Maker Pro
Maker Pro

Heathkit rain sensor problem

G

George Rachor

Jan 1, 1970
0
I've got an old heathkit weather computer where the rain sensor
has become very erratic. I've disassembled the rain collector and don't see anything
wrong with the mechanics.

The sensor seems to be recording much more rain than is possible during the
measured period. I noticed that the heart of this is a reed relay.

Can these relay's fail in such a way to be very bouncy?


George Rachor
 
R

Robert Baer

Jan 1, 1970
0
George said:
I've got an old heathkit weather computer where the rain sensor
has become very erratic. I've disassembled the rain collector and don't see anything
wrong with the mechanics.

The sensor seems to be recording much more rain than is possible during the
measured period. I noticed that the heart of this is a reed relay.

Can these relay's fail in such a way to be very bouncy?


George Rachor
*What* belongs to the relay?
DO NOT use the apostrophe, unless it is needed and required.
The *relays* either work or do not work.
If the drive to the relay has decreased a lot, then excess chattering
at pull-in is possible; but the relays are not at fault.
Use some logic.
What is used for a detector?
Is a float used, and a magnet is on the float, and the magnet
operates one or more relays?
Something else?
 
G

George Rachor From:

Jan 1, 1970
0
OK... Bad terms....

It isn't a reed relay but a magnet driven switch inside a glass bulb.

The switch simply closes a circuit generating a pulse that the weather computer reads everytime
enough rain enters the collector. The rain gathers in a cup that when it gets heavy enough it tips
the collector cup. This action causes a magnet to slide by the magnet driven switch.

The weather computer is seeing 4-20 inches of rain in a 24 hour period and I know that is just plain wrong.

George Rachor
 
R

Robert Baer

Jan 1, 1970
0
George said:
OK... Bad terms....

It isn't a reed relay but a magnet driven switch inside a glass bulb.

The switch simply closes a circuit generating a pulse that the weather computer reads everytime
enough rain enters the collector. The rain gathers in a cup that when it gets heavy enough it tips
the collector cup. This action causes a magnet to slide by the magnet driven switch.

The weather computer is seeing 4-20 inches of rain in a 24 hour period and I know that is just plain wrong.

George Rachor
Magnets do not go bad, and the reed relays that closes the pulse
generator(s) either work or do not work.
If the spacing between the magnet and the sensing reeds have not
changed, then there will not be excessive bounce from minimal magnetic
drive.
Check to see that all other associated moving parts are clean, free
of sticky oils, gummy deposits, etc.
Observe its operation carefully during a simulated rainfall, to see
if the cup fills properly before tipping.
Maybe it tips when only half full, which would double the reading.
 
A

Anthony Fremont

Jan 1, 1970
0
Robert Baer said:
Magnets do not go bad, and the reed relays that closes the pulse
generator(s) either work or do not work.

Magnets can and do lose their strength. I have also witnessed (first
hand) the failure of a magnetic reed switch after only a hundred cycles
or so.
If the spacing between the magnet and the sensing reeds have not
changed, then there will not be excessive bounce from minimal magnetic
drive.

Perhaps, but if the current thru the reed switch is excessive then
damage to the contacts is possible over time.

Since contact bounce is a way of life with any mechanical switch, there
must be some kind of debounce circuitry in the sensor or (more likely
IMO) inside the base unit. It is possible that it was barely sufficient
when the reed was new and is now unable to suppress the glitches. Or
even some component in the debounce circuitry has failed due to age or
lightning induced voltage surges.
Check to see that all other associated moving parts are clean, free
of sticky oils, gummy deposits, etc.
Observe its operation carefully during a simulated rainfall, to see
if the cup fills properly before tipping.
Maybe it tips when only half full, which would double the reading.

The readings the OP are getting are far higher than if the cup was
tipping when only half full.
 
R

Robert Baer

Jan 1, 1970
0
Anthony said:
Magnets can and do lose their strength. I have also witnessed (first
hand) the failure of a magnetic reed switch after only a hundred cycles
or so.




Perhaps, but if the current thru the reed switch is excessive then
damage to the contacts is possible over time.

Since contact bounce is a way of life with any mechanical switch, there
must be some kind of debounce circuitry in the sensor or (more likely
IMO) inside the base unit. It is possible that it was barely sufficient
when the reed was new and is now unable to suppress the glitches. Or
even some component in the debounce circuitry has failed due to age or
lightning induced voltage surges.




The readings the OP are getting are far higher than if the cup was
tipping when only half full.
1) Only the old iron magnets get weak over time. Alnico and ferrite
magnets DO NOT. However, if you put those magnets in a high temperature
environment, then the magnetism can be removed (look in the Handbook of
Chemistry and Physics about the Curie Point).
2) The "half full" statement was for illustration purposes only, and was
not to be taken literally or even exactly. I could have easily said "one
third" or some other fraction. Maybe your cat got curious every time
and tipped his cup...
 
K

Karl Uppiano

Jan 1, 1970
0
The weather computer is seeing 4-20 inches of rain in a 24 hour period and
I know that is just plain wrong.

Is it over-reporting rainfall, or is it registering rain on sunny days?

It is quite common for magnetic reed switches to bounce, yielding multiple
pulses per pass of the magnet. High speed counter logic will count each
bounce as a separate pass. The traditional way to debounce mechanical
switches is to use a Schmitt trigger with a small resistor/capacitor network
on the switch input. This works pretty well, especially for slow events like
rain sensor buckets tipping back and forth. Software debouncing is also
sometimes used. So you have some things you can look at, in order of
likelihood:

- The power supply could be getting noisy. All sensor inputs might have much
lower noise margins than they used to. If it gets bad enough the entire
system could become flaky.
- The reed switch might be going bad. If the internal contacts have
corroded, the debouncing strategy might not be adequate anymore.
- The Schmitt trigger IC (if there is one) might be going bad, but I think
that is somewhat unlikely.
- If the debouncing capacitor is electrolytic (unlikely) it might have dried
out. But usually they use small-valued caps, and those tend to be ceramic,
and not prone to failure.
- Circuitry exposed to the elements is much more likely to fail. You might
need to clean connectors and reflow solder joints, and possibly replace some
parts.
- Has anybody installed a new radio transmitter in your vicinity? RF
interference from anything bigger than a cell phone or wireless network
could be causing all sorts of problems. Shielding and bypassing RF is beyond
the scope of this post. :)

Given its age, I would check the power supply for ripple. Use a good DMM
that blocks DC on the AC ranges. Most "good" power supplies have less than 3
millivolts of AC. YMMV, but if you see anything approaching 1 volt of AC,
it's a pretty good bet that some of the electrolytic filter and bypass
capacitors have dried out, and it's probably time to replace them all.

If you have the Heathkit assembly manual, it probably has schematics,
voltage charts, circuit descriptions and layout diagrams which should help
tremendously in debugging this problem.

Regards,
Karl
http://mysite.verizon.net/karl_uppiano/
 
Top