Vegetable Oil SVO WVO waste cooking oil as biofuel

[SunWizard]

I am testing a Capacitive water % saturated meter I made and its giving interesting results so far. I am doing many calibrated samples and tests in the Science project thread which are also useful to test this device.
It cost $50, and gives data similar to a $3000 unit like the Eesiflo Online Water in oil Monitor I got a high accuracy (0.5%) capacitance meter (ebay Protek CM109 $45) and a honeywell humidity (RH) sensor HCH-1000-001 from digikey.com link for $5:
data sheet PDF & pic

I thought even if it didn't work immersed in VO, I could use it to measure and compare the humidity in the air in the headspace of stored barrels, as well as in the headspace of my dieselcraft CF rig to tell how the dewatering is going. But it is working great immersed in the VO!

Immersed in VO, it takes about 30-90 seconds to get a stable reading depending on the VO temp. It reads humidity of the VO just like it does in air. Humidity of the VO is basically % saturation, see Dissolved water and % saturation. Here is a chart of my 1st data points at 61F showing how linear its response is:


This is a much easier test than the HPT, and gives quantitative data for how far below saturation you are. Settling even for long periods is getting my SVO to ~70% saturated. The farther you get below 100% saturated is a very good thing and should be a goal for everyone. Because if you are near saturated, humidity brought in from your tank vent, or a temp. change could mean free water condensing in your tank.

With sample B, my good VO source, FFA=2% shown as the green line, capacitance rises much quicker, and fails the HPT much sooner <1100 ppm. Calibrating this meter I cover in a later post below. Then you take the meter reading, and compare to a chart like the above for your VO type to find % saturated and PPM. This is what is done internally in the expensive water in oil sensors.

For others wanting to make one, a high accuracy capacitance meter is needed since 0.5% of 300 pF = 1.5pF = 3% humidity. A cheap meter like the harbor freight and many others is only 3% accurate which is 18% humidity error. Don't touch the metal portions of the test leads since that throws off the capacitance. I leave the test leads and probe hang freely 8-10" below the meter in the WVO since since that seems to give a steady reading quicker.

Between samples, I wipe the VO completely off with a rag, which is easy since its a chip of glass, and it quickly goes back to measuring the RH of the air in my garage. You want to do this to ensure that no oil polymerizes on it, and to make the next reading accurate quicker.

Here is a link to buy the sensor:
http://search.digikey.com/scripts/DkSea ... 80-2903-ND

Here are some articles about water in oil and detection methods:
http://www.machinerylubricatio...ater-contaminant-oil
http://www.machinerylubricatio...ration-screening-oil
http://www.machinerylubricatio...tecting-water-in-oil
Quotes from the above research:
"Water can exist in oil in three states or phases."
1. "The first state, known as dissolved water, is characterized by individual water molecules dispersed throughout the oil.",
2. "Once the amount of water has exceeded the maximum level for it to remain dissolved, the oil is saturated. At this point, the water is suspended in the oil in microscopic droplets known as an emulsion.",
3."The addition of more water to an emulsified oil/water mixture will lead to a separation of the two phases producing a layer of free water as well as free and/or emulsified oil."

One trick is I will need to do temp compensation like shown on the datasheet link above, although its small at 0.16 pF/deg.C. These tests were run at room temp. 70F. (edit) I wrote a simple web page to do the temp. compensation calculation (which I find useful when I am running my centrifuge) here:Capacitive water in oil meter Percent saturated calculator
Much more testing to come...

[HoldOnTight]

The datasheets show the sensor is very linear. Your experiments show less linearity. So I wonder, why? Could you spike the samples to get more data points?

[SunWizard]

The datasheets show the sensor is very linear. Your experiments show less linearity. So I wonder, why? Could you spike the samples to get more data points?

Its probably minor errors in my cap. measurement, since it takes a while to reach a stable reading, and on these first tests I didn't even know if it would work at all so I may have not waited long enough. I still have the samples so I can re-test. The other thing that is likely is that on the high end of the 4.5 FFA line it went above saturated at a lower ppm than 1750, and the sensor maxes at that point so the line would go horizontal.

