[Last Updated: 2/9/2003]
The safety of every diver who dives a breathing gas other than air is dependent upon someone's oxygen analyzer to measure the concentration of oxygen in the gas. As we all learned in basic nitrox class, too much O2 can be potentially fatal at some depths, and sometimes too little for the dive plan can lead to a disaster. Because of this critical safety dependence, many divers may want their own O2analyzer to confirm the dive shop's analysis or to use in preparing one's own breathing mix.
The first plans I encountered for a home-built analyzer were published in Vance Harlow’s "Oxygen Hacker's Companion", a very useful and well-written book, which I highly recommend . The construction notes here represent my second "edition" which, although loosely based upon Harlow's original construction, is simpler, includes a few improvements, provides more background theory, and more detailed construction suggestions. I have also included a section on the construction of a simplified and improved flow restriction device, and its calibration.
Although much care has been taken with the design and construction notes for the oxygen analyzer presented here, the reader is warned that the accuracy of the meter should be checked with known sources. This information is presented for the education of the reader, but no guarantees are implied, nor liability assumed by the author. As with any critical element of life support equipment, the obligation for proper operation is solely the responsibility of the user.
Please direct all questions and comments to me via email.
Basic Elements of an O2 Analyzer: Every basic O2 analyzer has three fundamental elements: a sensor that senses the relative concentration of oxygen in the gas under test and produces a proportional electrical signal, an electrical meter that converts the electrical signal from the sensor into an appropriate readable output indication, along with some form of flow restrictor, which manages the flow of the gas over the sensor element. These three major components are shown in the following figure:
Figure 1. Basic Elements of an O2 Analyzer
The next few sections will address the desired characteristics of each of these basic elements, and present functional design and construction details needed to assemble a working, accurate, O2 analyzer. Additional technical discussions are provided for the interested reader.
II. Constructing the Oxygen Analyzer
There are several ways to realize these basic functions, as may be seen in the variety of commercial offerings. The two most common configurations are the handheld unit with the sensor external to the case (like the MiniOX), or the internally mounted sensor, where the gas flow is coupled to a connector on the meter case (like the Analox/OMS analyzer.) Because of the versatility of the external sensor design (easily replaceable sensor; can be used with different gas sampling techniques), the basic design presented here will focus on a portable, battery-operated hand-held meter, with an external sensor, as shown in the opening photo.
A. Sensor
The heart of an oxygen analyzer is the sensor itself. It is also the most expensive element of the analyzer, but represents a technology beyond normal homebuilt projects, and should be purchased ($50-$80). There are several manufacturers of gas oxygen sensors, among them Analytical Industries, Maxtec (previously Ceramatec), Teledyne, and Mine Safety Appliances Company (MSA, makers of the MiniOX analyzer). Important properties characterizing an oxygen sensor element are: output signal, linearity, lifetime, and operating conditions (especially gas flow requirements.)
Most conventional gaseous oxygen sensors are based upon some form of galvanic cell technology, which produces a voltage proportional to the O2 in the test gas. Actually, the voltage produced is directly proportional to the rate of oxygen molecules colliding with the sensing element, which is in turn proportional to the partial pressure of O2 (PPO2) in the test gas. Note that at 1 ATM (1 bar), the fraction of O2 (FO2) is the same as the partial pressure of O2 (PPO2). However, this condition also suggests that, as long as the pressure of the test gas is the same as the pressure of the gas used to calibrate the analyzer (e.g., air), the analyzer will read correctly, regardless of the actual absolute pressure (within reasonable limits.)
These sensors behave like a fuel cell, and typically develop approximately 50-150 mV (millivolts) for 100% O2 at 1 ATM. It seems that most common O2 sensors are designed to provide either 10-12 mV or 22-25 mV for air at 1 atm, although almost all sensors intended for external use are of the 10-12 mV design. (If you want to build an analyzer based upon a high output sensor, the meter customization becomes much simpler. See the note later on this.)
