LDC1101: L-Only Measurement

Hi John,

Can the stability improvement be quantified in any way?  Is the device using a current drive strength that is somehow superior for accuracy?  Would that be a higher or lower current?

If you plug in a LDC1101 EVM on your end, can you demonstrate any change in the behavior when you apply this setting?

On my end I tried to take a waveform measurement on the LDC1101 EVM after applying this setting, and it did not match my expectation.  I saw the decaying waveform - that made me think the device might still be measuring Rp, despite my attempts to disable the Rp measurement.

Arthur,

the questions so far cover a pretty broad area.

If we narrow the topic, we will most likely get the answers you are looking for more quickly.

First things first: What behavior and performance are you looking for?
What are your end goals?

Regards,
John

1.  Our application only uses LHR measurements, yet we have not disabled Rp measurements.  It would be nice to get improved sensing performance (metal proximity sensing) with only a software upgrade.  So far on the bench with the EVM, I have been unable to prove any sensing performance gains by changing the setting (such as standard deviation of the measurement with a static target present, for example).  What aspects of the performance are actually optimized when we "Optimize for L Measurements"?

2.  I am curious whether the Register Configuration section of the LDC1101 EVM GUI actually works. - I am able to write my changes to ALT_CONFIG and D_CONFIG and read them back successfully. But when I click back to the LHR tab I suspect that the sensor is re-initializing to some baseline mode where Rp measurements are again enabled, ignoring the changes I tried to make.

Arthur,

I will try to reproduce you findings with the GUI and will update this thread by COB on Thursday.

Regards,
John

Arthur,

Could you share a few more details about what you are seeing w/respect to item #1?
1.1 Are you using the EVM on-board sensor coil or an external one?
1.2 What are you using for a target and what are its dimensions?
1.3 What is the distance between the target and the sense coil in your measurement?
1.4 Are there other metrics - in addition to the measurement standard deviation - that are important for your application?

2. Could you share the register settings you mentioned in item #2?

There are some app notes that may be helpful:

1. Setting LDC1312/4, LDC1612/4, and LDC1101 Sensor Drive Configuration - the author provides guidance on amplitude settings and why automatic amplitude settings may not be best (sections 4 & 9).

2. Sensor Design for Inductive Sensing Applications Using LDC - a must-read if you are new to inductive sensing.
It talks about the practical aspects of layout, PCB parameters, parasitic capacitance, and their impact on a sensor design. Somewhere in there the author mentions the target-to-coil distance should be kept to 3% to 20% of the coil diameter for the best accuracy. Separation distances of up to 50% can be supported, but at reduced accuracy. 

3. Configuring Inductive-to-Digital-Converters for Parallel Resistance (RP) Variation in L-C Tank Sensors - it doesn't explicitly address the LDC1101, but it gives a great overview of the Rp impact on inductive sensors.

4. Optimizing L Measurement Resolution for the LDC161x and LDC1101 - it can be challenging to apply the contents of this paper to a design, but it does provide some insights on the factors that contribute to inductive position-sensing resolution.

Much more info is available at the [FAQ] Inductive sensing FAQ - Frequently asked questions.

regards,
John

Hi John,

1.1. on-board EVM sensor coil

1.2. A steel disc 32mm in diameter, 1mm thick.

1.3. 10mm

1.4.  Fast measurement speed is important, as we prefer to save power by taking one ~1msec measurement every 1 second.  If the measurement takes longer, it increases power consumption.

2.  The register settings are the same mentioned in the datasheet 9.1.11: 

ALT_CONFIG.LOPTIMAL(register 0x05-bit0) = 1 

D_CONFIG.DOK_REPORT (register 0x0C-bit0) = 1

Your app note #1, Section 9, referred me to that section in the datasheet.

So I am trying to ask about that #1 app note you linked, should I follow its advice and disable the automatic amplitude control?

Hi John, thank you for sharing the additional design resources.

My question from above in the thread:

 I am curious whether the Register Configuration section of the LDC1101 EVM GUI actually works. - I am able to write my changes to ALT_CONFIG and D_CONFIG and read them back successfully. But when I click back to the LHR tab I suspect that the sensor is re-initializing to some baseline mode where Rp measurements are again enabled, ignoring the changes I tried to make.

You replied that you would try to reproduce my findings with the GUI - did you find an answer?

John,

Thank you!  It's good to know it's not just me having the problem.

I am afraid it may not be realistic for me to expect updates to this old application.  Is there anyone supporting the app who can let us both know if there's a workaround for this issue?

Thanks again,

Arthur

John,

I should be able to write a script to manually perform register reads/writes over the COM port.  I assume even if the GUI is buggy, I can change the settings that way.

If I'm able to make the change, would I see a difference on the voltage waveform I'd previously measured?  Should the amplitude remain constant instead of decaying like this?  I can't find anywhere where it's clearly spelled out what is supposed to be happening.

Thanks,

Arthur

Hi John, I am concerned if I set my device to have RPmin = RPmax, it wouldn't be stable over all target distances.  I will not always have the target present at 10mm distance, sometimes the target will be missing.  So the LDC1101 may fail to start the oscillation in some cases and be unable to provide a measurement.

I still want to try the serial port method of setting the ALT_CONFIG and D_CONFIG LSB to 1, and I'll let you know if that works.

Steps below to try the disable Rp setting:

Follow the 3.15 EVM FW Protocol in the EVM Users guide https://www.ti.com/lit/ug/snou137/snou137.pdf
Use Coolterm to send serial strings:

020B01 - stop conversions, Start Config. the Scope waveform disappears
020501 - write register 0x05 to 1
020C01 - write register 0x0C to 1
020B00 - write register 0x0B to 0. This is START_CONFIG, when set to 0 it should start active conversions

The scope waveform now has a constant amplitude. The constant amplitude is too high. Per https://www.ti.com/lit/an/snoa950/snoa950.pdf, it should be between 1.0 and 1.25V peak (2 to 2.5V P-P).

020B01 - stop conversions, Start Config
020105 - Try setting RPMIN to 0x04, 3k ohms.
020B00 - oscilloscope waveform appears. Peak-Peak is ~1.

now close coolterm and launch ldc1101EVM GUI. The sensor gets reconfigured as soon as I launch the program without clicking any other buttons. So it won’t be possible to use the EVM GUI to gather measurements with these Rp disabled settings.

Arthur,

With regards to a ferrite shield, one of my colleagues tested an LDC1614 and an LDC3114 with ferrite shield.

Subjecting the ferrite to a fixed magnet made no observable difference in the behavior or performance of either device.

Regards,
John