Photos from travels. Sunrise over Maenporth.
Sunset.
Forgot to tell you Strava to keep me on the roads
Looks like public saunas are the trend this year in Cornwall, which is not a bad thing.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
]]>
I have been watching the Guinness Open Gate Brewery development in central Covent Garden for what seems like ages, so I was keen to finally get inside and take a look around. The opportunity came up this weekend when my sister-in-law managed to grab a couple of free tickets via an Irish travel email list she subscribes to.
I do like a pint of Guinness, and while I’ve been whiskey tasting in Raasay, wine tasting in the Barossa valley and olive oil tasting in Lefkada, I had never sampled a flight of beers! The visit was split into three phases; a tour around the brewery facility, a tasting, and a hands-on session covering how to properly pour a Guinness. While they make a wide selection of beers on-site, they don’t actually brew Guinness here.
A bit of brewing trivia that came up during the walkaround: Old Brewer’s Yard actually has brewing roots stretching back over 300 years to 1722, when it was the Woodyard Brewery (and later Combe & Co., one of the 18th-century powerhouses of the London porter boom). In a circular bit of brewing history, Arthur Guinness was originally inspired by those very London porters when developing the recipe for his stout in Dublin. Even later, around 1901, Guinness Extra Stout was bottled on this site. While the London location serves as an experimental microbrewery alongside sister Open Gate sites in Dublin, Baltimore, and Chicago, every drop of standard draught Guinness on tap is still shipped over directly from St. James’s Gate in Dublin.
Following the tour, we moved on to the tasting session, which consisted of four of their site-specific brews:
The Piazza Pale Ale was the favourite with the porter a close second.
The experience wrapped up with a practical, hands-on session covering how to properly pour a Guinness. The team running both the tour and the tasting were excellent—they brought loads of energy and made the whole process genuinely good fun. I would definitely recommend a visit if you are in the area, and I am already thinking about organising a work trip down there. For those interested, more details on the Open Gate experience can be found here.
As an extra treat from the sister-in-law we had a final pint printed by the Stoutie machine.
Full set of photos from the day over on Flickr.
In other news, got a couple more 101 Things I Learned books in prep for the new teaching year.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
Building on last weeks post about the energy neutral bio acoustic monitors we are developing, this week has been testing assemblies. We now have 4 units out in the Garden Lab testing and focusing on how to deploy these. The gif above shows one night of bat activity from one of the units in a north London allotment.
At the start of the week we were working out assembly requirements and ordering the little connectors and cables you forget about. One of our MSc students called into the lab to take some pictures of their dissertation project (deadline for submission was Thursday!) and asked if this was a PhD project, I said no this is our summer dissertation project!
Version one of a deployable prototype is now pinned down and assembly boards designed to fit the devices in the enclosure. The battery we ended up using was slightly wider than anticipated so extra enclosure boxes had to be ordered, but turns out the specs online were generous and the battery does fit the original box we hoped to use.
Simon has been having some fun working out how to mount the various components when we are not quite sure what and how we can make these “cow proof”, there will probably be lots of learning on the first few installs.
We now have a couple of devices in the GardenLab - one on cellular, one on Lora - and have managed to get the Multitech Lora GW also on cellular and battery powered. Need to do some analysis to see how sustainable that is through the winter but for current testing seems to be working fine.
The allotment is still looking green despite no rain for 8 weeks. Courgettes still going, the outside tomatoes seem to be doing much better this year than the greenhouse ones and the peppers are turning a lovely shade of red. Time to start harvesting.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
This week has been heads down on testing our energy neutral bio acoustic monitors we are deploying for a year in rural Somerset. Final design has been pretty much pinned down (still have Lora or Cellular decisions to make when on site) and the bulk of the ordering has been done. This is the setup we are going with.
An Arduino UNO Q with audiomoth mic running at 250kHz and either a LA66 Lorawan shield or A7670G multi network modem. All of which will be powered by a 50W panel and 100Wh battery mediated by a Victron MPPT controller. For the software we have been developing a new program using the Acoupi framework that records birds continuously and bats from disk until dawn. The Edge AI uses Birdnet for bird classification and Batdetect2 for bat calls. We have also optmised the messenger services in Acoupi primarily for the LoraWAN use case (where air time is minimal) which has also had a positive impact on the cellular workload.
More or less the setup! We are now using an audiomoth rather than the dodotronic mic and a slightly bigger battery with heating pads for the freezing winter days.
We have been doing lots of testing to understand the reliability of the acoupi processing pipeline under the constraints of energy budget, information volume and device cost. Am pretty happy with the sweat spot of sub £600, 10% recording time 24/7/365 and enough “spare” battery capacity to see us through 10 days of no sun in the middle of winter. We should also have enough on board storage to store about 200,000 archive recordings for off line analysis (files can be retrieved via an on demand WiFi hotspot).
Comparing results from 3 mics and cellular vs wifi energy use.
I have been impressed with the UNO Q. At the start we were also testing a RPi5 (great device, a bit more powerful, a bit more energy hungry, a bit more expensive), a RPi Zero 2 (not quite powerful enough to run acoupi at frequency we needed) and a Sony Spreesense (not developed beyond first week after realisation of integration work required to get to starting point). The UNO Q sat in a sweet spot just below the RPi5. Powerful enough to run the models and almost half the price with our build.
Looking at the Victron logs we are at about 40Wh per day with the current cellular setup. Still a few more test to run to see where else we can “turn things off”, but that is good enough for phase one deployment. This is only my second time using cellular (normally LoraWAN) since I have typically been worried about the high current spikes of sending messages. But this 12v battery setup is demonstrating a good use case scenario. This setup was a plan b since Lora had been the primary expectation. We are still working on that but still have some challenges around GW location and coverage.