There are many sensors available on the market, and
there are quite a few that are easily interchangeable and work equally
well Although my personal preference is for the PSR 11-39-JD from
Analytical Instruments, other external sensors such as the Maxtec
MAX-250E and MAX-13, the Teledyne RM17-D, and the MSA 406931 (as
used in the MiniOx) all function satisfactorily. The differences
among these sensors is shown in the following table (note: all parameters
quoted for air at 1 atm with a flow rate of 2-4 liters/minute over the
sensor.)
|
|
|
MAX-250E |
MAX-13 |
R-17D |
406931 |
|
Output (air)
|
|
|
|
|
|
|
Resp Time (90%)
|
|
|
|
|
|
|
Accuracy (FS)
|
|
|
|
|
|
|
Linearity
|
|
|
|
|
|
|
Oper Temp
|
|
|
|
|
|
|
Expected Life (in air)
|
|
|
|
|
|
|
Connector
|
|
|
|
|
|
|
Special Features
|
|
|
|
|
|
|
Approx Cost (5/'02)
|
|
|
|
|
|
|
Suggested Sources
|
|
Northeast Scuba |
|
|
|
Most of these sensors come with a coupler/diverter and sensor caps are available for most sensors (included from RC Dive Tech). Both RC and Oxycheq can supply a barbed sensor coupler (for $6-$12) - get one, it makes the analysis more accurate and stable (to match the flow restrictor, the coupler should couple to 1/8" ID tubing.) When installing the gas coupler, do not unscrew the white plastic diffuser - the gas coupler fits over it.
The Analytical Instruments
PSR 11-39-JD sensor with the barbed coupler attached is shown in the following
photo:
Figure 2. Typical Oxygen Sensor
B. Meter
With a sensor providing a reasonably linear electrical signal, any decent multimeter with a millivolt scale can be used to determine the percentage of oxygen in the test gas. For example, a new PSR 11-39-JD sensor will produce approximately 0.5 mV for each percent of O2 in the test gas. Therefore, if the meter reading is, say, 17mV, the oxygen content of the gas is approximately 34%. If you want to use a multimeter for oxygen analysis, the procedure is:
- Measure the sensor voltage in air (PPO2 = ~ 0.209, or 20.9 %.) Designate this voltage as Vair.
- Measure the sensor voltage for the test gas. Designate this voltage as Vtest.
- Determine the test gas oxygen percentage as: Pct O2 = 20.9 * Vtest / Vair
[For a technical discussion of the operation of the PM-128 and the ICL 7106 as used here, with notes regarding the adaptation of other model DPMs, click.]
Although the normal full scale range of the DPM is read directly in millivolts, we would like to scale it so that an input of around 10 millivolts (sensor in air) produces a reading of 209, or 20.9 with a decimal point! We do this by adjusting the meter's internal reference voltage so that the reading is about twice the input value, and inserting the decimal point in the proper place. This adjustment is accomplished by substituting a 62 kW resistor for R2 (30kW resistor) on the PM-128A circuit board which makes the effective gain of the PM-128A just over 2. The gain is then adjusted to the desired value by the on-board gain pot R4. [The circuit board for the PM-128A is shown below because it is the most common configuration now; the only difference is that the PM-128 uses the DIP IC package and discrete components, instead of surface mount ones.]
Figure 3. Connections to PM-128A Digital Panel Meter (Rear View)
To accomplish this resistor swap, using a pencil-tip low power (e.g., 25 watt) soldering iron, first heat one end of the SM (surface mount) resistor R2 until it melts (no more than 3-4 seconds, if the iron tip is hot enough), then quickly touch the iron to the other end, all the while gently prying under the center of the resistor with a pin, a wire, or something similar. The resistor will "release" and you can remove it. This may take a couple of tries, but should work. Be careful that the melted solder doesn't connect the pads for R2 to any other pads (if this happens, simply melt the solder and shake it off the board.) Then connect a 62 kW resistor across these two pads, or (and this is probably an easier way) connect one end of the 62 kW resistor to the R2 pad nearer the center of the circuit board and the other end to the V+ pad, as shown in Figure 5. [Electrically, the top or outside R2 pad is connected to the V+ pad.] The pads are quite small, and this mounting is much easier if the two ends of the resistor leads are well tinned (coated with solder) first - almost no additional solder should be needed. (For this, as all electronics projects, use only rosin core solder. Acid core solder will lead to internal corrosion, and must be avoided.)
