Waveshare D finishes charging or I guess it would be bricked. Cannot imagine that it stops charging when reaches 4.2 V - hope it stops sooner! Another thing I have not resolved is whether the batteries have to be well ballanced in term of their voltage behaviour. It might be, I suppose, that one badly different spec battery of the three might drag the other voltages down and either negate the extra capacity and/or make the solar power manager keep switching off the rak although there is plenty of juice theoretically available. I tested 6 batteries and took 3 with the most similar charging capacity curves and resistance. Comments welcome. I suggest a good testing period before final mounting. A real shame here is I have no such power data to be able to monitor this in this build and then compare to others…

USB C to USB C cable: https://amzn.eu/d/0i4C7llr but is too tight so am going to use this USB 2 x right-angled connector https://de.aliexpress.com/item/1005004357769641.html

The Waveshare solar power manager is mounted to an acrylic sheet cut to fit into the node box. Over the top of this is mounted the wisblock using the brass risers listed below The orientation is such that a USB cable can connect the two. The mounting is done using the M2.5 screws/bolts from the riser kit listed below, using drilled holes.

Acrylic mounting plate to mount node and solar manager and battery holder to node box: ASIN B08NP929SS https://www.amazon.de/dp/B08NP929SS?ref_=cm_sw_r_cp_ud_dp_9NZJ3ZCT1A97D9PWCZ4C cut to shape with hacksaw.

Screws for mounting acrylic plate to interior of box:, I currently do not need to screw down the acrylic plate into the node box as it fits tightly on its own and this is the last step. But am going to try these https://amzn.eu/d/0j2TGZrQ which might work. Most screws are too long or too thin to work.

Risers and spacers: M2.5 Spacer - Hex Spacer Sleeves PCB Board/Boards Male Female Thread Standoffs https://amzn.eu/d/00nBfm2E with M 2.5 screws/bolts.

Ideally one would use some o-rings to seal the glands to the side of the box but such o-rings would have to be very thin indeed. I did not manage to use any o-rings as this made the glands not fit properly because then there was not enough thread available on the inside side of the box to tighten the plastic nut that fixed the gland to the side of the node box.

The solar power manager needs to be off to the side so that the thick cable for the antenna has room to come in straight and be attached to the IPX to SMA antenna connector.

The antenna mount presses against the pole to which the whole thing is to be fixed. This means that the bolts sticking out the back of plate that mount the node box to the antenna mount must not be located so that the bolts touch the pole against which the whole construct is pressed. This means in practice keep all bolts that stick out of the back of the plate out at the side where the pole will not be located. This can depend upon a few mm only… so hold up the plate against the pole you intend to use to check if the location of any holes you drill for bolts results in the bolts touching the pole.

Improvements to follow: the solar panel cable going into the Waveshare D solar power manager will have an additional strain relief on the inside of the box using the acrylic plate.

The plastic cable-ties are temporary only. The final location will use metal ties that came with the holder.


Steps:

