New Nitrox Stick

[Blue Abyss]

I just installed a nitrox stick (oxygen hacker version) on my 3 cfm Bauer K13 compressor. I have installed two O₂ analyzers, one on the stick and one at my fill board. How much difference should I expect in the readings between the two analyzers? I am seeing between a 3 and 4 percent difference that seems like a lot but what is to be expected?

 

[boydski]

I'm suprised that you are seeing that much difference in your sensors. I have a similar setup and my Oxygen sensors are usually withing about a 1/2 % of each other.

When the garage is REALLY cold, they are off by about 1%, but that's likely due to one analyzer being temperature compensated and one is an El Cheapo.

Good Luck!

 

[oxyhacker]

Is the one on the board a proper downstream analyzer, fed by a regulator? If so you will be getting variances caused the differences in the moisture in the air, but in WA one would not expect them to be that great.

How are you calibrating them? Oh, and have you tried switching the analyzers/sensors back and forth between the two ports?

I've also seen this before with a very leaky setup, where extra air was getting sucked in downstream from the first analyzer.

The final test of course is what is in the tank after it has sat overnight. Then you can at least compare that with what your analyzer are telling you while mixing, and figure out which one to believe.

 

[pescador775]

Sometimes, there will be residual air in the compressor and filters. So, the O2 reading will climb more slowly in the downstream analyzer. As Hacker said, the output of the compressor is dryer than the intake air (in the stik). Thus, there will be up to a 1/2 % difference with the output gas reading lower O2 than the meter on the stik. Calibrate the analyzer located at the mixer stik using ambient air and a moisture compensation graph (avail on internet). Calibrate the analyzer on the output using compressed air from a tank. IMO, the downstream analyzer is useful for a newbie when attempting to understand the relationship between output and input results. Once the variances are understood they can be compensated for by manipulating the knobs during the fill process and the second analyzer is no longer needed. Knobology can be minimized by purging the compressor. This wastes a bit of gas but makes results more consistent. Just run some gas through with the drain open. Realize that a tank containing atmospheric air must also be compensated for by turning up the O2 by one or two tenths.

 

[rjack321]

My similar system has: 1) quite a lag time between when changes in the inlet show up on the panel and 2) some sensor drift over time and with changes in temp. Drift is especially bad if I don't allow the sensor to eqilibriate to ambient air for ~5 mins after removing the "sensor saver" caps. It takes a little while for the voltage to stabilize after storage IMO

After the compressor is warmed up I will oftentimes disconnect the post compression analyzer and recalibrate it to scuba air. I have also removed a stik sensor, taped over the hole, recalibrated it and reinstalled.

I have a 3cfm Alkins in my unheated Seattle garage.

How old are your sensors? How are you calibrating them? If you have el cheapo analyzers do you have any loose connections? I broke a soldered wire in the plug once and that took awhile to figure out.

Also what's your post compression flow rate and how are you measuring it?

 

[rjack321]

BTW mine are usually within 0.5% of each other or so - once they've stabilized and with good calibration.

I really don't need the post compression analyzer, but I have a CO meter online and need the reg to bleed off a few liters a minute for that anyway.

 

[Blue Abyss]

I think I must have a leak downstream of the stick sensor.
I have calibrated the stick sensor (which is hooked to an el cheapo) with atmospheric air and the sensor at the board (Oxycheq) I calibrate with air from a tank. The sensor on the board is hooked to a medical regulator and I have been setting the flow to 2.5 liters per minute. I ran the compressor for about 3 hours the other day and the output was stable on the down stream compressor right about 32 while the input sensor was reading about 35.5 (more variation up and down on this reading). My tanks analyzed within less than .5% of what the output sensor was reading.

I was able to control the system but I was surprised by the difference in the two reading.
Also it was really cold when I was filling, about 34-36 degrees in my garage.

Think that I may have a leak right at the input to the compressor.

Thanks for the input.

 

[nwdiver2]

Flip the analyzers and see if they show the same difference, if so it's your setup if not one of the sensors may need replacing.