While your soldering iron is hot, connect a short 1/4" copper jumper wire between the two pads labeled P1 - this enables the proper decimal point. Also, connect a couple of 6" wires (different color insulation) to the VIN and GND pins on the upper left corner, and two more 6" wires (different color insulation) to the V- and V+ terminals, as shown in the drawing above.
Now, cut a rectangle approximately 2 cm x 4.6 cm
(13/16" by 1 13/16") in the top of the case for the PM-128A
as shown in the photo on the opening page, with 1/8" (3mm) holes drilled
for the bezel mount. Mount the PM-128A, and connect the wires as
indicated in Figure 3 and also in the following schematic (or the 62KW
resistor between the "inner" pad and the V+ pad, as discussed in the last
paragraph.)
Battery Condition Test Switch: The PM 128A (and most other inexpensive DPMs) operates from a 9-volt supply, but the ICL7106 internally operates on a lower voltage, around 5 volts, and draws approximately one milliampere. Although the PM 128A is rated down to 7 volts, this internally regulated voltage actually permits stable operation from battery voltages down to around 6 volts. As the supply/battery voltage decreases below 6 volts, the meter reading increases. We will use this property to test the battery condition, by inserting a switchable voltage drop (here a 1.5 kW resistor which will drop the DPM supply voltage by about 1.5 volts) inline with the battery supply. In normal usage, the resistor is shorted by a normally-closed (NC) pushbutton switch. When reading the oxygen content of a test gas (other than 0.0 reading), just push the test button - if the reading does NOT increase, the battery is OK. However, if the reading increases noticeably, this indicates that the battery terminal voltages has dropped below 7.5 volts, and it is time for a new battery. [This is an optional part of the O2 meter, but is useful for testing the battery.] Incidentally, because of the low power drain of the DPM (~ 1 milliampere) and the normal capacity of an alkaline 9-volt battery (capacity ~ 500 mA-hours), a battery should last near its normal shelf life.
Figure 4. O2 Meter Schematic Wiring Diagram
Drill holes for and mount the other front panel components
- 1/4" diameter holes for the on/off toggle switch and the pushbutton switch,
a 5/16" diameter hole for the 10KW pot, and
a smaller 3/16" diameter hole for the input jack. If you are
using the recommended PacTec case, a 9V battery connector is supplied,
otherwise you will need one. Now solder the three fixed resistors
to the component terminals, and finally the interconnecting wires using
18-24 gauge "hookup" wire.
Figure 5. O2 Meter Assembled - Rear View
Initial Adjustment: Connect
the sensor to the input of the meter, using a cable with 1/8" phone plugs
on both ends, as shown in the opening photo. Insert and connect a
9-volt battery in the meter. Then, adjust the panel calibration pot
to the halfway point of its rotation, and install the knob such that it
is pointing straight up. Next, remove the cover from the sensor,
connect it to the input jack and wave it around a bit (to stabilize the
sensor in air). Now, open the back of the case, and flip the switch
to "On." Next, with a small flat-bladed screwdriver, adjust
the R4 pot (in the corner of the PM-128A circuit board, near the sensor
input pads, as shown in Figure 5, above) such that the meter reads 20.8-20.9
(the percentage of oxygen in normal air.) [If the meter reads
negative, the sensor leads are reversed; swap the leads coming from the
sensor input jack.] Close the case.
C. Flow Restrictor
For reliable readings from the O2 sensor, the flow of gas over the sensing element should be enough to continuously supply the sensor with new gas, but not so much that the pressure rises much above atmospheric. Essentially all galvanic O2 sensors (such as those listed in the table above) function best with a gas flow around 2-4 liters per minute. To achieve this flow, we use the first stage of a scuba regulator to reduce and stabilize the in-line pressure to around 130-160 psi. There are several schemes to restrict the flow to the desired rate, from needle valve metering devices to restrictive orifices. We have chosen the orifice method, which controls the rate of gas flow by inserting a severe restriction in the gas pathway. With a constant pressure difference across the orifice (e.g., 140 psi on one side and 14.7 psi on the other), a constant, predictable flow will occur, with a smaller orifice permitting a smaller flow rate. The diameter of an orifice needed to limit air flow to 2-4 liters/minute across a 125-psi pressure differential is between .006 and .0085 inches - a very small drill bit indeed!