I will be doing lots more data points since this appears to be the easiest measurement of water amounts. I will always be doing it before any other tests, even HPT. Its so easy I even went around to a bunch of my stored barrels today with it just for info so I know which old barrels I may want to further test.

[SunWizard]

This thing is GREAT! It is very handy for testing my dewatering with my CF or any other method, since it can be monitored real time, or as often as desired, watching the cap. number go down as it gets dewatered further. I will even use this dipped into my vehicle tank periodically as a very simple check. It might make me want to add a desiccant breather (I am a dry climate so I haven't), or could tell if your existing desiccant is still working.

I have done much more testing on it, and updated the chart in post#1 with some of the data.

Formula for %saturated= (cap. pF - 311) / .6

Calibration for % saturated
According to the data sheet, each sensor chip can vary in the pF reading at 0 % which for mine is the 311 number. Yours may be different, and you can find that number by testing several samples (low and higher water content) drawing a chart like I have to find where the line intercepts the Y axis. Then use that number in the above formula or on the % saturated calculator page.

(edit) I have found PPM is not as useful as % saturation since PPM varies hugely depending on the FFA. And % saturation easier to measure and compare between varied batches. I set a goal to get the % saturation below 60%. I find that 60% is easy to reach since my centrifuge runs have all been getting below 30%.

I am mainly reporting PPM as a means to compare to the tests that have measured PPM in the past. And I bought the expensive little tester so I might as well use the 50 test packets it came with.

There are 3 ways I have used to calibrate the results to PPM, outlined below. You can calibrate this meter for PPM without having the sandy brae tester ($250 + $2/test) Here is how:

Calibration to PPM without SB:
This device is so linear its easy to calculate PPM by doing a simple spiked sample as I have done. Save a jar of sample A, get a cap. reading, then dewater 1000ml of sample A using an evaporative method, even if its simply in a large pan on a hot stove for a while. Cool to at least 230F or it may hurt the sensor, and confirm its dewatered by a low cap. reading. Cool it down to room temp. so the readings don't need temp. compensation, write down the reading Acap. Then make a spiked sample (AS) by adding 0.5ml water to the 1000ml = 500 ppm and stir well, (I used an old blender on low for only a few seconds) and wait an hour or longer (I did overnight on some) for the water to become dissolved. I shake the sample in a jar before a test. These 2 samples then give you the slope of the line:
Slope = 500 / (AScap - Acap) in PPM per pF
Then to compute the PPM of any sample of oil similar to A:
PPM = (cap - 311) * slope.

For example, I did this PPM calc. on a group of my samples, then compared it to the SB readings, and the results are surprisingly close (within the error of the SB test):
Sample, computed, measured PPM
A 1026 1100
B 591 560
C 160 211

After you have done this cap. test on a few of your samples, run the above formula, and draw a chart like I have in post#1, then it becomes really easy since you don't need the formulas. Measure your FFA, measure the cap, look at the chart to get % saturated and PPM water.

Calibration to PPM with 1 SB or lab test:
With a SB or lab KF test, you only need 1 sample point to calibrate the saturated line shown on my chart for your oil. Test your sample with both SB and cap. meter.
Slope = SB PPM / (cap - 311) in PPM per pF
Then to compute the PPM of any sample of oil similar to your tested sample use the same formula as above:
PPM = (cap - 311) * slope.

The 311 number is the 0% RH for mine, which may vary for different chips you buy, but once determined it will drift only slightly over years according to the data sheet.

Calibration to PPM with 2 SB or lab tests:
With 2 SB tests, you get the best calibration of the saturated line shown on my chart for your oil. Test your sample 2 times, such as before (SB1) and after (SB2) dewatering, with both SB and cap. meter.
Slope = (SB1 - SB2 PPM) / (cap1 - cap2) in PPM per pF
Then to compute the PPM of any sample of oil similar to your tested sample use the same formula as above:
PPM = (cap - 311) * slope.

