Store & forward solar node on telescopic 9 m pole, a series of test builds.
Aim: 24/7 operation testing using off-the-shelf components making it easy to implement.
This build relies on the Jackery Explorer 100 Plus battery pack but had issues when plugging the USB C male connector from the solar panel directly into one of the USB C ports of the battery, as intended by the manufacturer, because the panel inexplicably drains the battery flat when there is no sun shining e.g. when indoors being tested. Jackery was contacted about this and they said that USB C is bidirectoinal and so one should unplug the pannel… OMG.
Therefore the simple configuration below initially had a disappointing flaw:
but I think there is a solution below.
So with that disappointment concerning the panel drawing power from the battery, I went down a rabbit hole with a different battery, the Jackery 240, which worked out to work even less well … immediately below.
=========== Jackery 240 fail
In previous versions of this post I was trying with the Jackery 240 and a keep-alive circuit and a different solar panel to the one above, to make this work.
These being the parts:
Battery: Jackery 240 https://www.cyberport.de/haushalt/outdoor-sport/camping-zubehoer/jackery/pdp/qx03-001/jackery-explorer-240-tragbare-power-station.html 56
Keep-alive load: USB Battery Pack Keep-Alive Load
This is shown below:
The Jackery 240 never stayed on for more than 48 hours and switched off bang on 12 hours later or 10 hours later or even faster. I used all the ports of the Jackery 240 (USB A, 12 V with PD USB A adapter, and even the AC outlet with a USB mains charger), all with maximum keep alive current of 0.26 mA., and nothing stopped the thing from switching off of its own accord. I did not see any such indication in the instructions of a timer but then perhaps I did not read them properly. So this post completely discourages use of the Jackery 240 for attempting a solar node for 24/7 operation. Many people has said it was a bonkers choice anyway, but I had it so I tried it while the problem with the alternative shown below was resolved.
========
Build details
Parts list
Node: Lilygo T3 S3 T3 S3 – LILYGO®
LoRa Antenna: Taoglas FW.86.B.SMA.M - Antenne, ISM, 850 … 890 MHz, 240mm, 3.5 dBi, SMA male, Verschraubung, Taoglas https://www.distrelec.de/antenne-ism-850-890-mhz-240mm-dbi-sma-male-verschraubung-taoglas-fw-86-sma/p/30285131 Distrelect Art. Nr. 302-85-131 Manuf. Part. FW.86.B.SMA.M
Bluetooth antenna and fixation parts KHIYQILO WiFi Antenne 2.4GHz 3dBi SMA Connector IPEX Antenne für WiFi Router Drahtlose Netzwerkkarte Sicherheit IP Kamera Videoüberwachung Monitor Bluetooth Antenne: Amazon.de: Computer & Zubehör
(Why a bluetooth antenna? Because the range is far better than using the built-in one and so the node, which is inherently far away, being on a pole, will remain in Bluetooth communication.)
IPEX to SMA adaptors: Hailege 5pcs IPX IPEX-1 U.FL zu SMA Buchse Pigtail Antenne Wi-Fi Koaxial Low Loss Kabel 6 Zoll (15cm) : Amazon.de: Elektronik & Foto with the fixation means
100 x 60 x 25 mm Project Box: https://amzn.eu/d/0x127xM
(This box is not really waterproof but with a bit of will
Update: box too small to include strain relief so needs to be deeper or mount LoRa antenna off to one side so box can still close.
…. Maybe this one would work too, don´t know HeiGroup - T4 ABS Kunststoffgehäuse mit Deckel, Mehrzweckgehäuse, Modulgehäuse, (LxBxH) 111 x 57 x 22 mm, Schwarz - Poliert : Amazon.de: Baumarkt )
Pole: Spiderbeam GFK mast, 10 m: WiMo: Größte Auswahl an Funktechnik SKU: 18362.10
Spiderbeam GFK Maste
Jackery Explorer 100 Plus battery: Jackery Explorer 100 Plus Tragbare Powerstation, 99Wh LiFePO4-Akku Schnellladung, 3-Port 128W TragBack-up-Akku für Reise Notfall : Amazon.de: Garten
Jackery SolarSaga 40 Mini Solar Panel Jackery SolarSaga 40 Mini-Solarmodul, tragbare Solarmodule mit USB-C & USB-A Ports, faltbares Solar-Ladegerät in Buchgröße für Telefone, kompatibel mit Jackery 100Plus/300Plus, IP68 wasserdicht : Amazon.de: Elektronik & Foto
USB C to USB A adaptor e.g. CLEEFUN USB C Kabel 15M, Langlebig USB 2.0 USB A auf USB-C Schnell Ladekabel für Handys, Tablets, Security Überwachungskamera, PS5 Controller, andere Geräte mit einem USB-C Ladeanschluss : Amazon.de: Computer & Zubehör to get the USB A male plug of the 15 m USB cable into the USB C port of the Jackery to power the node.
