Angles Mort

I saw this on a truck (non UK plates), and they do puzzle me a bit.

For a start, it is not clear - are the black bits the death angles - in which case why the big black areas in front left and right where there is clear diver visibility? Or is it the yellow bits with the warning triangles, in which case why the bits left and right which are also clear driver visibility. It makes no sense!

But the bigger issue is how we seem to find it quite acceptable to allow death machines to be driven on the public roads like this. I mean that is what it is saying - angles mort - death angles - places where you could die even though you might be a legitimate road user legitimately in such a place on the road.

Surely if there is a dangerous machine, a machine that can kill people, we need to make it safer, including mirrors or even cameras if necessary, and driver training, to ensure there are no "death angles".

To be honest this seems a lot like victim blaming.


Pet hate number 0̷

I have mentioned before, but one of my pet hates is crossed zeros.

I have gone in to the history a bit (here), but I still encounter these. The latest is a new calculator.

I know it is perhaps weird to own a calculator, let alone buy a new one. I previously used an hp 12c Platinum which hp re-launched, but it is a tad more "accountant" than "scientific" for my usage. So when I found Swiss Micros series of new versions of retro hp calculators I ordered their DM42 which is based on the hp 42-S.

It is impressive... It also has fun screen savers as you can see. Nice touch, well done.

Someone reported they did not like the buttons, but they seem good to me.

But I only really have one gripe. All of the fonts (of which there is a choice) have crossed zeros.

A crossed zero is a bad choice - something that is pretty much always displaying numbers, like a calculator, has no excuse to not show a zero as a round shape, but something like this where it has a high resolution display has really go no excuse. It can show a 0 and O distinctly with no problem.

This is a massive usability issue - a crossed zero, and even a "dotted" zero, can be easily confused with an 8 by anyone with poor eyesight. There is no excuse. What is worse is it has settings and font styles and sizes, and they did not think to include choice of crossed, dotted, or neither. Why?

To be clear, this is a zero, as you know...

What is especially weird is the Swiss Micros webs site shows it with the slightly better dotted zero.

But for some reason, what I have, does not have this, but has a crossed zero.

So basically I have not been supplied with a product as described, which is always annoying.

The good news is that it looks like this is using open source code, so it may be possible for me to re-make it with a new font. I'll see what I can do.

Update: You have to love open source.

I have not actually managed to find where the fonts are though, I just changed 0 on the display, which is not ideal as I would like a visually distinct 0 and O, but it is a start. Nice that you can set a sane date format though.

Recording multiple cameras on a Mac

I wanted to know how to do this, and as usual, with anything video related, all I found were youtube videos. Some times I'd just like to know then answer without watching a video. So this is my effort to explain in simple text.

The scenario is simple. I have two studio monitors, which include camera and mic. I wanted to record on both cameras at once so I could cut from one to the other (see example video here). This was basically an "interview with myself" video. Obviously this could just as easily be multiple web cams. Computer performance is also a concern, and more so if 4k recording.

  1. It seems Final Cut Pro does not record multiple cameras at once, shame.
  2. It seems there are commercial applications that will do this.
  3. OBS can merge multiple cameras in to one virtual camera, but that is not the same and not then easy to make a multi-cam clip. It seems a plug-in may help.
  4. It seems Quicktime Player can record video, but only one camera/mic.

But there is a simple answer: The trick, which I was dead impressed with from this video, was you duplicate your Quicktime Player application. Literally select it in the Applications folder in finder, and select duplicate. You can then run two separate instances of it at once, and set each to record separate camera/mic, start both, and sorted.

Saving both files, you can then load them in to Final Cut and make a multi-cam clip synchronised by mic, edit views, and bingo. And obviously you could put a bit more effort in to is that my example video, but it shows the principle.


Weird science

LED panel (when switched on)
I have LED lighting panels in my bedroom, two of them. Nothing complicated.

They each have a power supply behind them in the ceiling, and connect to the power with just live and neutral (no earth).

The live goes via a switch, obviously. It's a Shelly, so a clean relay contact.

So, when they are off, they should be off. Simple. They only have the neutral connected at that point.

The weird bit

However, during the night, I can see that they are not quite off. They have a faint glow some of the time. It can only be seen in the pitch black of night, and when your eyes have adapted to the dark. So if I wake up in the night I can see it.

At first I did wonder if I was imagining it, it was so faint, but no, I have seen them glow turn on and off, not fading in/out but sudden on/off. So very real. They are not glowing all the time, and not at the same time necessarily - sometimes one or the other. But most of the time as far a I can see, i.e. if I wake up in the night they are glowing. And frankly they are annoying me.

But with only neutral connected I don't see how this can happen at all!

The weirder bit

This only started happening the day my battery was installed. I have had these over 18 months and never seen then glow before, but they do now, every night I see them.

How can my battery install have triggered this?

The even weirder bit

The point the battery stopped powering the house
As I say, I now see this every night, well, except for last night. Last night I turned off the hot tub and allowed the house to run on battery all night for the first time. To be fair, I kept checking the battery charge in the night. Sad, I know. But yes, the battery lasted not only until sunrise, but until enough solar to run things and start charging it again. That means I have not used any power from the grid for over 24 hours, and in fact exported 9½kWh, even using the hot tub in the day yesterday. Amazing.

