LDC1101: Using LDC for high speed precise distance sensor

Hi Michael,

First of all, thank you for doing so much research and providing so much information! If you don't mind, can you share how you found the collateral you used? We're always looking to improve our collateral content and organization.

I've copy-pasted your questions and answered them below:

1. What is optimum sensor diameter and other parameters? Should I used wire wound coil, should be there ferrite core? I can fit even diameter of 20mm if needed.

Unless your target is smaller than your sensor, the optimum sensor diameter is as large as you have space. I recommend using a PCB coil, which have much tighter inductance tolerances than wire wound coils. Increasing your coil diameter would increase your sensitivity.

2. Inductance should be high or it does not really matter? I understand it that higher the change in frequency higher resolution I get. Also it is better to have lower LC sensor frequency because I can fit more cycles of reference clock in it and it makes measurement more precise. 

From a resolution standpoint, there is an optimum range of sensor frequencies (400kHz to 6MHz). That's discussed in more detail in this application note. It is possible to have a sensor frequency that's low enough that the limited number of sensor oscillations reduces the resolution. The inductance primarily matters because of its relationship to the sensor frequency. Within the optimal frequency range, the inductance does not matter much. However, higher inductances will reduce the power consumption of the system.

3. Does parallel sensor coil configuration have any positive effect on LHR measurement? 

Yes, we recommend this configuration for both L measurements and Rp measurements. This reduces the coil's Rs, which increases the Q value.

4. One of the limitation is f_CLKIN which is now 12MHz. It can be changed up to 16MHz which should decrease conversion time by 1/4 with keeping same performance. Is default reference clock precise enough or is it only for evaluation purposes? Precision of reference clock is also one of the main component which can influence overall performance. 

Actually, I recommend using an external chip oscillator with defined jitter and temperature drift specifications. This could make a big difference in your resolution. The LDC1614EVM uses an external oscillator that may work for you, though you'd need to order it in 16MHz instead of 40MHz.

5. I read manual snoa941a, 3 heading: In addition, the LDC1101’s LHR (High-resolution inductance mode) measurements have an effective reference frequency of 32 MHz (when the external reference frequency is set to the maximum supported 16MHz). Does the LDC1101 support even higher reference frequency then 16MHz? 

No, the maximum supported reference frequency is 16MHz. 

6. Unfortunately at this configuration I did not get performance I need. Repeatablity rapidly decreases with increase of distance and sample rate which I understand but is there still some way how it could be improved? Or am I on the edge of performance possibilities? Would custom made sensor coil improve overall performance? 

Increasing the overall diameter of the coil and using a difference reference clock could both improve your resolution. However, it will be challenging to achieve the resolution you need at a high sample rate. Why do you need a 20kHz sample rate?

Best Regards,

Hi Michael,

Thanks for the feedback. If you have time to upload any suggestions for improvement we'd welcome them. If you haven't seen it yet, we have the inductive sensing FAQ (linked in my signature) that has some additional tools that should be helpful once you start designing a coil.

In answer to your questions:

1. Would the sensor be more sensitive to change of position if I have a ferrite cup core behind the the PCB coil? Or should be there at least some shielding around?

You could add a layer of ferrite backing, which would both increase the dynamic range and provide shielding between the coil and any metal on the other side of the ferrite. We have a blog post about that here: https://e2e.ti.com/blogs_/b/analogwire/archive/2014/10/08/inductive-sensing-how-to-shield-from-metal-interference

2. This means to have oscillation frequency 990.0 kHz → 1.0 MHz but it outputs maximum unique codes for LDC161X subfamily. Is it optimum to LDC1101 as well?

Yes, this is optimal for the LDC1101 as well. 

Ideally the sensor should be tuned to have in minimum distance (0.2 mm) 6MHz and maximum distance (2 mm) 400kHz. Or just to be "somewhere" in 400kHz to 6MHz range?

The sensor frequency just needs to be within those bounds. It's very unlikely that your sensor frequency would shift as much as 5.6MHz. Also note that the minimum sensor frequency that the LDC1101 can drive is 500kHz, so there's a slightly higher minimum frequency for this range.

4. Are there any limitation in connecting external oscillator or I can just cut through pins in breakaway section and connect external oscillator? Will all other functionality as GUI work?

As long as the LDC1101 has a valid CLKIN input the EVM should still work. You definitely don't want to cut the CLDO trace shown in your diagram though. That capacitor is needed for the internal LDO; the chip won't operate correctly without it. If you can clip just the CLKIN connection between the MSP430 section and the LDC1101 section of the board, then you can just connect your external oscillator. You may need to break apart the sections and reconnect the SPI and power connections.

Do you have any recommendation for already set up external oscillator? I did not find any unit in LDC1614EVM which I can use. Does TI offer such a device? Or should I just use oscilloscope as reference clock generator for lab tests? 

We don't make a suitable oscillator, but I do recommend using the jitter and drift specs of the CTS oscillator in the LDC1614EVM to guide your oscillator selection. You could use a function generator for lab tests with the EVM prototype, but you should expect your custom boards to behave differently when you use a chip oscillator. 

Best Regards,