I doubt I will even be using my Sandy Brae unit much anymore since its $2 per test, and uses hazardous hydride so I have to put on gloves, goggles, mask, protective shirt, etc. This meter is much quicker and simpler in every way. I have found the colder the oil, the longer you need to leave the sensor sit there before it stabilizes at the final reading. At 60F and high % saturation it can take a couple minutes. At 80F, its 30 seconds.

This would also be very handy for the Biodiesel makers. They don't have to account for the widely varying FFA as we do.

I will be curious to see some others data points. I know some of you will want one of these right away. I wish I made one long ago. If you already have a digital voltmeter that reads capacitance, its only $5 for the humidity sensor to try it!

Check to see if your meter can read in the 300pF range. Mine is a Protek CM109 Capacitance Meter. Source: ebay. Another tip is keep the wires short, the high accuracy ones like mine have leads that are only 6" long since length affects the cap. measurements.

[SunWizard]

Some smart folks out there may ask, why not simply buy a ready-made RH meter? Most of the ready-made RH meters use a different type of sensor (resistive not capacitive) which isn't nearly as good, and might not work at all immersed in VO. The other reason not to get a commercial RH meter is that the sensor is usually enclosed so you can't wipe the VO off it after each test to prevent a polymerization build-up.

Another question is, how do you have the sensor attached? Presently, I have it rubber-banded to the non-ball end of a used-up BIC ball point pen. Now that I know how well it works, I may make a huge upgrade and epoxy it onto the pen. Maybe even run the wires up inside a hollowed out pen body, which would make wiping the VO off after a test easier.

[SunWizard]

For temperature compensation, its 0.16 pF/ degree C.
Here are some calculations I did for those of us wanting to read hot VO like in my dieselcraft centrifuge while processing. To convert readings at 160F (71C) to 70F (21C), add 8 pF, then I can use the room temp. chart or formulas I posted above.

I do this F to C conversion so often, I wrote a little page here:Fahrenheit to Celsius Converter

Making that little page was so easy, it made me think of this which is really handy since it computes % saturated from any cap reading, and I added temperature compensation too:
Capacitive water in oil meter Percent saturated calculator
If anybody likes it, I could make it compute PPM if you input the slope. Maybe even enter FFA, it computes slope and PPM if I get enough data showing the slopes for various FFA samples.

[John Galt]

Keep up the good work. Your research is appreciated.

[HoldOnTight]

I am testing a Capacitive water meter I made and its giving interesting results so far. I am doing many calibrated samples and tests in the Science project thread which are also useful to test this device.
It cost $50, and gives data similar to a $3000 unit like the Eesiflo Online Water in oil Monitor I got a high accuracy (0.5%) capacitance meter (ebay Protek CM109 $45) and a honeywell humidity (RH) sensor HCH-1000-001 from digikey.com link
for $5:
data sheet PDF & pic

I thought even if it didn't work immersed in VO, I could use it to measure and compare the humidity in the air in the headspace of stored barrels, as well as in the headspace of my dieselcraft CF rig to tell how the dewatering is going. But it is working great immersed in the VO!

Immersed in VO, it takes about 30-90 seconds to get a stable reading depending on the VO temp. It reads humidity of the VO just like it does in air. Humidity of the VO is basically % saturation, see Dissolved water and % saturation. Here is a chart of my 1st data points at 61F showing how linear its response is:


With sample A=4.5%FFA shown as the red line, as I got above 2000 ppm, the VO was saturated, and it started failing the HPT, and the sensor maxed at 367. Low FFA samples of VO fail a HPT before 100% saturated, and this is a much easier test than the HPT, and gives quantitative data for how far below saturation you are if you use a dewatering method that is capable of getting you below saturation. (Settling and motor driven CF will not get much below 100% unless you add an evaporating method) Getting below 100% saturated is a very good thing and should be a goal for everyone. Because if you are near saturated, humidity brought in from your tank vent, or a temp. change could mean free water condensing in your tank.