But, guess what, the lights did not glow last night, not once did I catch them doing so.

So they only glow, once the battery was installed, but only when it is in standby, not powering the house.

I repeat, there is no earth on these lights. And the extra earth rod for the battery install was done 6 months ago anyway. So all that changed when the battery was installed was, err, the battery and gateway box.

I may play with switching the neutral as well, but wow, why would I need to do that.

Explanations welcome!


Usually the switch for a light like this would be L/N to the ceiling, the L going all the way down to a light switch, and all the way back, in the same wire/sleeve, to the ceiling, and then in to the light with N. This long wire acts as a capacitor allowing small amounts of AC current to flow.

But sorry. My lights are L/N to a shelly, L via the relay in shelly, and the L(switched)/N on to the light. No cable to act as a capacitor. Nice try twitter.

Further: We are planning to take up floor boards tomorrow (Friday) to confirm wiring is as I remember it, and try some things - I'll post more details here when I know. Thank you all for the ideas on this.

Final answer

Firstly, there is no switch lead, so capacitive pick up L to Switched-L was not the cause.

One person sort of got it right, a lively neutral. But not quite what you may think. In fact it was L/N tails to CU reverse, so all neutrals were very live and unprotected. The cable to the light therefor had a live (on neutral wire), and an earth (not used at light, but obviously connected to earth), and next to the earth was the neutral (on the switched live wire), which capacitively picked up from earth and power the lights.

Yes, there was a big flash and a bang in discovering this. No, I am not going to point fingers (was not me, obviously).


Winter is coming

The air-con is great for cooling in the winter, though that obviously has a cost, it is not as bad as I expected, and I have systems set up to control when it is on, and in which rooms, etc.

The air-con could be used for heating in winter, as it works out more than 100% efficient. The issue is that even with the new silly prices I think gas is still a lot cheaper. I do need to do the sums, which will depend on the tariffs I can get for my electricity and using a second battery. But even that is not simple, even with a second battery, and charging when very cheap at night, because that only works up to the capacity of the battery. If the air-con used for heating exceeds that capacity then we are back to expensive electricity compared to gas.

So my next project is improving the gas central heating. First step is a smart meter with half hour stats via MQTT using an in home display that connects to my MQTT server. It is on order now. This will help track the cost and usage more accurately.

But I also need finer control of individual room heating. At present we have two heating loops (up/down) and some thermostatic radiator valves. This provides limited options.

Per room control

A key factor is per room control. I already have my own temperature monitoring which I use to control the air-con for cooling. But for heating my only real control is heating on or off for whole floor.

My plan it to fit each radiator with an "actuator". I was originally thinking a shelly thermostatic radiator valve with wifi, but to be honest that is not what I really want, and an actuator and simple shelly 1 to control it is easier. I can then turn the radiator in any room on or off based on my temperature sensor which is located near the bed (for bedrooms) rather than next to the hot radiator itself.

I can then set controls for target temperatures during day and night so that the room is heated only when needed, and only to the needed temperature per room.

I already have logic to link the room control status to an aggregate for the heating itself, ie. any room heating on a floor means the floor level gas boiler heating is turned on.

Room occupancy

One of the key reasons for this is not just that different rooms need different targets (my wife and I prefer quite different temperatures), but also that not all rooms are always in use. We have a couple of guest rooms, for example.

At present I include a room manually in the controls and adjust the radiator valves manually, but once I have an actuator I can be smarter.

Just to be clear, this is not quite the same as "occupancy" for, say, lights coming on and off - for which there are various sensors. This is occupancy along the lines of "the guest room is in use today". So I can ensure it is heated (or cooled) sensibly for the day/night, and ready for when someone goes to bed - that means knowing hours before and keeping things going all night, etc.

My current thinking is any use of the light switch in a room marks the room for occupancy for next 24 hours or some such, maybe 18 hours...

That way a room that is not in use is not heated (or cooled when we get back in to summer), but a room that is in use gets heated (at night for bedrooms, based on a temperature profile).

It also means that when my wife goes in the room during the day to make the bed because we are expecting a guest, the room will be ready when they arrive, etc.

Anyway, once I have this working, I'll post some more details and pictures.

Note: these actuators come in a variety of fittings, modes, and voltages. Take care to order the right one. Pictured is Danfoss Actuator TWA-A 230V NC (RA) Danfoss manifold 088H3112. This is 230V activated so ideal for using with a Shelly, NC (normally closed radiator), and a fitting that works to replace a normal Danfoss thermostatic radiator valve fitting. Note, this seems to use about 2W to hold "open".


Battery, part 3

Slightly misleading as battery is charging!
I have just been going over the final handover and testing with Green Park for the Tesla Battery.

Part of it was making a power cut. This was interesting.

So a few facts that are useful to know.

