![lubrificant[1].jpg](https://i.imgur.com/g4HSm6f.jpg "lubrificant[1].jpg")
So I want to make a multi sensor to measure the following things in engine oil.
>diesel
>water
>anti-freeze
>soot
>metal particles
Wat do?
How do I detect the percentage of each contaminant?
I need ideas.
>>949259
a spectrum device on a glass tube or something similar, but this is hardly diyable
>>949284
>this is hardly diyable
That's why I want it to be a multi sensor.
It's easier when you're just measuring one unique property of each contaminant.
>>949291
No dude, the way you detect minuscule contaminants in oil is with a spectral analyzer which is stupid expensive lab equipment. It's soooooo much cheaper just to pull an oil sample and send it to get tested that "diy-ing" a spectral analyzer into your project (I'm assuming your truck) is objectively a bad idea.
>>949259
Have you considered a turbidity sensor?
It looks at a fluid flow and measures the opacity.
>>949297
Interdasting.
So I could detect diesel with that.
>>949301
It would tell you viscosity, but there is no way to tell how much or which contaminant you have.
>>949301
You will detect when your liquid is more opaque than it was before
>>949297
The automotive manufacturers thought of OP's idea decades ago, and this was the best they could do. It measured 'oil life' by checking opacity as oil collects various byproducts of combustion and engine operation and becomes progressively more opaque.
What OP wants to do is ridiculously difficult and impractical. It depends on knowing the exact chemical composition of the engine oil (which differs considerably across brands/preparations) and comparing the chemical composition of the working fluid to it. It would take a scientist with a laboratory, a mass spec, and an NMR hours to come up with one satisfactory analysis.
An interesting solution and is fun to diy is set up a usb microscope to look at a pumped glass tube, with a screen somewhere to monitor the real color of the oil, and at a microscopic level you will be able to detect how dirty by seeing the particles flying pass.
But you will have to set up the exposure and color tempreture constant and have the thing sealed (so no light can get in) to have accurate lighting.
>>949310
I do have a lab and a mass spec.
But I'm looking for other ways.
>>949309
So does almost everything else.
>>949314
I suggest a synchrotron light source
>>949259
>water, and diesel
conductivity analysis, if the conductivity is too high then there is too much water, the conductivity sensor also has the advantage of igniting any diesel fuel thus alerting the operator to the presence of said fuel
>anti-freeze
the "barking at dog at 3am that mysteriously killed over" have been known to collect copious amounts of anti-freeze however the saturation level various wildly between breeds
>soot
santa clause used to be an effective indicator of soot but since the advent of gas fireplaces this method has gone to the way side. however science is always advancing and now the method of choice is the "fine honey walking down the street that's about to get coal rolled on that ass" scientific video related https://www.youtube.com/watch?v=iRu3xtTH9ng
>metal particles
magnets.... how do they work
>>949314
>I do have a lab and a mass spec.
Then why are you asking us? You are more set up to handle this than 99% of this board.
Seems to me you are just hoping some genius is lurking and will suddenly post a billion dollar idea for you.
Now for my 2 cent idea, everything evaporates if you get it hot enough, so rig up a device that will boil the oil and measure the temperature. So long as nothing in the oil boils off at the same temperature, you should get a load of data that basically shows the boiling points of all the various substance in the oil. Figure out what boils of when and your set.
Only downside is you will constantly be replacing the oil required for testing, so you're not doing anything that an oil change won't fix.
>>949345
>Seems to me you are just hoping some genius is lurking and will suddenly post a billion dollar idea for you.
Didn't know this was a billion dollar thing.
It's just a pet project.
>>949347
>a device that can tell you the exact composition of a coolant/lubricant in real time as the machine is being used
>not a billion dollar idea
M8, what you are describing would change it from "maybe she needs an oil change" to "holy hell look at that, how do we fix that"
From a research point it's a game changer (especially if it can monitor more fluids than oil), and commercially it would be a huge improvement over anything we have now, assuming it isn't cost prohibitive. It goes way beyond trucks and cars, it would be useful everywhere from robotic assembly to monitoring nuclear plants.
>>949353
Is it being researched then?
You also have to remember that it's only a billion dollar idea if you actually manufacture and sell the sensors that can do this.
Just coming up with solution to this problem won't make you rich.
>>949354
I would assume someone somewhere is researching it, but I have no idea who or where. What's wrong with just changing you oil when the light comes on like everyone else on the planet?
>>949360
Because you could save some money if the oil is still good for another good while.
>>949361
You could just go by what's in your book instead of what your mechanic or your dad told you. Or clean it out and switch to synthetic and get like 10,000+ miles off it.
Ultimately you do still jut need to change your oil.
This would be more useful in an oil recycling place to verify it was clean or if it should go for another cycle.
>>949361
>build $5000+ rig
>save $10 a year on oil
What's the point?
>>949360
>changing you oil when the light comes on
>when the light comes on
In 99% of cars, the light coming on means you waited way the fuck too long to change your oil. Because it's normally a check engine light.
>>949259
good idea, but you dont.
like everyone else has stated you need a spectral analyzer, there is companies that you mail an oil sample off too and they will test for all kinds of stuff and give you a health status of your engine. its not too expensive.
