Maximum number of samples from AD7730

[dmccric]

Hi All,

I am using the AD7730 in 24-bit, bi-polar mode and require to have more than 100samples/s.

At the minute I get around 60samples/s when my motor is not running but when I run the motor it can
go as low as 20 samples/s.

I have designed my PCB with the data sheet in mind, decent grounding, earthing, ect.

Has anyone managed to get several hundred samples/s?

 

[FvM]

It's completely unclear, how the ADC data rate should be influenced by "running a motor". I sounds like you are talking about an unsaid software problem or an unsuitable SPI interface method. With appropriate hard- and software, there should be no problem to achieve the data rates specified in the AD7730 datasheet.

Please clarify about the involved hard- and software.

 

[dmccric]

I have a timer running that measures and displays the number of pulses in a second, this changes when I press the button to enable and run the motor.
The AD7730 is running with at 4.1952MHz (as specified) and the SPI bus is running at 1.125MHz connected directly to my Microcontroller running at 3V3.
I have attached a PIC of the schematic.

View attachment Load Cell Schematic.bmp

As you can see from the schematic I have seperate grounds for analogue and digital supplies and earthing on the load cell connector to reduce noise.
The motor is a 400W AC servo motor controlled through RS232 commands.

Any more suggestions are much appreciated.

Thank you

 

[FvM]

I have a timer running that measures and displays the number of pulses in a second

Which pulses? At the #RDY output? What's the programmed ADC word rate?

Unfortunately, I'm unable to download a readable schematic, it seems to be broken somehow. Can you provide a usual compressed image format instead of *.bmp?

 

[dmccric]

Hi,

Sorry the number of samples/s, I wait for the /RDY line to go LOW, read my 3 bytes for 24-bits and then wait for the /RDY to go HIGH and then repeat this again.
The Filter Word is set for 400Hz with Fast Mode and Chop Mode both enabled which should give me a conversion time of 7.5ms which should give me about 133 samples/s.

I have also tried this for 600Hz but still get the same number of samples.

I have included a pdf of the schematic.

 

[FvM]

Sounds like a software problem where the processor can't keep track with the output data rate. If you suspect any hardware issues, you should check the signals with an oscilloscope.

 

[dmccric]

That is not the case as I am running a 72MHz processor and have disabled all code apart from reading the load cell and I get the same results.

 

[FvM]

Fine. How about the #RDY waveform?