Re: WiFi range ...

On 05/08/2026 06:52, c186282 wrote:
> On 8/4/26 06:37, The Natural Philosopher wrote:
>> On 04/08/2026 01:20, c186282 wrote:
>>> On 8/3/26 07:22, The Natural Philosopher wrote:
>>>> On 03/08/2026 05:12, c186282 wrote:
>>>>> kind-of EXACT humidity readings are a sort
>>>>>    of black art.
>>>>
>>>> The presence of a tree exuding transpired water can totally distort 
>>>> a reading
>>>> As cam a crowd of sweaty people
>>>
>>>    Immediate environment is indeed a complicating factor !
>>>    Similar issues with getting accurate temperatures. Sun
>>>    hits a dark or concrete patch 50 yards away at a certain
>>>    time of day and it may send a plume of hot air at yer
>>>    device.
>>>
>> My oil tank monitor regularly shows temps over 40°C in its black case 
>> with the sun shining on it
> 
>    Never had to monitor an "oil tank".
> 
>    But, if exposed to sun/wind/etc, your ARE gonna
>    get weird figures all day.
> 
The ultrasonics delay that measures oil level changes with tempertaure 
as does the depth of the oil due to expansion., I WANT to know how hot 
the tank is.

>    As said, Temp/Humidity is a ROUGH figure always.
>    Never try to get TOO technical on those figures.
> 
>>>    As such, most such measurements are going to just be
>>>    "(maybe) Good Enough".
>>>
>>>    There are standards for professional instruments like
>>>    the Weather Service might install at airports and such.
>>>    A certain kind of housing, placement, height and such.
>>>    That doesn't mean the readings are 100% accurate, just
>>>    that they're done in a standard fashion that can more
>>>    or less be compared to similar installations.
>>>
>> Yes.
>> The whole concept of 'global average temperature' falls apart when you 
>> realise most of it is interpolated from stations thousands of miles 
>> apart mounted in less than ideal positions.
> 
>    This "science" falls apart once you pay even the
>    slightest attention.
> 
Certainly to the depth of precision and alarm its stated with.


>    WHAT the fuck is going on - today or this year - is
>    always kind of a mystery. Butterflies and more
>    butterflies .........
> 
Less butterflies bees and all sorts of insects year upon year. Just 
infestations of ladybirds, fruit flies and house flies.

>>>    I've used the olde-tyme mechanical drum recorders. It's
>>>    amazing how much "jitter" you could outright SEE on the
>>>    paper. Even opening the instrument box caused the figures
>>>    to jit up or down a little right before your eyes. Means
>>>    "accuracy" here is always a somewhat soft figure.
>>>
>> Agreed
> 
>    From those devices - hell YES. Even tiny shit
>    would, visibly, influence the figures.
> 
I always laugh at the people who say 'solar panels work under cloud'.
Some of us did photography and tow stops in cloud or near sunset - 4 
times less light - was the order of the day.

Its all part of the Liberal non scientific brain that doesn't do math, 
quantitiy or any kind of gradation. It's all black and white.

V8 bad. Solar panel Good


>>>    Did a one-off climate gadget for a pro greenhouse once.
>>>    They had data saying the crop qual would be best at a
>>>    very specific temperature and humidity. Spent something
>>>    near $400 (mid 80s dollars) on something that'd be
>>>    accurate enough to make them happy. It was a wet/dry
>>>    system (thermocouples I think but can't remember) and had
>>>    a little fan to suck air past the sensors. You had to
>>>    re-fill it with water every so many days. Used an 8051 uC
>>>    for recording AND actuating 'mist' squirters (just a
>>>    solenoid). The clients sold to the 'elite organic vegan'
>>>    crowd and could get top dollar for like mini tomatoes
>>>    and weird salad green shit I'd never eat. Pretty simple
>>>    programming with just a little IQ added to note the
>>>    RATE of humidity change and react accordingly.
>>>
>> Nice fun bit of coding
> 
>    Really wasn't difficult at all. Close to 'fuzzy logic'.
>    Easy to do with even a '51 so long as you boosted vals
>    to integers.
> 
>    Did have a big tome on fuzzy logic - dumped it when
>    I retired. It DOES have its practical uses though.
>    Adds some crude "AI" to things with minimal CPU
>    consumption.
Yes.

