Oh dear!

[mcrossley]

My web site went offline yesterday afternoon, no response from the server at all. I wonder what awaits me when I get home, a dead laptop I suspect.

If it comes to the worse, has anybody got good (recent) links for a cheap Windows box?

I was looking at these - they seem pretty good value at £120 after cashback, but is there a better buy out there?

[water01]

I bought one of those Mark, very good value and a very good computer. Only problem I had was I wanted to install W7 but Acer only have W8 drivers on the installation CD. I found all the W7 drivers though and they work perfectly and I can package them up in a zip file if you would like them.

I searched for a while and could not find a similar speced machine for such a good price. Looked at Gigabyte Brix, Foxconn, Shuttle and Zotac but this was the best for the price, and the Cashback works!!

[AllyCat]

Hi Mark,

You do realise that the linked item only comes with "free DOS"? Bundling Windows 8.1 adds about £80 to the cost. Still not a bad price, but you might find a "refurbished" Netbook/Laptop at a similar price, which would include a screen.

Cheers, Alan.

[mcrossley]

Thanks Mike.

Hi Alan, yes I do, I have a full MSDN subscription, so an OS isn't a problem.

[mcrossley]

I had hoped Steve would have the Pi version running before this laptop have up but the disk drive was a bit noisy when I left earlier this week.

[steve]

Work is a bit crazy at the moment, and has been for a while; I did a 14 hour day on Tuesday for example, and I'm spending so much time writing software at work and working around problems in third-party software that I really don't have the energy for Cumulus development. This is one reason I've had even less patience than usual for people on the forum being 'difficult' with me. I'm supposed to be on leave at the moment, but I had to cancel it.

However, I am taking three weeks off at the end of the year, and hopefully I'll be able to get back to the raspi version of Cumulus. It's actually not very far off being usable (with a FO station) for someone as technically able as you. It doesn't actually have any UI, so data can only be viewed via web tags etc.

[mcrossley]

No problem Steve. I'll have to wait for the VP2 code anyway. A bit busy myself at the mo. Currently stood in Euston awaiting a train home.

[water01]

I have a raspi as well and I am reasonably techie competent so if you want any other Alpha testers I am willing!!

[mcrossley]

Aye, same here. I have the old FO kicking around too. I think the only thing my website uses from Cumulus (apart from the web tags) is the trend graphs, and they could be created on the webserver anyway as I store the real-time data in a SQL table.

[steve]

As this version doesn't produce graphs itself, I modified something you did, to produce the graphs for the web site: http://nybbles.co.uk/test/trends.htm.

[mcrossley]

Panic over, it was powered off. Powered it up and it booted fine, nothing in the event to indicate a shutdown, so it looks like a hardware power off.

Those graphs are fine for a web page Steve, but I use images for the steel series gauges. I did have some server side graph generating routines at one time, I'll have to see if I can find those and get them to save to file periodically, or when requested, after checking for a recent cached copy first....

Hmm, this is the first time I have had to do a catch up from the VP2 logger, it looks like it plays havoc with the wind speed records I'll have to sort that out later this evening.

[steve]

Those graphs are fine for a web page Steve, but I use images for the steel series gauges.

There is a charting package available which produces graphs vaguely similar to the ones that Cumulus 1 produces, but I was really hoping that all the graphics in the next version could be done in the browser.

Hmm, this is the first time I have had to do a catch up from the VP2 logger, it looks like it plays havoc with the wind speed records

In what way? The same information isn't available from the logger compared to running live, so the average wind speed calculations do tend to go slightly awry (and it depends on the logger interval). Our power tends to go off when the wind gets really high, so the average wind speeds afterwards do get a bit skewed.

[mcrossley]

Hi Steve, yes it's the average speed. I haven't had a chance to examine it yet, but the average seemed to only be a little below the gust during the catch up and it broke my all-time record.

Bad news is the laptop powered off again. The last event both times in the system log was MSE updating the virus definitions.

[steve]

Hi Steve, yes it's the average speed. I haven't had a chance to examine it yet, but the average seemed to only be a little below the gust during the catch up and it broke my all-time record.

Yes, it's because the station-calculated average isn't available in the logger data, and the Cumulus calculation has very few samples to work with - it depends on your logger interval, e.g. with a logger interval of 10 minutes it only has one sample to calculate it from. At best it has 10 samples, with a 1-minute logger interval.

Bad news is the laptop powered off again. The last event both times in the system log was MSE updating the virus definitions.

I'll get working on the Pi version

[mcrossley]

Got it sorted - I hope - there was a thread associated with Windows Updates (not the update service itself) running at 100% CPU, eventually the laptop shut down due to overheating. I suspended that thread, allowed the monthly updates to run, rebooted and it all seems OK now.

I removed the wind records generated today (oddly the max wind speed - 31.2mph - was higher than the max gust - 30.0 mph). Ah! As I typed that I realised why, I have a slighter higher multiplier on the windspeed than gust (1.3 v 1.2), and as I have a 10 min logger interval, the raw values will be the same.

More good news is I should soon be able to drop those multipliers as I bought a secondhand anemometer wireless transmitter today so I can get the anemometer up above the roof level.