• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 20 subscribers
  • Views 80 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LMX2694EPEVM: EVM Queries

Kazuki Itoh
Genius 9175 points
Part Number: LMX2694EPEVM
Other Parts Discussed in Thread: USB2ANY, LMX2582

My customer is using LMX2694EPEVM.
Could you please answer inquiries below?

1. Phase noise
The calculated and measured phase noise value (loop bandwidth ~100kHz to out of the loop) doesn't match in VCO frequency range from 8GHz to 14GHz (VCO core 1 to 6).

　　VCO core1-2：Calculation matches the measurement
VCO core3　 ：Calculation doesn't match the measurement (~10dB in 1MHz detuning)
　　VCO core4　 ：Calculation matches the measurement　　　　　 　
VCO core5〜6：Calculation doesn't match the measurement (~3-5dB in 1MHz detuning)

Could you let me know why the measurement doesn't match the calculation?
The calculated VCO phase noise value uses PLLatinum Sim VCO model.
Is it different from the actual VCO phase noise? Is the power supply related to this difference?

2. Spurious
My customer sees ~80dBc spur close to carrier frequency +/-1.5GHz with the several tens of MHz frequency interval.
Is this spur expected to operate the PLL on the EVM?
Is it coming from the EVM, or the customer's measurement set up?

3. Frequency switching command

My customer is considering to use full assist mode for the VCO calibration.
For the frequency switching, it is needed to send four commands (VCO three parameters + frequency setting).
Is the wait time needed between each commands?
If so, what is the recommended wait time?

4. Frequency switching response time measurement
When the full assist mode is used and measure the frequency switching time, does the below two method has difference in the response time?

(a) Send the command (VCO parameter and frequency setting) in the same time, and trigger it by the CS signal rising edge after sending commands. (Clock frequency 40MHz is assumed)

(b) VCO three parameters and frequency are manually set, and then the trigger it by the CS signal rising edge after sending frequency setting.

Currently, it is measured with (b) method using TICS Pro.
However, the measurement is much different from the measurement (~60us difference).
So they want to identify the cause.

Please see below for the measurement setup.
- Power supply: +3.3V coming from the external DC power supply.
- Loop filter uses R and C with the constants canclulated with PLLatinum Sim
- The EVM is controlled with PC via USB2ANY

Best regards,
Kazuki Itoh

  • 0
    TI__Genius 12961 points

    Hi Kazuki, 
    1. What is the reference they're using? 

    2. Can you provide their config (.tcs) file ? You say their are spurs near the carrier, it's possible they're encountering a integer boundary spur. 

    Whether is comes from their setup or due to their config is hard to say at the moment. Section 8.1.1 of LMX2582 DS explains a bit on how customer can calculate most common spurs.https://www.ti.com/lit/ds/symlink/lmx2582.pdf?ts=1705679598108&ref_url=https%253A%252F%252Fwww.google.com%252F

    Can customer see spur on the reference only? 

    If customer changes power supply, do they see same spurs still? 

    3. Refer customer to SPI write interface requirements, as long as this is being met they should not have a problem. 

    The way I typically measure this is I measure the time from the last clk write to when VTUNE voltage settles which indicates lock (~1.2V)

    Another better method, if customer has access to equipment that can plot frequency vs time is lock PLL to one frequency and then program a new frequency and measure how long it takes to change to that new frequency. 

    I don't quite understand what you're trying to say in part a unfortunately. 

    Please note USB2ANY write speeds are very slow (125kHz), you can get a much faster lock time using faster SPI bus speed. 

    Refer to this appnote for information regarding how to reduce lock time. https://www.ti.com/lit/an/snaa336a/snaa336a.pdf?ts=1705011436322&ref_url=https%253A%252F%252Fwww.google.com%252F

    Section 3.1 also explains why SPI bus speed matters. 

    Regards, 

    Vicente