Week of 4th September, 2017

Last Week

Next Week

Monday, 4th September

I had occasion to wander down to the local shops, today. One reason for this was to collect some eyedrops for Jane. And, of course, I got the wrong ones. Luckily, I kept the till receipt, so I should be able to exchange them for the correct variety.

The main reason for the outing was to bespeak my tame clock repairer, who is back from holiday. The French wall clock needs more attention to its pendulum, so that I can regulate it properly (it currently gains about 5 minutes a week) and also investigation of why it only goes for 3 days. I suspect that this is because the strings that support the weights are incorrectly wound on the barrels, and do not unwind fully. Anyway, he will attend on Wednesday evening, and probably take the clock away for attention, under warranty, since this was Mum's clock, and he fettled it literally weeks before she went into hospital for the last time, this time last year.

Once that's done, the Gustav Becker needs attention for its tendency to stop at about 10 minutes to the hour, and also to get the hands in sync with the strike (currently it strikes when the hands say 3 minutes to the hour, and I don't know how to adjust that)

My parents' Korean clock is going a treat. It loses a minute or so a week, but I hope to have adjusted that out at last winding time - yesterday.

I'd like to regulate all 5 currently going clocks properly, and start collecting stability statistics for them. And while I'm at it, stability statistics for the mains frequency are another item of interest. I can't justify the cost of one of Brian Mumford's MicroSet timers for the amount of use it'll get, even though they're the gold standard for rating clocks, but Allan Deviation statistics are easier and cheaper

Back to Daynotes


Tuesday, 5th September

The Register brought to my attention the fact that asteroid 3122 Florence made an approach to Earth over the weekend. Not a close approach - about 7,000,000km at its closest - but a rock 4.5km wide would seriously ruin your day if it hit us. There's no danger of that for Florence until some time after 2500AD.

Just to point out the size, radar observations from Arecibo seem to indicate that Florence has two moons, or asteroidlets, of between 100 and 300 metre across, with orbital periods of some few hours.

And, of course, someone made a Magic Roundabout joke in the comments to the article.

The NASA Cassini probe will make its death dive into Saturn next Friday, after some aggressively risky final orbits around the planet. This is because the probe is almost out of manoevring fuel, and the project scientists don't want to risk an impact with one of the Saturnian moons at some indeterminate point in the future. Everybody will be tracking it as it makes that dive, particularly NASA's Canberra DSN tracking station, which will be optimally placed to see the final moments.

Back to Daynotes


Wednesday, 6th September

One of Jane's childminding clients had to nominate another person to collect her son this evening. Of course, he forgot, which meant that Jane missed tonight's Prom concert - Prom 71, with the LPO under Vladimir Jurowski, playing Shostakovich' Symphony No 11 among other works. Jane was not pleased, because she couldn't leave me in charge, and had to miss that work (one of her favourites) So I was instructed to take Jenny's sandwich and ticket to her, and then chauffeur her to the local Tube station, from whence she would excurse to South Kensington, meeting Sarah at the Albert Hall.

And while I was doing that, the clock repairer attended, and took away the French clock for a little remedial work.

So, barring Jane's inability to attend the concert, it all went well enough.

Back to Daynotes


Thursday, 7th September

I collected last night's concert from the Radio 3 transmission, via the Pure Evoke 3 DAB receiver. There was another Prom concert immediately afterwards, and the Pure receiver cannot do 2 overlapping recordings at the same time, so I had to grab one long file to cover both concerts, and then split it in audacity, which I've just done. Had to save as .mp3 rather than the original .mp2, but that's no great problem.

That said, the follow-on concert was aggressively modernist, of which I am not a fan, so I only collected it for completeness.

It's been some time since the BBC broke get_iplayer on torrentpi, by removing the metadata upon which V2.99 and earlier depends, so I've been running V3.0 on inspiration. Today, I decided to see if there's an updated build of V3.0 or higher. You will find it here, which contains a one line shell command that will install the latest version -

wget http://packages.hedgerows.org.uk/scripts/install.sh -O - | sh

which I copy-pasted into a terminal window. It proceeded to update get_iplayer, in-place, to V3.01. in less than a minute. This is good, but I have background security worries about effectively piping an unknown script into a shell, over http, yet! This is not, in the general case, a good idea, so don't do it unless you trust the source.

The problem now, of course, is to avoid redownloading material I already have, which means I'll be getting material via index numbers, or PIDs, for the next month, rather than from a search listing directly.

