Ti LDC1000EVM A Question on Inductance

Question

Inspired by shabaz I purchased a TI LDC1000EVMLDC1000EVM Evaluation Kit and set it up If you want more information on the kit itself check out Shabaz's blog as it is very interesting Getting Started with the LDC1000 - Small Metal Detection!

After setting up the board I wanted to confirm some of the things I thought I understood about inductors and so I performed a couple of experiments and my results leave me with a question that I am hoping you can help me with.

My first experiment was to look at the inductor in free air.

The GUI for the Evaluation board was reading 19.496 uH with an open air core.

For my next experiment I inserted a small ferrite core into the inductor.

My expectation was that this ferrite core would increase the inductance of the coil. I was not disappointed as the inductance rose to 30 uH.

My next experiment was to replace the ferrite core with a soft steel rod of similar length to the ferrite one.

To my surprise and confusion the inductance of the coil dropped to 16.811 uH which was almost 3 uH lower than open air. I am hoping that someone can explain this behavior to me. I did not expect the steel core to do as well as the ferrite but I did expect it to do better than open air.

I have my physics book open right now and I am reviewing inductors to see if I can better understand why a steel core would lower the inductance of a coil instead of raising it.

John

shabaz · Accepted Answer

Hi John and Everyone, Chris Oberhauser from TI replied : 
 At the higher frequencies (>500kHz) that LDC devices typically operate, the eddy currents on the target surface overwhelm the permeability and hence the inductance drops. If you change the sensor to one that operates at a much lower frequency (~40kHz), then you will see an increase in inductance when the steel is brought into the field. There is a transition region where you will see a minimal change in inductance due to a steel target position; this frequency range will vary based on the type of steel. 
 
 The Rp measurement will still show a shift due to target presence.

So, our initial hypothesis that inductance ought to go up was true, but only at lower frequencies in the case of steel. According to Chris, at higher frequencies the presence of steel causes sufficient eddy currents that will have a greater effect than permeability. The eddy currents cause a magnetic field which opposes the field from the inductor, and this reduction in flux density results in a reduced inductive reactance. Hence we see the resonant frequency of the LC circuit to go up. Some sources ( such as this one http://https//www.nde-ed.org/EducationResources/CommunityCollege/EddyCurrents/Instrumentation/impedanceplane.htm) state that for steel, permeability has a greater effect than eddy currents, but I guess that the information on those sites is presuming a lower frequency measurement which is not the case for us. There is a great 'Eddy Current Applications' document here (PDF) . (Perhaps the book is aimed specifically at eddy current testing at frequencies relevant to industry convention for aeronautics). Figure 8-47 shows some graphs at a fixed frequency , where they have a coil, and they show on the X and Y axis the impedance (resistance on x-axis, inductive reactance on y-axis) when the coil is in free air (marked the 'Air Point'), and when a material is brought closer to it. The arrow shows the path the impedance takes (all the time at that fixed frequency) as item is brought closer. Diagram (c) in that figure is the scenario for ferrite, which is as we expect (bringing ferrite closer will increase the inductance). Diagram (b) is what we thought should occur with steel, and apparently it does at lower frequency. Diagram (a) is something like copper ( - and from what we're seeing is also the behavior of steel at the higher frequency). So, your question and experiment appears to have an answer which is inconsistent with some books/resources because they assume a lower frequency for test (perhaps because they are related to particular traditional industry tests/equipment), compared to the higher frequency that the LDC1000 is capable of!

shabaz · Answer

Hi John, Very interesting results! It is impressive that the LDC1000 circuit can reveal so much detail about inductance and the effect of cores. However, interpreting this is hard I think.. : ( I also think the same, the steel has caused the inductor to become lossy, i.e. it now has significant resistance at operating frequency. According to the LDC1000 datasheet, it computes inductance by observing resonant frequency. With the lossy inductor, a higher current is needed to maintain a particular amplitude oscillation, and this can be observed by calculating Rp which is the parallel resistance of the resonant circuit. There is a formula on page 12 of the datasheet which shows how to see the Rp value, I put that formula in Excel, and this was the output (x-axis is sample number, y-axis is calculated Rp in kohms) - the value dropped from 11kohms to just over 7kohms. The graph above initially shows nothing close to the coil. As the steel was brought close to the coil, the Rp dropped. I then removed it, hence the graph went back up to 11kohms. The graph below shows the result with ferrite, where the only relevant bits are at the beginning and at the end of the graph. At the beginning, there was nothing near the core. At the middle, I was busy securing it to the coil using tape. The end of the graph shows the actual result, and it is at a higher value (about 12kohms) than at the beginning (11kohms). I only used a tiny piece of ferrite (ferrite type 61 - there are various mixes of ferrite possible). Another way we could dig deeper is by subjecting that coil to a vector network analyzer (VNA), which would basically show us the equivalent circuit of the inductor with the steel or ferrite. I do have a (home-made) VNA but it needs calibration (and some practice) to be useful, and I still need to do that : (

michaelkellett · Answer

I suspect that the problem is that the steel core is much more lossy than the ferrite at the operating frequency of the chip, or possibly that the inductance has increased too much and is out of range. What happens if the steel slug is close to the coil but not inside it ? MK

Ti LDC1000EVM A Question on Inductance

Top Replies