Just throwing out there that I did spend a little time playing with reducing charge pump current / gain (via LMX2592 register settings -- cp_idn, cp_iup, cp_icoarse). I was able to eliminate the spurs that were bouncing up and down by trading off some phase noise. I'm not going to use those changes for now, since changing loop parameters like that would require a lot of testing over frequency, temperature, etc. I only wanted to mention that tweaks can potentially be made to help deal with some of these issues.
Thanks, Jim ________________________________ From: David Raeman <da...@synopticengineering.com> Sent: Wednesday, November 2, 2022 2:05 PM To: Jim Palladino <j...@gardettoengineering.com>; Marcus D. Leech <patchvonbr...@gmail.com>; USRP-users@lists.ettus.com <usrp-users@lists.ettus.com> Subject: RE: [USRP-users] Re: N320 LO stability problem Jim, thanks for the feedback and I’m glad it helped. Spurious phase noise tends to impair performance for our projects as well, though in that case just fixing the locking allowed me to move on. My current project has shifted my focus to E320s instead of N320s, but I’m very interested in following any further research/improvements that reduce phase noise on either radio architecture. Best, David From: Jim Palladino <j...@gardettoengineering.com> Sent: Wednesday, November 2, 2022 1:33 PM To: David Raeman <da...@synopticengineering.com>; Marcus D. Leech <patchvonbr...@gmail.com>; USRP-users@lists.ettus.com Subject: Re: [USRP-users] Re: N320 LO stability problem I ended up manually applying David's "patch" to UHD 4.1.0.5 (manually edited the files and rebuilt UHD). This did fix the issue with the N320's not locking based off of the limited tests I've performed so far. Thanks for pointing to that, David. However, I think there is still more not quite right. Even though I seem to consistently achieve lock at these "trouble frequencies", I still see spurs increasing and decreasing in power like I described in the first email. Also, I occasionally see what looks like quick jumps in overall phase noise. David, I saw you made a comment in the original UHD issue that you posted regarding not using the oscillator doubler and multiplier at the same time. If/when I have time, I might look more at this and other LMX settings. For now, the fact that the synthesizer seems to lock consistently will allow us to keep moving forward. Also, I did briefly update one of the N320s to UHD 4.3.0.0 and saw the exact same overall behavior. Thanks, Jim ________________________________ From: Jim Palladino <j...@gardettoengineering.com<mailto:j...@gardettoengineering.com>> Sent: Wednesday, November 2, 2022 8:52 AM To: David Raeman <da...@synopticengineering.com<mailto:da...@synopticengineering.com>>; Marcus D. Leech <patchvonbr...@gmail.com<mailto:patchvonbr...@gmail.com>>; usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com> <usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com>> Subject: [USRP-users] Re: N320 LO stability problem David, that is great, I'll look into that. I'm working on updating to 4.3.0.0 right now, but I'll try your double tune solution first. If this has been solved, that will save me a lot of time poking around in the LMX code. Thanks, Jim ________________________________ From: David Raeman <da...@synopticengineering.com<mailto:da...@synopticengineering.com>> Sent: Wednesday, November 2, 2022 8:49 AM To: Jim Palladino <j...@gardettoengineering.com<mailto:j...@gardettoengineering.com>>; Marcus D. Leech <patchvonbr...@gmail.com<mailto:patchvonbr...@gmail.com>>; usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com> <usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com>> Subject: RE: [USRP-users] Re: N320 LO stability problem Hi Jim, I ran into what sounds like the same issue using N320s operating at nearly the same center frequency. I was able to isolate a fix (some tweaks in the LMX2592 configuration), and my PR was folded into UHD 4.2. I believe another, unrelated fix for N320 tuning was also committed in UHD 4.2.0.1. I’d strongly recommend trying that release. For more context, my PR here<https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_EttusResearch_uhd_pull_572&d=DwMFAg&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=XUEEtUEfpaAEGxRI-WGuqHauOvsPdD2NZkfwDnwpYx0&m=CglWmCZNAv75-zoj2CB9dpWj4DRHkMpB8nWcwWdM2mk&s=m4xlgwVRVSi2686M6zZkXh8oqKef561kyaZzbOzO0kI&e=> has details. When I was using UHD 4.1, I discovered that I could bypass this problem by tuning to the same frequency twice in a row. That might be worth a try on your end if upgrading UHD is onerous.. Cheers, David From: Jim Palladino <j...