LMX2595EVM: Was locking yesterday, not locking today for the same config

Sagar Achar

Part Number: LMX2595EVM
Other Parts Discussed in Thread: LMX2595,

Hello TI Community,

I am working on the LM2595EVM + Reference Pro with TICS Pro. My connection is exactly as shown in the diagram below (minus the 50 Ohm Termination):

Here is my TICS Pro Config File: Config File

I was successful in locking the PLL and ramping it between required frequencies yesterday (video). I come back today to work the same, but the EVM doesn't lock.

Observations:

The Lock LED flickers on Ctrl+L / Write All Registers and goes off less than a second later
Current draw is steady at around 172mA (before and after write)
On Reading Registers, most registers are filled with 1s
Sometimes on toggling the OUTA_PD / OUTB_PD, the lock LED turns on but the output doesn't show any locking

Here is what I have tried:

Directly use the saved TCS file: Doesn't work
Recreate the entire configuration from scratch: Doesn't work
Clock Verification: I am getting a stable 100 MHz clock from the Reference Pro
Remove Load: Doesn't lock
Try Default Config and Test Mode: Doesn't work / lock
Disconnect and reconnect everything: Performed twice, still doesn't work
Check Vcc_TP and VccRF_TP: Steady 3.3V
Alternate between Out A and Out B: Doesn't work
Switch from 100MHz to 125MHz input clock: Doesn't work

Here are some other details:

Supply: 3.3V 600mA CV with OVP and OCP via a makeshift SMA power cable
Load: tinySA Ultra Spectrum Analyzer (single ended)

Please provide me with a solution to fix this issue and get my setup back to working condition.

Thanks in Advance!

over 1 year ago

0 vicente flores prado over 1 year ago

TI__Mastermind 29402 points

Hello Sagar,
You mentioned yesterday you were able to achieve lock, if you probe the SPI lines do you see traffic when you attempt to write/communicate to LMX2595?

Regards,

Vince

0 Sagar Achar over 1 year ago in reply to vicente flores prado

Prodigy 10 points

Hey Vince,

I probed the SDI_TP and SCK_TP using my Rigol Oscilloscope and the SPI traffic seems normal. Hence plugged in my logic analyzer to the output of the Reference Pro and the write operation looks like this. The data is written in 16 register intervals and the waveform values match the values in TICS Pro.

But, if you notice between the 5th and 6th burst, the data line is at Logic HIGH. Can you please confirm if this is normal?

Also, I have tried using multiple Reference Pros I had lying around and yet, my LMX shows no signs of working.

Today, my LMX is drawing around 212mA in steady state. What else can I do to ensure working?

Regards,

Sagar

0 vicente flores prado over 1 year ago in reply to Sagar Achar

TI__Mastermind 29402 points

Hi Sagar,
You mentioned current stays the same before and after programming
Before programming my EVM draws about 143mA of current and the moment I load the default configuration I see current draw rise to ~380mA.
So something is definitely up.

Lets find out if the LMX2595 is responding - what happens if you toggle reset or Powerdown?
If I enable Powerdown I see current draw drop to 2mA after having programmed the EVM with the default configuration.

Regards,

Vicente

0 Sagar Achar over 1 year ago in reply to vicente flores prado

Prodigy 10 points

Hey Vincente,

Well this is strange, not only is the current not dropping when POWERDOWN, but increasing! The RESET does nothing.

My supply shows the following readings (consistent on multiple trials):

POWERDOWN OFF: 209mA
POWERDOWN ON: 212 mA

This is both confusing and concerning...

Regards,

Sagar

0 vicente flores prado over 1 year ago in reply to Sagar Achar

TI__Mastermind 29402 points

Hi Sagar,

Something is definitely up. It appears the chip might be dead, especially given that current doesn’t drop when you enable power down.

But lets ensure that the issue isn’t communication related. You mentioned you had a few ref pro boards laying around - This indicates to me you have other EVMs. Can you communicate/control a different EVM using these ref pro boards?

Regards,

Vicente

0 vicente flores prado over 1 year ago in reply to vicente flores prado

TI__Mastermind 29402 points

Hi Sagar,
One more thing.
Please probe bias/LDO lines on LMX2595 to confirm supply voltages are as expected.
https://e2e.ti.com/support/clock-timing-group/clock-and-timing/f/clock-timing-forum/914390/lmx2594evm-defective-with-instabilities-and-spurs

Refer to this thread, that I was made aware of from my colleague, to get a ballpark estimate of what supplies should be.

Regards,

Vicente

0 Sagar Achar over 1 year ago in reply to vicente flores prado

Prodigy 10 points

Hey Vincente,

Thanks for the LDO suggestion. Unfortunately, the voltages I measured had no correspondence to the expected values.

Here are my observations, I observed the same results on multiple trials, even after different conditions (power after programming, power before programming, power down):

Also, my VTune_TP reads 1.329V.

So, what can I conclude from this?

Regards,

Sagar

+1 vicente flores prado over 1 year ago in reply to Sagar Achar

TI__Mastermind 29402 points

Hi Sagar,
Unfortunately It appears internal LDOs have failed.
You will need a new LMX2595.
Do you have a local FAE or field office you can reach out to get a replacement?

Regards,

Vicente

0 Sagar Achar over 1 year ago in reply to vicente flores prado

Prodigy 10 points

Hey Vincente,

We don't have a local FAE or a field office. So, we are thinking of buying a new LMX2595EVM. Can you recommend some good practices so that we can operate the EVM safely and avoid such problems again?

Also, can you help us with some good design practices for boards involving the LMX259x?

Thanks,

Sagar

0 Noel Fung over 1 year ago in reply to Sagar Achar

TI__Guru** 109470 points

Hi Sagar,

There is no onboard LDO, so make sure the input supply voltage is less than 3.4V.

Never modify the board before you confirm the board is working properly.

When you rework the board, follow ESD precaution and practice.

Did you mean pcn board design for synthesizer? Just follow general rule of thumb such as we need a solid ground plane underneath the synthesizer layer. Follow datasheet land pattern and thermal via recommendation. Vcc decoupling should be placed as close to the Vcc pin as possible.

FYI, some appnote available in ti.com

https://www.ti.com/lit/pdf/scaa082

https://www.ti.com/lit/pdf/spraar7

https://www.ti.com/lit/pdf/swra367