The stuff youre monitoring for isnt that big of an issue right away. the if the metal particles are that big of an impact Id imaging the engine would blow up almost as soon as your sensor would detect them. Id say if you are really worried about this then have a sample mailed off every year or whenever.
>>949388
>check engine light directly relates to oil
most shit that triggers the check engine light is not related to oil in any way.
also I think he meant the oil light in particular
>>949360
>What's wrong with just changing you oil when the light comes on like everyone else on the planet?
the oil light means your oil is low and maybe the oil pressure is fucked up.
the oil light in no way means "about now is when you change the oil".
nobody that deserves to have a car waits for an oil leak to change the oil. running a car with low or no oil will fuck it up bad.
>>949388
>>949399
You guys are a little behind. There actually is a message displayed in many newer vehicles that says "change oil soon," or something to that effect. They claim to use an algorithm to determine when you need it, I think it's bs and don't totally trust it myself. I think it's mostly just milage. I know it's not every vehicle, but it's way more than 1%. Also nice dubs to you both.
>>949401
I live under a rock, stop trying to keep me updated. have you ever though about how your actions and "modern information" can damage my lifestyle?
really though, how does the timing get reset? is there a button you push that says "ok I changed my oil" or do you need a bullshit scan tool
>>949259
>detecting diesel in engine oil
>oil in oil
That's going to be a tough one.
>>949410
on chevys you turn the key to accessory and push the gas pedal three times or something.
In old volvos there is a button by the ECU you push in a magic sequence.
That's all I have experience with, I think a lot of brands might expect you to have an ODBII meter.
My wife's car has an 'oil life' text read out, and my volvo will send me a message when it's time.
I'm scared to wait for the volvo though, even though the manual clearly says 10,000km/7500 miles, but I can't do it, I'm too scared to let it go so long because I was raised on 3000 miles, so I never see the oil change message
You're talking about (aliphatic and simple) hydrocarbon mixtures in a hydrocarbon solvent.
For the organic substances on a small (and cheap) scale FTIR is your best bet. For the metals the industry uses either AAS or ICP.
There's no such thing as a "sensor" for this thing. There's a reason ana chem is a billion, if not trillion, industry. Find yourself a book online on Analytical Chemistry.
Now if you have any experience in this field and you want to play around with relatively new detection techniques. Look around for "Lab on a Chip" inventions which are disposable and look up OpenQCM, might help you with analysing flow rate cheaply.
Source: Bsc in Chemistry
>>949314
Bwahahaha what a pile of bullshit. You have a mass spec for yourself. Do you happen to have a few GCs as well? Not to forget the SEM and AAS for the real diying.
>>949553
now that you mention it my moms car bitches to change the oil every time Ive driven it, odometer says in english "change oil", the computer physically beeps, and the sticker in the windshield clearly says shit needs changed. last I checked it was past the 10,000 mark.
>inb4 get it changed for her
>>949259
Like I said in the thread you started on /sci/, the only reliable way to do all of these simultaneously is some kind of miniature mass spectrometer. Good luck.
>>949307
Add some kind of conductivity or resistance test function. That could indicate the presence of water, antifreeze, and possibly metal particles.
>>949410
Basically yes, you hit a button that says "I changed my oil." Of course, since I kind of live under a rock myself and only peek out on occasion, first you Google "how to reset change oil interval" or some such. Then you get your teenaged son to do it for you.
>>949689
Stop stalking me.
I think I will make myself an impedance analyzer.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.568.7291&rep=rep1&type=pdf
>>949309
Synthetic oil is almost clear when it goes into a brand new engine, and almost entirely black the next day. And water coolant is clear when it goes into a brand new motor, and who knows what disgusting color it turns when it's next changed. Ever seen a cooling system get flushed? There's all kinds of nasty shit in there.
OP I don't think we have the technology for you to easily do what you want :/
Are you retarded?. Look diesel burns in the oil galleys. Coolant can be readily detected by looking at the dipstick by giving the oil a halo/ rainbow effect. Soot is a byproduct of diesel combustion and not a problem. Metal particles can be trapped with magnets attached to the oil filter and oil pan. On top of that diesels have very large tolerances compared to the working tolerances of 4stroke gasoline motors. But if you want to waste a fortune on something dumb be my guess.
>>949664
>mom
not using it as a reason for the food and leftovers. Show up, open the drain plug. Go inside sit down for dinner. Go back out while the moms paper plates your leftovers.
I do this for a couple of my friends' moms. Its like thanksgiving dinner every 3 months.
>>949799
>food and leftovers
must be nice living in the 1960's
my mom and her husband neglect that car. I'm waiting for it to fall apart and its not even close to being 10 years old yet. some how the brake fluid is only half full, and it is as black as his heart.
there is a bubbling noise every time you use the breaks.
fast food left in there so much the car permanently smells like old mcdonalds.
oil hasn't been changed in over 10k miles.
few other things not worth mentioning
>>949812
>must be nice living in the 1960's
You're like 30 years too early
>>949819
>must be nice living in the 1960's
>You're like 30 years too early
so 1930's or 1990's?
>>949812
he probably topped it up with water or motor oil or some shit. you should fix his brakes anon. or get them to get the leaking seal replaced.
drive to where there is a secluded gentle hill. then get them to use the brakes with the motor off. explain to them that this is what it will be like when their brakes fail.
>>949842
probably.
>>949259
Dream on, kid.
>>949259
Mass spectrometer
