LMS TxLO noise

historical

Hi Andrey,

Sorry I did answer sooner. The E4406A arrived so I have been setting
that up. I have UmTRX hooked up right now, which I will be testing
very shortly.

  Thomas

On Fri, Jul 20, 2012 at 1:14 PM, Andrey Sviyazov
<andrey.sviyazov at fairwaves.ru> wrote:
> Hi Thomas.
>
> I couldn't wait your reply and start to implement 10kHz BW of PLL.
> You can find here two pictures with results.
> First of all I found that PLL tuning algorithm doesn't work properly at low
> PLL BW.
> VCOCAP register (r0x19) should contain higher value at least for +3 units
> (to make CAP lower), otherwise freq's above 950MHz never locked.
> Note, that you should read and change register 0x19 after autotuning.
> Second, you can see some noise difference for the 925MHz because of used
> DIV=8 of  the VCO, instead of DIV=4 for 942 and 960MHz.
> Third, you can see PLL noice dependance with the charge pump current
> (r0x16).
>
> On the other two pictures you can find which components have to be changed.
> Thomas, please make one more measure of LO noise and jitter at PLL BW=10kHz
> by your instrument.
> We need to know, is it real to reach modulation accuracy of 1.5 degrees RMS
> or impossible, just because of LMS PLL have bigger jitter.
>
> Best regards,
> Andrey Sviyazov.
>
>
>
> 2012/7/19 Andrey Sviyazov <andrey.sviyazov at fairwaves.ru>
>>
>> Thomas.
>> I've calculate few PIF filters for PLL by the ADIsimPLL and found that RMS
>> phase jitter in case of 5kHz and 50kHz only 0.33 and 0.46 degrees.
>> It is very small difference, so are you sure that I should implement 10kHz
>> bandwidth filter as in USRP?
>> Actually it is absolutly not an problem for me to implement it, but I have
>> to be sure.
>> I've calculated it for 10kHz bandwidth, noise plot will be around
>> 75-80dBc/Hz @ 10kHz and 105-110 dBc/Hz @ 100kHz offset.
>>
>> Best regards,
>> Andrey Sviyazov.
>>
>>
>>
>> 2012/7/19 Andrey Sviyazov <andrey.sviyazov at fairwaves.ru>
>>>
>>> Alexander.
>>> I think that few pulses of LMS hardware reset (LMS_NRST-1 and LMS_NRST-2
>>> pins) will be much easily than all registers checking.
>>> BTW, are there reset pulses after power up or not?
>>>
>>> Best regards,
>>> Andrey Sviyazov.
>>>
>>>
>>>
>>> 2012/7/19 Alexander Chemeris <alexander.chemeris at gmail.com>
>>>>
>>>> Guys, please keep track of all LMS configuration values which are not
>>>> set to default during power on and we'll add them to our manual
>>>> initialization script.
>>>>
>>>> On Thu, Jul 19, 2012 at 6:56 PM, Andrey Sviyazov
>>>> <andrey.sviyazov at fairwaves.ru> wrote:
>>>> > Sergey.
>>>> > I'll play around GPS power and RTC supply, thank you for your help.
>>>> >
>>>> > Thomas.
>>>> > About r0x16 please be sure that sometimes it isn't set to default
>>>> > value
>>>> > after power up (I saw it few times).
>>>> > Also I think that possible those bad "power on reset" gave us unstable
>>>> > result of LMS autocalibration.
>>>> >
>>>> > Best regards,
