LDC1101: Compensating for distance variation.

Ali Goname

Part Number: LDC1101

Hi,

Background information: I would like to determine the dopant density (concentration of boron or phosphorous) in a silicon ingot. Due to chemistry I do not fully understand, when a block of it is cast, more non-metals at the top than at the bottom. The change in composition will have a significant and measurable impact on the resistivity according to literature I've read and some preliminary tests.

My proposed set-up: I will have a coil facing down at a fixed height. Below it, I will have the block on a motorized platform. I will laid the block on its side, such that the variation in composition, and thereby in resistivity, is now on the x axis. As the platform moves left or right, I will be able to observe a change in Rp.

My problem: I believe in order for this to work with any accuracy, the distance between the block and the coil has to be uniform. As the blocks are not perfectly flat, there can be up to 1.5mm of variation in the distance between the block and the coil.

My questions:

Silicon, boron and phosphorous all have a relative permeability of 1. I believe that would mean the inductance measured at a certain distance to the ingot will be the same regardless of the composition. Is that correct?
If so, is the Rp vs distance curve always the same?

I am hoping that if the answer to question 1 is yes, that I can create an instance lookup table that will give me the absolute distance from the sensor, regardless of composition.

If the Rp vs distance curve is always the same, I hope then to compensate for any distance changes, as deduced from the inductance.

Does this seem like a valid idea?

over 7 years ago

0 Kristin Jones93 over 7 years ago

TI__Genius 17440 points

Hi Ali,

The basic principle of inductive sensing is that when a conductive target is exposed to an AC magnetic field from an inductor, eddy currents will form on the conductor's surface, creating a magnetic field that opposes the inductor's magnetic field and reduces its inductance. The strength of the response is heavily based on the conductivity of the metal. The easier it is for the eddy currents to flow, the stronger the opposition to the inductor's magnetic field. This means that if the different regions of your ingot have different conductivities, you will see different responses in the inductor. This means that the Rp vs distance curve would also vary. It may be impossible to differentiate between changes in conductivity along the x-axis and changes in distance between the ingot and the sensor coil.

It's possible for this setup to work if you have a very level ingot, but I don't think you will be able to use the inductance measurements to compensate for changes in target distance. In addition, you may not see strong responses from the ingot, because the doped silicon is likely much less conductive than the targets we usually recommend (copper or aluminum). I would suggest operating at high frequencies to minimize the skin depth as much as possible.

Regards,

0 Ali Goname over 7 years ago in reply to Kristin Jones93

Prodigy 70 points

Hi Kristin,

Thank you for your reply - Makes perfect sense. I had been thinking of the mutual inductance only. I have a few other questions I hope you can help me with.

1. What steps, if any, can I take to minimize the effect of the change in distance on the Rp value?

2. If the effect of the distance change on Rp is proportional to the change in distance as a percentage of the coil diameter, would it be true to conclude that a larger coil will be more tolerant to changes in distance?

The graph below shows the resistivity and calculated skin depth for different frequencies vs the expected dopant density range.

The LDC Target Design document states:

“With a conductor of 3 skin depths thick, 95% of the total current will be induced…To achieve best performance, a good rule of thumb is to that the target thickness should be at least 2 to 3 skin depths.”

For the dopant density range I would like to analyse, a minimum frequency of 2MHz is required so that the target material depth (150mm) is greater than 3 times the skin depth.

I have created the following graph that shows the LC tank design restraints.

3. Are there any design constraints I have missed?

4. From this window, how should I proceed in choosing an LC combination?

P.S Does clicking the "this has resolved my issue" button lock this thread? I would like to give credit where it is due for your earlier reply which fully addressed my question, but don't want the thread locked before getting an answer to the questions above.

Thank you again,

Ali Goname

0 Kristin Jones93 over 7 years ago in reply to Ali Goname

TI__Genius 17440 points

Hi Ali,

1. The distance between the coil and the target is one of the primary variables that changes the coil's Rp. There is no way to minimize the effect of the change in distance on the Rp value without diminishing your sensitivity so much that the measurements are no longer useful.

2. Actually, a larger coil would improve the measurement resolution. I recommend reading section 2.4 of LDC Sensor Design, which discusses both this topic and the previous question to some extent.

3/4. This is a very helpful graph. The only other design constraints I would consider are the sensor frequency and the inductor Rp (or Q) value when the target is present. The conductive target will decrease the inductor's Rp and increase the sensor frequency. I would recommend choosing LC values that give you a sensor frequency of about 5MHz, then test to make sure that the sensor characteristics are within the recommended operating conditions when the target is the minimum distance from the sensor coil.

Finally, clicking the green button will not lock the thread. It will only lock if it has been inactive for 30 days.

Regards,

Data converters

Data converters forum

LDC1101: Compensating for distance variation.