Fortunately, there is a simple and inexpensive way to build such a device that is easily adjustable in the desired range – we build a larger orifice and then gradually close it down. This method is used to produce a fixed orifice for about $6, including the $5 quick-connect coupler for the scuba low pressure BC inflator hose. Here, the orifice is actually the space between the threads of a brass screw and its nut! Except for the Quick Connect (QC) coupler, all parts are brass and may be obtained in the plumbing department of any reasonably well-equipped hardware store, such as Lowes or Home Depot. The flow restrictor assembly is made up of a brass barbed nipple (1/4" NPT Female to 1/8" barb), a 4-40x1/2" brass screw, and a quick connect coupler (for the BC inflator hose). The only modification is to tap the inside of the barbed nipple with a 4-40 tap, then simply screw the brass 4-40 screw into the nipple. Tighten the screw about halfway between when you first encounter resistance and fully "tight." [If these particular components aren't readily available, you should be able to assemble an equivalent device - the idea is straightforward, and easily adapted to available parts.] These three components are shown in an exploded view in the following photo:
Figure 6. Parts for Flow Restrictor
Flow Restrictor Calibration: To calibrate, you will need a tank of compressed air (at least 300 psi should do it), a deep (4"-6") bowl of water, about 3-5 feet of 1/8" ID plastic tubing, a watch with a second readout, and a 1-2 liter plastic soda container. Set up the scuba tank with a first stage regulator in the normal way and connect the flow meter to the BC inflator hose quick connect fitting. First, we will try to set the screw to get the flow in the ballpark of 2-4 liters/min. Check the flow by placing the nipple barb in your mouth, which should "fill up" in about 2 seconds if the flow is about right. If the flow is not right, disassemble and loosen or tighten the screw accordingly and reassemble. When the flow is about right, connect the tubing to the "output" barb (the tank valve should be off.) Fill the soda bottle with water, and while holding your thumb over the mouth of the bottle, place the mouth below the level of water in the bowl. Now, run the tubing (connected to the flow restrictor) below the water line and into the soda bottle. This experimental setup is shown in Figure 7, below.
Figure 7. Setup for Calibratin of the Flow Restrictor
To calibrate the flow restriction device, turn on the tank valve and start the stopwatch at the same time. When the soda bottle has emptied and the first bubble emerges from mouth of the soda bottle, stop the timer. The actual flow rate may be calculated by:
Flow Rate (liters/minute) = Volume (liters) of the soda bottle / Fraction of minute until bottle is empty
For example, if your soda bottle is 2 liters, and the first bubble appeared in 40 seconds (2/3 minute), the flow rate would be (2 liters)/ (2/3 minute) = 3 liters/minute. Now, if the flow rate is too fast, tighten the 4-40 screw, then run the water experiment again. If the flow rate is too slow, loosen the screw somewhat. When the flow is acceptable (2-4 liters/minute), disassemble the two pieces, add teflon tape to the threads of the QC coupler, and reassemble. You now have a dandy $6 flow controller!
III. Operation
The operation of the O2 analyzer is pretty straightforward:
1. Connect the scuba first stage to the tankPrior to checking an unknown gas, however, one should first calibrate the sensor and meter with a known gas (for most nitrox mixes, air is a pretty good calibration gas, because it’s always pretty close to 20.9% O2, although this fraction falls with increasing temperature and humidity - see the Analox table.) Simply follow the above sequence using the known gas, and adjust the calibration knob on the analyzer so that the reading agrees with the oxygen content of the known gas. Then, change the regulator first stage to the test gas tank, open the tank valve, and watch the readout until the reading settles down (at least 10-15 seconds.) If the reading is not stable after 30 seconds, recheck the connections to the sensor, the analyzer, and all gas connections. [Of course, if the analyzer cannot be calibrated to the known gas, the R4 coarse adjustment on the circuit board must be reset. See the Initial Adjustment paragraph in Part II.B, above.]