>>>> > Andrey Sviyazov.
>>>> >
>>>> >
>>>> >
>>>> > 2012/7/19 sergey kostanbaev <sergey.kostanbaev at gmail.com>
>>>> >>
>>>> >> Which gives ~1.5khz cut-off, that's strange
>>>> >>
>>>> >> Also I'd check GPS VDD and AVDD nosies
>>>> >>
>>>> >>
>>>> >> On Thu, Jul 19, 2012 at 6:41 PM, Andrey Sviyazov
>>>> >> <andrey.sviyazov at fairwaves.ru> wrote:
>>>> >>>
>>>> >>> 1kOhm resistor and 0.1uF capacitor already exist between DAC and VC
>>>> >>> input.
>>>> >>> Of course I'll increase capacitance or resistance to be sure about
>>>> >>> VC
>>>> >>> pin.
>>>> >>>
>>>> >>> Best regards,
>>>> >>> Andrey Sviyazov.
>>>> >>>
>>>> >>>
>>>> >>>
>>>> >>> 2012/7/19 sergey kostanbaev <sergey.kostanbaev at gmail.com>
>>>> >>>>
>>>> >>>> No :) But it may cause.
>>>> >>>>
>>>> >>>> I'd check all the line from FPGA to DAC to VCTXCO because VC pin is
>>>> >>>> really sensible to any noise.
>>>> >>>>  - caused by power supply of DAC.
>>>> >>>>  - output noise of DAC.
>>>> >>>>  - algorithm of changing values
>>>> >>>>
>>>> >>>> At first I'd inspect VC pin at VCTXCO and try to filter it with
>>>> >>>> cut-off
>>>> >>>> frequency 1-0.1 Hz
>>>> >>>>
>>>> >>>>
>>>> >>>> On Thu, Jul 19, 2012 at 6:29 PM, Andrey Sviyazov
>>>> >>>> <andrey.sviyazov at fairwaves.ru> wrote:
>>>> >>>>>
>>>> >>>>> Sergey.
>>>> >>>>> Is it really so frequently (32k/s updates at 1pps reference)?
>>>> >>>>>
>>>> >>>>> Best regards,
>>>> >>>>> Andrey Sviyazov.
>>>> >>>>>
>>>> >>>>>
>>>> >>>>>
>>>> >>>>> 2012/7/19 sergey kostanbaev <sergey.kostanbaev at gmail.com>
>>>> >>>>>>
>>>> >>>>>>
>>>> >>>>>>
>>>> >>>>>> On Thu, Jul 19, 2012 at 6:10 PM, Andrey Sviyazov
>>>> >>>>>> <andrey.sviyazov at fairwaves.ru> wrote:
>>>> >>>>>>>
>>>> >>>>>>> Hi all.
>>>> >>>>>>>
>>>> >>>>>>> I've found spurs on the LO noise plot at  ~33kHz and ~66kHz
>>>> >>>>>>> offset
>>>> >>>>>>> when GPS antenna used and position locked.
>>>> >>>>>>> This spurs begin grow up when GPS just near to lock position and
>>>> >>>>>>> after locking spurs stopped to grow as you can see at picture.
>>>> >>>>>>> If thereafter GPS disconnected then noise coming back to normal
>>>> >>>>>>> plot.
>>>> >>>>>>> I think that it is result of 32768Hz clock in the GPS module
>>>> >>>>>>> EB-230,
>>>> >>>>>>> but can't understant how it can impact to 26MHz clock or VCO.
>>>> >>>>>>
>>>> >>>>>>
>>>> >>>>>> As another idea it can be caused by continues GPS correction lead
>>>> >>>>>> to
>>>> >>>>>> DAC changes to the VCTXCO.
>>>> >>>>>>
>>>> >>>>>>>
>>>> >>>>>>> Please tell me something who know.
>>>> >>>>>>>
>>>> >>>>>>> Thomas.
>>>> >>>>>>> Was GPS antenna connected when noise measured like on the
>>>> >>>>>>> picture
>>>> >>>>>>> which you sent us (also attached here)?
>>>> >>>>>>> I would to know because of anomal peak near to 30kHz offset too.
>>>> >>>>>>>
>>>> >>>>>>> Best regards,
>>>> >>>>>>> Andrey Sviyazov.
>>>> >>>>>>>
>>>> >>>>>>>
>>>> >>>>>>>
>>>> >>>>>>> 2012/7/19 Andrey Sviyazov <andrey.sviyazov at fairwaves.ru>
