Using the ADS7884EVM

Hi Chris,

Thanks for your quick response. Here is some of the information we highlighted from the datasheet:

"The serial interface for each ADC is controlled by two signals: chip select and the serial shift clock"

"The EVM provides users the option to power the ADC from either the analog supply voltage applied at TP4 or from an onboard 2.5-V reference chip"

"The analog input circuit is optimized for a 100-kHz input signal; therefore, users may need to adjust the resistor and capacitor values to accommodate higher frequencies and different test systems"

"The analog, power, and digital connectors can be made to plug into a standard 0.1-inch breadboard or ribbon cables to interface directly to FPGAs or processors."

The processor we have been working with is the ATMEGA328P from the arduino uno board, unfortunately, its built-in A/D (the arduino takes care of the clocking) only samples at about 62ksps, that is why we decided to use an external A/D (ads7884) that samples at least 2.5MSPS to sample a 1MHz input signal.  We are trying to figure out how to program the atmega328p through the arduino's own IDE so that two pins on the arduino serve as clock inputs to the CS' and SCLK pins of the ads7884, although we are not sure if the arduino is capable of doing this.

Could you suggest a microprocessor or FPGA that we could try instead? and also, can you please give us more insight on "the newer EVMs interface with ADCPro, a Window's-based software program to immediately show you the data the part" that TI offers?

Thanks!

Pavel

Hi Pavel,

It looks like your board has an SPI peripheral. I came across this document which may help you with SPI Setup:  www.atmel.com/Images/doc2585.pdf

I looked into other ADCs that you could use with ADCPro, however, there are only a few ADCs in the Precision Data Converter category that have sampling rates of 2.5MSPS or greater and all of them used the 5-6k interface card with DSK kit, so they will not run with ADCPro.

I would use your current processor since you already have it and are probably most familiar with it.

Regards,
Chris

Hello Chris,

Sorry for my late reply, I read about interfacing the ADS7884 and the ATMEGA328 on the Arduino via SPI but, as we understand, the arduino's speed seems to be very slow to read 2.5MSPS, so we decided to use a 8-bit shift register to latch the ADS7884 outputs to make the arduino read them at its slower speed, "It takes about 100 microseconds (0.0001 s) to read an analog input, so the maximum reading rate is about 10,000 times a second", which means the arduino's reading rate it's 200 times slower than the ADS7884 output. We think using shift registers such as the 74HC595 to latch the readings momentarily so that the arduino is able to read all the samples.

Is there a TI shift register that you might recommend or an alternative component that you think could be useful to reach our goal? Any hints would be appreciated. Thanks!

Best Regards,

Pavel

Hi Pavel,

I don't understand how you plan on implementing a shift register; you should not need one. The ADS7884 is an SPI slave device which requires a fast SPI master clock to output data. To get 2.5MSPS throughput from the ADS7884 you will need to apply a clock with a frequency of at least 35MHz. The ADS7884 will accept clock frequencies as high as 40MHz (VDD = 3V) or 48MHz (VDD = 5V). If you use a slower clock then you will not be able to get the 2.5MSPS data throughput of the ADS7884 that you want.

What is the fastest SPI clock you can generate with the ATMEGA328 Arduino microcontroller? If it is limited to 10 MHz then you may only get about 500kSPS throughput. 

Regards,
Chris

Hello Chris,

We came across this article on the Arduino SPI Control Register (SPCR), that can be found here http://arduino.cc/en/Tutorial/SPIEEPROM

It says that "SPR1 and SPR0 - Sets the SPI speed, 00 is fastest (4MHz) 11 is slowest (250KHz)", so 4MHz being much more lower than 40MHz, we concluded that SPI will not work between the ADS7884 and the Arduino to effectively collect all the samples (are we correct?)  This is why we think we cannot use SPI and need to just read the samples directly from the Arduino analog pin.

Best regards,

Pavel

Hi Pavel,

Yeah, you will not be able to get the throughput and sampling rate you want with this part combination.

What about looking into an ADC that uses a parallel interface, such as the ADS7881? Parallel interface ADC's typically have internal clocks, so you would not need to generate it with the microcontroller. The draw-back of a parallel interface is that you will need to use a lot more pins on your microcontroller to read in the data. I don't know how many pins are available to you but if you have enough this may be a possible solution.

Regards,
Chris

Hi Chris,

Hi Pavel,

Our breakout board, (we call the evaluation modules), for the ADS8411 was made obsolete. I would recommend going with the ADS7881 because it has a faster sampling rate that will meet your 2.5MSPS requirement, it has an internal clock, and an evaluation module (EVM) is available for this part.

Check out the ADS7881 device page at: http://www.ti.com/product/ads7881

and the ADS7881EVM page at: http://www.ti.com/tool/ads7881evm

Regards,
Chris

Hi Chris,

We'll go ahead and check out these boards and documentation.

In regards to our application, we need to be able construct a portable unit to sample a sine wave up to 1MHz, store the samples in digital storage of our choice (we had chosen a SD card that works effortlessly with the arduino...) and later reproduce the original signal from the sample data, either on an oscilloscope or on a computer.  Since we are basically at square 1 again, we don't know yet the kind of storage we'll be able to interface with the new microcontroller.

Thanks.

Pavel