Assuming I'm working with a UBlox LEA-M8F, and a Stanford PRS-10: The LEA-M8F has a 30.72 MHz output, along with pulse per second and another configurable output. Are there any advantages to be had doing anything unconventional rather than simply applying the PPS to the PRS-10? Priorities would be fast convergence to GPS second, without adding phase noise to the PRS-10. Within reason, power consumption and component count need not be minimized. Is there anything to be gained from delaying the 30.72 MHz output from the UBlox a few nanoseconds if necessary (coax or shielded differential pair) and using logic to AND the PPS output and the 30.72 MHz output to pick off a narrow pulse that would be biased to the trailing edge rather than nominally centered about the jitter? The PRS-10 could have negative delay on the PPS to eliminate the propogation delay. Similarly, is there a "better" frequency the second output could be programmed to put out that tight have less jitter? What I'm after is shortening the time constants, at least at startup, at the cost of complexity, without negative effects on phase noise. The purpose of the hypothetical device is to provide a master 10MHz to my bench, along with a reference for Sync-E clocks to transfer 10MHz optically to a tower (back channel of voice with software defined radios in a weatherproof enclosure. Somewhere along the way, it will also need to provide a 1.544 MHz BITS clock. I know...sounds sort of silly for home use, but with just the right gear, the network hardware costs less than a used PRS-10. Then again, I am sending this to time nuts.

On 4/2/25 3:19 pm, Ed Marciniak via time-nuts wrote: > Assuming I'm working with a UBlox LEA-M8F, and a Stanford PRS-10: > > The LEA-M8F has a 30.72 MHz output, along with pulse per second and another configurable output. Running from memory here as it has been about six years since I played with it, but the M8F can discipline an external oscillator on a variety of frequencies, I'd intended to use 27MHz for a project, and built all the prototype elements, but never actually did the testing. The 30.72MHz is _ok_, but nothing great, to the extent that it's not really worth using the M8F in any actual product, you're better served by using the underlying chip and just configuring it to manage the oscillator you want ... assuming uBlox will sell you them, which was an issue back then. > Are there any advantages to be had doing anything unconventional rather than simply applying the PPS to the PRS-10? > > Priorities would be fast convergence to GPS second, without adding phase noise to the PRS-10. It's worth specifying what "fast" is here, some folk here would consider that to be sub-minute, some would think sub-week. > The purpose of the hypothetical device is to provide a master 10MHz to my bench, along with a reference for Sync-E clocks to transfer 10MHz optically to a tower (back channel of voice with software defined radios in a weatherproof enclosure. Somewhere along the way, it will also need to provide a 1.544 MHz BITS clock. I'd be tempted to use CWDM and do things as separate frequencies, but see previous posts here with me complaining about synchronous networks, although as long as it's just a local link most of my issues don't apply. Do you need the BITS clock up the tower or only at the bottom, I'm sure _someone_ sells something that'll generate one from syncE, that's designed for outdoor use, but it's probably unreasonably expensive, and I tend to be dubious about the actual quality of implementation (this goes back to my ranting about synchronous networks) largely because I doubt most customers actually measure the quality of them (or even could). Another option for the BITS signal would be to find a Spectracom 8140 tap already configured for them, if you keep an eBay search going one will probably pop up cheap within a few months. I'll drop the obligatory mention for the White Rabbit work, https://ohwr.org/project/white-rabbit/-/wikis/home

Hi There is no advantage to running anything other than a PPS into something like the PRS-10. It’s going to do the “same thing” with an RF signal as with the PPS. In terms of occasionally *calibrating* the PRS-10, the PPS input works great. Run things for a couple of days and you probably will be pretty close to “tuned right”. In terms of replacing a GPSDO, the firmware / filters in the PRS-10 are not set up to handle the strange things GPS does long term. They designed it for quick calibration rather something fancier. Bob > On Feb 3, 2025, at 11:19 PM, Ed Marciniak via time-nuts <time-nuts@lists.febo.com> wrote: > > Assuming I'm working with a UBlox LEA-M8F, and a Stanford PRS-10: > > The LEA-M8F has a 30.72 MHz output, along with pulse per second and another configurable output. > > Are there any advantages to be had doing anything unconventional rather than simply applying the PPS to the PRS-10? > > Priorities would be fast convergence to GPS second, without adding phase noise to the PRS-10. > > Within reason, power consumption and component count need not be minimized. > > Is there anything to be gained from delaying the 30.72 MHz output from the UBlox a few nanoseconds if necessary (coax or shielded differential pair) and using logic to AND the PPS output and the 30.72 MHz output to pick off a narrow pulse that would be biased to the trailing edge rather than nominally centered about the jitter? > > The PRS-10 could have negative delay on the PPS to eliminate the propogation delay. > > Similarly, is there a "better" frequency the second output could be programmed to put out that tight have less jitter? > > What I'm after is shortening the time constants, at least at startup, at the cost of complexity, without negative effects on phase noise. > > The purpose of the hypothetical device is to provide a master 10MHz to my bench, along with a reference for Sync-E clocks to transfer 10MHz optically to a tower (back channel of voice with software defined radios in a weatherproof enclosure. Somewhere along the way, it will also need to provide a 1.544 MHz BITS clock. > > I know...sounds sort of silly for home use, but with just the right gear, the network hardware costs less than a used PRS-10. Then again, I am sending this to time nuts. > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com