2. Connect the flow restrictor to the BC inflator hose connector
3. Connect the cable to the sensor and the meter/readout
4. Turn on the meter
5. Turn on the tank valve
6. Read the meter
The Analytical Instruments PSR 11-39-JD sensor is stable across temperatures from 32°-122° F. I found that my sensor output increased as the temperature dropped below 30° F, peaked around 18 deg F, and then decreased back as the temp went to -2° F. The LCD display faded (not permanently) entirely at around -10° F. Accordingly, although I think the meter should work properly if it is calibrated at the test temperature, the meter should only be trusted at operating temperatures above 40° F and relative humidities less than 80%.
IV. Maintenance
The O2 analyzer requires very little maintenance. The sensor is really the only part of the analyzer requiring any special attention. The sensor should be protected from reactive gas exposure, but this is easily taken care of by replacing the sensor cap when not in use, or storing the sensor in a simple plastic sleeve. [The sleeve need not be airtight, however. According to Maxtec, the sensor should be stored in air and at "room temperature."]
If the panel calibration is unable to bring the meter into the desired range, try opening the case and adjusting the trim pot R4, shown in Figure 5. If this adjustment cannot correct the range, or if the readings start to rise beyond calibration, or the display is not clear, it may be time to replace the battery. If you didn't build in the battery check switch, replace the battery at least annually.
V. References/Sources
Harlow, Vance, "Oxygen Hacker's Companion"
Airspeed Press
79 Old Denny Hill
Warner, NH 03278Maxtec, Inc.
2425 South 900 West, Suite B
Salt Lake City, UT 84119
(800) 748-5355
Maxtecinc.comNortheast Scuba Supply
539 N. Trooper Rd
Norristown, PA 19403
(610) 631-2288
John@Northeastscubasupply.com
northeastscubasupply.com
OxyCheq
34929 Sweetwater Drive
Agua Dulce, CA, 91390 USA
sales@oxycheq.com
(661) 268.0182 (ask for Patrick)
oxycheq.com
RC Dive Technology
422 Salem St Suite 144
Medford Massachusetts 02155
781-367-5515
oxygenanalyzer.com
BC inflator QC coupler to1/4" NPT fitting [available from a dive shop or mail order for ~ $5]
Brass barbed nipple – 1/4" NPT Female to 1/8" barb
Brass screw – 4-40 x 1/2"
1-2' 1/8" ID poly tubing
Handle and 4-40 machine screw tap for tapping inside of 1/8" barbed nipple
For calibration: 4' 1/8" ID poly tubing, 1-2 liter soda bottle, deep pan
PM-128A - the DPM may be obtained from any of the following (prices shown valid Jan 2003 exclusive of shipping):
Case is Pactec HP9VB ($5.49 + s/h)Web-Tronics (~$11)
All Electronics (~$12)
RP Electronics (~$11)
JDR Microdevices (~$9.50)
2' 18-20 ga hookup wire
9-Volt BatteryRemaining parts from Radio Shack or Web-Tronics :
|
|
|
|
| 10 KW Linear pot |
($1.99) |
($1.20) |
| 1.5 KW Fixed Resistor |
($0.99 Pkg of 5) |
($0.10) |
| 62 KW Fixed Resistor |
|
($0.10) |
|
100 KW Fixed Resistor |
($0.99 Pkg of 5) |
($0.10) |
| SPST Toggle Switch |
($2.99) |
($1.02) |
| SPST Pushbutton Switch (Normally Closed) |
($2.99 Pkg of 4) |
($0.58) |
| Mono Phone Jack 1/8" (3.5mm) |
($2.99 Pkg of 2) |
|
| Knob |
($1.49 Pkg of 2) |
|
| Shielded Cable w/ 1/8" plugs (for Sensor) |
($2.99 - 6' ) |
|
If you get the PacTec case, a battery connector is provided. Otherwise, you will need one (Radio Shack part # RS 270-324; $1.99 for a package of 5)Correspondence
If you are interested in this project, and especially if you build it, I would very much like to hear from you about your experiences, comments, and suggestions via email!