The following USB C to USB A adapters and USB C male to USB A male cables were used: USB C 3.1 Adapter , USB C Female Adapter , USB C Male Adapter , Compatible with MacBook Pro, Tablet, Samsung Galaxy, etc: Amazon.de: Electronics & Photo and
USB C Adapter Pack of 4, USB C to USB 3.0 OTG Adapter, Micro USB to USB C Adapter Female to Male Compatible with MacBook Pro, Samsung Galaxy, Mobile Phone PC: Amazon.de: Electronics & Photo
with the following cable: TECHGEAR USB C Cable, [30 cm] USB C 90 Degree Right Angle Cable Fits Samsung S24, S23, S22 S21 S20 FE/Plus/Ultra S10, S9, A12 A13 A14 A22 A23 A32 A33 5G, A02s, A03s, A52 A53 A5354 5G, etc : Amazon.de: Computer & Accessories or a higher ampere one 1m USB C Kabel 2 Pack, Nylon 3ft USB A auf USB C 3A Ladekabel Typ C Schnellladekabel Kompatibel für Samsung Galaxy S22 S21 S20 S10 S9 S8 Plus Note 10 9 8, Huawei, Google Pixel, Xiaomi, Sony : Amazon.de: Computer & Zubehör , tests to follow and in progress
USB C cable to go up the long pole, 15 m CLEEFUN USB C Kabel 15M, Langlebig USB 2.0 USB A auf USB-C Schnell Ladekabel für Handys, Tablets, Security Überwachungskamera, PS5 Controller, andere Geräte mit einem USB-C Ladeanschluss : Amazon.de: Computer & Zubehör
M22 cable gland: sourcing map 10Stk.7mm-12mm Kabelverschraubung M22x1,5 Wasserdichtes Kontermutter mit Unterlegscheibe Weiß : Amazon.de: Baumarkt
A cable tie
O-ring kit: use 2 x O-rings about 18 mm outer diameter and 3 mm thick.
Heat shrink tubing of high diameter for USB C to USB A adapters mentioned: damoguin 90mm Extra lange Schrumpfschlauch-Set mit Kleber, 4:1 Wasserdicht Schrumpfschläuche, Heat Shrink Tube, Isolierschlauch-Sortiment mit 5 verschiedenen Größen（schwarz, 100 Stück） : Amazon.de: Gewerbe, Industrie & Wissenschaft though that one seems to have adhesive and is not intended to be removed again ever, so a different sort might be preferred.
Heat gun for heat shrink tubing Steinel Heißluftpistole HL Stick, 350 W Heißluftfön, 400°-500°C, LED-Arbeitslicht, Standfläche, Basteln und Modellbau : Amazon.de: Baumarkt
Tools
String to pull the USB C cable out of the Spiderbeam mast
Step drill (Stufenbohrer) such as Flintronic HSS Stufenbohrer, Sechskantschaft-Stufenbohrer-Set, Kegelbohrer Holz, Professioneller Stufenbohrer 3-teilig 4-12/4-20/4-32 mm Für Metall, Messing, Holz, Kunststoff, Silber : Amazon.de: Baumarkt
Electric drill and 6.5 mm drill bit
Cutters
Steps
Remove the thinnest two rods from the spiderbeam mast, as these are too small in diameter to pass the USB C cable connector. This reduces the length of the pole by about 1 m. Feed the USB C cable into the collapsed Spiderbeam mast from the screw cap end and out through the thinnest mast portion: with the help of a piece of string, pull the USB C cable USB C jack out through the thinnest mast portion so that it sticks out.
Drill a hole in the screw-on cap of the Spiderbeam to allow you to feed the USB A cable portion out of the Spiderbeam. To do this I used a mixture of step drill, 10 cm drill bit, filing and swearing. There is a foam insert in the screw-on cap, and I made a hole in that and pushed the USB A cable end of the cable through that and out through the screw-on cap. Feed the 15 m of the cable through the cap, and then screw the cap back on with the USB cable sticking out of it. Then you have a collapsed Spider beam with 15 m USB cable running through it.
