LMH6517: SPI problem

JPL

Part Number: LMH6517

Hello!

I am trying to program the gain of the LMH6517 via SPI using a beagleboard, but that does not seem to work. I already checked the communication lines on the oscilloscope and everything seem fine, so I would like to know if anyone can help identify what the problem is.There is some fixed gain there because I get an amplified signal out of it, but it does not change. The latch pin should not matter when spi is in use so I wonder what could be happening.

This would be the python code:

from Adafruit_BBIO.SPI import SPI

spi = SPI(1,0) #configure device and bus at beagleboard

spi.msh = 10500000 #clock
spi.bpw = 16 #bits per word

spi.open(1,0) #opens it

spi.xfer2([128,0]) #transmits data

Thank you,

JPL

over 5 years ago

0 Sean Cashin over 5 years ago

TI__Genius 14166 points

Hi JPL,

What is the voltage on the ENA and ENB pins? If they are grounded instead of floating, it would explain the behavior you describe.

Best regards,

Sean

0 JPL over 5 years ago in reply to Sean Cashin

Prodigy 90 points

Hi Sean,

The configuration pins are set to:

ENA = +3.3V

ENB = +3.3V

LATA = GND

LATB = GND

MOD0 = GND

MOD1 = +3.3V

Best regards,

JPL

0 Sean Cashin over 5 years ago in reply to JPL

TI__Genius 14166 points

Does either parallel or pulse mode work for this device? Also, is the device powered with 5V?

0 JPL over 5 years ago in reply to Sean Cashin

Prodigy 90 points

The supply is +5V on the designed board and I also tested it on a modified evaluation board LMH6517EVAL-R1 which works for pulse/parallel. There it didn't work either.

0 Sean Cashin over 5 years ago in reply to JPL

TI__Genius 14166 points

Did you try a different device to see if a damaged part is the issue?

0 JPL over 5 years ago in reply to Sean Cashin

Prodigy 90 points

I have two parts here that I used for testing and I get the same issue. Maybe the way I am programming it is wrong?

0 Sean Cashin over 5 years ago in reply to JPL

TI__Genius 14166 points

How are you programming it?

0 JPL over 5 years ago in reply to Sean Cashin

Prodigy 90 points

I am using a beagleboard black and programming in python which I am not very experienced with. Signals for clock, cs, sdi and sdo were checked on oscilloscope and apparently seem ok. The basic code is the one I send above and I wonder if the problem lies on the way I send the bits over spi.xfer2 or some other configuration that could be missing.

0 JPL over 5 years ago in reply to JPL

Prodigy 90 points

Oscilloscope images when sending spi.xfer2([1,254]), maybe it can help. I cannot understand the issue yet.

Here I am using a slower clk than the maximum 10.5MHz specified in the datasheet.

CS and SCLK:

MOSI and SCLK:

MISO and SCLK:

And these are the signals at input (1) and output (2). This gain never changes when set to SPI mode:

0 JPL over 5 years ago in reply to JPL

Prodigy 90 points

I noticed the high level of the MOSI is actually low. This could be it, I will try: e2e.ti.com/.../1137694

0 JPL over 5 years ago in reply to JPL

Prodigy 90 points

It was just a probe issue here, the signals are correct... I am back to the start.

Since I get the signal transfer if I use just the beagleboard and connect MISO/MOSI together, the problem would not have been there I guess.

On page 13 here the SPI settings are given, I tried changing the mode before and it didn't help. Just to be on the safe side, it should be mode 3, am I right?

0 Jacob Freet over 5 years ago

TI__Expert 7875 points

Hi JPL,

Based on your images, it looks like you aren't addressing the SPI properly. It looks like your MOSI (SDI) signal is just kept high for the first 8 bits. In order to write the SPI for the LMH6517 the first 8 bits need to be set according to table 3 in the datasheet. The first 8 bits are the command bits that tell the SPI if it is a read/write, and what address to write to. By default the addresses are 0h for channel A and 1h for channel B. So to write the channels, your first 8 bits need to be...

00000000 to write channel A
00000001 to write channel B

Hope this helps to explain.

Regards,

0 JPL over 5 years ago in reply to Jacob Freet

Prodigy 90 points

Hi Jacob Freet,

Even when sending like that the chip does not respond, I get the same constant gain when sending these commands for example:

I redesigned my board to use the LMH6517 on pulse mode, I don't know what is happening with the SPI. I tried programming a different PGA here that I also designed a circuit using SPI and for the other one my code worked, so I guess the LMH6517 requires something else that I was not able to find out.

Thank you,

JPL

0 Jacob Freet over 5 years ago in reply to JPL

TI__Expert 7875 points

Hi JPL,

I'm glad you were able to get it working with the pulsed mode. If you ever need to switch to the SPI please contact us and I'm sure we can get it to work successfully.

Regards,

Amplifiers

Amplifiers forum

LMH6517: SPI problem