LMX2492EVM: Evaluation board cannot change starting frequency or create ramp signal

Dear Noel,

The 100MHz reference signal can be detected at pin 9 of LMX2492. Meanwhile, the MUXout pin still has nothing.

Thanks,

Shuo

Hi, Dean

I powered board down and then up through USB2ANY. I did see the current change after all registers written into board. So, the connection shouldn't cause the issue. Furthermore, I performed the MUXout measurement as you suggested, I still couldn't see the waveform on the scope. The XO generated 100MHz reference which can be detected from both the output pin of CWX-813 and input pin of LMX2492 (OSCin).

Thanks,

Shuo

Dean,

The MUXout pin could be toggled when we change from PP to inverted PP in TICS PRO. The output voltage of MUXout is between 0V to 3.14V. I guess the interface works fine.

Thanks,

Shuo

Shou,

OK, so definitely it looks like SPI programming is working.

Other signs suggest that there is is some issue with the OSCin input signal:

1.  Can't see R/2 output on MUXout pin

2.  Frequency is going to lower end of the VCO

I know you say you see the OSCin signal going to our device, but perhaps there is some issue with matching, level, or DC bias.  For instance, this input should be AC coupled, otherwise the device might not be able to properly detect the signal.

Because the XO is not hooked up by default, this means you had to do some modification to hook it up, so potentially there could be an issue, such as not being AC coupled.  

For diagnostic purposes, you might also want to test this MUXout pin by trying to look at the N/2 output to make sure something comes out for that.

Regards,

Dean

Dean,

This makes me confusing. We didn't apply any input to the OSCin port. All the test are done with the on board XO, and we saw the 100MHz at the input pin of LMX2492. With all the default setup, we could even get 4.8GHz at RFout/2 port of very first experiment listed on the manual. 

Thanks,

Shuo

Shou,

I'm not sure if this is a question or a statement.

In any case, there is something suspicious about the input reference.

The default for the LMX2492 board is to have the XO disconnected and on the board.  The reason for this is that the XO does not have very good phase noise and we wanted to ensure that users would not blame the lMX2492 for close in phase noise that is due to the XO.

On the other hand, we did want people to be able to evaluate the device and an XO is certainly better than a 100 MHz signal generator.

In any case, if you are running this board with no XO hooked up, then the VCO will still oscillate, but at a lower frequency corresponding to the VCO tuning voltage being zero.  That might be around 4.6 GHz or so (rough guess) with divide by 2 mode. 

Now it sounds like a modification to the board was made to provide power to the XO and then was hooked up to the device as it is claimed that there is a signal at the OSCin pin.  But perhaps somethign was hooked up wrong.   In the previous response, I said it might be DC coupled instead of AC coupled.  Or perhaps the GND/OSCin* pin is not handled properly.  For instance, if the same signal went to OSCin and GND/OSCin*, then there would be no output as this is a differential input.

Regards,

Dean

Dean,

Based on the manual, the default setup is having XO connected to the PLL. I took two screenshots from the TI manual of LMX2492EVM, and one of them shows the experiment we are doing now (Chapter 2 setup) and the other one shows the schematic of the evaluation board. I use red line to indicate the path from XO output to PLL OSCin pin. From the schematic, they are connected without any modification.

Again, we did nothing to the board, except powering it up (with required 5V) and measuring the output with oscilloscope.

Thanks,

Shuo