Disclaimer: I mounted the LoRa antenna IPEX connector to the small IPEX connector next to the SMA connector on the node, but found no documentation describing whether that is the LoRa connection. Similarly I attached the Bluetooth antenna to the IPEX connector near the built-in bluetooth antenna, but found no documentation on this. So far this seems to be OK but if that is wrong and the node fries slowly over time, then so be it, it was fun while it lasted. Update: seems totally OK, node is fine after more than several weeks.
Drill a large hole centrally in the box for the node, with the step drill, in order to take the fat cable gland. Fix the cable gland to the box with the thin plastic nut of the cable gland. Remove the rubber O-ring from the cable gland. Place the cable gland external dome-shaped part over the distal end of the Spiderbeam mast section. Feed the distal section through the gland a little and tighten the dome-shaped cable gland part really tightly to fix it, and so the box too, to the distal part of the spider beam mast section. The Spiderbeam mast is slippery so this technique might prove to be fishy in the long run. May need some strain relief at the node in box due to weight of the wire. I tied a knot which pushed down the mast section and passed a cable tie through the knot to make the knot larger than the mast section and thus hold… in theory
Drill two 6.5 mm drill holes into the box to take the SMA connectors of the IPEX to SMA adaptors. I used the lid to mount the LoRa antenna; but doing it the other way up with the lid downwards might be better from the point of view of water ingress. Mount the LoRa antenna off to one side, not centrally as in the current images, else with strain relief (see above) also in box, the lid won’t close. I chose to mount the Bluetooth antenna downwards because this is one less entry point for water.
! The above shows the LoRa antenna SMA connection in the wrong place - is needs to be to the side of the lid (not central as in picture) else the box does not close properly
Warning: the box is actually a bit too small, squashing the contents of the box somewhat, especially if the LoRa antenna is mounted centrally in the lid. Mount it off center. The mast terminal section can push on the node if it is not attached well. Care required. Might not be a fab solution. A different box would have advantages, but the O-rings mentioned below seem to help. Different attachment designs may also probably be better, but I used what I had.
The cable gland is not very tight on the distal end of the terminal Spiderbeam mast section, so the node box is a bit wobbly on the end. An o-ring or two squeezed in the gland between mast and gland edge, on the inside side of the box around the mast section, helps a lot to tighten how the mast section sits in the gland and stops any wobbling. The box sits very firmly indeed with the two O-rings installed. Time will tell.
The image above shows using a screwdriver to push the O-rings down into the gland.
The box is not actually watertight, as it is merely a snap fit, but this can be improved with some last-minute tape around the edges of the lid. No build is complete without duct-tape, after all. The watertightness of the antenna entry point to the box is of course not perfect and maybe some waterproofing goop there would be good.
The Jackery Explorer 100 battery does not seem to allow a small power drain on the USB A output and switches off if you plug the node into that output of the battery, so I needed a USB A to USB C adaptor to provide power to the node via the USB cable from the USB C output of one of the two USB C sockets of the said battery.
Temporary mounting: the base of the Spiderbeam is well supported by the empty slot of a full crate of beer and a workmate, where cable ties can help too.
Pushing this out of a Velux (roof) window is easy and is great fun on a bank holiday.
Permanent mounting: working on it, may involve my balcony railings and cable ties, or a tree to act as a support with bungee cords.
Do not make it into a lightning conductor - keep it away from buildings where people are inside - hence the tree.
Field testing to follow……. (the bank holiday was not long enough, clearly, but that is always the case)
The Explorer 100 Plus battery is the only battery pack I have found (available in Europe) that does not switch off if the power (current) drawn is low so should hold for many days. This makes the device fine even without permanent solar charging. (Voltaic Systems seem to make similar batteries but I cannot get them in Germany.)
This Lilygo T3S3 node uses typically 0.1 A at 5.1 V
which means that the battery should last in theory 12 days (31 000 mAh / (100 mA*24 h)) on the
Jackery Explorer 100 Plus
but I suspect more like 5, but this will be tested.
Field test
Prior to launch put duct tape around lid of box.
An old broken roof box of a car serves as a case for the battery, and a drain pipe can protect the USB cable powering the node.
Used bungee cords to fix spiderbeam to tree trunk. I will put the panel onto the side of a ping-pong table and point it south. Like shown below for another panel (not the Jackery Saga 40 Mini but the one listed above as used with the Jackery 240).
This node is called MS&F
Testing using USB C input via USB A output from panel to prevent drain of battery by the solar panel
USB C indoor Test 1: the above-listed USB C to USB A parts are used as shown in the images below.
because a test over many days appeared to show that the panel, using this adapter, did not drain the battery, and I heat shrunk the adapters together for use in the field:
using a straight USB A plug (1m USB C Kabel 2 Pack, Nylon 3ft USB A auf USB C 3A Ladekabel Typ C Schnellladekabel Kompatibel für Samsung Galaxy S22 S21 S20 S10 S9 S8 Plus Note 10 9 8, Huawei, Google Pixel, Xiaomi, Sony : Amazon.de: Computer & Zubehör ) with higher 3 A rating, rather than the angled one which was only to 2.4 and would not have been easy to heat shrink.
