Token's lvl1 blog- edit -- Token's rantings

Had to revert to SLZB OS v 2.8.7 AND Zigbee Coordinator 20240710 to get restored service.

At time of post SLZB is up to v 2.9.8 and Coordinator is at 20250325.

SLZB-06 updates has been really awesome until this so I hope they are not getting too fragmented and hairbrained on updates. I hope the SLZB’s are not already past a golden age.

NOW the other thing I noticed via looking at the logs:
I think my unifi controller container is way outdated and an API is upset causing log events (https://redacted:redacted/v2/api/site/default/trafficroutes received 404 Not Found)

and

my ESP32 garage device keeps having a heap overflow. Looks like in ESPhome builder I need to add a line to either increase buffer size, change idle time or some kind of RF filtering- or all of the above…

I was randomly doing a Tasmota vs. ESPhome list in my head and had Tasmota’s only “gotcha” win that it hosts a webUI so that if your HA + MQTT is down you can still navigate to an IP address and still control it (say you are trying to open a garage door in a pinch).

So then I wondered… does ESPhome offer something simular?

It does, and it was one commented out line and one line of code to do it. Wow.

So now I have that redundancy that helps me sleep at night.

1 Like

For what it’s worth I have settled on a diverse ecosystem of zwave, zigbee, and esphome to try and spread out the failure modes.

  1. My initial plan was to migrate away from zwave but now I’m going to keep both, especially when realizing there is one door I thought to keep both a zwave and a zigbee sensor on it to have redundancy both in two sensors but also for two different radio systems and associated hardware (zwave usb stick with passthrough to a VM, zigbee via ethernet connected controller).

I’m really looking forward to the official release of the SLZB-MRW10 to get rid of zwave being a usb stick.

Home Assistant came out with their own device that by all accounts seems to be the best on the market. Downside is that it is USB.

I ended up going with this guys product he builds and have been very happy if you want to stick with PoE:

If it is wired into mains or in the walls I try to stick with zwave as I believe their quality standards are higher and the protocol is more resilient to wireless interference. If it can plug in or is outside the walls then I don’t really care as much. I still try to stick with zigbee/zwave to keep traffic off my WiFi but some stuff requires esphome like mmWave and garage door stuff.

1 Like

That is sick! And I love supporting people’s small side hustles like that.

When is the next lunch meet going down?

I know I’m overdue for a trip up your way. I am in Illinois this week attending a training. I’ll have to see what my schedule looks like next week.

1 Like

Got some new hardware and happy it seems 100% nix compliant as I really want to be 100% Windows free as 11 becomes mandatory (with an exception to a win VM or laptop just for work).

Inspiroy H950P drawing surface/tablet.

1 Like

@PhaseLockedLoop @THEkitchenSINK

Airframes meshtastic filter batch 14 seems to be shipping out now. Woot.

2 Likes

I should really get a HAM license… I cant legally use an outdoor transmitter without one…

Not even meshtastic? Where are you loacted?

Nice

Hey I posted this in several mesh communities to help them out with charge board issues so I’ll repaste my PSA:


Hey folks, I have been running a long-term experiment on one of my nodes and comparing with two others and I think I figured out what causes the strange behavior with these MPPT charge boards.

The problem is capacitance on the battery rail. If you add too much bulk capacitance directly across the battery input the MPPT can get stuck hunting between CC (constant current) and CV (constant voltage) modes. When that happens it never charges fully. It just oscillates and sits around 30 to 50 percent state of charge without climbing.

Why does this happen? The MPPT assumes it is only looking at the battery. If the load is tied directly to the battery rail the MPPT also sees the device through the cell along with the extra capacitance inductance and resistance. That throws off its algorithm and confuses both charge regulation and the fuel gauge. Instead of smoothly moving into CV mode it keeps backing off and the battery stalls halfway.

The fix is simple. Put your bulk capacitors on the load rail, not on the battery rail. For example if your device runs at 5 V but your battery is 3 V on an Adafruit MPPT, place the capacitors on the regulated 5 V load side. The MPPT then sees the battery in a clean state while your device still gets the smoothing benefits.

If you cannot move the caps off the battery line you can add a small series resistor around 0.22 to 0.47 ohms between the battery and the load. This isolates the MPPT from the added capacitance and helps it regulate normally.

So keep big caps on the load side or isolate with a resistor and you will avoid the false hunting behavior. I have tested this over time and it fixes the charging stall.

2 Likes

Philippines.

Law says I need a license for an outdoor antenna. Realistically I don’t think it is enforced because I am in the middle of nowhere and yet I am still in the vicinity of the provincial capital. I’d rather not push it and be caught without a license.

Nice!

A friend was just telling me how he used CB (or maybe HAM bands) a lot over there and I’m 99% sure he wasn’t licensed haha.

1 Like

I wonder how heavily policed is the RF airwaves there in the US. How often do people get caught and fined with having unlicensed outdoor transmitters.

According to notarubicon and ham radio crash course there seems to be only a few documented cases over the span of what must be years if not over a decade.

And those ones are really egregious, like using 1000watt setups jamming or messing with nets really badly.

1 Like

Freq51 community is going to be doing some testing of the faster modes with our routers. we are trying to establish worst case link budget connectivity . People will lose connectivity this weekend but oh well. It’s a necessary evil

LF52 switch at ~8AM tomorrow morning. (24 hours of testing)
MS51 switch at ~8AM the following day (09/26)
MF51 switch at ~8AM (09/27)
SF51 switch at ~8AM (09/28)
Switch back to LF51 at ~8AM (09/29)

I trully do believe we have the infra for shortfast and if that all works out we will have a far more robust long distance network soon ™

Thank you for your attention in this matter

1 Like

Im just excited these filters are arriving. They are so highly reviewed and i can finally build out my “main” node that will have the most LOS of all my available locations.

1 Like

These things are heavy. But now finally i can work on enclosure sizing, install and testing

ROFL @Token @THEkitchenSINK @SgtAwesomesauce … your gonna die of laughter. We are cooking up a hilarious easter egg only for our tower mounted infra nodes that have SDRs that can transmit…

We are going to do something identical to this in the ISM band but with the following change

Adding a tone:

Male Voice: Me
“At the tone 0 hours 51 minutes Coordinated Universal Time”

Female Voice (Someone from the hackerspace or my little woman… who knows)
“What follows is a scientific modulation test. For more information visit Freq51.net/onboarding

Just for giggles since we can do it legally under 1 watt LMAO

4 Likes