>>>> >>>>>>>>
>>>> >>>>>>>> Sorry, that was meant to be sent to the mailing list :)
>>>> >>>>>>>>
>>>> >>>>>>>> Best regards,
>>>> >>>>>>>> Andrey Sviyazov.
>>>> >>>>>>>>
>>>> >>>>>>>> ---------- Forwarded message ----------
>>>> >>>>>>>> From: Andrey Sviyazov <andreysviyaz at gmail.com>
>>>> >>>>>>>> Date: 2012/7/18
>>>> >>>>>>>> Subject: LMS TxLO noise
>>>> >>>>>>>>
>>>> >>>>>>>> Hi Thomas.
>>>> >>>>>>>>
>>>> >>>>>>>> Here forwarded my last e-mail with noise plots when I stopped
>>>> >>>>>>>> work
>>>> >>>>>>>> around it at first time, please see below.
>>>> >>>>>>>>
>>>> >>>>>>>> Please try to play around Tx PLL charge pump current (register
>>>> >>>>>>>> 0x16)
>>>> >>>>>>>> for better RMS phase stability.
>>>> >>>>>>>> I think we should reach 1 degree or below.
>>>> >>>>>>>>
>>>> >>>>>>>> Alexander gave me the second UmTRX board and after checking and
>>>> >>>>>>>> fixing all known hardware issues I've got roughly the same LO
>>>> >>>>>>>> noise plot.
>>>> >>>>>>>> Possible Robin had no time to fixing all of our issues, so
>>>> >>>>>>>> check
>>>> >>>>>>>> them all please.
>>>> >>>>>>>> And also check please what type of TCXO installed on your
>>>> >>>>>>>> board.
>>>> >>>>>>>>
>>>> >>>>>>>> Best regards,
>>>> >>>>>>>> Andrey Sviyazov.
>>>> >>>>>>>>
>>>> >>>>>>>> ---------- Forwarded message ----------
>>>> >>>>>>>> From: Andrey Sviyazov <andreysviyaz at gmail.com>
>>>> >>>>>>>> Date: 2012/4/13
>>>> >>>>>>>> Subject: Re: LMS TxLO noise
>>>> >>>>>>>>
>>>> >>>>>>>> Hi all.
>>>> >>>>>>>>
>>>> >>>>>>>> There is progress with LMS PLL :)
>>>> >>>>>>>> Pictures are attached here.
>>>> >>>>>>>> t was discovered that 80 kHz spurs come from Ethernet, or
>>>> >>>>>>>> rather
>>>> >>>>>>>> from the ET1011.
>>>> >>>>>>>> I unknowingly put the choke between transistor of 1V regulator
>>>> >>>>>>>> and
>>>> >>>>>>>> analog power 1V.
>>>> >>>>>>>> As a result, the regulator has become unstable and oscillated
>>>> >>>>>>>> 80 kHz
>>>> >>>>>>>> with amplitude of 200 mV, which climbed into the LMS PLL.
>>>> >>>>>>>> To correct this problem L46 should be replaced by jumper on all
>>>> >>>>>>>> alfa
>>>> >>>>>>>> version PCB's.
>>>> >>>>>>>>
>>>> >>>>>>>> Also I just played with current in the PLL loop, shown on the
>>>> >>>>>>>> picture for clarity.
>>>> >>>>>>>> Proved to be the optimal current 1,9 mA (you should write 0x93
>>>> >>>>>>>> in
>>>> >>>>>>>> the register of 0x16).
>>>> >>>>>>>> But, I think, for the RxPLL will be better use of the current
>>>> >>>>>>>> 2.4
>>>> >>>>>>>> mA, because the nearest noises more important for Rx (you
>>>> >>>>>>>> should write 0x98
>>>> >>>>>>>> in the register 0x26).
>>>> >>>>>>>>
>>>> >>>>>>>> Best regards,
>>>> >>>>>>>> Andrey Sviyazov.
>>>> >>>>>>>>
>>>> >>>>>>>>
>>>> >>>>>>>
>>>> >>>>>>
>>>> >>>>>
>>>> >>>>
>>>> >>>
>>>> >>
>>>> >
>>>>
>>>>
>>>>
>>>> --
>>>> Regards,
>>>> Alexander Chemeris.
>>>> CEO, Fairwaves LLC / ООО УмРадио
>>>> http://fairwaves.ru
>>>
>>>
>>
>