Electronic Antifreeze Tester?

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
Hi,

Anyone ever try to build one of these? I got one recently from an auto parts store but it's the turkey baster type, and the 'cold' fluid measurements are not very informative as shown here (and note 3 balls floating with pure distilled water already):

All fluids room temperature, about 75 to 80 degrees F.
Starting concentration with distilled water: 0.000 percent
After 0 pure antifreeze additions: 00 percent, 3 balls floating
After 1 pure antifreeze additions: 20 percent, 4 balls floating
After 2 pure antifreeze additions: 33 percent, 4 balls floating
After 3 pure antifreeze additions: 43 percent, 4 balls floating
After 4 pure antifreeze additions: 50 percent, 4 balls floating
After 5 pure antifreeze additions: 56 percent, 5 balls floating

The test concentration is given above in percent, and the number of balls floating with that concentration. Granted the water should be hot, but i cant get it hot in my car because the only access to the water is through the overflow tank, which doesnt get hot.

So i was thinking an electronic one would be interesting. If we had a way to measure something about the water mixture, we could correlate that to the degrees of protection, and calibrate it with pre mixed known concentrations of pure antifreeze.

I dont want to buy another tester because it's probably just as inaccurate. Maybe an electronic one would be more accurate and reliable.

Any ideas?
 

jpanhalt

Joined Jan 18, 2008
11,087
Use refractive index. There also might be something based on conductivity, but it won't be as accurate given the variation in water used to dilute the antifreeze..

John

Edit: Actually density/specific gravity is pretty accurate for that mixture. Do you have a scale/balance accurate to 10 mg? If so, just weight a known volume.
 
Last edited:

OBW0549

Joined Mar 2, 2015
3,566
Maybe you could test for dielectric constant, and correlate the results with water/glycol ratio?

EDIT: this site gives dielectric constants for various materials. The values for both water and ethylene glycol vary with temperature, so temperature would have to be taken into account in calculating the water/glycol ratio; but there's enough difference between the two that this might be feasible.
 
Last edited:

GopherT

Joined Nov 23, 2012
8,009
Maybe you could test for dielectric constant, and correlate the results with water/glycol ratio?

EDIT: this site gives dielectric constants for various materials. The values for both water and ethylene glycol vary with temperature, so temperature would have to be taken into account in calculating the water/glycol ratio; but there's enough difference between the two that this might be feasible.
Each manufacturer has their own brew of amines, nitrites and other anti-oxidant/ oxygen scavengers and anti-scale. Some ionize more easily than others and increase conductivity. Conductivity vs concentration will be difficult to calibrate between brands of antifreeze.
 

tcmtech

Joined Nov 4, 2013
2,867
On top of all of that also adding metallic leaching that will also have a substantial influence on the mixtures conductive and other properties that simple floating ball tester won't pick up under normal conditions.

Also as a side note, the fluids in a typical engines cooling system should be drained out and the system flushed every 5 years or so many miles /running hours being antifreeze just like every other fluid in a vehicle has a finite working life before it starts to break down and loose ability to do its job correctly.

IF your fluid is more than 5 years old and of questionable mix value its time to do a dump and flush of the whole system.
 

OBW0549

Joined Mar 2, 2015
3,566
My suggestion was to measure dielectric constant, a.k.a relative permittivity, not conductivity. I would expect conductivity to vary significantly for the reasons you both cited, which would indeed make it a rather unreliable indicator of water/glycol ratio; but I would expect the dielectric constant to depend much less on those factors and give a fairly reliable indication once temperature effects are taken into account.
 

OBW0549

Joined Mar 2, 2015
3,566
Dielectric constant can also be used, but the percentage change for change in concentration is not as great as RI:
The way I'm reading those charts, I get RI varying from 1.333 to 1.43 as %Glycol goes from 0 100, a ratio of 1.073 to 1. Dielectric constant, on the other hand, appears to vary roughly 2:1 from 0% glycol to 100% glycol. Wouldn't this mean that the dielectric constant percent change per change in concentration is MORE than the percent change in RI? Am I reading these curves wrong?

(It's quite possible RI might be a better method regardless, because of its inherent simplicity: just shine a light beam into the fluid at an angle, and measure the deflection.)
 

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
Hi,

Yes thanks, these are all good ideas, thanks. I'll have to think about this a little bit more.

The RI means a little more difficult to read due to the equipment needed, while the dielectric constant is well, an electrical property so should be easier to measure. I'd have to know how much the impurities in the mixture affect the dielectric constant first though.

