How Can I Measure Dosage of Ferromagnetic Powder in Kerosene?

[mendes]

Hello,

I am looking for a way to measure the dosage of a powder (made of metallic particles (ferromagnetic) that are 6 microns large) in a liquid solution that is kerozene. When mixed with the liquid a suspension is obtained that takes about one hour for the powder to settle down. Once settled down, the powder covers around 5% of the liquid volume. I am looking for a technique to measure the dosage of the powder in the liquid (its percentage in the liquid volume.)

I tried U tube liquid variation but it was not conclusive...

Any idea?


Thanks.

 

[Baluncore]

This is the type of problem encountered in the monitoring of dielectric quality in die sinking EDM.
I would consider monitoring the frequency of an inductor tuned resonant circuit immersed in the contaminated liquid. The magnetic particle concentration will change the inductance of the coil. Magnetic particles will lower the frequency, non-magnetic electrically conductive particles will raise the frequency very slightly. Take care that the dielectric level does not significantly change the capacitance of the circuit.

 

[rollingstein]

Why not filter n weigh? Or does it have to be online?

 

[Baluncore]

The particles could be collected on a magnet.

The 6um particles may pass through any filter with a realistic fluid flow rate. Fuel filters on diesel vehicles are rated at about 10um. If the filter was finer then it would clog when faced with such high concentrations as 5% by volume.

If the fluid needs a particular particle density to function, then a non-destructive assay is needed.

We need more information on the application to better assess the possibilities.

 

[mendes]

This is the type of problem encountered in the monitoring of dielectric quality in die sinking EDM.
I would consider monitoring the frequency of an inductor tuned resonant circuit immersed in the contaminated liquid. The magnetic particle concentration will change the inductance of the coil. Magnetic particles will lower the frequency, non-magnetic electrically conductive particles will raise the frequency very slightly. Take care that the dielectric level does not significantly change the capacitance of the circuit.

Thanks for your answer.

I would like to have a link on how the tank circuit is used to measure the dielectric constant of liquid in die-sinking EDM.

Some time ago I used a tube with mixed kerozene and powder that I inserted in a coil to see if it changes its inductance but there was no measurable effect.rollingstein: i want a quick technology, online, to do the measurement.

 

[Baluncore]

You need to measure the magnetic permeability of the powder while avoiding the dielectric constant of the kerosene.

An LC oscillator has a frequency set by the values of inductance and capacitance.
http://en.wikipedia.org/wiki/LC_oscillator
http://en.wikipedia.org/wiki/Electronic_oscillator
The capacitance is controlled by the dielectric constant. The inductance is changed by the presence of magnetic particles. http://en.wikipedia.org/wiki/Ferrite_core

An inductance can only be measured to a few percent but the frequency of an RF oscillator can be measured to at least 6 digits when referenced to a crystal oscillator. An LC network that sets the frequency of an oscillator can operate at 1MHz without problems.

The critical thing here is that the size of the magnetic particles that make up the magnetic core must be smaller than the skin depth in that material at that frequency. 6um or smaller is well into the MHz range. http://en.wikipedia.org/wiki/Skin_effect

Unfortunately kerosene has a dielectric constant of about 1.8 and probably varies with temperature. Air and space have constants of 1.00 while water is about 80. So the capacitance of the LC circuit must be independent of the kerosene temperature and composition. That will require a temperature stable reference capacitance, such as silver mica. Because the inductor will be immersed in the dielectric there will be some capacitance between the end terminals of the inductor that will be added to the fixed reference capacitor. That could be a problem.

It would be possible to build two identical oscillators, one with clean kerosene, the other with kerosene-powder mix. The temperature of the two could be equilibrated. The frequency difference would then be a strong function of magnetic particle content.

A frequency difference can be generated by using an RF mixer (non-linear) or by a simple exclusive OR gate to generate the difference or beat frequency. That system is usually called a Beat Frequency Oscillator, BFO.
http://en.wikipedia.org/wiki/Beat_frequency
http://www.geotech1.com/pages/metdet/info/bfotheory/bfo.pdf

You will need to design an enclosed inductor in a flow cell that does not precipitate or clog with the powder. Avoid magnetic materials or you will have powder separation problems. Two of those cells in close thermal contact then make the kerosene-powder sensor as part of two oscillators.

 

[mendes]

You need to measure the magnetic permeability of the powder while avoiding the dielectric constant of the kerosene.

An LC oscillator has a frequency set by the values of inductance and capacitance.
http://en.wikipedia.org/wiki/LC_oscillator
http://en.wikipedia.org/wiki/Electronic_oscillator
The capacitance is controlled by the dielectric constant. The inductance is changed by the presence of magnetic particles. http://en.wikipedia.org/wiki/Ferrite_core

An inductance can only be measured to a few percent but the frequency of an RF oscillator can be measured to at least 6 digits when referenced to a crystal oscillator. An LC network that sets the frequency of an oscillator can operate at 1MHz without problems.

The critical thing here is that the size of the magnetic particles that make up the magnetic core must be smaller than the skin depth in that material at that frequency. 6um or smaller is well into the MHz range. http://en.wikipedia.org/wiki/Skin_effect

Unfortunately kerosene has a dielectric constant of about 1.8 and probably varies with temperature. Air and space have constants of 1.00 while water is about 80. So the capacitance of the LC circuit must be independent of the kerosene temperature and composition. That will require a temperature stable reference capacitance, such as silver mica. Because the inductor will be immersed in the dielectric there will be some capacitance between the end terminals of the inductor that will be added to the fixed reference capacitor. That could be a problem.

It would be possible to build two identical oscillators, one with clean kerosene, the other with kerosene-powder mix. The temperature of the two could be equilibrated. The frequency difference would then be a strong function of magnetic particle content.

A frequency difference can be generated by using an RF mixer (non-linear) or by a simple exclusive OR gate to generate the difference or beat frequency. That system is usually called a Beat Frequency Oscillator, BFO.
http://en.wikipedia.org/wiki/Beat_frequency
http://www.geotech1.com/pages/metdet/info/bfotheory/bfo.pdf

You will need to design an enclosed inductor in a flow cell that does not precipitate or clog with the powder. Avoid magnetic materials or you will have powder separation problems. Two of those cells in close thermal contact then make the kerosene-powder sensor as part of two oscillators.

Many thanks for the detailed reply.

In terms of capacitance because kerosene and iron oxide (powder) have very similar dielectric constants, this property can not be used to measure the powder concentration, I think. So, if we use the inductance based principle we do not expect a considerable change in the capacitance, which i presume is a good thing.

In terms of inductance, using the mixture that I put in a thin tube as a core inside a coil, this did not lead to a measurable change in inductance. But, you are saying that measuring the resonance frequency change of an LC circuit immersed in the mixture may be more conclusive. Is that correct?

 

[Baluncore]

The dielectric constant of iron powder may be very high but should be irrelevant if you are measuring the magnetic permeability. The measurement of inductance is fickle because it is frequency and flux sensitive. There is no such thing as a wide-band reference inductor or meter. Your failure to measure a change in inductance when placing a thin tube in a coil is not surprising. I have no idea of the geometry of the tube and coil, or the frequency, resolution or type of instrument used to measure the inductance.

An LC oscillator has a frequency determined by an inductor in parallel with a capacitor.
Stray capacitance between the ends of the inductor must be added to the capacitance.
Stray inductance in the leads of the capacitor must be added to the inductance.

Good temperature stable reference capacitors must be used, hence silver mica.
The stray capacitance variation across the inductor must be minimised.
The dielectric constant of kerosene is probably temperature sensitive.
The inductor connections are immersed in kerosene, hence variable stray capacitance.

The presence of iron powder in the inductor will increase it's inductance, lower the frequency.
By building two identical oscillators, everything is canceled except the iron powder effect.
By beating the two oscillators together gives a very sensitive inductance measurement.The world of magnetic measurements is complex enough without unknowns. There are too many subtleties. It really would help to know what you are using the iron powder for and why it could not simply be scrubbed from the kerosene with a magnet. If the iron powder needs to be present in the correct ratio then the process it is part of should help identify the optimum assay technique.Another way to monitor powder could be with IR transmission from an LED to a phototransistor.
IR transmission measurement of oil quality. 1986, US Patent 4,570,069 must have expired by now.

 

[mendes]

The dielectric constant of iron oxide is around 1 (very low), quite similar to the one of kerosene. I thought about using a simple capacitor to measure the powder concentration, as it is usually done in moisture meters (water dielectric constant is high (80), is quite different from the one of the material where water may be present as moisture like paper bales for example). But, because kerosene and powder have similar permitivities the capacitance method obviously can not work.

I will consider the LC circuit method.

IR absorption may work but it has a high inaccuracy (around +- 5%). My powder is supposed to be present in kerosene with contents no more nor less than around 0.3 %! So as you can see it is quite a very low concentration so both high sensitivity and high accuracy are required for the measurement method.

 

[Baluncore]

Once settled down, the powder covers around 5% of the liquid volume.

My powder is supposed to be present in kerosene with contents no more nor less than around 0.3 %!

That now suggests “precisely 0.3 %” which would imply a reagent rather than a contaminant. Is that by weight or could it be by volume?

You have now moved the goal posts again. Magnetism is a chemical bonding property. There is a huge difference between iron powder and the forms of iron oxide. Iron oxides can take on many forms but Haematite (Fe₂O₃) or Magnetite (Fe₃O₄) are most common, but have quite different magnetic properties.
What really is the powder you have here?

Are you trying to detect sub-filter sized fines contamination of the electrolyte used in Ram/Die Sinking EDM? If so, why are the metal particles in the form of an oxide when in kerosene, I would have expected any oxygen present to make a thick hydroxide sludge that would block the filters. Maybe, if it is a contaminant that settles out, you should use an inline centrifuge to clean the dielectric.

If you insist on holding all your cards so close to your chest that I cannot help you, then you are also wasting your time.
You must be new to this Physics Forum game.

 

[mendes]

That now suggests “precisely 0.3 %” which would imply a reagent rather than a contaminant. Is that by weight or could it be by volume?

You have now moved the goal posts again. Magnetism is a chemical bonding property. There is a huge difference between iron powder and the forms of iron oxide. Iron oxides can take on many forms but Haematite (Fe₂O₃) or Magnetite (Fe₃O₄) are most common, but have quite different magnetic properties.
What really is the powder you have here?

Are you trying to detect sub-filter sized fines contamination of the electrolyte used in Ram/Die Sinking EDM? If so, why are the metal particles in the form of an oxide when in kerosene, I would have expected any oxygen present to make a thick hydroxide sludge that would block the filters. Maybe, if it is a contaminant that settles out, you should use an inline centrifuge to clean the dielectric.

If you insist on holding all your cards so close to your chest that I cannot help you, then you are also wasting your time.
You must be new to this Physics Forum game.

Its quite simple actually, the suspension is used in magnetic particle testing. The magnetic powder is needed in a liquid carrier (kerosene or sometimes water). A magnetic field is applied on the metallic part that is to be inspected against flaws, and the suspension dropped on the part. The magnetic powder is attracted to areas where there is a magnetic leakage.

Concentration should be around 0.3% in volume.

I don't know what kind of iron oxide is the powder, but it is magnetic (weak ferromagntism). There are quite different types of powder. One example is here:

http://www.envirocoustics.gr/products/pt_mt/pdf/Fluorescent Wet 14A.pdf



As far as the LC resonant circuit is concerned, because the resonance frequency is proportional to the square root of the LC product, I am afraid that if the change in inductance is small and not measurable in my case then the change in the resonant frequency is even smaller.