hello, I have some questions about LDC1612 that I need to ask.
I am using ldc1612 with external inductance, read the value of the recount register after booting, this value has been in an upward state until twenty-thirty minutes before it is stabilized near a value.
why is it ?
Hello,
As a point of clarification, you mentioned that your sensor has the following characteristics:
L = 386µH, C = 22pF
This puts the expected frequency around 1.7MHz, however you are reading DATA_MSB_CH0=0011, DATA_LSB_CH0=CD32 for CH0 and DATA_MSB_CH1=0014, DATA_LSB_CH1=1D19; for CH1 which calculates to about 0.173MHz and 0.196MHz respectively (assuming CLKIN = 40MHz, CHx_FIN_DIVIDER = 0x01, CHx_FREF_DIVIDER = 0x001). Therefore it's not clear if these values are proper or if there are divider settings being used. Note to convert to frequency you can follow equation 8 from the LDC1612 datasheet.
The other thing to point out is that the discrete capacitor is only 22pF which is relatively small given that the pin parasitic typical capacitance is around 4pF. If a C0G capacitor is being used then it provides relatively good temperature coefficient and stable oscillation. However, parasitic capacitances are not as stable. For example, assuming that you have an inductance of 386µH and total capacitance of 26pF including parasitic to give a nominal oscillation of 1.5887MHz. Then a 0.5pF shift decrease of parasitic capacitance increase the frequency 1.604MHz or a 1% frequency shift. By increasing the nominal capacitance to 100pF you can reduce the overall impact of the 0.5pF parasitic shift to 0.25% frequency shift.
Now, examining the data it looks like the CH0 sensor is drifting about 700Hz and CH1 is drifting about 100Hz.
Given that the data convers to the following frequencies: Ch0=0.173MHz and CH1= 0.196MHz, we can convert this plot to % frequency shift.
We can see that the drift is on the order of 0.5% which means you could be seeing as little as 0.2pF parasitic shift in the sensor capacitance which sounds like it could be the culprit here. Note that some other sources of parasitic capacitance can come from the wiring that connects the inductor to the LDC pins and windings of the inductor itself which creates inner winding capacitance.
I would first recommend to confirm your sensor parameters, including the dividers to eliminate the discrepancy between the calculated frequency and the measured. Then you could try to increase the sensor capacitor to help mitigate parasitic effects.
Regards,
Luke LaPointe
Hello,
The sensor frequency is determined by both the inductance value of your circuit and capacitance (whatever capacitor value you're connecting). You can then compute the sensor frequency as I mentioned above using equation 9 in the LDC1612 datasheet.
For example, if you have L = 386µH, C = 22pF, then we expect the sensor to oscillate around 1.7MHz.
To convert this frequency to what the LDC1612 DATA reads you can use equation 8 from the LDC1612 datasheet:
Notice that this is also a function of the divider settings and the reference clock. For example, assuming CLKIN = 40MHz, CHx_FIN_DIVIDER = 0x01, CHx_FREF_DIVIDER = 0x001, We would expect the LDC to read 11408507 in decimal or equivalently AE147A in hex.
The other parameters you mentioned such as RCOUNT or SETTLECOUNT don't have an impact on this calculation, they merely effect the conversion time and the sample-to-sample noise in the data reading. A higher RCOUNT value yields a slower but less noisy measurement.
Regards,
Luke LaPointe
