2019-03-20

Air quality and CO₂ levels

I saw an interesting report that a few people in a conference room can quickly result in CO₂ levels such that there is impaired decision making. Wow.

So I thought, I wonder what CO₂ levels I work in normally.

Of course a good start is understanding what CO₂ levels are sensible. Wikipedia has some answers. It seems below 1000 ppm is acceptable, but there safety levels set way higher than that.

I went and got a meter off Amazon (where else!). It seems to be easy to use, and gives two particulate values and a CO₂ value as well as temperature and humidity.

So, first off, my man-cave where I work...

This is a small room, under 5m x 5m, with windows (that I keep closed), two doors (also closed most of the time) and no air vents as such. It has air-con which seems to filter quite well, so the day starts with this - very low particulates and sensible CO₂ well under 1000 ppm. It is only myself in the room normally though.

Once I am in here a while, CO₂ gets to over 900, but it depends what I am doing.

I have a treadmill - if I use that then the CO₂ gets over 1000 in a few minutes.

So maybe I don't need extra ventilation in here - if I open a window when using treadmill perhaps. This time of year a window is not problem but in summer or winter it is not ideal, obviously.

For reference, outside this morning varied between 520 to 560 as I walked through town to Tesco for my Costa breakfast. At Costa it was 700 ppm, and particulate levels of 3. Impressive.

Where next - well, my bedroom. I have air-conditioning in there as well, and as such do not normally have a window open. The bathroom has a window slightly open but the door to bedroom is normally closed. I was quite surprised that during the night the levels were over 1300. I suspect I do need some ventilation! I tried a window but it is behind curtains and did not result in much in the way of extra air circulation or a noticeable drop in CO₂!

I also checked the office, which was around 1200. They have a fan but normally off, so I think we can improve working conditions there. I also wonder what a plant or two would do.

Do I really need to do anything? I am thinking a bit more ventilation perhaps in some cases, maybe. In practice, I am planning to install some almost silent low power alternating external air vent/fans with heat exchangers. I'll report back on how well they work in due course.

Assuming I improve the CO₂ levels I won't be able to say if that makes any difference as no real objective way to tell without huge biases creeping in, I don't think.

However, for one final bit of fun - I went in to the kitchen while there was much cooking (so much that shortly afterwards the smoke alarm went off), even with a couple of windows open...

That was impressive - particulate meters literally off the scale (100 is a level that is bad), and CO₂ at silly levels.

If you can't stand the heat, etc, etc...

And yes, I do like my gadgets.

2019-03-09

Serious WiFi case study (a house in Wales)

Two of my staff work from their home in Wales some of the time. Working for an ISP you expect that they have good Internet, and perhaps even good WiFi. Well, they have good Internet with our standard dual bonded FTTC offering, but WiFi posed more of a challenge.

The problem is the house! Whilst old buildings with thick walls are not an issue unique to Wales, they are a problem for WiFi. The outside walls are around a metre thick, and the middle of the house has huge fire places and chimney breast making an even thicker wall. Whilst the basic layout is two rooms per floor (one front, one back), it has three stories and high ceilings.

What this means is one WiFi access point in the front room does not working the back (kitchen). Similarly WiFi on the ground floor struggles on the 2nd floor, if at all. Basically, there is no one place to put a WiFi access point (AP) that will work sensibly for the whole house, or even most of it.

This is a big problem for the traditional arrangement of ISP provided combined router and WiFi AP. Even solutions using WiFi repeaters would struggle, so the best bet is to have multiple wired in APs.

The solution, for some time, was to have several Apple Airport Express APs, around 4 of them. This is very much "on the cheap". For some reason this was not a good solution. Much of the house was not well covered, and even in the same room as an AP a speed test would rarely show more than 1Mb/s and usually the WiFi was unusable - so much so that they took to turning off WiFi on phones and tablets and using mobile data instead, in there own house!

We recently added a new package to our order forms at A&A, a "Serious WiFi" package which includes two WiFi-5 APs, a PoE switch and selection of cables and couplers. The idea is that this gives you the kit to deploy two WiFi APs in such a premises. It is far from cheap.

