Thanks, Josh and Mike. I was hoping you'd chime in. The best practices for repeater config is especially appreciated. I was describing a scheme with two repeaters here, the one relying on the other to escape the neighborhood.  And thanks, Mike for the pointers to support tools.

Pete

On 8/25/26 16:26, Josh Wyatt wrote:
Greetings Pete,

Looks like you've made great progress!  Some notes/things that jumped out while reading your thorough and informative email:

1. Transmit power from the repeater is important for your repeater to be able to reach other repeaters and others' companion nodes, not just your own companions; unless you're in an area flush with other repeaters, and/or want your repeater to be "household only" (AKA not interacting or contributing to the existing mesh beyond your location), you'll want to run power at the maximum possible.

2. The byte hash size should be set to 2 or 3. I think what happens is, a hash is calculated using N bytes from the key, which is then used as a unique identifier for the repeater/node. A single byte hash results in collisions when trying to uniquely identify contacts/nodes. The caveat is that older firmwares can only do single byte hashes, so it's somewhat of a backward compatibility thing. I don't think any nodes in the rdumesh are running such old firmware.

3. I can't offer advice on the telemetry bits; I've also tried making it publicly available without much success.

4. Two excellent resources, if you're not already aware, are the rdumesh.org <http://rdumesh.org> website, and a couple of discord servers, "RDUMesh" at https://discord.gg/GBUyuk2H6, and "NC Mesh" at https://discord.gg/jsxevk7zp . There are other discords for meshcore, LoRa, etc, but these are regionally apropos for us and pretty active.

5. Do note that another set of recommended repeater configurations for our area are region settings. They're meant to curb wayward traffic over time. Details here: https://rdumesh.org/meshcore-region-scoping-setup-guide/

Happy Meshing!
Josh

On Tue, Aug 25, 2026 at 2:55 PM Pete Soper via TriEmbed <[email protected]> wrote:

    Hi Glen,

       I got as much config as I could up front and suffered with
    making changes after hoisting the node. It blows my mind that
    there is no "help" command in CLI mode. That's so 1960s astronaut
    in a world with megabytes of available flash.

      I set an admin password and put a copy of the keys in a safe place.

      Putting latitude and longitude in place and making this
    available is important too unless you aren't comfortable with
    folks knowing where you are. I didn't bother to hook up a GPS but
    configuring this up front is also on the check list.

       I swear I configured my repeater to allow folks to see my
    telemetry (DC supply voltage and ESP32 chip temp), but it appears
    to require admin login to access, even with a "blank" password. I
    obviously didn't understand what I was doing and am left wondering
    how to fix this with some other definition of "blank". I think I
    missed this step doing the initial configuration and "fixed it"
    with a CLI session. But you can't put the set command in with no
    characters after the resource name: it flags this as an unknown
    command (IMO it's silly to call this "unknown"). I eventually
    found the settings interface and was able to change the password
    without typing anything into the text box but that didn't change
    anything.

      I set minutes between to "adverts" to 240 minutes (longest
    interval allowed) and "flood advert" every 25 hours so the advert
    time of day would wander around the clock over time.

      I set the radio to 50% transmit duty cycle: the default. This is
    a regulatory issue, by the way and 50% is max in the US. The
    allowed duty cycle is much less than 50 in Europe, for instance
    (10% I think). This is to cap potential interference and give
    other stations a fair chance to use the band.

      I set power to 22dbm. With the Heltec 32 V4 there is a 17db
    attenuater between the radio and power amp but I was too lazy to
    look at specs and see the most efficient way to get full power
    output. That's on my list. A repeater that is inside your house
    should not need anywhere near the max. My "companion-driven mode"
    is set to 10dbm but I suspect it could hit the repeater with less
    than that. Note that vertical antennas with gain have attenuated
    signal/receive capability "underneath". Picture a cone with a
    varying angle to the "skirt". So my repeater's skirt encompasses
    some major fraction of the house. It would have been better to
    have it further away because of this.