  1. Cut acrylic plate to size to fit snugly into the node box
  2. Note that the battery holder has to be to one side as far as possible as it only just fits, so there needs to be as big a gap as possible on the right to allow the battery holder cables to feed through to the Waveshare solar battery manager.
  3. Drill one hole for the battery holder, holding the battery holder in its correct place (see 2)
  4. Screw the battery holder down at that hole
  5. Drill other holes for the battery holder, using the holder in situ as a “tempate”
  6. Mount the Waveshare solar battery manager using its little brass legs and M2.5 screws but to one side so that there is still room for step 7 below where the Wisblock straddles the Waveshre solar battery manager. This step required drilling of 4 holes.
  7. Mount the Wisblock over the top of the Waveshare solar battery manager, on brass legs. Make sure the legs are firmly screwed together. This step required drilling of 4 holes.
  8. Make holes for the cable glands using the step drill. This can go wrong so go slow. Or practice on another sheet first. So not make the hole larger than needed, else the waterproofing that the gland is supposed to ensure, will not be present. Drill and check, then drill a bit and check.
  9. Fix the plate, for mounting the node box to the antenna mount, to the antenna holder. I did it so that the plate is at the back, against the post, and there is then a step at the “front” up to which the box is set. The fixing is done with just 2 bolts. Do not do it too high, as then the bolts get in the way of the nuts for the u-shape clamps that clamp the antenna holder to the post.
  10. Attach the node box to the plate. This required drilling into the plate. I ended up with the box a bit to the left, which is not good as this means the bolts that stick out the back can interfere with the mounting to the post. The more symmetrical the better. The bolts should not stick out the back too much as this interferes with the mounting to the post. I cut them off flush at the nut.
  11. The biggest pain was the strain relief needed at the antenna gland. The cable going in is not held sufficiently tight by the gland alone. The strain relief was simple in the end. The strain relief is a 60 mm M5 bolt which is fixed using three nuts to the base plate. On the rest of the thread of that bolt is mounted a strain relief clamp (see link). Two bolts, one each above and below the clamp, with washers between clamp and nuts, can be adjusted to compress the clamp and have the resulting compressed clamp be at the right height for the antenna cable going into the gland and box.
  12. Consider putting heat-shrink on the antenna cable before feeding it into the solar node box.
  13. When the antenna and box are joined you can move to constructing the solar panel mounting arrangement.
  14. The solar panel frame already has four holes in it which allow you to fix plates (solar panel adaptor plate, see below) to the panel. This fixing requires four 10 mm M4 cross-head screws. The screws are passed through the plate and frame holes, and need to be screwed into M5 nuts. This is fiddly because the access to the hole under the frame, where the bolts go, is really tight. I needed to use a very long pair of scissors to act as a tong / tweezers to hold the nut in place while I turn the screw. Took a lot of fiddling because the plates that you are mounting also restrict the access further to the place where the nut has to go.
  15. To fix the solar panel’s adaptor plates to the angled mount, I used M4 bolts and both an M4 and M5 washer to stop the bolt head from slipping into the cut-outs of the mount. The part of the mount with the crossed shaped slots goes nicely onto the solar panel adaptor plates on the panel - a bit of a fluke ! The part with the circular cut-outs goes onto the solar panel post mount (see below). This part might have needed a crazy amount of more drilling to get it to mount to the solar panel adaptor plates, but it worked by chance without extra drilling, or due to some sort of norm I am not aware of. Again, as for fixing the solar panel to the solar panel adaptor plates, to fix this solar panel angled mount to the solar panel adaptor plates required quite a bit of dexterity with the fingers because the location where the nuts have to go was cramped underneath the solar panel adaptor plates.
  16. Fix the angle mount to the solar panel post mount using as short a set of bolts as you can, in case you need to bend the post mount for the width of post you are using to mount the node. The post mount has horribly sharp edges, so don’t get blood on your kit…
  17. Set the bluetooth to fixed pin in case you need to connect via bluetooth again and cannot access the device.
  18. Consider setting up an admin channel to allow the node settings to be changed via the mesh. At the time of writing, it seems that only on iOS devices can you in the meshtastic app add a single extra channel to an existing set.
  19. If the device is going somewhere where there is no lightning protection, then take care not to create a lighting rod node and endanger any inhabitants connected to the mounting location.
  20. It might be worth having some sort of metal cover / shield over the box to stop it getting really hot in summer, but I will try without.
  21. I thought of fitting a vent like thishttps://www.distrelec.de/en/pressure-compensating-plug-m12-12mm-ip68-ip69-polyamide-black-wiska-ltd-evps-12-black/p/30158346 but in the build there is an air gap in the solar input gland so no need.
  22. The last step is to fix the solar panel to the screw terminals of the Solar Power Manager. The size of the hole to put the cables in is so small that if you tin the ends of the solar panel cable (after having cut off the jack), then the chances of it fitting are low - so the tinning (putting on solder onto the exposed wire) has to be super thin.
  23. I put more heat shrink on the solar panel cables at the screw terminals and over the antenna connector in the box.
  24. Used some plastic adhesive furniture floor protection pads and foil tape to make box less hot in direct sunlight and made box face north. https://amzn.eu/d/05k9y1Mt and https://amzn.eu/d/0aDtMN4N The bottom of the box has no tape on it - and with the bluetooth aerial inside the box arranged as shown on the bottom right, there was still bluetooth signal from the box.
  25. A last step but one was to make sure that the bluetooth pin was set to fixed and something simple like 123456.
  26. Final step was to establish an admin channel and make sure that the node was able to have its settings changed via this. How this is done is described a little here: Functioning admin channel via android / iOS app? I read somewhere that the admin channel key has to be a certain number of bytes if the channel is to work…
  27. The outstanding question is then how to do updates of the firmwear when needed or desirable. I would have to remove the foil (if the wind and rain has not already done so). I will in fact have to have access to the location of the node in oder to do that. One consolation is that is is really easy to update the firmwear - all the settings remain as they were in that process.
  28. Brief comment about the expense. The batteries and wisblock and antenna and mount and panel came in in total at about 200 euros. The bolts and plates and box brass risers and so on probably at another 100 euros, but the links above should give a more accurate tallly. Such a cost is a bit OTT but the aim was to not to set fire to anything and to provide a 24/7 node. Testing shows at least 13km coverage radius in town if there is a line of sight. Tesing with a station G2 in the car reveals a range of coverage that is vast. The node is not just on a tall building but on a raised platform on the top, so there appears to be a clear 360 degree coverage. This the expense was easily worth it considering the benefits shown thus far.

More build images:











5 Likes

I had already marveled at this in the Whatsapp group and hope you find a top location for your node.

1 Like

How is the Waveshare managing those shutdowns and restarts exactly? Is it just cutting the supply at a certain voltage?

Cheers.

Apparently it cuts the voltage when the voltage of the battery is below a certain level, and when it exceeds another level switches cleanly back on.

1 Like

The coverage of this node appears to be a bit like that shown below (but not entirely sure of height to use for the building it is on, so took 36 m or 120 feet:

map of Munich and surroundings showing horizon lines and regions in red with direct line of sight

from https://www.heywhatsthat.com/

An example of a path in a traceroute from this node.

I further tested using the Waveshare D in a similar setup to the above but using a T-echo plugged into the USB output of the Waveshare D (and with a larger solar panel), as shown in the image, which is purely a test rig to test nodes plugged into the Waveshare D for restart after brownout.




However although the T-echo sprang to life after all the batteries connected to the Waveshare D and its own were flat and the T-echo showed no signs of life, after some running of the node the T-echo stopped being able to send or receive (despite having enough charge) and needed a button press to bring it back into life. So this setup does not appear to be suitable for running a T-echo in solar mode from a Waveshare D. To be fair, I should say my T-echo has been giving me all manner of problems (screen crashed permanently, frozen a few times) so perhaps the manual button press would not have been needed if the T-echo were behaving perfectly in all other respects.

Another odd aspect of this test rig is that the batteries of the Waveshare D do not indicate as charging beyond 25 % even when the T-echo is unplugged. I am still investigating this. I am a bit horrified at how difficult it is just to get the solar aspect working as expected and needed for a node project. Any suggestions for better solar power managers are most welcome.

I found this nice mount on-line for these components: