Ti LDC1000EVM A Question on Inductance

John;

Any chance that it's stainless steel?

Hi Michael,

I set up the LDC1000 EVM again so that I could see the effect of a Stainless Steel core. As a reminder here is a screen shot of the GUI with an air core:

We begin with 19.497 uH in air. Now I will put a stainless steel screw through the coil.

Here is a screen shot of the GUI with the change as affected by the SS screw:

We are seeing an indicated decrease in the inductance though it is only a 1.5 uH decrease which is about half that was encountered with the soft steel core. The experiment that I did at the end of this section would indicate that we should not be seeing any change in the  inductance as the SS Screw is paramagnetic and should not affect the inductance.

As a side I did a little research and I learned that not all stainless steel grades are paramagnetic. The screw that I used for this experiment was paramagnetic but apparently there are exceptions. Here is a short article about magnetic permeability in stainless steel.

http://www.azom.com/article.aspx?ArticleID=1140

For the fun of it I hooked up a small inductor and measured it with my Cheap LCR meter. Then I put a brass screw, a stainless steel screw, a steel screw, and finally a ferrite inductor core, into the coil to see what effect they each would have.

Base line open air core 3.3 uH coil.

Brass screw with no magnetic permeability has no effect on the coils inductance.

Stainless screw with no magnetic permeability also has no effect on the inductance.

Ferromagnetic steel screw causes the inductance to more than double to 6.9 uH

Small Ferrite core causes the inductance to quadruple to 13.9 uH

Remember that the ferrite when used with the LDC 1000 EVM also caused the GUI to show an increase in the inductance of the test coil but the steel core had the opposite effect.

John

Hi Michael,

I set up the LDC1000 EVM again so that I could see the effect of a Stainless Steel core. As a reminder here is a screen shot of the GUI with an air core:

We begin with 19.497 uH in air. Now I will put a stainless steel screw through the coil.

Here is a screen shot of the GUI with the change as affected by the SS screw:

We are seeing an indicated decrease in the inductance though it is only a 1.5 uH decrease which is about half that was encountered with the soft steel core. The experiment that I did at the end of this section would indicate that we should not be seeing any change in the  inductance as the SS Screw is paramagnetic and should not affect the inductance.

As a side I did a little research and I learned that not all stainless steel grades are paramagnetic. The screw that I used for this experiment was paramagnetic but apparently there are exceptions. Here is a short article about magnetic permeability in stainless steel.

http://www.azom.com/article.aspx?ArticleID=1140

For the fun of it I hooked up a small inductor and measured it with my Cheap LCR meter. Then I put a brass screw, a stainless steel screw, a steel screw, and finally a ferrite inductor core, into the coil to see what effect they each would have.

Base line open air core 3.3 uH coil.

Brass screw with no magnetic permeability has no effect on the coils inductance.

Stainless screw with no magnetic permeability also has no effect on the inductance.

Ferromagnetic steel screw causes the inductance to more than double to 6.9 uH

Small Ferrite core causes the inductance to quadruple to 13.9 uH

Remember that the ferrite when used with the LDC 1000 EVM also caused the GUI to show an increase in the inductance of the test coil but the steel core had the opposite effect.

John

I looked into the design of the coil on the sensor board:

This isn't what I would exactly call a standard coil. The coil you are comparing to on your "cheapy" meter is a different architecture. Try winding a coil in a spiral and measuring the inductance with and without the steel core?

Yes you are correct the architecture is different. Incidentally now that I am getting into the data sheet it is becoming clear that the design is for sensing a distance between the coil and a conductive plate. The coil induces eddy currents in the conductive plate and then senses the distance to the conductive plate by the counter EMF that eddy currents create in the coil. I believe that the inductance of the coil as measured by the LDC 1000 will be quite accurate as long as the conductive plate is used. When I started putting different metals through the center of the coil I was outside the design parameters of the device.

You suggested that I create a coil with a spiral architecture and rerun my tests on the cheap LCR meter. I have no doubt that a change in architecture will change the inductance and even if I can wind a spiral that starts at 3.3 uH it will not react to the same degree as the linear coil that I used. However the basic physics should still hold. The Brass and the Stainless with not change the open air value. The steel will have an effect to increase the inductance and the ferrite will have the greatest effect to increase the inductance. What results would you predict for this experiment?

I hope all is well in CA. Watching the News makes me wonder if I should return in January or not. I hope all is well with the family.

John

After some reading, this spiral inductor should have it's inductance increasing with a higher permeability core. It's gotta be something else. I started a new job in August, so now I'm living in Los Angeles (Panorama City).

Hi Michael,

Congratulations on the new job. Panorama City is only 25 miles from Torrance on the 405. If I decide to come out in January we could try to get together and discuss electronics if you have time this year.

John

I would look forward to that very much

I snapped off that resonant circuit and applied it to the VNA, this is the result I get (it could be inaccurate, the VNA is freshly built and could have issues. I'm still learning how to use it).

I've overlaid the results for the resonant circuit with it's air core, and then with ferrite placed close to it, and also with a (guessing it is steel) M4x6 pan head screw placed close to it.

All the time taking care not to move or change anything else.

It is very interesting. As expected, the ferrite increased the inductance, and the Q dropped too (but very slightly). We can see the inductance increased because the resonant frequency has reduced (from about 3.71MHz in air [peak of the red curve], to 3.56MHz with the ferrite [peak of the blue curve]).

With the steel screw however, the resonant frequency increased (to 3.84MHz [peak of the brown curve]). I can't explain it : (  But I'm no expert on inductors nor cores!

And again the Q decreased slightly (as expected - I think anything placed close to that fixed number of turns will decrease Q).

Hi Charles gervasi do you know what is going on? Or is this something that needs simulation to analyze? It looks like magnetic flux density dropped when the steel was close to the spiral (I didn't place it inside, just close to it).

The inductor in this circuit is a spiral trace on two sides of this PCB (LDC1000 eval board), the photo below shows what it looks like:

Basically, I did a reflection measurement, I'll post a photo of the testbed later today.

Thanks!

" We can see the inductance increased because the resonant frequency has reduced (from about 3.71MHz in air [peak of the red curve], to 3.56MHz with the ferrite [peak of the blue curve])."

I do not know what's happening, why the resonant frequency is going up with steel.  I've never experimented with this stuff.

My only thought is that the lower resonant frequency does not necessarily mean increased inductance.  I see increased inductance as appearing in the higher Z at lower frequencies.  My crude understanding is the inductance peaks at the resonant frequency because of the effects of the unwanted capacitance of the inductor beginning to predominate over the desired inductance. 

It's suprising to me, although I've never experimented with this, that the resonant frequencies are all clustered in the 3.5MHz range. 

I wonder if the issue is core losses acting as a short as in this diagram from Trilogy of Inductors:

Hi Charles,

Thanks for examining it!

Regarding the 3.5MHz range, sorry it may not have been clear, but there is a 100pF cap on that board, so the analysis was of the complete resonant circuit (LC tank), not just the coil.

I'm thinking, maybe we should also ask TI, since they had the tools to design the inductor in the first place.

I'll post the question on their forum, to see what they respond with.

Unrelated, I also found this low-cost set of metals (it is only six metals, but at least they are all the same size) http://shop.miniscience.com/navigation/detail.asp?id=METCYLSET6

that may be useful for experimenting with the LDC1000.