      I temporarily set coding rate to 8. The default 5 gives "5-4 =
    1" correction bit per 4 payload bits, 6 gives 2, 7 gives 3, 8
    (max) gives 4s correction bits per 4 payload bits.  My bias with
    this stuff is to give myself maximum chance of success until the
    node is actually handling traffic properly and needs optimization.
    But this extends "on the air time" significantly and ties up other
    node's receivers, so that's not very neighborly while you're
    gathering data for optimization. One of a gaggle of changes I'd
    like to make to the firmware is per destination optimization of
    the coding rate, TX power level, etc. If anybody is doing their
    own firmware builds I'd love to hear about any relevant build
    tips. IMO there shouldn't even be fixed parameters but instead for
    stations fixed in place the system should adjust to every node it
    has "conversations" with over time. Notice that our WiFi networks
    don't blast at full power all the time.

      I selected one byte hash mode without having a clue what this
    means for the chances of joy or sorrow. The choices are one, two
    or three bytes and this also affects "time on the air".

    On 8/25/26 13:25, Glen Smith via TriEmbed wrote:
    Pete,

    Any thing specific that needs to be done with configuration re
    repeater vs a companion node? I think I'm going to do the same -
    deploy a repeater and see how things go. According to the map I
    have other repeaters closer to me than you do. Some careful
    branch selection may allow me to put the trunk shadow Mike
    mentions in an optimal position.

    On Tue, Aug 25, 2026 at 9:55 AM Pete Soper via TriEmbed
    <[email protected]> wrote:

        I put a Heltec 32 V4 configured as a repeater near the top of
        a tree
        roughly 60-70ft above ground yesterday and initially the
        results were
        very discouraging: I could no longer hit the Carolina
        Backhaul in Chapel
        Hill I'd been clinging to earlier; by setting coding rate to
        8 aka "max
        error correction bits" I could ping it and get an ack one
        time out of
        three. But I left everything running over night and this
        morning the
        contacts in the companion driven node connected to my
        repeater had maxed
        out. One of them jumped right off the screen of my phone: DH
        Hill
        library at NC State 14 miles away. I  tried pinging it and
        had immediate
        and repeatable success! So I have at least one reliable path
        out of my
        "tree farm near Apex" with this setup and hopefully there
        will be others
        in the companion node's contact list and data structures
        inside the
        repeater.  I'm fully expecting this link to break down on
        rainy days
        until I can get a node above most of my trees. Still, this
        success will
        keep me going. And this setup is a hack, running with a 10AH
        battery.
        I'll keep it up as much as I can after protecting the guts
        from rain and
        as time allows swap in solar power and hopefully get it near
        the top of
        a very tall Tulip Poplar here. The immediate priorities are
        to turn
        attention back to gathering info and publishing it on the
        TriEmbed web
        site and proposing some fun-oriented tutorial meetings.

        Pete



        _______________________________________________
        Triangle, NC Embedded Interest Group mailing list

        To post message: [email protected]
        List info:
        http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
        TriEmbed web site: https://TriEmbed.org <https://TriEmbed.org>
        To unsubscribe, click link and send a blank message:
        mailto:[email protected]?subject=unsubscribe
        Searchable email archive available at
        https://www.mail-archive.com/[email protected]/


    _______________________________________________
    Triangle, NC Embedded Interest Group mailing list

    To post message:[email protected]
    List info:http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
    TriEmbed web site:https://TriEmbed.org <https://TriEmbed.org>
    To unsubscribe, click link and send a blank 
message:mailto:[email protected]?subject=unsubscribe 
<mailto:[email protected]?subject=unsubscribe>
    Searchable email archive available 
athttps://www.mail-archive.com/[email protected]/
    _______________________________________________
    Triangle, NC Embedded Interest Group mailing list

    To post message: [email protected]
    List info:
    http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
    TriEmbed web site: https://TriEmbed.org <https://TriEmbed.org>
    To unsubscribe, click link and send a blank message:
    mailto:[email protected]?subject=unsubscribe
    Searchable email archive available at
    https://www.mail-archive.com/[email protected]/
_______________________________________________
Triangle, NC Embedded Interest Group mailing list

To post message: [email protected]
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: https://TriEmbed.org
To unsubscribe, click link and send a blank message: 
mailto:[email protected]?subject=unsubscribe
Searchable email archive available at 
https://www.mail-archive.com/[email protected]/

Reply via email to