PGA450-Q1: PGA450Q1EVM-S wrong distance sensing

Hi GD,
Yes, you should be able to read back the FIFO data before optimizing the threshold times and levels. Reading back the FIFO will help you to set the EEPROM based threshold values. Based on the results shown in your spreadsheet, the long range measurement command appears to be working well for >70cm measurement. The blanking time is maximized in the hard-coded long range command, and prevents <70cm object detection. The short range measurement will suffice for the purpose of sub-70cm ranging.

I recommend that you re-purpose the EEPROM space to use command 5, which refers to the driver and receive setting set in EEPROM (which are copied to the ESFR when run). This will enable you to use a single command to cover both the short (30cm) and long (1.5m) range.

I suspect your inconsistent results are currently likely due to an un-optimized threshold configuration. I recommend that you continue adjusting the thresholds to check and improve performance. Configuring the thresholds is typically the most time consuming process of the PGA450 development since the values are based on the system's required range and object type (which vary from customer to customer).

8780.PGA450-Q1 THRESHOLD CALC 150cm.xlsxHi Akeem,

Thanks for the quick response.

I still have some questions for you:

1) from my understanding of the way distance reading works, the sensor reports the TOF of the first point in which the FIFO data crosses the threshold. Looking at my 150cm experiment, for example, the FIFO location should be around 175 (calculated as FIFO_location = TOF/downsample), but when I plot the data on the Excel file it seems that the FIFO location should be in between 111 and 116 for Long Range Reading.  This is my main source of confusion: apart from un-optimized threshold configuration, when the reading is correct then I should see the correct FIFO location in the plot. Am I missing something important in the way the sensor works? I attached a new spreadsheet with the 150cm data plotted in the graph. 

2) In addition, when I run the command 6 to read threshold values stored in the sensor, it returns 

      09 ff 20 84 40 74 60 54 80 4c a0 3c e0 34 19 1e

Starting from your THRESHOLD CALC for UART OTP EXAMPLE.xlsx, I expect thresholds equal to (hex values): 84 74 54 4C 3C 3C 34 1e. If I analyze the values returned by the sensor, I can find a correspondence: 09 ff 20 84 40 74 60 54 80 4c a0 3c e0 34 19 1e. I notice that a 3c is missing. Am I analyzing data in the correct way? Is problem 1 related to this fact? 

3) why did you advice the use of command 5 instead of command 1?

Thanks in advance,

GD

Hi GD,

1) Are you accounting for the blanking time offset? The full equation to convert the FIFO Time of Flight to Distance is:
Distance (m) = [BlankingTime(us) + (FIFOSampleNumber * Downsample * ADCSampleRate)] * SpeedOfSound/2

2) Command 6 reports the threshold values, and is decoded as follows using your return as an example:
(Note: there are seven threshold values. Command 6 identifies the range across the FIFO's 768 points where the threshold level changes. The level is an 8-bit value from 0-255.)
For example:
"09 FF 20 84 40 74 60 54 80 4C A0 3C E0 34 19 1E"
Byte1 = x09 = from FIFO 0 to 9, the threshold levelInit is Byte2 (255) (based on EEPROM addr 0x09 as initial ignore count defaulted to 9. Ignore count means threshold level is set to 255)
Byte2 = xFF = 255d level
Byte3 = x20 = from FIFO 10 to 32, the threshold level0 is Byte4 (132)
Byte4 = x84 = 132d level
Byte5 = x40 = from FIFO 33 to 64, the threshold level1 is Byte6 (116)
Byte6 = x74 = 64d level
Byte7 = x60 = from FIFO 65 to 96, the threshold level2 is Byte8 (84)
Byte8 = x54 = 84d level
Byte9 = x80 = from FIFO 97 to 128, the threshold level3 is Byte10 (76)
Byte10 = x4C = 76d level
Byte11 = xA0 = from FIFO 129 to 160, the threshold level4 is Byte12 (60)
Byte12 = x3C = 60d
Byte13 = xE0 = from FIFO 161 to 224, the threshold level5 and level6 is Byte14 (52) (in this example, level5 and level6 are the same values, lumped into a single return value, not the double interval of 224-160=64, which is twich the typical interval of 32)
Byte14 = x34 = 52d
Byte15 = x19 = from FIFO 225 to the end of the FIFO, the threshold level7 is Byte16 (30)
Byte16 = x1E = 30d
This explanation will be added to the user's guide in the next revision.

3) Command 5 is a fully configurable version of the burst/listen command. The EVM-S User's Guide ( www.ti.com/.../sldu019b.pdf ) Table 3. " EERPOM Register Map Values" shows how the EEPROM is re-purposed as an ESFR reserve. This allows you to update the EEPROM with the ESFR values for customizing the transducer frequency, burst strength, and receive filtering/coefficients on the fly. When command 5 is run, it first reads the EEPROM, loads these values into the actual ESFR register space, and then runs a burst/listen command. This fully configurable burst/listen profile allows you to optimize your performance based on a custom range, while commands 1 and 2 are fixed / hardcoded to only work at one frequency with permanent settings for short and long only modes respectively.

I should also note that the PGA450-Q1 is an older device. I highly recommend the PGA460-Q1 and its small form factor equivalents (Design files: www.ti.com/.../technicaldocuments ) for improved performance and reduced software complexity. We will be selling the PGA460 small form factor equivalent on the TI eStore before the end of this year. In the meanwhile, the PGA460 EVM is already available ( www.ti.com/.../boostxl-pga460 ) for you to begin evaluation.