Thanks for the nice graphs John. That one about the specific gravity tells me a lot about the SG change with temperature which is one issue that come up with this particular meter i have because it wants the water 'hot' and it is a lot easier to test it 'cold' in this car.
 

hp1729

Joined Nov 23, 2015
2,304
Hi,

Anyone ever try to build one of these? I got one recently from an auto parts store but it's the turkey baster type, and the 'cold' fluid measurements are not very informative as shown here (and note 3 balls floating with pure distilled water already):

All fluids room temperature, about 75 to 80 degrees F.
Starting concentration with distilled water: 0.000 percent
After 0 pure antifreeze additions: 00 percent, 3 balls floating
After 1 pure antifreeze additions: 20 percent, 4 balls floating
After 2 pure antifreeze additions: 33 percent, 4 balls floating
After 3 pure antifreeze additions: 43 percent, 4 balls floating
After 4 pure antifreeze additions: 50 percent, 4 balls floating
After 5 pure antifreeze additions: 56 percent, 5 balls floating

The test concentration is given above in percent, and the number of balls floating with that concentration. Granted the water should be hot, but i cant get it hot in my car because the only access to the water is through the overflow tank, which doesnt get hot.

So i was thinking an electronic one would be interesting. If we had a way to measure something about the water mixture, we could correlate that to the degrees of protection, and calibrate it with pre mixed known concentrations of pure antifreeze.

I dont want to buy another tester because it's probably just as inaccurate. Maybe an electronic one would be more accurate and reliable.

Any ideas?
So y'all are saying the results won't be any better than the turkey baster?
 

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
Hi,

Well keep in mind that with the particular 'turkey baster' type tester i got from AutoZone for less than 2 dollars (USD), there were THREE balls floating with PURE distilled water!
That means it was way off, and the only way to use it would be to test it yourself first to find out how many balls SHOULD really be floating for a 50 / 50 mixture.

The other day i picked up a different type and i will be testing that soon. It does not have floating balls it has a 'needle' that rotates when there is fluid inside which gets sucked up using a rubber ball type squeeze thing similar to the turkey baster type.
 

jpanhalt

Joined Jan 18, 2008
11,087
@MrAl
Did you try that with a little liquid detergent in the water? The balls, being plastic, may not be wetted by pure distilled water. Any air on their surface will cause an incorrect reading.
 

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
@MrAl
Did you try that with a little liquid detergent in the water? The balls, being plastic, may not be wetted by pure distilled water. Any air on their surface will cause an incorrect reading.
Hi John,

Good point, but when i added just a LITTLE actual antifreeze to the mixture (less than 20 percent) i got 4 balls to float which is just silly.

Dont you think the detergent would change the specific gravity? How much we talking here, say 1 percent or something like that?

Thanks.
 

GopherT

Joined Nov 23, 2012
8,009
Hi John,

Good point, but when i added just a LITTLE actual antifreeze to the mixture (less than 20 percent) i got 4 balls to float which is just silly.

Dont you think the detergent would change the specific gravity? How much we talking here, say 1 percent or something like that?

Thanks.
The detergent is for a quick rinse to wet the balls. Also, they are extremely sensitive to any micro bubbles so avoid any turbulence as you draw fluid up into the baster. That goes for all types of gravimetric testers.
 

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
Hello,

Thank you for the advice. So i can test it again after washing it with detergent, i guess that is what you are saying, so the balls will only have a very thin coating of soap. Sounds good :)

One thing though, do you think this is what happens when you go out to the car with no detergent in hand and dip it in, take a sample, and then read the number of balls that are floating? Maybe this is an inherent problem with these testers.

But i will try that soap thing and see if it makes a difference. Will be very interesting. I should be able to try that today, and also another tester i have now to test. Wow, testing the testers, isnt that fun :)
Maybe if this new tester actually works i wont bother with the electronic one, although that would be a very handy instrument to have. As long as it doesnt take a year to build one :)
 

hp1729

Joined Nov 23, 2015
2,304
Hello,

Thank you for the advice. So i can test it again after washing it with detergent, i guess that is what you are saying, so the balls will only have a very thin coating of soap. Sounds good :)

One thing though, do you think this is what happens when you go out to the car with no detergent in hand and dip it in, take a sample, and then read the number of balls that are floating? Maybe this is an inherent problem with these testers.

But i will try that soap thing and see if it makes a difference. Will be very interesting. I should be able to try that today, and also another tester i have now to test. Wow, testing the testers, isnt that fun :)
Maybe if this new tester actually works i wont bother with the electronic one, although that would be a very handy instrument to have. As long as it doesnt take a year to build one :)
Even if you get a different reading could you tell which reading is correct?
 

Thread Starter

MrAl

Joined Jun 17, 2014
11,389
Even if you get a different reading could you tell which reading is correct?
Hi,

Yes, as Gopher says you can mix up your own little sample batch of 50/50 antifreeze/water and test the tester that way. With the antifreeze type i have, if it reads exactly -34 degrees F then the tester is very accurate.

I just tested my newer antifreeze tester and i found that it reads -20 degrees F with a perfect 50/50 mixture, and -45 degrees F with a 56 percent antifreeze mixture. It reads what looks like +30 degrees for pure distilled water, which is a good sign, although it is not made for that high.
So it too is a little off, but better than the one with the balls in it. The main thing is it says that you can use it cold so the water does not have to be, nor should it be, hot. So with that i know that if i read -20 degrees F on the actual car then i must have a 50/50 mixture.

Just for reference, between 0 percent concentration and 60 percent concentration and temperatures of 50 to 250 degrees F, the specific gravity matches very close to a tilted plane in 3d space x,y,z.
 
Top