Lightning protection - SPD

Installed a ADEE  VAR15GT5EF T2 in a separate 2 module enclosure adjacent to the main (500mA) trip. 15kA/275V. This was partly to meet the 0.5m rule on phase wiring, and partly as there wasn't really space in the existing board.
The device is internally fused, so doesn't require a separate isolator to meet NF15-100. This does mean the supply will have to be isolated to swap it. It's wired with 16mm incoming, but the earth was just unmanageable, so that's in 10mm. 6mm minimum. The instructions show the main earth to be looped through the spare earth terminal. The thin bit of wire you can see linking the terminals (they might be linked in the potting too, but why do the job twice?) would represent a significant impedance under a overload. Earthing stayed at the earth strip. There was, in any event, not a chance of getting 2 x 16mm cables in; they would have gone into the terminals ok.
Green led is lit, and can be detected with a LDR.
Hopefully this will be enough to reduce future spikes to low levels.

Frying tonight

What goes flash, <pause>, crack, <pause>, rumble, <pause>, bugger that sounds expensive.
Well,
flash - lightning
crack, router plus a ton of other stuff arguing about whose voltage is bigger
rumble - thunder in the distance
bugger that sounds expensive - me.

We had a thunderstorm on Sunday night, 12/8/18. We get a good number of thunderstorms and today's was average. Lots of rumbling and flashing, but not close to us. Tonight we'd had 2 blackouts of ~2s, then there was a flash outside and load crack from the main distribution where the VDSL router and IT stuff lives. The thunder suggested the lightning was not near by!

18/8
re-fitted the phone line protection; this was installed with the ADSL router in 2010, but reduced the speed too much, so got removed again. We got upgraded to VDSL, and I never tried the filter. Today, I have just tested the speed with and without. We get a 97/12.3M connection to the local box, which was reduced to 90/12.0 Using speedtest.net, the speed went from 12/42 to 11/40. This change is probably within the bounds of variation during the test; we're usually throttled back to 12/30, so it looks pretty good today.  The filter can stay until they give me back the bandwidth to 97M !
I do need to separate the phone line wiring from the CAT5, and put an earth connection directly from the earth rod to the phone socket, rather than going via the dis board (normal install) - done.

The death toll in active components is quite impressive, as is some of the damage (splash marks)

Cabled weather station (PIC16f884 working, but comm port dead)
Main processor (CPI) (PIC32MX775F512H and power monitor ADE7953. PHY + EEPROM ok)
Main processor IO (opamp for water meter, 1W i2c conveter ds2482-800, 1W mux ds2408 74HCT244, BMP180)
Caller ID (ESp8266 + dsPIC33FJ128GP, opamp. EEPROM 25l256)
12V to 3.3V/5V regulator AC, 5 + 12V fuses, plus regulator chip+fets 
12V battery backup. 2  Fets that switch 12V.  The other 2 that operate battery undervolt were energised at the time and ok. Both 1W devices
5V relay for garage door power - tracks fried from live to +5, plus protection diode to 0V. Works ok after clean. Relay believed off at the time.
Garage 1W dual IO DS2413.
mains SMPS for all devices  - both main diodes, both output caps swollen, a zener crowbar
LF356 on burst controller. pic ok.

All 1W temperature devices (2 wire) all survived
Rpi connected via ethernet cable, but unearthed psu ok
New burst controller / power monitor running direct off mains ok.

Roof leak

A stain appeared on the lounge ceiling, east side 1m from the doors. This is in a tricky spot that is difficult to get to from  either the inside or the outside. There is a gully that had 2 clumps of leaves, and the wet ceiling is between the 2 clumps. We had just had 45mm of rain. From the south guttering the leak is 3rd row up, under the gully. 

chainsaw service

cleaned all dust out, new air filter, cleaned plug.
Removed grease plug from head and put ~2ml of GP grease in with a syringe. No obvious wear.
Plug is a bit knackered and is probably T27, although a T30 has been jammed in, but doesn't bottom out. Was glue locked. Needed a hammer wrench to remove. Blob of blue glue put back.