Using V3.0 on inspiration, the download speed ranges from 20-odd Mbit/s bursty, to 55Mbit/s or so, depending on which of the BBC's CDNs I hit. This appears to be a random choice, which doesn't really matter, since they all work. It remains to be seen what torrentpi will be able to do.

Back to Daynotes


Friday, 8th September

I fired up get_iplayer on torrentpi today, and found that download speed was up to about 45Mbit/sec, but the post-processing in the 3-stage automated pipeline was noticeably slower. Also, there's a difference between the Window$ and Linux implementations, which I'm sure is because of OS differences - on Window$, ffmpeg, the second stage that strips the .ts wrapper off the downloaded file, shows every status line as a scrolling list, whereas on Linux, each status line overwrites the last one on screen. The 3rd stage, atomicparsley, behaves the same on both OSs.

Still, it works - speed is not of the essence here.

After the snafu on Wednesday when I wasn't here to meet Dariusz, my tame clock repairer, when he collected the French clock in order to fettle a couple of minor niggles, I attended his workshop, to brief him on the short going time. And he'd already noticed, and fixed, it. It's on the wall there, ticking away nicely. I should get it back next week, properly regulated. Then it'll be the Gustav Becker clock's turn - to get the strike problems fixed.

I've commented before about the inconsistent dating evidence for this clock. Becker stopped using weight-driven movements in about 1880, whereas the serial number and logos on the back plate of the movement suggest a date in the mid-1920s, and Freiburg provenance -

As you can see, it's a P64 movement. The stamped logos say Freiburg, and their relative positions suggest later construction. The serial number is (1)882547 (the 1 is hidden behind the pendulum crutch) which dates to the mid-1920s, according to online sources. But it's weight-driven, which online history says is pre-1880. I'm stumped.

The photo is a Nexus 5X shot, cropped and resized to show the relevant detail.

While I was visiting the clock shop, I collected my latest repeat prescription, so I'm good for another month with the blood pressure caps.

Yesterday evening, in preparation for the Steam Dreams trip tomorrow, Jane insisted on vetting my clothing plans - Steam Dreams dress code for Pullman Style Dining, the class we are travelling, is "smart casual", so sports jacket, good trousers, shirt, with tie optional. Definitely not Milty Travel Vest.

She spent about an hour exchanging messages and photos with one of the girls (Sarah, as it turned out) and finally settled on a combination - pretty much what I'd have chosen anyway, but it was more a case of, "You will wear this..." Henpecked, moi?

Given the dress code, a MOLLE-equipped shoulder bag is probably not the best idea. So I looked out, and packed the big Fossil leather messenger bag. I only really needed that because of the FiloFax, which was unnecessary.

Back to Daynotes


Saturday, 9th September

Settle and Carlisle Steam Rail Excursion

This is a trip from London Euston station, up the West Coast Main Line to Carlisle in Cumbria, returning along the Settle and Carlisle Railway to Preston, and back home from there. Vintage electric traction - 86259 Peter Pan - south of Preston, steam to the North - ex-LMS Jubilee-class 5690 (45690) Leander.

This is almost an exact remount of the abortive December, 2015 trip, except that on that occasion, our steam loco would have been 5699 (45699) Galatea, another LMS Jubilee-class 4-6-0.

So meet Sarah and Jenny at the Tube station, for the 06:05 Central Line departure, to arrive at Euston at about 06:40 for the excursion departure at 07:10. A quick stop for coffee, and we boarded - Pullman Style dining car all complete, 3 of us around a table for 4, with the 4th seat filled by a charming gentleman named John. The table for 2 on the other side of the aisle, which we think would have been occupied by Jane and Katy, remaind unused.

So off we set. Curiousity about loco speeds, particularly steam-hauled speeds, had persuaded me to bring my old Garmin GPSMap76, which has a nice GPS-driven speedometer display. I set that up against the carriage window - east side of the train - and initially had trouble getting and maintaining GPS lock. But that settled down, and I could see that Peter Pan was maintaining 80+ mile/h speeds as we travelled north. She is limited to 90mile/h by Network Rail fiat.

During the trip North, a full English breakfast was served, together with coffee and tea on request. And very good it was, too, especially under high-speed mobile conditions.

At Carnforth, north of Preston (further than we managed in 2015) we stopped to change locos. This was not in a platform, so we were not allowed off the train to watch. And then off we chuffed, behind Leander, leaving Peter Pan behind, which slightly surprised me - I was expecting Peter Pan to couple to the rear of the train as a backup. But then it dawned on me - the Settle and Carlisle is not electrified.

