Member postings for Tony Jeffree

True...and I dread to think of their development cost. However, if you are running an Arduino or similar, there's nothing terribly difficult about the protocol & you should be able to get synchronisation down to within a few milliseconds, especially if you don't have to deal with any other kinds of traffic on the network.

I agree that power consumption is a problem, but protocol latency need not be. For example, IEEE Std 802.1AS, a standard I worked on back in the day, defines a time synchronisation protocol that allows such problems to be taken into account - it was originally defined for use in Audio/Video networks so that several nodes in a network could broadcast their media simultaneously, but the engineers at CERN implemented the protocol in such a way that it was possible for them to time-synchronise nodes around the LHC to within a couple of nanoseconds.

You make the 1 hour adjustment when you first fit the hands to the movement - simples. When you buy the movement, it has a register pin inserted in the back that locks the movement at its "home" time - usually 12:00:00 - so you fit the hands to 12:00:00 before pulling out the pin and inserting the battery. If you know it is tuned to Frankfurt you set the hour hand to 11 rather than 12.

Sounds like the module is misbehaving. The 3 radio clocks I have are all on the "wrong" side of the house for the signal (I am in Mull and they are on the North of the house) and one is mounted on an aluminium clock face, which is less than ideal, but all three function perfectly.

A colleague of mine used to work at Rugby before al the ULF stuff was moved. Apparently, the amount of radiated RF in the control room was such that the fluorescent lights continued to fluoresce after you switched off the power!

When they are working correctly, with a god signal, they update the time hourly or thereabouts, so the deviation is likely to be very small. I have a Casio radio controlled watch that only updates once a night.

There seem to be plenty of radio clocks on Ebay these days that claim to be MSF - I've ordered one up to take apart & see if I can extract usable pulses from it.

Yes, I had spotted that module on Amazon too - looks very interesting!

Apparently the Micro:bit radio is good for around 300 feet in clear air, so it might just be workable to distribute time signals in a house.

That would work! Actually, the BBC Micro:bit processor that I am using in my current clock is capable of broadcasting to other Micro:bit processors so that might be a simple thing to achieve - not sure about range though, I suspect it is similar to the range of Bluetooth so may not be good for the whole house.

I have 3 wall clocks dotted around the house which used to contain "standard" Quartz clock movements, but a couple of weeks ago I remembered that I had three MSF radio controlled clock movements squirreled away doing nothing, and have replaced the "vanilla" Quartz movements with them. For the benefit of those that are not familiar, these movements use the radio time signal from Anthorn in the Lakes (broadcast at 60 KHz) to keep the clock accurate, and they revert to a conventional Quartz movement if they lose radio signal for any reason. The great advantage of these movements is that they are always accurate and automatically compensate for GMT and British Summer Time changes, so you fit and forget.

So far, so dull. However, given my recent experiments with a free pendulum clock where I used a gutted Quartz movement as a cheap means of providing motion work (see https://www.model-engineer.co.uk/forums/postings.asp?th=171748 ), it occurred to me that these movements could be useful in other projects, given that it should be possible to extract useful pulses from the drive system. The movements I have here seem to have separate drives for the minute/hour hands (pulsed every 15 seconds) and the second hand (pulsed every second), and these could be used to drive other things, for example:

- Use the 1-second pulse in place of a pendulum to drive more conventional and visually interesting motion work;

- Use the 1-second pulse to drive a hit-and-miss synchroniser to accurately synchronise a mechanical pendulum & associated motion work;

- etc.

There are also receivers/decoders around that can extract the whole of the data stream transmitted by Anthorn, which includes year, month, day of the month, day of the week as well as time of day, so potentially one could produce a mechanical clock with perpetual calendar, taking the data from the radio signal.

Anyone ventured down these (or similar) paths?

You can certainly do that - apply two sinusoids, 90 degrees out of phase, to the two coils of a conventional stepper and it should give you continuous motion or near enough. In effect that is what happens in a brushless motor.

I bought the book too - makes an interesting read!

Talking of resonances, we have a long case clock that occasionally strikes a "duff" note - never the first of a strike sequence, and never more than a couple of times when (for example) striking 12. Nothing wrong with the striking mechanism as far as I can tell; I eventually concluded that sometimes the hammer strikes the bell at just the right point in its oscillation for the movement of the bell and the hammer to effectively cancel each other out.

Ugh - that one's a bit rough to say the least - some nasty resonances interacting by the look of it. Definitely in need of some damping somewhere. There's an awful lot of inertia in that drive system which isn't helping.

You could probably improve that somewhat by reducing the rotor mass and increasing the coil current...

Interesting - thanks. I had already started experimenting with magnetic gears...

I converted my Taig CNC mill some years ago to make use of DeskCNC as the controller software/hardware (writeups were in MEW at the time). The mill has been gathering dust since we moved to Scotland in 2014 and I have only recently been resurrecting the PC side - the original laptop that I used had a serial (COM) port, which is long gone from modern spec PCs, so a major hurdle was to find a suitable USB-to-serial adapter that would do the job. I had 2 of these lying around that I spent much time & effort on, unsuccessfully - mostly due to hardware in the adapters that was too old to have any drivers that would run on Windows 10. So, after much frustrating effort, including managing to get one of the adapters to sort-of work with an old driver, I bit the bullet and bought a new adapter, one of these:

<USB adapter>

Which seems to work nicely with the in-built driver options that Win10 provides. I thought it would be worth mentioning in case others have come across a similar problem.

Now to see if I can remember how to run the mill and cut a few parts...

There may be a short (or long!) delay while I get my Taig CNC Mill back up and running - looks like I will need to do some major surgery on the controller side as my Desk CNC software/hardware doesn't seem to want to play with a modern-spec PC. So currently looking at alternatives...

That is very sad.