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.

LMX2572LP: LMX2572LP best software for design and simulation.

Part Number: LMX2572LP
Other Parts Discussed in Thread: LMX2531, LMX2572, LMX2582

Hi,

I'm a little new to the LMX2572LP series of synthesizers and would like to know what is the best way to begin designing with these parts other than goinf trough the datsheets etc. I notice there are various software packages to assist in this task but there seem to be three or so ... So which ones are recommended to start with.

Also it may be that the LMX2572LP, even though it appears to meet the requirements for low phase noise, it may not be the lowest power of them all. So was wondering if other similar parts may do as well with lower power consumption ? any suggestions right of the bat ?

With regards to choice of TCXO reference is an LVDS source recommended or will a standard TCXO sinfle-ended output be ok ? The application is for DMR in the ISM 868 and ISM 915 bands.

Will the simulation software be able to assist me with the choice of best reference input frequency (TCXO) for optimal phase noise and spurs at the 800 to 1GHz frequency band ? Otherwise whta method is recommended for this choice ??

regards,

mike

  • Hi mike,

    TICS Pro can help you setup your configuration and generate all the register values. Some registers have critical restriction, this tool will alert you if you have violated the restriction. You could also mouse-over the boxes to see the description of that register.

    PLL Sim is a design tool, you will need this tool to design the loop filter, check the phase noise and spurs.

    For DMR application, I think LMX2572LP is the best choice. The lowest power consumption synthesizers we have is the LMX2531 series, but they are narrow band devices, you may need several parts to support all the frequencies you want.

    Again for DMR application, you will need to use TCXO, so the output format is never LVDS. In order to get better PLL noise, I suggest use as highest frequency as possible. Higher frequency returns higher slew rate, this could compensate the slew rate loss in clipped-sine wave format.

  • Noel Fung said:

    Hi mike,

    TICS Pro can help you setup your configuration and generate all the register values. Some registers have critical restriction, this tool will alert you if you have violated the restriction. You could also mouse-over the boxes to see the description of that register.

    PLL Sim is a design tool, you will need this tool to design the loop filter, check the phase noise and spurs.

    For DMR application, I think LMX2572LP is the best choice. The lowest power consumption synthesizers we have is the LMX2531 series, but they are narrow band devices, you may need several parts to support all the frequencies you want.

    Again for DMR application, you will need to use TCXO, so the output format is never LVDS. In order to get better PLL noise, I suggest use as highest frequency as possible. Higher frequency returns higher slew rate, this could compensate the slew rate loss in clipped-sine wave format.

    Hi Noel,

    Thanks, will take a look at the software in due time. But just a question regarding the initial tuning and the loop filter. These settings are set-up once at boot-time and then will not require any further adjustment regardless of the frequency tuned, or will some adjustment be required during the use of the synthesizer. The reason I'm asking is that it is intended for an application that includes FHSS hopping at say 50ms intervals...

    With regards to the TCXOs we have come across some devices from Abracon, the ASGTX5 Series of XOs that do come with an LVDS output option besides PECL or HCSL etc , and also meet Stratum 4 requirements. Their designed with a 100 Ohm differential load in mind, and with a typical 300mVp-p output. I think these might be OK for the LMX2572 ? Any thoughts on the suitability of these devices for the LMX2572  ?

    Also, I notice the highest clock input frequency is 250MHz without the doubler. What is better using the doubler, or bypassing it and going straight in with say a 200MHz differential output XO ??

    Regards, mm

  • Hi Mike,

    The ASGTX5 is a XO, not TCXO, are you sure you can use it in DMR application? DMR is a narrow band system and it has quite tight frequency accuracy spec.

    Since there are many channel in a certain DMR band, you may need to adjust some of the loop parameters in order to meed spurs and phase noise requirement. 

    The doubler does not add noise, but it is required only when the input clock frequency is low. For example, if the input clock is just 20MHz, then I would suggest use the doubler to make fpd = 40MHz. This will reduce the N-divider value and therefore reduce PLL noise.

  • Noel Fung said:

    Hi Mike,

    The ASGTX5 is a XO, not TCXO, are you sure you can use it in DMR application? DMR is a narrow band system and it has quite tight frequency accuracy spec.

    Since there are many channel in a certain DMR band, you may need to adjust some of the loop parameters in order to meed spurs and phase noise requirement. 

    The doubler does not add noise, but it is required only when the input clock frequency is low. For example, if the input clock is just 20MHz, then I would suggest use the doubler to make fpd = 40MHz. This will reduce the N-divider value and therefore reduce PLL noise.

    Hi Noel,

    The application is proprietary DMR in the ISM bands and not subject to the same standards and classic requirements. 50 KHz channel separation can accommodate more relaxed specifications on frequency stability, however given that, I notice that Abracon also have the ASGTX-D series of TCXO, albeit x2 more expensive. It could be a contender ...

    With regards to the adjustment of loop parameters, can the software simulator be used as a good enough guide to the final expected spurs and phase noise resulting with the change of one channel to another (assuming all other factors such as board layout etc meet the performance required for the LMX2572) ? What are the parameters that generally need to be tweaked to obtain such fine adjustments that you refer to ?

    To some extent the stringent channel separation issue is not as severe in the proprietary application we are considering where 50KHz up to 100 KHz separation is acceptable. Close-in spurs and phase noise requirements could be more relaxed due to the wider channel separation possible.

    Ok for the doubler, if it’s bypassed, I guess using a 100 – 200 MHz LVDS oscillator with 2 - 3ppm will meet the requirements.

    Thanks in advance.

  • Hi Mike,

    Great, I think a 100MHz LVDS reference clock is good enough. Doubler is not necessary, 100MHz is an ideal phase detector frequency. 

    Spurs, especially Integer Boundary Spurs (IBS) is the most headache problem you will encounter. You may need to use the Pre-R, MULT and/or Post-R divider to change the fpd to a number which is not divisible by 100 (or 50 or 25 or ....) to avoid IBS.  

    There are good examples in LMX2582 datasheet section 8.1 on how to predict the spurs frequency. Alternatively, you can use PLL Sim to check both the phase noise and spurs.