Mercury Hourmeter / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Mercury Hourmeter

1 2 next

Move Thread

LAN_403

klugesmith

Thu Oct 01 2009, 10:46PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Not sure if this thread belongs in Projects, but you will see why it begins on the first day of a month.
An alternate title, inspired by a contemporary thread on 4hv, is Ionized Mercury (aq).

I have a small collection of coulometric elapsed-time meters.
These have a glass capillary tube marked with a linear scale, like a very short thermometer. The entire length is filled with mercury, except for a tiny gap filled with an electrolyte solution. As the device conducts DC current, mercury is electroplated from one column to the other. Over a period of hundreds of hours (in normal service), the gap moves along the tube.

It's a simple and elegant invention, but can be destroyed by any mechanical shock that makes the two columns touch and merge. They were (and still are) often used as accessories to valuable objects that need periodic replacement and gentle handling anyway. Examples include record-player styluses and deuterium UV lamps.

So I made a little demo, mounting a specimen on black cardboard along with a 9V battery and a LED wired in series, all inside a clear plastic box.

Showed it to my co-workers, let it run for a month or two, reversed connections and ran it back, then removed battery & packed it away for 5 years.

Today the unit has a new battery and is running forward again. You all can share my excitement as we watch it go. While we wait for a visible change, I'll present some related measurements and analysis.

This specimen came from a friendly ebay seller in Lithuania called kwtubes, who has plenty more of the same for sale today. See a listing for good closeups and description. Full scale is supposed to be 2500 hours with 12.6 VDC applied. (Ñ‡Ð°Ñ is the Russian word for hour, and -10Â°C must be the minimum temperature for proper electrolytic behavior.) "The meters are non-resettable," but I can attest that they are reversible.

So...
Between -12.6 V and +12.6 V, my meter's I/V behavior looks just like a 102 kâ„¦ resistor. Nominal current is 123 ÂµA. Now running 73.5 ÂµA in the demo, with battery/LED/meter voltages of 9.156, 1.621, and 7.533 V.
Even at 0.07 mA, this LED can project a circle of red light visible on white paper without turning the room lights off.
And on 2009 October 1, 21:06 UT, the middle of the gap stands at about 98 hours. (I love fine reading of analog scales)

Wolfram

Thu Oct 01 2009, 11:25PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Ever since I saw one of these devices on some webpage many years ago, I've been fascinated by them. I didn't realize they were available and cheap on eBay, so now I'll buy some of these gems for myself.

Anders M.

Proud Mary

Thu Oct 01 2009, 11:25PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I've bought so much good stuff from Gintaras at KW Tubes - he has the finest selection of close tolerance silver mica capacitors I've ever seen, going right up to unbelievable values like 50,000pF 350V.

He's got buckets full of those Hg time elapsed counters so cheap that I thought of allocating them on an individual basis to some of my more costly thyratrons and X-ray tubes. I recall Gintaras saying that they all started out at 100 hours for some reason which was not clear to me, though he speaks very good English.

Electra

Fri Oct 02 2009, 10:59PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

I once removed something similar to one of these but smaller, from an old pcb,
And tested it in the same way, though I let it run off the scale and no amount of current would reverse it. I was hoping there would be some change of resistance that could be detected when it got to the end that could be used to trigger a circuit to reverse the voltage, I couldnâ€™t find any difference or it was too small, and I guess it must have shorted out or stopped working at that point.

Tried to think of a way of sensing the gap in the mercury at each end of the tube, the tube was very thin, and would probably break if I tried to remove it from the plastic part, to make like fitting it to an opto sensor or something, and not sure if any sort of inductive way of sensing would work either. So the way to build the worldâ€™s slowest oscillator has yet to be found.
All this was a while ago now but it reminded me I should get another.

I thought they were fairly low resistance devices and had to be operated with a large series resistor to limit the current, or are there different typeâ€™s, or just down to a difference in physical size? I ran mine at 50ua, I remember reading they also integrate current fairly well if you have a pulse or other wave shape presumably it has to have an average dc component to move in one direction.

klugesmith

Thu Oct 08 2009, 02:28AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

I opened up another unit, from Curtis Industries. The mercury tube is mounted in a reversible bezel with pin connectors, and it looks like the hour scale can also be moved for zero adjustment.

Electra wrote ...
...I thought they were fairly low resistance devices and had to be operated with a large series resistor to limit the current, or are there different typeâ€™s, or just down to a difference in physical size? I ran mine at 50ua, I remember reading they also integrate current fairly well if you have a pulse or other wave shape presumably it has to have an average dc component to move in one direction.

Looks like you were right. The base part of this one has a hidden resistor measured to be about 1.5 megohms, which is consistent with the normally hidden inscription "5 VDC 3.21 uA". When I connected the mercury tube in series, its voltage drop at the nominal current was just a couple of mV.