With sample B, my good VO source, FFA=2% shown as the green line, capacitance rises much quicker, and fails the HPT much sooner <1100 ppm. Calibrating this meter I cover in a later post below. Then you take the meter reading, and compare to a chart like the above for your VO type to find % saturated and PPM. This is what is done internally in the expensive water in oil sensors.

For others wanting to make one, a high accuracy capacitance meter is needed since 0.5% of 300 pF = 1.5pF = 3% humidity. A cheap meter like the harbor freight and many others is only 3% accurate which is 18% humidity error.

Between samples, I wipe the VO completely off with a rag, which is easy since its a chip of glass, and it quickly goes back to measuring the RH of the air in my garage. You want to do this to ensure that no oil polymerizes on it, and to make the next reading accurate quicker.

Here is a link to buy the sensor:
http://search.digikey.com/scripts/DkSea ... 80-2903-ND

One trick is I will need to do temp compensation like shown on the datasheet link above, although its small at 0.16 pF/deg.C. These tests were run at room temp. 70F. Much more testing to come...

Sunwizard, since the saturation point is strongly dependent on temperature as well as the presence (or absence) of polar molecules (dirt, water, soap?), how do you eliminate these false actors to get an accurate water content reading?

Also, I received my sensor today. I already have a capacitance meter. Have you got a good probe setup yet? How about a Bic Pen case for a probe starter, if they even make those any more... I'm looking for ideas... Pics are always appreciated and my request is one of the best forms of flattery!

If you used Excel to generate the chart, I'm asking you to make it available on this forum as a download. I'll PM you with my email address otherwise. Thanks!!!

[SunWizard]

Sunwizard, since the saturation point is strongly dependent on temperature as well as the presence (or absence) of polar molecules (dirt, water, soap?), how do you eliminate these false actors to get an accurate water content reading?

For temp, I apply temp. compensation as described above.

With this sensor, only the molecular water gets through the polymer layer which is where the cap reading comes from, so its accuracy in % saturation shouldn't be affected by the other polar materials (FFA, soap, etc.) This makes it a more useful number than PPM. The PPM computed from the cap. reading can vary widely as shown by my chart, with FFA being the biggest polar material by far that affects the PPM for the same % saturation. That's the reason I made the chart, to see the large effect that FFA is having on my VO.

There might be some contaminants (soluble acids?) that could affect the sensor over the long term, causing a slow drift from its previous range. This is something to test for, maybe periodically do a calibration check by dipping into a sample that you measure regularly and know where it should read.

Also, I received my sensor today. I already have a capacitance meter. Have you got a good probe setup yet? How about a Bic Pen case for a probe starter, if they even make those any more... I'm looking for ideas... Pics are always appreciated and my request is one of the best forms of flattery!

Good to hear you got one, its a great tool. I am using it rubber banded to the non ball end of a bic ball point pen. I will soon make a big upgrade and use silicone or epoxy to mount the sensor on the end of a hollowed out pen body. This will make it easier to wipe off after each dip into the WVO. I am waiting to post a picture of that.

If you used Excel to generate the chart, I'm asking you to make it available on this forum as a download. I'll PM you with my email address otherwise. Thanks!!!

I used Microsoft works since it was simpler than excel to get the chart I wanted. Here is a link: http://www.burnveg.com/cap_water_ppm.xlr

[HoldOnTight]

Good to hear you got one, its a great tool. I am using it rubber banded to the non ball end of a bic ball point pen. I will soon make a big upgrade and use silicone or epoxy to mount the sensor on the end of a hollowed out pen body. This will make it easier to wipe off after each dip into the WVO. I am waiting to post a picture of that.

If you used Excel to generate the chart, I'm asking you to make it available on this forum as a download. I'll PM you with my email address otherwise. Thanks!!!

