Paulius Velesko

Important changes to forums and questions

All forums and questions are now archived. To start a new conversation or read the latest updates go to forums.mbed.com.

10 years, 9 months ago.

SPI with 2 slaves - Logic levels gone crazy

Hello. I'm running an mbed NXP to drive to Phase Locked Loop eval boards(ADF4350) I can successfully and consistently drive one board and I have managed to program two of these, but the results are very inconsistent.

The code I'm running selects a board by bringing a specific LE pin low, wriiting and pulling it back up again and it works perfectly from software point of view. Originally I ran SPI at 8Khz(max for ADF4350), then I slowed it down to 1 Khz and now I'm down to 500Mhz.

I'm probing with my probe grounded to Mbed GND and probe at CLK.

All the voltage levels I see do not make any sense to me.

I'm attaching my oscilloscope readout and a picture of my setup

/media/uploads/pvelesko/20130807_141515.jpg

/media/uploads/pvelesko/20130807_140545.jpg

2 Answers

Vo Xuan
10 years ago.

Hi Paulius Velesko,

did you solve the problems?

I am working with the board like yours (but only one) using PIC microcontroller. Could you please tell me how to write the ADF4350 registers?

Thank you very much. Xuanvn

Are you having problems using the SPI or the general method for writing the registers?

posted by Paulius Velesko 09 May 2014

Hi Velesko,

Thanks for your reply. I will check my SPI connection again. But, can you tell me the general method for writing registers and, can you give me some sample code of your project?

Thanks in advance. Xuanvn

posted by Vo Xuan 10 May 2014
Martin Smith
10 years, 9 months ago.

Have you connected the GND of both boards together?

As of right now those ADF4350 boards aren't even powered. So technically these SPI lines aren't driving any loads/boards. But under normal operation I power them both off a power supply and connect ADF4350(1), ADF4350(2), and mbed grounds together.

posted by Paulius Velesko 07 Aug 2013

Its not a good idea to measure signals without powering the slave boards. There are probably some protection diodes etc on the input pins that could result in shorts between data, clock and chip select lines. The input signals could also intermittently power up the slave circuit. Better to measure signals when slave is powered as specified.

posted by Wim Huiskamp 07 Aug 2013