Sod’s law. 3 hours after heat shrinking - battery totally flat. FAILED Test ? The USB C still apparently wants to “power the panel”. Well actually I am starting to think the display is not showing the right battery level so perhaps in fact the node simply drained the battery.
I used a T-beam Supreme ( T-Beam SUPREME Meshtastic – LILYGO® ) rather than the T3 S3 node in the tree, as this is an indoor test that avoids me having to take down the pole and node.
The Supreme appears to use about 10% of the capacity of the Explorer 100 Plus in 24 hours, according to its display, but I do not believe the display data.
The string of USB A adapters (listed above), plugged together as shown in the images, do allow for charging of the battery, as shown below where on a foggy day at noon 1 W is pumped into the Jackery. Note that 1 W more than what the Lilygo T3 S3 used on average, so this augers well.
=====
The Jackery Explorer 100 Plus display
The display of the Jackery Explorer 100 Plus shows a percentage. But what is this number?
I tested to what percentage the panel alone brings down the battery, from 100 %, in 48 hours. This was the result:
At the same time I tested T-beam Supreme for 48 hours to see what the total power x time was that the node drew. The measurement device used is this USB Tester Messgerät USB C Power Meter Strommessgerät 2-in-1 Digital Multimeter, Spannung Voltmeter, Strom, Ladestrom Leistungskapazität, Arbeitszeit-Detektor Amperemeter mit Typ-C Adapter (A3) : Amazon.de: Baumarkt . This was the result:
The bottom of the Jackery Explorer 100 Plus contains this information:
The battery has a capacity of about 100 Wh
It is listed as 31000 mAh 3.2V in the image.
The percentage shown is NOT the capacity remaining but the percentage of 31000 mAh used, which is a meaningless representation of the percentage capacity remaining as the voltage is not known.
Thus after 48 hours, the node (plus meter, which uses some power due to the display) used up 30 Wh
and
after 48 hours the panel used up 10 % of 31000 mAh, which is 3.1 Ah.
My testing indicated that when both the node and panel were plugged in to a 100% full battery, it went flat after about 3.5 days. That was the mystery, as shortly before that the battery showed about 67 % and it made no sense.
So I assumed 12 V for the panel when it is being “driven” by the battery, and then one comes out with a drain to the panel in 48 hours of 3.1 Ah x 12 V = 37 Wh approx. This explains why the battery was empty after about 3.5 days when both the node and panel were plugged in: because the Wh used in 3.5 days, consisting of the panel “dark drain” plus node drain, then comes to about the battery capacity of 100 Wh.
I note that the panel “dark drain” is even more than the node itself uses. This “dark drain” has been the problem from the outset. But now at least it is quantified (and clearly my fancy the adapter has no effect at all). The problem in understanding what was going on was to believe that what the display was showing was the remaining capacity.
The conclusion: the Jackery Explorer 100 Plus with the SolarSaga 40 Mini can only power the node about 3.5 x 24 hours when it is dark all that time, and this relatively poor performance is due to the solar panel “dark drain” draining the battery when dark. As I said above, Jackery responded to this issue by saying one should unplug the panel when it is not charging.
Of course outside of the house the panel will have sun during the day and will hopefully stand a better chance of staying on 24/7, even if there are stretches of many days in winter where there is extremely little panel illumination. It is the a matter of luck though if the node stays on 24/7, as this would depend on the weather.
This grossly ridiculous design of the panel’s USB interface, meaning that it drains power from the battery when the panel is not generating power i.e. at night, has to be countered if this off-the-shelf setup is going to perform as well as possible over a relatively dark winter, and thereby stand a chance of keeping the node running 24/7. I do not want to waste all that that power due to the “dark drain” at night.
So the next step is to use a switch to cut, during the night, the electrical contact of the battery with the solar panel. I wish this project to be off-the-shelf and so easy for anyone to copy. Therefore I am going to test the off-the-shelf USB device shown below to prevent the panel from draining the battery at night:
The results will follow once I have the device, which is on order.
(Using a diode is the obvious thing to do, but that would take the project away from being off-the-shelf, because it would require splicing of the USB cable and testing the effect on the USB C interface e.g
will the interface still negotiate correctly and accept the power from the panel).
