• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 30 subscribers
  • Views 1389 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LDC1000: problem with data ready bit and SPI data

Oleg King
Prodigy 90 points

Part Number: LDC1000

Hello!

I'm trying to implement inductive sensor using LDC1000, it is connected to STM32 MCU, gets 8MHz clock from MCO output (3.3v) of MCU and connected to its SPI. I'm using 3.8MHz frequency in LC.

I'm using middle part of LDC1000EVM (with LDC1000 itself), that connected to my own coil by one side, and with short (~5cm) wires to Olimex STM32-H405 board - by other side.

Start sequence:

0x00 -> 0x0B // power down

0x00, 0x3F, 220, (CONFIG_AMP_4_VOLT | CONFIG_RESP_TIME_6144), 0x01, 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x01 -> 0x01..0x0B in one sequence // configure and power up

0x04 -> 0x0A // set INTB to DRDY

After some hard time spent finding errors in datasheet, I can finally program it, can read values, it reacts to metal parts and I can see all this with JScope (J-link tool, that can show graphs based on real-time values from MCU) and in debug window of my IDE.

I have tried to increase polling frequency and met limit at ~1.8kHz, exactly matching formula from datasheet. At 1.9kHz I see approximately each second read from LDC1000 marked ad "Data Not Ready", what is absolutely expected behaviour.

But when I have decreased polling frequency, I met a problem - from time to time I got 0x70 in status register @ 0x20 (instead of 0x30 as usual). Sometimes proximity data also can be read as 0x0000 or as its maximum value. This happens approximately 1-5 times per second, I can not find any correlation with anything.

I tried to use slowest SPI speed, insert delays between changing CS and actual exchange, make polls even less frequently, up to 100Hz, connected 8MHz clock with shortest wire with 22 Ohm resistor in series - and this does not changed anything. Still errors (but between these errors I still can see that LDC1000 is working, proximity data is changing). МСС is 5 volts, VDDIO is 3.3 volts.

Logic analyzer connected to MCU side or LDC1000 side shows me same results - 0x70 from LDC1000 instead of 0x30, so this is not MCU SPI issue.

Can you give me some advice with this problem?

Thank you in advance! :( 

p.s. I have tried to solder extra wires to all SPI pads and tried to touch them by hand or place them in different order - and frequency of errors was increased depending on position, from common 1-5 errors per second up to ~50-80. But I cannot understand how this additionally wires can be a reason for LDC1000 "Data Not Ready" bit - I have tried to lower SPI speed by 16 times - and this not changed anything. After I have desoldered that wires, everything was returned to its "stable" state - 1-5 errors per second.

p.p.s. I have tried to turn on DRDY signal on INTB pin and found something strange.. you can see it here: 

p.p.p.s. I made test pin to go high when error happened (status byte read as 0x30) and tried to catch this error. Experiments showed that it returns 0x70 instead of 0x30 when DRDY pin pulses high (conversion finished) at the end of first byte of read command (0x20, read from 0x20). I have attached a file when you can see that DRDY pin goes high at end of first command byte just before status byte will be returned. BTW, before this moment 5 DRDY pulses was already (without any read from device), so I caught 6th conversion end here.

  • 0
    TI__Intellectual 2725 points
    Hi Oleg,

    Would you be able to provide screenshots of the INX pins to ground? What are the coil characteristics.

    If you are trying to implement an application other than metal identification the LDC1101 or LDC1x1x devices maybe a better device choice. If you can comment on your target application I can help you guide towards the correct part.

    Regards,

    Varn.
  • 0 in reply to Varn Khanna
    Prodigy 90 points
    I will try to make screenshots today at evening.

    LDC1101's and EVM's was arrived yesterday, I will try them, but as I understand, it still cannot be synchronized between two chips with different probe frequencies.

    I can hope just that this is a bug in chip design that was removed in next generation (i.e. 1101). I will try this in two days and will post results here.

    I'm using de-synchronized approach now:
    - get info from two LDC1000 just after conversion was finished (on INTB) into memory
    - get current values from memory each X ms

    This is not-so-good approach (i wish I can start conversion manually each time I need it), but it works at least.
  • 0 in reply to Oleg King
    Prodigy 90 points

    I have tried to get screenshots (I have two-rays oscilloscope only).

    Here you can see everything at all:

    A little bit closer:

    More closer:

    With SPI clocks instead of coil:

    Each time I catch this situation, this happens when I'm trying to get data immediately before conversion was completed, so 8th clock of my read command fits with DRDY. I always getting 0x70 from status register in this case.

    p.s. Just tested LDC1101, it has no such errors, just reports data with status register 0x2C (no 'data not ready' cases was caught).