LMX2581EVM: Unable to produce RF output

Hi Chris,

Can you share your TICS Pro programming? You can save a .tcs file that has all the settings and post it to E2E for us to examine.

Regards,

Derek Payne

Chris,

It occurs to me you are referring to CodeLoader, which is a deprecated platform. I strongly recommend you use TICS Pro, since it is currently the supported software platform and it will be much easier for us to debug issues with currently supported software.

Our typical procedure is to reset the device by toggling the RESET bit in R5 0->1->0, then write all registers in descending order. If this is not producing an RF output, we would need to see the register settings. You should be able to export the register settings from CodeLoader and we can import them into TICS Pro...

Just double-checking, are you sure you have the OUTA_MUX or OUTB_MUX configured to the right setting (typically Fin), the MODE field set to External VCO, and the OUTA_PD or OUTB_PD bits cleared?

Regards,

Derek Payne

Hi Derek,

Thanks for your message.

I'm in the process of getting TICS Pro setup right now (I was indeed using CodeLoader) and will try out the startup procedure in your message.

And yes, I have the OUT_MUX's set to Fin, MODE set to External VCO, and the OUTA_PD bit cleared as I'm trying to measure from the RF A+ port.

Best,

Christian

Hi Derek,

I switched over to TICS Pro (from CodeLoader) and was able to get RF signals out of the device - definitely a step forward!

A couple of follow-up questions:

1. In order to get a signal, I had to manipulate the registers. I tried to reset R5 per your instructions but couldn't find a way to actually "reset" R5. So, instead I clicked read and then write for R5, followed by write for all of the registers. I'm unsure if this is the proper way to complete this sequence, but after doing this a signal did appear on my spectrum analyzer. Is "writing" R5 and then all the other registers the proper way to initialize the registers?

2. I have been getting strong signals at the desired frequencies, but I'm also seeing multiple (~6) higher harmonics above the noise floor on my spectrum analyzer. Is this expected based on the design of the device or might it be a sign that I've configured it incorrectly?

3. Can the device take commands via the command line rather than relying on the TICS Pro interface?

Thank you.

Best,

Christian

Hi Chris,

1. In the "User Controls" page, under the general settings pane, there is a toggle for the RESET bit specifically. The act of toggling and untoggling this checkbox will behave as a live programmer to the LMX2581. Per the datasheet section 8.5.2, writing the reset bit before other registers is a part of the recommended programming sequence. While the datasheet doesn't explicitly say it, the expectation is that the subsequent register writes will be performed, including R5 once again, this time with the RESET bit cleared. By toggling the RESET bit 0->1->0 before programming this is effectively equivalent. The subsequent writing of all registers is the next step in the programming sequence, so you're doing everything correctly.

Also note that TICS Pro has a macro for writing all registers: Ctrl + L.

2. It's possible that there's a misconfiguration. If you want, you can save a .tcs file and upload it to E2E so we can analyze it.

3. At this time there's no explicit command line interface. Technically there is a 32-bit ActiveX interface for TICS Pro, and in the near future we will be releasing a version of TICS Pro with a simple asynchronous socket server, to perform certain operations such as modifying GUI fields or reading/writing registers. Maybe the closest thing we could describe as a command line interface would be directly controlling the USB2ANY. For an overview of ways to automate control of TICS Pro or the USB2ANY, please see the post below.

https://e2e.ti.com/support/clock-timing-group/clock-and-timing/f/clock-timing-forum/1025934/lmx2594evm-serial-control-through-python

Regards,

Derek Payne

Hi Derek,

Thank you for the guidance.

I followed your instructions for resetting the registers. However, I'm still seeing many harmonics and now the main RF output signal is not even at the right frequency (off by over 100 MHz).

I'm trying to upload the .tcs file that I saved but the "Insert" --> "Image/video/file" option isn't letting me upload the document. It's saying that I can't upload the file after I type in the file name, including file path. It also won't let me drag and drop the file into the pop-up box. Is there another way to upload the file here?

Best,

Chris

Chris,

Can you try saving the tcs file as a .txt file? Under the hood it's just an INI.

If that doesn't work, you can use the Export Hex Register Values option in the file menu and upload the text file produced by this. As long as I know what frequencies you're trying to use, I should be able to work from just the register settings.

Regards,

Derek Payne

TICS Pro Configuration File.tcs

Hi Derek,

I managed to get the file uploaded here.

Best,

Christian

Thanks Christian.

I checked the configuration, and I see that the MODE mux is configured as full synthesizer. If you're using an external VCO, this needs to be set for external VCO; in which case, the maximum frequency for external VCO is 2200MHz.

Next thing to check, is your VCO control pin polarity positive or negative? The default is negative for LMX2581, because our internal VCO is negative polarity; if you need to swap it, you can clear the CPP bit to 0 (R2[27] = 0). I checked the GUI and it seems we somehow overlooked giving this a visually represented control, so I put that on the list of things to fix in the next release - for now you should be able to set the bit from the raw registers page by selecting bit R2[27] and typing '0', if needed.

After all this, I think you still have one other potential issue: your loop filter may be unstable, owing to the low reference and phase detector frequencies. We have a tool called PLLatinum Sim which can be used to analyze the loop bandwidth and the phase margin. I can almost guarantee that the loop is unstable with the default EVM values for your lower phase detector frequency. For first-time use PLLatinum Sim can feel like drinking from a fire hose, so my recommendation is to use intermediate mode and program your own Kvco based on the tuning range of the VCO you want to use. With the default loop filter I calculate around 30° phase margin, but 50° should be achievable with higher phase detector frequency and different loop filter values.

If you have a VCO datasheet, I could also go a little deeper on PLLatinum Sim and offer some updated guidance.

Regards,

Derek Payne

Hi Derek,

Thanks for the information.

I think I had the Mode set as External VCO previously, but had switched it to Full Synthesizer because it wasn't generating any signal when it was set to External VCO. I swapped it back to External VCO and, as before, didn't see any signal.

I also made the change to R2, by setting R2[27] = 0, but also didn't see a signal after making this change.

I downloaded PLLatinum Sim and am taking a look at it, but I'm wondering whether I should have seen any signal at all by this point based on the modifications made as described above.

Best,

Chris

Chris,

I would have expected by this point you would have a signal, and definitely you should have seen something by activating the internal synthesizer, so I'm not sure what's going on.

Try toggling powerdown/powerup for the device using the PWDN_MODE field in the general settings of the User Controls page in TICS Pro... If you don't see a noticeable reduction in power consumption during powered down state, there may be a problem communicating with the EVM preventing programming.

Assuming you can still talk to the device, you may also want to set up the MUXOUT_SELECT field on the same page to output the R-divide or N-divide signals, just to see if anything is actually happening inside the device.

Regards,

Derek Payne