... let it run off the scale and no amount of current would reverse it. I was hoping there would be some change of resistance that could be detected when it got to the end that could be used to trigger a circuit to reverse the voltage, I couldnâ€™t find any difference or it was too small, and I guess it must have shorted out or stopped working at that point.

If there were an enlarged bulb at the end of the tube, we'd expect the gap to short out when it gets there. Not sure what would happen with this specimen:

The end farther from the gap today, which might or might not be where the gap started, seems to be sealed with organic resin instead of a glass-to-metal seal. It just occurred to me that these devices could be filled at the factory by direct injection, using hypodermic tubes thin enough to catheterize the glass capillary!

Tried to think of a way of sensing the gap in the mercury at each end of the tube, the tube was very thin, and would probably break if I tried to remove it from the plastic part, to make like fitting it to an opto sensor or something, and not sure if any sort of inductive way of sensing would work either. So the way to build the worldâ€™s slowest oscillator has yet to be found. ...

A fun idea! If I had to do it, would consider a high frequency measurement of capacitance between the mercury columns and a conductive sleeve[s]. How 'bout TDR with 20 picosecond risetime, to toggle a multivibrator with period of 20 megaseconds?

At Thursday October 8 00:40 UT, the demo with Russian meter had voltages of 8.976/1.620/7.356 and the gap was at 182 hours. Since original post it's logged about 10.7 mAh (39 coulombs).

Wolfram

Thu Oct 08 2009, 09:06AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Here

is a patent for a very similar device to the one you dismantled, complete with the movable scale.

Anders M.

klugesmith

Tue Nov 24 2009, 11:01PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Thank you for the patent hint, Anders. It kicked off a search that found many interesting variations -- nonlinear scales, color change at end of scale, optical sensing, even a film-based radiation dosimeter that captures the times of exposure events. I think it's a coincidence that inventor Curtis Beusman worked for Curtis Industries.

No apology for the slow reply; this thread is partly about patience. It is moving faster than the conduction electrons in indicator tube: two per Hg atom, so their drift velocity equals the mercury gap velocity -- on the order of 1 centimeter per month.

The original experiment ended on November 1, after one mean month (730.5 hours, which you should be able to figure in your head). Voltage and gap position were checked every 6 days. The gap moved 422 "hours" with time-weighted average of 7.26 volts at meter terminals. That's consistent (within 1%) with the stated nominal voltage of 12.6 VDC. Total charge through meter was 52 mAh = 186 C, making a significant indentation into the capacity of the 9V battery.

Then I recognized a great discrepancy between the current in my Curtis meter (3.21 uA on label) and the Russian meter (measured to be 123 uA) for similar nominal speeds: 12.7 mm/kh and 16 mm/kh. The capillary bore diameters are similar, so we can conclude that most of the current in Russian meter flows in a shunt resistor, and only a few percent go through the mercury tube. No surprise -- here is another meter, with 1 ohm shunt resistor and 86 K series resistor, that came with a 300 mA ultraviolet lamp.

Now (review would be welcome) it takes 13.03 coulombs to plate 1 mm^3 of Hg at its room temperature density.
So plating current density must be 57.9 A/m^2 for Russian scale and 46.0 A/m^2 for Curtis scale.
From scaled photo, Russian bore diameter appears to be 0.35 mm, corresponging to nominal 5.6 uA. But the actual bore is probably much smaller before optical magnification by the glass tube.

Curtis meter bore diameter is figured (from known current) to be 0.298 mm, but in photo it looks like 0.77 mm!

Next step: Run Curtis meter & try to clear up the "bubble" to right of gap. Concurrently (no pun intended) run the lamp meter (which has gone off scale) backwards in attempt to recover/recreate a gap.

hasta la vista
Rich

Wolfram

Wed Nov 25 2009, 07:19AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Excellent information.

I can confirm that there are shunt resistors inside the russian units. There is a voltage divider hooked up to the 12V input, and the mercury tube is connected, in series with two (!) resistors, from the midpoint to the negative terminal.

Any ideas on if it would be a good idea to speed these up by running a higher current through the tube?

Anders M.

IntraWinding

Wed Nov 25 2009, 05:51PM

Registered Member #2261 Joined: Mon Aug 03 2009, 01:19AM
Location: London, UK
Posts: 581

I love these!

I got one from somewhere years ago, built just like a UK domestic plug fuse with the usual metal end caps.

It's such a clever idea, but when I expressed my enthusiasm to my boss at the time he gave me a lecture on polluting the environment with mercury. I guess they're history now but the story of the leap of imagination that led to their invention would make an interesting read.

(Anyone know how they do the electrolysis without releasing any gasses?)

Bored Chemist

Wed Nov 25 2009, 06:17PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

"Anyone know how they do the electrolysis without releasing any gasses"
Yes.
Rather than oxidise water to oxygen and reduce it to hydrogen, they oxidise mercury to Hg++ ions and then reduce those ions back to mercury.
In principle the voltage needed to do that is zero so there's not much potential to split water.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob