Nate, I appreciate your suggestions and comments, even though we have differing opinions about accuracy. In my area, starting an ARES or MARS net a minute or two early or late is not acceptable. We pride ourselves at beginning the net exactly on the second. After all, we're Hams, and we have access to WWV, don't we? I realize that some folks on this forum may be rolling their eyes at that statement, but hey- if sloppy operating is okay with them, let them do their thing!
My obsession (yes, that's probably what it is!) with time accuracy began when I was Chief Engineer at WLRW, an FM station in Champaign-Urbana, Illinois, back in the late 60's. There was an IGM (International Good Music) automation machine that played music and ran commercials and IDs during the periods when live talent wasn't on the mike. The machine was designed to join the ABC Network News feed every hour on the hour, and being a minute early or late was not an option. The problem was that the AC power was locally generated, and was not synchronized to the national power grid as it is today. Even though the timer in the IGM controller made preparations to join the ABC Network exactly on the hour, the small variations in the AC power frequency caused the connection to be made several seconds off, either early or late, and the station owner was on my case constantly. He didn't want to spring for a Favag or Western Union precision time service, so I cooked up a crystal-controlled power oscillator to drive the IGM schedule timer with a TCXO-stabilized power source. It used a Hewlett-Packard oven time base at 10 MHz as a standard, making it easy to synch to WWV. It was a kluge, to be sure, but it worked. With this background information, perhaps you can understand that all I really need is some signal that occurs at exactly some point in time, every day, that can be used to synchronize a repeater controller automatically. Most real-time clock chips, including those made by Dallas Semiconductor, have sufficient short-term accuracy to "flywheel" through one day without getting more than a second off. If I can tweak such a clock once a day to bring it to the exact time, that is enough. I really don't want to add phone lines, IRLP links, wireless networks, or anything else to make this happen. It would be great if the next-generation repeater controllers had a BNC or TNC connector on the back labeled "GPS antenna" or "WWVB antenna" and all I needed to do was install one simple antenna, and the controller would know the time! 73, Eric Lemmon WB6FLY -----Original Message----- From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of Nate Duehr Sent: Monday, November 12, 2007 11:57 AM To: [email protected] Subject: Re: [Repeater-Builder] Dallas Semiconductor Real-Time Clock (Was RC-96 Controller Problem) Eric Lemmon wrote: > Don, > > The "on-the-hour" tone is an 800 ms burst of 1500 Hz. I have built a PLL > 1500 Hz tone detector into a Hamtronics WWV receiver, and it works fine- > giving me a relay contact closure exactly on the hour. Unfortunately, that > would only allow me to jam-set the minutes and seconds to zero, and would > not correct an hour error- such as when DST starts and stops. Eric, There are a number of "easy" WWV and GPS projects to drive things like Nixie clocks, etc... from simple microcontrollers like the Microchip PIC and Atmel AVR. Those are a good starting point for a project to set a controller's time remotely. Adding code to the microcontroller to then drive a DTMF encoder (or even an R2R ladder for sine-wave output from multiple digital outputs if your micro is fast enough) to set a controller's time, is fairly simple. One of the local clubs here in town has had such a system for a long time, but hasn't published anything about it. From talking with their techs, they receive WWV at a ham's house, set the clock in the Atmel, and then it has a transmitter on a common control receiver frequency for all of their machines. They had DST hard-coded to specific weeks of the year in their microcontroller code, and had to modify their code during the great DST mess that Congress created (with little to zero impact on energy use, which was supposedly their goal) the last couple of years. My club never built such a gadget, we just go in and bump the time around as necessary and don't get too wigged out if it's off by a minute or two. Everyone has network-synced cell phones in their pockets these days, and worrying about the repeater time just doesn't seem "worth it" at this point. We get it close and then have to deal with DST. We also got rid of the hourly chimes/announcements/etc. The only time you hear the time announced is after an autopatch, and that's really just in case we had a need to record the autopatch calls for abuse, etc. Building an auto-time set device and having another transmitter just seems like it breaks the KISS principle. As someone else mentioned, an IRLP node that is properly NTP synchronized can also handle sending DTMF time-set commands easily. Nate WY0X
