[chrony-dev] Frequency synchronization without phase

2016-11-28 Thread Dan Drown
I've been thinking about NTP servers without easy access to a window or rooftop. They wouldn't be able to use GPS, but they could have some sort of local higher quality frequency standard to rely on. They would still need phase from a network source (and thus be Stratum 2+), but the

Re: [chrony-dev] Frequency synchronization without phase

2016-11-29 Thread Dan Drown
Quoting Miroslav Lichvar <mlich...@redhat.com>: On Tue, Nov 29, 2016 at 09:54:36AM +0100, Miroslav Lichvar wrote: On Mon, Nov 28, 2016 at 01:38:35PM -0600, Dan Drown wrote: > Additional information: > > https://dan.drown.org/stm32/adev-mix.png > ^ comparison of various ade

Re: [chrony-dev] Frequency synchronization without phase

2016-11-29 Thread Dan Drown
After letting it settle down after starting, the initial results covering 14 hours: https://dan.drown.org/odroid/run5/tempcomp.png Shows the tempcomp.log(purple, left axis) vs tracking.log(green, right axis) The tempcomp factor has moved 200ppb (with some high frequency movements), while

Re: [chrony-dev] Frequency synchronization without phase

2017-01-06 Thread Dan Drown
Quoting Miroslav Lichvar <mlich...@redhat.com>: On Wed, Dec 28, 2016 at 05:57:15PM -0600, Dan Drown wrote: That looks like this (green=Linux system's clock, purple=TCXO's measurement of Linux system's clock). https://dan.drown.org/odroid/run7/tcxo-acc-1024.png I've adjusted the starting

Re: [chrony-dev] Frequency synchronization without phase

2016-12-07 Thread Dan Drown
Quoting Miroslav Lichvar : I setup a quick benchmark of heating the CPU and on-board crystal with a bunch of cpu wasting processes. CPU core went from 51C to 70C (as measured by the on-CPU temperature sensor): https://dan.drown.org/odroid/run5.7/tempcomp-cpuheating.png The

[chrony-dev] disable RX timestamp adjustment patch

2020-05-25 Thread Dan Drown
This adds the "disablerxadjust" config option to hwtimestamp. This uses RX timestamps at the 1588 point (after SFD/before ethernet header) instead of the NTP point (after ethernet CRC). At 1G, this changes the timestamp (and RTT) by 752ns. To be symmetrical, you'd want this enabled on

Re: [chrony-dev] Frequency transfer in NTP

2021-01-28 Thread Dan Drown
Quoting Miroslav Lichvar : I guess most people here don't follow the NTP WG list. There is one feature proposed for NTPv5 that I think would make a big difference for chrony: https://fedorapeople.org/~mlichvar/ntp-freq-transfer/ Very interesting, thank you for gathering this data. I'm