Date: Mon, 3 Sep 2018 21:41:02 -0400
Reply-To: David Beierl <dbeierl@ATTGLOBAL.NET>
Sender: Vanagon Mailing List <vanagon@gerry.vanagon.com>
From: David Beierl <dbeierl@ATTGLOBAL.NET>
Subject: Re: Westfalia Water Tank Level Indicator Issues
In-Reply-To: <CAMOH8LKG2Mmt89ZGhBH5RADUCCRUAE43Y9UyaWFJaNAWi1_5hA@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Note that the magic box total resistance must be at least a megohm and a
half to get the LEDs to shut off when the level falls below the last bolt.
That means that at that point there's only two and change microamps flowing
in the sender circuit, so your input impedance needs to be quite high. The
three LM324 op amps in parallel take about 0.015 microamps, so effectively
that's over fifty megohms input.
Yrs,
d
On Mon, Sep 3, 2018 at 9:01 PM, David Beierl <dbeierl@attglobal.net> wrote:
> Argghhhhh.....left out the one megohm feeding the sender which becomes
> part of its network.
>
> So B+-----1MΩ------sender wire+top bolt---/\/\/\/(black box)----next
> bolt----/\/\/\/---third bolt----/\/\/\/----B-
>
> Yrs,
> d
>
> On Mon, Sep 3, 2018 at 8:44 PM, David Beierl <dbeierl@attglobal.net>
> wrote:
>
>> I've been playing with arduino off and on for a while and this seems like
>>> an application for that. it could probably be configured to read the
>>> water
>>> contacts such that there would be more leeway as far as all the
>>> connections
>>> being perfect and variations in water source.
>>
>>
>> The resistor network that the panel compares against is:
>>
>> B+ --------150K---100K---R---100K---Y---100k---G---150K--------- B-
>>
>> That makes the green light at 25% of B+ or less, yellow from 25-42%, red
>> from 42-58%, and no LED above 58%.**
>>
>> The network for the sender is similar:
>>
>> B+ is fed through one megohm to the green sender wire at the top bolt,
>> and the bottom bolt is grounded.
>>
>> So B+-----sender wire+top bolt---/\/\/\/(black box)----next
>> bolt----/\/\/\/---third bolt----/\/\/\/----B-
>>
>> As the water rises it first touches only the ground bolt, for no LED.
>> Then it successively shorts the bottom bolt for red, middle for yellow, and
>> top for green through the water to the ground bolt. I don't have a magic
>> box to measure the values, but they have to be such that the resulting
>> voltage divider with the tank water in parallel with the bottom n segments
>> will cause the output voltage to be between the thresholds.
>>
>> Since both networks are connected to B+ the absolute voltage doesn't
>> matter -- it's the proportion that counts.
>>
>> The voltage lights behave similarly except they compare against a network
>> fed by a 5.1v zener diode.
>>
>> And then you could also set
>>> the voltage level lights to switch at specific voltages.
>>
>>
>> You need proportions rather than absolute voltages unless you supply a
>> regulated voltage to the sender.
>>
>> The existing board is simple and works well, sometimes with a tweak
>> resistor added to the comparison network.
>>
>> The later float wand of course gets rid of the uncertainty of water
>> resistance at the expense of mechanical fragility/leak susceptibility and
>> losing the lowest (no LED) level indication.**
>>
>>
>> **Note that the '80-81 panels with four water LEDs G Y R R and no flame
>> detector will light the bottom red in this case, rather than showing no
>> LED. They will also shut off the water pump at that time.
>>
>> Yrs,
>> d
>>
>
>
|