> 
>>>    Anyway, a home-grown wet/dry setup is very easy to
>>>    realize using ds18b20s and something even as small
>>>    as a Uno or Pi0. Adafruit(?) does or did sell a
>>>    tiny uC unit (ard compat) with a couple of A/D pins
>>>    and I think I2C and/or 1-wire connections. "Trinket" ?
>>>    Bought one long back but never got around to using it.
>>>    Really small and low power though.
>>>
>> I've just used TMP36s on the PCB coupled to a PICO Ws ADC. If the Pi 
>> pico W is permanently powered up I subtract  four degrees to 
>> compensate for onboard heating
> 
>    PICOs seem pretty good - although the instructions
>    on-line seem WAY too complicated. Pref'ed PICS back
>    in the day. A little ASM mixed with 'C' and ....
> 
>    Do HAVE a PICO ... haven't yet found a use for it.
>    Will eventually.
> 
I found more trouble in setting up the development enviroment (C SDK) 
than in coding the little buggers

Although some of the networking libraries are rdacically non Unix in 
their layout


>    I'm gonna still boost the Ard Uno for 'small' stuff.
>    Massively supported, a lib for EVERYTHING. Cheap,
>    simple, easy to program. Even has low-power libs
>    that can put it to sleep between Needs - making
>    solar-powered apps work good.
> 
>> In practice it seems good to about half a degree °C
> 
>    The 18b20 ??? It IS pretty damned good - and,
>    with one-wire libs, easy to deal with. Used a
>    lot of them with uC's, Ards, and lower-end PIs.
> 
No. The TMP36.
>    More than "good enough".
> 
>    LAST one I did, for a Pi3, monitored temps in
>    the "server room". Worked great. Little client/
>    server app (python) to deliver the info. The
>    "server room" was an old CLOSET ... an A/C vent
>    was added but it STILL kinda ran hot. Six
>    'servers' in there.
> 
>> But as you say, issues can be summarised into two camps
>>
>> 1. How accurate is your equipment?
>> 2. How representative is it's siting of the environment you want to 
>> monitor?
> 
>    Don't forget "How Accurate Do You REALLY Need".
> 
>    Temp/H ... always kinda "fuzzy".
> 
>> I may do something with humidity too. Experience suggest you need 
>> about 1°C extra heating for every 10% reduction of in-house humidity.
> 
>    If you WANT ... I'd suggest the little box with two
>    18b20s - one set with a "wet bulb" ... and a tiny PI
>    fan that sucks air over the sensors. This is CHEAP
>    and EASY - you can fab something from junk easily
>    in an afternoon.
> 
Nah. I'll juts buy one of those modules that do it all with resistance 
or capacitance

>    One of those 'travel-sized' Tylenol plastic
>    bottles are good for yer wet-bulb. Arid
>    climate, refill once a week. Humid climates
>    you might get by longer.
> 
>    Time ... start fan 30 secs early. Take readings.
>    Turn off fan. 5/10/30 minutes later, repeat.
> 
>    PICs ... long back MicroChip sold a nice all-purpose
>    serial programmer box. Fit anything from 12xxx to
>    all the other 8/16 chips. Great. THEN they Went
>    Bad ... NOW expensive "ICE" units and you have
>    to buy pretty dedicated adapters for almost EVERY
>    kind of chip. A horrible rip-off. USED to use a
>    lot of PICs - you could even use them to emulate
>    'smart' 4000 series logic chips if you didn't need
>    hyper speed - CHEAPER than many 4000 series. Have
>    not used them for a good decade now ... they just
>    made it too expensive/complex.
> 
>    Idiots.
> 
The model of simply downloading a file over USB to code the Picos is 
nice and simple.

-- 
In theory, there is no difference between theory and practice.
In practice, there is.
-- Yogi Berra