@gardettoengineering.com<mailto:j...@gardettoengineering.com>> Sent: Wednesday, November 2, 2022 8:38 AM To: Marcus D. Leech <patchvonbr...@gmail.com<mailto:patchvonbr...@gmail.com>>; usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com> Subject: [USRP-users] Re: N320 LO stability problem Hello, Thanks for the responses. Yes, I have the same issue with tones off center -- we initially noticed the issue with some wide-band modulated waveforms. I did just try another flowgraph with a tone at 100KHz off of DC and confirmed that the issue does still persist. I'll try to experiment with other UHD versions today to see if that makes a difference. Thanks, Jim ________________________________ From: Marcus D. Leech <patchvonbr...@gmail.com<mailto:patchvonbr...@gmail.com>> Sent: Wednesday, November 2, 2022 8:31 AM To: usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com> <usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com>> Subject: [USRP-users] Re: N320 LO stability problem On 02/11/2022 08:11, Jim Palladino wrote: Hello, We have about ten N320s and almost all are exhibiting issues regarding the LO stability. It appears the LO is not locking at certain frequencies, or if it does, it barely maintains lock. We can see this with either a gnuradio flowgraph consisting of nothing more than a usrp_sink and a constant driving it's input, or using the included UHD example "tx_waveforms". The problem frequency I've been focusing on is 1112MHz. So, the following command demonstrates the issue: ./tx_waveforms --freq=1112000000 --wave-type=CONST --wave-freq=0 --rate=1000000 --gain=40 Some of the N320s seem to lock ok, and you can see a reasonable tone at the output. However, other N320's don't lock -- we will see a several MHz-wide "blob" about 4 MHz lower than the requested frequency. Note that they aren't reporting that the synthesizer isn't locked, but that is what it looks like. On units where the LO appears to lock, if I look closely on a spectrum analyzer, the phase noise often looks horrible, or I see large spurs around 106KHz off of center that slowly move up and down by 30 to 40 dB. It looks like it's barely maintaining lock. This issue varies by N320 and also by channel ("A:0" vs "B:0") on the N320. It doesn't matter if I use an internal or external reference -- the behavior might be very slightly different, but not much. Gain settings, sample rates, etc. don't seem to matter -- it appears to be an RF/synthesizer issue. I also tried enabling "spur_dodging", since that changes LMX loop parameters and that seemed to help in some cases (units/channels) but hurt in others. I've been focusing on the TX path, but someone else in my office was mentioning that they have seen the same type of thing when receiving -- the result was that 1 out of several N320s he was using to simultaneously receive a signal showed the same signal several MHz off of what the other N320s saw -- and it looked very distorted. I'm guessing that what he saw was the result of the LO not locking properly. We're using UHD 4.1.0.5 and associated filesystem, FPGA image. Any thoughts on this? Thanks, Jim _______________________________________________ USRP-users mailing list -- usrp-users@lists.ettus.com<mailto:usrp-users@lists.ettus.com> To unsubscribe send an email to usrp-users-le...@lists.ettus.com<mailto:usrp-users-le...@lists.ettus.com> Also, if you try sending a tone at other than DC (perhaps 10kHz), do you get different results? This might just be the DC-offset removal algorithm producing results that look like lack of synthesizer lock.
_______________________________________________ USRP-users mailing list -- usrp-users@lists.ettus.com To unsubscribe send an email to usrp-users-le...@lists.ettus.com