I used Microsoft works since it was simpler than excel to get the chart I wanted. Here is a link: http://www.burnveg.com/cap_water_ppm.xlr

Sun,
That is exactly what I was thinking. Body of a pen or sturdy straw, with each wire separated as much as possible to minimize extraneous capacitance, epoxy to support wiping. But my recent thought is to use a flat strip of plastic (something like a kitchen ladel handle or a .5 in. PVC Tube), so I can separate the two wires more, and seal it in epoxy. How to measure half way down a barrel if the extension is only 1 foot long? You have more experience using it and can more easily try different configurations. Thoughts/Ideas?

I'll have another week to think about this since I'm going out of town for another week and little time to play this weekend.

Okay, I tried right clicking on the link and the download option wasn't available. I also tried loading the address in bitTorrent and it couldn't find th e file. Maybe you could send it to me using gmail: inspironator at... Thx.

[SunWizard]

Sun,
That is exactly what I was thinking. Body of a pen or sturdy straw, with each wire separated as much as possible to minimize extraneous capacitance, epoxy to support wiping. But my recent thought is to use a flat strip of plastic (something like a kitchen ladel handle or a .5 in. PVC Tube), so I can separate the two wires more, and seal it in epoxy. How to measure half way down a barrel if the extension is only 1 foot long? You have more experience using it and can more easily try different configurations. Thoughts/Ideas?

It may not be good to dip that deep, since the long wires may affect capacitance even when separated. Mine only has 6" leads so I haven't tried depth. I pump out a sample from depth. Here are ways to test effect of depth: 1 .stir the barrel well, then read the top, read at depth and they should match. 2. take off the sensor and dip just your long test leads and see what capacitance you read. This is how you zero adjust my meter, but the leads are so short it made little difference.

Okay, I tried right clicking on the link and the download option wasn't available.

With my Firefox browser, all I do is right click and pick Save Link As...

[SunWizard]

To make the probe, I soldered 6" insulated 22 guage wires onto the sensor leads, then slid them through an old hollowed out ball point pen body, then sealed both ends with silicone caulk. If you don't have a soldering iron, this isn't needed since I have been using it just fine for the past week with the sensor rubber banded to the pen body. Putting the wires inside the pen body makes wiping it off after each sample easier.

The 1st pic of my meter and probe with sensor on the right side:

[HoldOnTight]

Yup, pumping out from a given depth works.

I hoped to use longer leads and zero out the extra capacitance, but after thinking about it, that wouldn't work without introducing error, got it!

Thanks for the pics. I am now convinced that the sensor support & leads only need be long enough to dip into a sample jar.

Now I can proceed. I'm processing the asian source of oil today and I may just measure the RH of oil before and after dewatering. My before sample is in the prefilter barrel.

I'll start saving the oil to the spreadsheet so I can build a baseline for comparison. I will measure FFA (before and after processing) and now, RH (before & after).

[SunWizard]

I added a chart of % saturation versus temperature (created with this meter) to this thread:
Dissolved water and % saturation in WVO

Here are links to commercial products that measure water in oil using a capacitive sensor like mine. These are the models I studied before I made mine. Useful if you want to buy a commercial version:
Eesiflo Online Water in oil Monitor
Pall WS09 Portable water sensor
Doble Domino Moisture Sensor
Water Saturation as a Screening Method for On-site Oil Analysis
ANALEX moisture sensor

[SunWizard]

I discovered another nice way to use this meter today. I made a 6" diameter circle of plastic cut out from an old VO bottle (a cubie would work.) Then stuck the probe through a 1/4" hole drilled into the plastic. I measure the humidity in the headspace of stored barrels by opening them and quickly sticking the sensor in there and blocking the opening without letting in ambient air. The reading in the air has matched the reading in the WVO near the top in 3 barrels. This is handy since the sensor responds quicker than when its dipped into the WVO, and it doesn't need wiped off.

Here are the results from a centrifuge run I did today. Average quality WVO source, FFA=2.5%, settled for 4 months. Began at 700 PPM and end 270 PPM with sandy brae, also measured every 30 minutes with my capacitive water in oil meter to make this chart, and temp. compensated. (anyone with a CF should make one, its great, much better than sandy brae!)