They finally asked me to come and sort the WiFi, and I thought this was a great chance to test this new "package" which we have started selling in a typical situation. I took the two APs, a PoE switch, and cables. The APs are Aruba IAP-305s which are "WiFi-5" (802.11ac).

The first step was to work out where to put the APs for good coverage. This was a combination of looking at the rooms and the thick walls, and also considering where the mobile devices will most commonly be used. We decided high on the wall on the ground floor at the front facing in to the house, and high on the wall in a back room, on the 1st floor facing in to the house. There are a number of other places we could have set up the APs I am sure.

We unplugged the old APs, as the new ones do a frequency scan to decide on best channels anyway, and we used a double sided sticky pad to stick the new APs to the walls in the selected locations with a loose trailing network lead. This allowed us to test the positioning easily.

We looked at mobile signal strength and speed tests in each room, the results were good. Having decided on the position, this meant a masonry drill to fix the access points properly and cable clip the cable in place. Yes, I am not a decorator, I was only there to "make it work" - there are plans for some white trunking in the near future I gather :-)

We were quite pleased that the solution worked and did not need a third access point, which was always a possibility.

Once done, more testing, and they are over the moon with the result. Phones can seamlessly roam between the two access points. And the speed tests are somewhat better than the 1Mb/s they could manage before.
To be honest I was surprised how much difference this has made. The Apple Airports may be a bit dated, but they should basically work. Obviously the new access points have 2.4GHz, and 5GHz, and multiple radios and antenna, and newer (faster protocols), so this clearly makes a difference.

It does rather prove the worth of the new "Serious WiFi" package though, and I am pleased with the result. Expensive, but compared to "unusable" WiFi before, I suspect it was worth it - and would be well worth it for a business premises.

P.S. We had a bit of fun as well...

2019-03-04

Daikin Air-con WiFi control

One of my air-con units was sufficiently ill that we gave up and changed it, and now I have a nice new Daikin one with WiFi control via a phone App.



The WiFi sticker that came with it had a QR code which was oddly not the WiFi login, even though iPhones understand such things, but I got it on to my WiFi (only 2.4GHz by the look of it) and all working with the app on the phone - nice.

What is nicer is poking it using curl. It has a noddy TCP stack and http interface (not https) which makes it very easy to script stuff. Several people have done this, but I have not found quite what I was looking for, so some poking around.

So, here goes, what I have found so far (subject to updates).

Sensor info

A simple get of /aircon/get_sensor_info gets :-

ret=OK,htemp=20.0,hhum=40,otemp=9.0,err=0,cmpfreq=26,mompow=2

Which is nice as it has room temp and humidity and external temp to 0.1C precision.
  • ret: A return status, with OK being good, it seems
  • htemp: Inside temp in C
  • hhum: Inside humidity, I assume in %
  • otemp: outside temp in 
  • err: I assume an error setting
  • cmpfreq: I am guessing compressor or fan frequency
  • mompow: Not sure, was 1 when idle and when heating, 2 now we are cooling

Control info

This is where it gets useful, a simple get of /aircon/get_control_info gives

ret=OK,pow=1,mode=4,adv=,stemp=30.0,shum=0,dt1=21.0,dt2=M,dt3=18.0,dt4=30.0,dt5=30.0,dt7=21.0,dh1=0,dh2=50,dh3=0,dh4=0,dh5=0,dh7=0,dhh=50,b_mode=4,b_stemp=30.0,b_shum=0,alert=255,f_rate=A,f_dir=0,b_f_rate=A,b_f_dir=0,dfr1=B,dfr2=5,dfr3=B,dfr4=A,dfr5=A,dfr6=A,dfr7=B,dfrh=5,dfd1=0,dfd2=0,dfd3=0,dfd4=0,dfd5=0,dfd6=0,dfd7=0,dfdh=0,dmnd_run=0,en_demand=0

