The time it takes droplets on my rain lamp to drip the full 13.75 inches of strand length.
How long does it take for all the stands to start dripping completely?
First of all, my rain lamp is very sensitive to being level.
If my rain lamp is not very level, some strands will never drip,
presumably because the oil level at the elevated end of the upper basin
will then not reach the most elevated spouts.
When my lamp was level, though,
I timed the difference between the time-to-first-drips to the time-to-complete-drips as
6 minutes 15 seconds, and 5 minutes 15 seconds, in two different trials.
Which strands drip first, or isn't there a predictable pattern?
For a fixed position of a rain lamp on a stand, I've found that the first strand to drip is
very predictable. In my current lamp placement, strand #11 (next to the drain post)
always drips first.
To a lesser extent, the subsequent strands to drip are also predictable.
In my case the strand order for dripping is: 11, 9, 12, 10, 8, 7, 6, 14, etc.
Unpredictably, however, sometimes adjacent strands begin dripping so close in time
that it is nearly impossible to discern an order.
I haven't yet experimented with rotating the lamp on the stand to see if the drip order
can be changed.
How many droplets are on a strand at a time?
On my lamp,
as far as I have been able to determine by photography so far (12-29-05),
the number of droplets per strand varies from 17-26,
with an average of about 21.7.
How long does it take for the droplets to slide to the bottom?
On my lamp, the average drip time is 7.45 seconds,
though it ranges from 6.56 seconds to 8.42 seconds.
See the section below for more details.
What is the drip rate?
By calculation on average values, 21.7 droplets / 7.45 seconds = 2.91 droplets per second.
This is nearly three droplets per second--pretty fast!
What is the absolute speed of the droplets?
I measured 13.75 inches for each strand length,
so on the average, 13.75 inches / 7.45 seconds =
1.85 inches per second =
(1.85 inches/second) x (60 seconds / 1 minute) x (1 feet / 12 inches) = 9.23 feet per minute.
Therefore on a 20-foot rain lamp like at Disneyland,
it would take over 2 minutes for a droplet to reach the bottom, if all other values are the same.
Using a stopwatch, I measured at least three drip times per strand,
averaged those values for each of the 32 strands,
entered those 32 average times on an Excel spreadsheet, and plotted the results.
I numbered the strands from 1-32, starting with spout #1 at the pumping post,
and proceding clockwise when looking down at the lamp.
This means the three posts are located roughly as follows:
Some observations:
The same graph converted from time to speed looks like the following.
Summarized data:
If the tilt of the lamp were affecting drip speeds, then there would tend to be faster speeds at one portion of the lamp, which is apparently not the case. Likewise, if proximity to the pumping post or drain post were influencing the speeds, then the speeds would tend to be faster in one area, which is also apparently not the case. (There is a brief but debatable pattern of increasing speed just before the drain post, though.) Therefore the variations in speed are apparently just due to random influences on the spouts and strands.
Preliminary results (12-31-05) on the number of droplets per strand are:
So far the results indicate that speed and number are negatively correlated, or equivalently, that speed and frequency are positively correlated. The meaning of this relationship is not clear to me, and it's not yet clear if this correlation really exists.
Unlike the speed measurements, which required no more than a stopwatch, the count measurements were more difficult. A digital camera worked great for freezing the state of a strand to allow droplets to be counted, but this worked well only for strands not visually obscured by posts. The way I successfully handled the obscuring post problem was to use the Paint image editor to put count labels, extending lines, and fixed length rectangles along the obscured strand length in the photographs to reconstruct the missing length of strand. This was accurate enough that I did not need to take perspective scaling into account. I also found it helped to cover the statue with a taped cylinder of white paper so that the droplets were more visible against a light homogeneous background.
It takes a few minutes for the strands to stop dripping after the power is turned off. For many hours afterwards, tiny droplets a fraction of their normal size can still be seen coursing down the strands, though slower, unevenly distributed, and more sporadically.
The drip rate is clearly influenced by lubrication because usually the first drips on a given line move slower than the drips behind them, with the result that the upper drips keep running into the lower drips until the entire strand is lubricated. Afterwards, the drips become very regular in timing and spacing.
When I first received my vintage rain lamp, I ran it with new oil but without cleaning, and the droplets on each strand were initially blue-green from the patina of the brass, although they soon turned clear. After cleaning the lamp and changing the mineral oil, however, all the droplets were clear right from the start.
Sometimes droplets always seem to vibrate or shimmer at a certain part of a certain strand. This means there's a rough region on the strand at that location, which can be verified by running a thin paper towel pinched around the strand, for the length of that strand past that point. Interestingly, that roughness doesn't appear to influence droplet speed at all, only the droplet shape mometarily. Also interesting is that sometimes a speck of dust will get stuck to a strand, and the droplets slide right past the dust speck without dislodging it or even moving it. Maybe that is because such dust specks are on the upper side of the slanted line whereas the oil droplets hang underneath the slanted line.