The Supreme node alone plus current meter seem to have drained the batter in about 65 hours, noting that I left the display of the meter on all this time.
=====
USBC indoor Test 2: the following arrangement could also be tested but reduces the length of orange cable available to run to the battery:
This might theoretically stop the “dark drain” in view of the potential effect on a USB C interface of being connected to a USB A female power source.
Testing this arrangement: after 10 hours the display still says 100. This might be the way to go, and tap off the solar power from the USB A female port. The tests shows the USB A socket drains no power at all from the panel. However the charging via that socket is limited to 5 V 2.4 A so wastes 30 W of power from panel and charges more slowly requiring sun for longer.
Testing in the field how long the Lilygo T3S3, when outside up on the 9 m pole, runs on the Jackery Explorer 100 Plus at 100 % when it alone is plugged in:
Why test all these crazy adapters ?
Part of this test was to use USB A connectors with the panel, to present to the Jackery Explorer 100 Plus USB C port a USB A male cable plugged into a USB A female port. This tests to see if that way the USB C protocol and adapter cables used can make the Jackery not think the panel is an item that needs to draw power, but only one that is to provide power. I was inspired by this chat : Reddit - Dive into anything
and so I decided that testing it was better than trying to understand the whole USB C with USB A protocol behaviour. (Side remark: I tried performing tests using an artifical MILYN LED Pflanzenlampe mit Timing und hängendem System, 9H/12H/15H Timer, 150LEDs Weiß Vollspektrum Grow Light mit 16,4ft Netzkabel für Gewächshaus, Innengärten, Zimmerpflanzen, Hydroponik : Amazon.de: Beleuchtung but that does not drive the panel at all.) The outcome was that the adapters did not have any effect at all, and the battery is still persuaded that the attached panel is to be powered! But it was worth a try to see if the USB C protocol could be made to do my bidding using simple connectors. This effect, if not fantasy, might come into play in “USBC indoor Test 2” above.
I can still run the device on a yet different solar setup I have in mind, which is so OTT that I deleted it from here but keep it as a crazy backup brute force solution (100 W panel with 100 Ah LiFePo4 battery that has USB output that never ever switches off just because the current drawn is low, and this charger Victron Energy SmartSolar MPPT 75V 15 Amp 12/24-Volt Solar Laderegler (Bluetooth) : Amazon.de: Gewerbe, Industrie & Wissenschaft controller or similar.)
Store and forward node and channel configurations
S&F seems not to work on nodes with firmwear 2.5, at least not if the secondary channel uses AQ== as PSK (the default key, see here **S&F - Not available on this channel). If you try to use 2.4 nodes to talk to 2.5 S&F nodes, good luck. The upshot is actually that if you want to use S&F reliably you had better move to firmwear 2.5 in the S&F node, and all nodes that what to retrieve stored messages from the S&F node should be 2.5 and need to have the same fully encrypted private channel as the S&F node. This topic is still being worked out in Munich, where we are setting up a “private” encrypted channel for everyone to use i.e. the PSK is made public, so that everyone can benefit from S&F by having that channel in their channel list. The S&F topic is a bit out of scope of this post, which is about the build and getting it working in principle. The settings and so-on could be another topic, but thought I would just mention the above issue as a heads-up because once the node is working, the next issue is the settings in view of firmwear changes.
IF the solar setup works and the device can run on solar power 24/7
the the next issue is:
Coping with freezing conditions
A next step is to switch off the solar generation when power bank is below 0 °C and to perhaps heat the power bank enclosure using solar power when the box enclosure is below 0 °C and the sun is shining. A friend pointed me to this beauty which heats the battery: EREMIT 12V 100Ah LiFePO4 Batterie für Camper Wohnwagen Boot CB-Funk but father Christmas is not going to play ball on that one. The supplier of that battery, when I asked, was positive on it working in combination with the following solar manager . This solution is expensive but still off-the-shelf.
BTW: This node sort name is: MS&F
Why go to this length? Well the sorts of nodes that currently provide Store & Forward (S&F) are ones that need quite a lot of power, and making them solar is a bit of a challenge. Also such nodes, if they are to be most useful, need to have a good location so that most missed messages can be retrieved from them.
Store and forward works on Android like this:
provided you have a secondary channel set up that the S&F node and you both have set in your node settings, as well as set in the the message sender’s node, you can retrieve messages you missed from the sender by sending the S&F node the direct message SF. Then your messages are re-sent by the S&F node and appear in your node provided the S&F node, due to it being on and radio-reachable by the sender, was in a traceroute from the sender to you. This is an extremely handy service.