What I have worked out so far :- 
  • pow: Power 1=on 0=off
  • mode: 1=auto, 2=dry, 3=cool, 4=heat, 5=?, 6=fan, 7=auto
  • adv: blank normal, 2=powerful, 13=streamer, 2/13=both
  • stemp: Set temperature
  • shum: Set humidity
  • dt1/2/3/4/5/7: Target temp for each mode
  • dh1/2/3/4/5/7/h: Target humidity for each mode (and h?)
  • alert: ?
  • f_rate: Fan rate A=Auto, B=Quiet, 3 to 7=speeds, 
  • f_dir: Fan direction 0=fixed, 1=vertical, 2=horizontal, 3=both
  • dfr1/2/3/4/5/7/h: Per mode something, not sure
  • dfd1/2/3/4/5/7/h: Per mode something, not sure
  • dmnd_run: Not sure
  • en_demand: Not sure
  • b_setting: Not sure

Control setting

Setting is a simple get /aircon/set_control_info?pow=1&mode=4&stemp=29&shum=0&f_rate=A which just responds with ret and adv.
The settings are as above.

So a simple cron to turn off at 06:35 is :-


35 6 * * * curl --silent 'http://x.x.x.x/aircon/set_control_info?pow=0&mode=7&stemp=21&shum=0&f_rate=B' | grep -v OK

2019-03-03

Bleeding time and motion

I have a blood test every year, sometimes more often, and that means going to Heatherwood phlebotomy clinic.

They open at 08:30, and at that point there are typically 30 or more people waiting already (which is about capacity for the waiting room). They have numbered cards you take, and they call a number. If you are lucky then there are two of them taking blood.

First off, they really should have a second set of cards (maybe just for the first hour of the day), perhaps in red or something, for fasting blood tests. I was feeling hypo as it was, but had to wait over an hour to be seen. Luckily my test was not a fasting one, but had it been I could well have been collapsing. Normally a fasting test is not a problem for someone, but I have daily insulin, and that can mean some times I have to eat - such as when I have not eaten for 12 to 14 hours for a fasting test. Even though, in my case, it would have put people ahead of me in the queue - a priority queue for fasting would make sense. The blood test form says if fasting or not, so not like people could game the system and take the wrong card.

They are pretty efficient, check your name and DoB, strap arm, clean, find vein, take blood, tape dressing over it, and then spend about 50% of the overall time, or more, writing your name, and details (about 4 lines of text) on each of the blood sample containers (in my case, two).

It strikes me that the system could be massively better with a simple barcode readers and label printer. Even with nothing needing to be on-line, just a QR code on the blood form the doctor sends that when read provides the lines of text to print on the sample label, just that. Such a device would not be expensive (well, not compared to staff time over its operational life) and could mean processing at roughly twice the rate, by my estimate. A simple fall back to writing means not building in a dependancy on technology.

Of course the printed label could also have a QR code which probably then saves time when the samples are processed later, as well as reducing transcription errors.

Don't the NHS have people whose job it is to think of things like this?

P.S. I am surprised someone does not make a small label printer with QR reader that literally just prints the QR label content on a label on each scan. Must have loads of applications just like this!

2019-03-01

Over kill?

A friend of mind dropped dead of a heart attach recently, and he was a year younger than me. He was at home, but so easily could have been at work.

It makes you think.

Well, more, it makes you shit scared!

As usual one's assessment of risk is skewed by the most recent experience.

Even so, I do think that the potential life saving benefit of a defibrillator way out weights the cost. More to the point, I can afford the cost. I can't afford to drop dead (IMHO).

We took advice from Red Cross, and we now have one at the office, and I have one at home too. I am surprised more offices do not have one. Our first aiders at work are trained on it as well (thanks to Red Cross), but it is pretty idiot proof.

I hope we never have to use it, either of them.

P.S. From the comments, one thing is getting one that can go outside, and so be accessible to others (e.g. others in my road) if needed. The ones we have are not for outside installation, sorry, but certainly something to consider if you are getting one installed.