  • A real power cut means a couple of seconds of no power. It is not seamless.
  • A deliberate "off grid" (done on the app) does not, so if it is a planned power cut then you can seamlessly go off grid if needed. Good.
  • Going back on grid is seamless, a few seconds while it sorts phase match and back on. Good.
  • It will take solar during a power cut, and charge the battery if needed. Good.
  • If the solar exceeds its 5kW charging capacity it will shut down the solar (done using frequency change, apparently).

The last point is a pain as the solar can do 10kW. More importantly the solar can control what it makes by de-optimising the panels, so it if knew it was a problem it could reduce power. There seems no way to link these intelligently, sadly. But once I have a second battery we will be fine at it will match the battery, at least until the battery is full. What I don't know is if there is any way it tries to turn the solar back on in such cases, maybe trying it for a few seconds every now and then? Must find out.

And as for the first point, a power cut causing a blip - I need a very small UPS for my comms rack to last a few seconds, that is the easy part.

Second battery

One battery can handle 5kW in/out, and charge to 14kWh. This is fine for most cases for balancing the solar and usage and reducing export. It is also perfect for getting on the Octopus tariff for this where they manage the battery and I pay for net usage.

However, one battery does limit my options. It does not have capability to run the house for 24 hours, meaning that I cannot make use of "cheap charge at night and run of battery for the day". If I could, I would save even more, but without that I could be using more expensive electricity later in the day, especially in winter when less sun.

One battery is also likely to have the same issue if I went on an agile tariff, in that trying to match charge at low price and usage when high would be hard if I do not have a whole day's capacity.

So for these various reasons, a second battery is on order. For now, I am trying to sort the Octopus tariff, which is fun.

Off backup

The other work still planned is an additional small consumer unit that has a few circuits which will not be on backup. The Tesla can only do 5kW, so if I have the tumble drier on, or the car charger, it won't work. There are about 4 circuits to move to that. All good fun.


You learn something every day (law)

By my age people usually have a reasonable grip on how things work, in this case basics like contracts and law. No, I don't know every law (even though I am, in theory, expected to!).

But one principle which I thought I knew was that it is the person that breaks the law that gets punished, not someone else. Yes, there are situations where another person may also break a law by "conspiring" and so on, but I mean a case where one person is punished for another person's wrongdoing, when the first person played no part in it, did not encourage it, and was not even aware of the wrong doing having happened.

Being the "keeper" of a motor vehicle has various responsibilities. These include things like paying Vehicle Excise Duty, ensuring the vehicle is insured or SORN, etc, etc.

It is not the same as being the "owner" or the "driver", although in a lot of cases all there are the same person.

But even as a "keeper", and even with those responsibilities, you are not liable for the actions of the "driver". If a "driver" speeds, your responsibility is to say who the driver is when asked. That is what being "keeper" requires. This fits with the notion that you are not responsible for someone else's wrongdoing.

Even if you don't say who was driver, unless the vehicle was taken without consent or some other excuse, you can be punished (fined), but your are punished for not saying who the driver is. You are not punished for what the driver did. Indeed, in some cases, such as where a vehicle has a company as a keeper, it can be a lesser penalty (just a fine) to deliberately not say who the driver is. Doing so deliberately could be perverting the course of justice, but hard to prove.

This even applies to things like "contracts" for parking. The driver enters in to the contract, not the keeper. The keeper plays a part in tracking down the driver - the parking firm can ask who the driver was, but it is the driver that entered in to a contract (and presumably did not follow the agreed terms, creating a charge).

However, to my utter surprise, there is an exception.

It seems you can be liable to a penalty for something which someone else does, that you did not do, did not condone or encourage, did not even know about. And indeed, the person that did the wrongdoing is not liable in these case.

Congestion charges

Yes, that seems completely crazy to me, it seems to be a broken law. Indeed, I would hope one day someone challenges it. It breaks an important principle that you are not another persons "keeper" (even if you are a keeper of a car).

For congestion charge penalties the keeper is liable, expect for a few special cases. Not being the "driver" is not one of them.

Now, some have argued that congestion charge is not a penalty, it is a road usage fee, much like VED. Which is fair enough. And a responsibility to pay for that being with the keeper is not perhaps that unreasonable. You pay to allow the car to be used in the congestion zone, much like paying VED.

The issue is when someone drives in to a congestion zone without paying the congestion charge. It is the "driver" that did the wrong thing here, very clearly, not the keeper, I feel.

Yet, surprising, the penalty for this wrong doing falls on the keeper, and not the driver. The keeper cannot pass this on by reporting who the driver was. That is crazy.

How is that not a seriously broken law?

No consent

There is a way out, but not nice. There are special cases for hire companies, but not simply lending your car to a mate. For that you could make it clear they have to ensure congestion charges are paid as a condition of "consent" to drive the car. Perhaps a card on the dash board or a recorded conversation so you can prove it. If you can show the driver was driving without consent then the keeper is no longer liable. The catch is that immediately makes the driver liable to criminal charges for driving without consent. This is very different, you don't give a driver "consent" to speed, for example and it is not assumed you did, so the driver is liable. Why would driving in to a congestion zone be any different?

Angles Mort

I saw this on a truck (non UK plates), and they do puzzle me a bit. For a start, it is not clear - are the black bits the death angles - in ...