This part of the trip includes two of the most severe climbs on the UK rail network - the Northbound ascent to Shap Summit up a 1-in-75 gradient on the West Coast Main Line, and the 15 mile "Long Drag" southbound ascent to Ais Gill Summit on the Settle and Carlisle, up a 1-in-100 gradient. I was particularly interested in what speeds Leander could maintain up these gradients, wirth the usual 12-coach rake of vintage BR stock.

I was not disappointed - Leander managed to reach the Network Rail mandated speed limit of 75 mile/h for steam for considerable periods on the level, and even on the Shap ascent she never dropped below 20 mile/h. The corresponding performance up the "Long Drag" was 30 mile/h at the summit. Ais Gill is 1169 feet AMSL, Shap is only 916 feet, but steeper.

Of course, between these two climbs was a break of a couple of hours in Carlisle. I didn't feel like traipsing around looking at the sights, so I settled myself in a cafe while Sarah and Jenny did the rounds. The I returned to the station to await Leander's return to the train after turnaround. This let me get a better photo of her, taken from Platform 3, as she stood at Platform 5.

And these are from just after arrival. Note the crowds.

During the stop, I confirmed that the original Nexus 5X still will not do mobile data. So I fired up the Huawei Wingle, and discovered that Three's mobile network is not having problems with data. OK, it's the Nexus. So I researched fixes - one of which was "restart the device". I thought that was only for M$, but it's broken, so this can't do any harm.

Ta-daa! Fixed.

But monitoring train progress on a moving map had caned the battery on the newer Nexus 5 - the Vodafone one. This is my comms terminal - I need this working. So I put it on charge, from the TeckNet 12000maH USB battery. It took roughly forever to charge, which suggests that the battery is not really fit for purpose - that purpose being recharging my phones on the road. It does run the Wingle fine, though.

The other main feature of the trip is the Ribblehead Viaduct, near Settle, which has 24 arches, is 440 yards long and 165 feet high, and is unimpressive from a train travelling on it - you really just notice the ground is rather far below you. The viaduct itself is virtually invisible, even if you are on the left (east) side of the train. Because of the left curve in the track, the parapet is only visible from that side. I posted Jenny's pictures of this here.

We needed two water stops to refresh Leander, even on this short a trip.. Network Rail are evidently paranoid about steam engines running low - they mandate a massive reserve at all times. I suspect that the reserve amounts to half a tender-full of water. The first top-up, at Appleby, went fine, but we ran into a problem at Helliwell Goods Loop - a points failure (Merkin: switch - I think. Anyway, it's a track device to allow changing tracks - as it might be, into a separate parking area) This resulted in about a half-hour delay.

After the loco change, back to Peter Pan, who had been patiently waiting for us just south of Preston, we departed south, at speed, and were served a full 3 course evening meal of roast lamb and trimmings. Again very good.

We made up time on the trip south, to the extent that we were only about 10 minutes late into Euston, and after an uneventful Tube trip, Jane collected us from the station.

So a very good day. Leander performed superbly, and service on the train was excellent. I don't think we'll make many more Pullman Dining excursions, but as a one-off, it was very good. Made even better by the fix for the Nexus 5X mobile data problem, and a swift download from BBC iPlayer afterwards.

Back to Daynotes


Sunday, 10th September

Pretty much finished with the BBC Proms recordings for 2017, after a quick burst with get_iplayer to collect the final TV coverage. Next up, collecting Simon and family from Gatwick, this afternoon.

But before that... It has been announced that Dr. J.E. "Jerry" Pournelle, science fiction author, blogger, and scientist, has died, peacefully, in his sleep, soon after his return home after a visit to DragonCon. I have been a fan of his writings, both fiction and blog, since the days of Byte magazine, wherein he wrote the User's Column, which later became "Computing at Chaos Manor", as he whimsically called his home. Rest in Peace, Jerry, you will be missed.

The Korean clock appears to now have a minimal rate error - less than a minute a week. Which is good. Past experience suggests that it will keep this rate for the next several years, at least.

And I've discovered why the TeckNet battery was slow at charging the Nexus 5X. The battery's two USB ports are not identical - one of them, the one I was using yesterday, of course, is marked with one lightning bolt, the other has two lightning bolts, and does charge faster. Not enough to cause a Nexus 5X to say "Charging rapidly", but at least it's not "Charging slowly".

Later: The trip to Gatwick to collect Simon and family went well, albeit there were jams on the M25 (so what else is new, I hear you ask) Took about 3 hours for the return trip, door-to-door. And I timed it well - they walked out to the agreed pickup place at Gatwick North just as I was driving past the entrance to Gatwick South.

Last Week

Back to Daynotes

Next Week