DIY gas sensors?

[gsk3]

I've got a DIY O2 sensor which works great, but I'd like to get a He and CO sensor in the lineup (ideally with one hookup so gas runs the circuit past all of them).

Does anyone have DIY plans for either type of sensor? I have family members who are electrical engineers, so even a pointer towards circuits which would do the measurement (even if that measurement needs substantial conversion to be useful) would be helpful....

 

[CuNanoTube]

Helium is not very easy to detect. Unlike Oxygen which is reactive, helium cannot be sensed using a galvanic method. The easiest way to measure He% is to use thermal conductivity, and that requires very very precise temperature measuring and high quality materials with constant thermal conductivity. It may be possible to build such a system, but I would not know how. The other option for measuring helium content is spectroscopy, but that brings other problems into play. For He you are best off going commercial for the sensor.
Katharometer - Wikipedia, the free encyclopedia
Carbon Monoxide is reactive, and there are some technologies that can detect it, however I have not found any good documentation on how to build one. Since CO can be detected with a fuel cell type sensor it should be possible to build one from scratch as long as you have the facilities to calibrate. That would require a source of CO, a properly calibrated detector, a fume hood, and accurate flow meters. A Typical inorganic chem lab would have these. (like the one I work at.)
Anyways, sorry for not having any concrete information for you, I will see if I can find out more next time I am at a lab.

 

[gsk3]

Thanks. That's a lot of helpful info. I figured I'd have to go with a pre-made sensor, and I'm fine with that. Calibration is also a given (at least for CO).

My understanding of He is that at least one of the methods is based off the speed of sound as compared to a reference cylinder of He. Seems like that might be more possible to homebrew the sensor itself. But since He sensors have been around for a long time and still cost $1k or so, I suspect that's non-trivial.

CO on the other hand is new, and given how cheap sensors for non-diving applications are, I have a hard time believing they should really cost $300+ for the unit. I suspect that's just a matter of finding a cheap sensor and doing something similar to the homebrew O2 analyzers.

 

[gsk3]

Digikey has a cheap CO sensor but it isn't sensitive enough (>20ppm).

It also has complete units by TPI and Amprobe which are around $200 and do >0ppm.

There's one that hooks up to a DMM for $120 that looks like it could be a good bet. With a DMM the kit would be unwieldy, but if you built it into a case which used a panel meter it could be a nice setup.

The trick with most of these is getting them set up with proper plumbing. Otherwise you're stuck putting the tank gas in a bag along with the sensor, which is a royal pain.

 

[CuNanoTube]

I should show you some pictures from my DIY O2 unit. I have in a UK hard case with the flow meter panel mounted (plastic panel) in the lid and the meter securely held in foam. I built it with the Oxycheq El Cheapo 2 kit.
Plumbing high pressure gas is something I have some experience with. Look into "Swagelok" fittings. They make SS, plastic (the plastic is lower pressure), and Titanium hardline that can be used at 5000psi with appropriate fittings. You could probably mount a CO tank in the case and use a needle valve to feed it for calibration.

One thing you will need is a flow meter or Rotometer, you can get them from omega engineering. Omega makes good stuff. Their cheapest model is $40 + tax.

 

[gsk3]

I was thinking it would be LP gas...could put an old reg on the CO cylinder, and use a BC fitting from the tank to be measured to provide a nice even flow.

The issue with the CO sensors is they seem to require ~15 secs of dwell time before they give a good read. That means parking some gas in front of the sensor is necessary. Options include enclosing the unit in a bag full of tank gas, or having tubing directly connect the tank to a dwell chamber which encloses the sensor.

Ideally I'd wind up with something like the three Oxycheq Expedition (O2, CO, He) sensors in a row, where it looks like they could be plumbed in serial. Then you'd only need to crack the tank valve once and the same gas would pass by all three sensors.

 

[gsk3]

Current best solution for CO:

The TPI A771 costs $110-130 and puts out 1mV per 1ppm of CO. It comes with DMM leads.

The $25 Martel V125 panel meter reads 200mV full scale, 3.5 digits, and comes with a bezel and mount. Spec sheet here http://www.martelmeters.com/pdf/V_125.pdf . It also has a pot in the back for calibration.

Rig this up in a medium-sized Pelican next to a small CO calibration cylinder, and you could have two panels next to each other, one for O2 one for CO, with a common set of plumbing linking them.

The A771 reads from 0-2000ppm, so I'm not sure how tolerant the panel meter would be to a very high CO reading which fed it 2V. But I suspect a tank >200ppm of CO would be rare indeed--rare enough that I'd be willing to invest in another $25 panel meter if it saved my life. The only issue would be whether you'd be able to tell if it was reading high before it crapped out. Otherwise you might assume the panel was broken rather than the gas too high. But the A771 has a built-in alarm which provides redundancy anyhow.

Still have to figure out how to get gas at the right flow rate through the A771.

All 3 (the Oxycheq, the panel meter, and the A771) take 9V sources, so you could rig them up through a single power source or just have individual 9V's to provide redundancy.

 

[gsk3]

Have to get the impedance matching right. Not sure what the panel's is, but it might need some extra basic circuitry there. Plus the panel actually reads in V not mV, so it may require some conversion or just a different panel. OTOH, the A771 manual specifies to set the attached DMM to 2-4VDC scale, and the V125 lists a set of resistors for 2V full scale reading, so with four digits available that could read in mV.

A771 Specifications
Carbon Monoxide (CO) Measurement Range 0 - 1999ppm
Accuracy ±5ppm or 5% of reading whichever is greater
Resolution 1ppm
Output 1mVDC / 1ppm CO into 10M Ohm impedance
IR Printer Port No
Flue Probe No
Internal Pump No
Combustible Gas Leak Detection No
Backlight N/A
Alarm Audible
Power Supply 9V Alkaline Battery
Dimensions 6.5” x 2.1” x 1.2”
(165 x 55 x 30 mm)
>Weight 7.23oz (205g)
Warranty 3-year limited warranty
2-year sensor warranty

Manual: http://tpi-thevalueleader.com/products/carbonMonoxideAnalyzers/A771/a771 manual.pdf
Product sheet: http://tpi-thevalueleader.com/products/carbonMonoxideAnalyzers/A771/LA771-1104.pdf

 

[HenrikBP]

Looks like an interesting build Ari. Please keep us posted on your progress.

Henrik

 

[CuNanoTube]

Nice.
For the CO sensor you could consider putting the sensor in a PVC housing (to which you can easily tap in fittings) to allow for calibration. As for splitting the gas feed it seems like you will need a much lower flow rate over the CO sensor compared to the O2 sensor, so you could use 2 flow meters and have the gas in come to a t-joint. It might be worth considering a check valve for the co testing line to prevent contaminating the O2 line (I think CO throws off galvanic O2 sensors).
You might also need a "purge" for the chamber to get a stable reading at such a low flow rate after calibration.
Converting from mV to V is as far as I know harder than the reverse, so it might be worth looking into another panel. The panel I got in my El-cheapo II kit was stock, it read in mV, unfortunatly I am away at university at the moment, so I cannot check.

This seems like it can work.
Good work man.