Wheelbarrow

Changed the tyre 4.0/4.8-8 (~16" diameter) on the old wheelbarrow and the inner on the newer one 3.50-6 (tyre ~12"). Couldn't find the hole int he old tube, and had changed the valve.
Old trailer tyre fitted on old barrow but the width meant it rubbed the wheel protector

UMIDIGI portable wireless stereo speaker UMI-BTS1

This speaker is really well made with a good attention to detail. All joints are sealed well.

To open, remove the rubber foot, free the antenna. Slide guitar picks to release the 4 tabs see pic of ali case. Slide large plastic part out, being careful to not kill the antenna. 6 screws to remove the back. No battery connector. You can now remove the front grill, but it doesnt give any more access.

2 speakers and 2 ports.

It tells you what it is doing at full volume. Really annoying. Also, if you power it up with aux plugged in, it doesn't recognise it, thus you have to jack in/out again. A small microprocessor (pic10f220) was added to address most of the problems; it detects the aux jack and mutes the voice. It mutes at power up, and when the off button is pressed. It asserts the 'jack present line' at power up for 6s then re-connects it; thus if there was a jack plugged in, it is now detected; if not it enters BT as normal.

I'd buy another, but not to modify. I'd bought it to fit a small pcb inside with an internet radio. Unfortunately the aluminium case will limit the wifi range; the BT antenna is not ideally placed currently being on the base under the rubber foot.

A 5V port was added to the rear; 5V as I don't know what the internal 5v to 3v3 regulator is rated for - it was only just giving 3.1V. The 5V comes from a mt3608 boost from the 3.7V li-ion cell which is good for 2A.

Pics showing the mods. A new board with a pic10f220 on the left to mute the full volume voice prompts, plus forcing the unit into AUX on power up.

Deck maintenance

The deck will have been down 7 years in a few months. For the last year, bits have been a bit firmer, or a bit saggy. A check last year showed a few joists had rotted out, which is pretty normal for untreated Douglas Fir, particularly when it gets wet from on top, and the bottom and sides have increasingly been filling with leaves and clay.

Spot the carefully placed flower pot.

It seems that the spring rain the last 3 years and mediocre summers, plus the northern part of the west face rarely sees sun, has taken its toll. Originally I left quite a lot of clay below and in between the joists, which liquefied and leveled out. This is slowly being removed to reveal a much higher water table than 2011 (now about 18cm below DPC). The weed suppressing matting also wicks water, so any in touch with wood was rotted. The hop had spread its roots a long way (mainly to the low spot where the new drain is) and were chopped out; the roots near the original plant were put back under the matting. A few bits died, but it's started sprouting again.  The vine didnt seem to have many roots. The honeysuckle had done a good job of keeping clear of my digging, and seems ok.



This is about 56 tiles up, out of 159.
All this clay is being barrowed uphill to fill an old ditch. The rain continues to fall which limits the load, as you just cant get grip. About 2m3 in total. 

The tiles also needed cleaning and oiling. They were jet washed first, but hand prep prior to oiling. 56 tiles prepared and painted in 4 hours. Paint is 1L cold pressed linseed oil, 200ml turps, 30ml siccatif (drying agent banned in some countries due to toxicity. Speeds drying)
Stacked on end to dry. For all the tiles, I used just under 3L of oil for 1 coat, nearly 2L of glue D3 glue and almost 800 25 x 4.5 screws.
All joists where covered in damp proof membrane. Time will tell if this is a bad idea; previously the joists soaked up the water. Now the water sits on the membrane and may rot out the tile bases.

Tiles then relaid temporarily on new joists, and another 50 taken up. More temporary joists, and the remaining tiles removed under the covered area (different look to that batch - several marked Burmese teak) and replaced with the batch from the kitchen end. The house isn't quite square, so there is a bit of a twist from the covered area to the main area. A reference line of tiles is screwed down at joist 10 in the long run, but the deck just isn't square towards the kitchen.
9 joists replaced, and some cut down and swapped. 1 joined as I couldn't face buying more wood.
A gravel filled trench runs the length of the deck and drains from the middle to the hedge (in line with the rose). This will join up to a new drain slightly further uphill; a job that has been on the list for a couple of years. The trench finally went dry about a week after the sun came out; it nicely filled when it rained and flowed well.


D:\photos\Garden\2018\Deck Renovation