Orders for 20 off devices have been made so next week will be building and testing ready for first deployments the week after. The yeo.cetools.org website shows work in progress observations - mainly from an allotment in North London.
In other news, the bike build continues - now have tyres and handlebars.
And the eclipse was fun - but my photos were bad!
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
Almost there with the installation of sensors on the UCL200 Greenwall. Lovely hot weather to be install sensors in the full sun. Weather station up, 3 devices running bioacoustics (acoupi bird, acoupi batdetect2 and birdnet-pi) plus a camera looking at the greenwall. Soil sensors installed and reporting to The Things Network and people counters installed under the canopy to measure dwell times.
Sensors are alive and sending data to our MQTT broker. A website with access to the data feeds will follow shortly.
People Counters under the canopy
Screenshot from camera view of green wall.
Birdnet-pi with camera and weather GW
Acoupi-bird on bottom, Acoupi-bat on top
Weather Station
Managed a few nights away in Eastbourne to walk the Seven Sisters coast path. A lovely walk if a little windy!
Seven Sisters
Birling Gap Ride
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
]]>
Managed to get around to photographing all the CASA CE 2026 exhibition projects this week. Some lovely projects again this year and great to see the students work come together over a frenetic two weeks leading up to the install week.
Leah and Simon did a great job again of curating the space and was lovely to show a few visitors around this past week. Each time I was in the space there was a few people looking around the work.
We deinstall on Monday so unfortunately the short run is over and the students focus on their dissertation write ups. We do have a few archives of the work:
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
Mostly photos this week. Have been getting some UltraLED sample’s working for potential use in Manifold, harvesting fruit - seems like the redcurrant are early this year, I almost missed the gooseberries and the blackcurrants provided no meaningful human harvest. The “plop, plop” sound of apples hitting the floor and greengages hitting the decking have also started to get much more regular. Am sure this is a few weeks earlier than last year.
The UltraLED’s are a next level up from the usual LED strips we use in the lab. Still RGBW LED’s but in an architectural enclosure making them feel 10x more professional. Had a bit of a struggle trying to work out which NeoPixel or FastLED library / settings I could use to get them running but in the end they look pretty impressive.
Planted pumpkins in the allotment.
Did a bit more frame prep - tried the hand held clothes steamer to get the flux off. Seemed to have some impact but think I am going to need to do some filing for final prep.
I do like this particular brazing joint where you can see the little rods we inserted through the frame.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
]]>
Fulfilled one of my bucket list items and made my own bike frame. I completed the frame building 101 course at the Stayer workshop over in Leytonstone. The process involved working out what kind of bike I was making, getting a bike fit to make sure the dimensions I was anticipating were correct, a design meeting with Sam to confirm the BikeCAD measurements and then a week of sizing, filing and brazing.
During Covid I turned 2 rusty old mountain bikes into one new retro bike. I had planned to try a home respray, but the time taken to strip down the bikes, merge them into one and source the parts needed meant that just getting the rebuild done became the goal. It felt like a half completed project. A chance encounter with the Stayer folk at Bespoked and seeing one of their customers bikes from the 101 course and I was putting a deposit down.
Photos from the week are below as a reminder to myself of the process. Work in progress notes are in the captions of the flickr album (68 photos). If doing research around what I wanted to build was part 1, this is very much part 2. Next I will be entering into part 3, prepping for coating the frame (I am going for clear coat) and sourcing all the parts to build the bike.
Big thanks to Sam and Willow at Stayer for making the week a perfect balance of relaxed and getting things done. They are both really impressive instructors and would recommend them in an instant. The week flew past way to quick and was great to meet fellow builders Matt, Doug and Andy.
Stayer Workshop - home for the week. Sam makes a great cup of coffee using beans from Perky Blinders roasters across the way.
Lots of measure twice, cut once.
Practice brazing first.
First bit of brazing on the real thing - BB first.
Pretty happy with the brazing.
Brazing complete.
Willow TIG Welding the rear drop outs.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
The week started with a hot visit to Yeo Valley with the WAM PAM team Kate, Santiago and Mark. We were joining Lara and Elliott who were collecting in Audiomoths, and meeting Tom and Patrick who work at Yeo Valley. Was pretty cool to experience first hand how they are turning “nature-positive” farming from a buzzword into a practical, landscape-scale system. Tom was pretty passionate about the work they are doing to change farming supply chains - was pretty infectious. I learned all about rotational grazing, Agroforestry and mixed crops.
The farm use a structured, adaptive rotational grazing system for their cows (Dairy and Beef). Rather than letting livestock wander freely across massive fields for weeks (set stocking), they rotate them daily in subsections of pastures. Known as mob grazing, livestock are moved frequently to fresh paddocks. This allows the grass in resting paddocks to grow much taller, which drastically improves above-ground biomass and allows deep root structures to develop. Some of the fields we were walking through were chest high with grasses and flowers. The photo below shows one that had probably been left for about 40 days. This equates to about 6500kg of dry matter per hectare compared to the “usual” 2800kg that farmers typically would introduce cattle. (When grazed it is reduced down to about 1500kg).
Photo of Santi walking through a field to collect an Audiomoth.
They actively view their cattle as mobile bioreactors. As the animals move, they trample organic matter into the ground and spread diverse microbes through their manure, jump-starting nutrient cycling and feeding the soil microbiome.
Where ash dieback has cleared gaps in their organic woodlands, they have introduced conservation grazing. The cattle manage weeds and brambles naturally, benefit from the shade, and help cycle nutrients directly back into the woodland soil without the need for heavy machinery.
Photo of Beef Woodland Pasture - cattle graze under the canopy of dispersed trees.
It was also really interesting to see the mixed crops in their arable and pasture systems which rely heavily on diverse crop rotation and multi-species mixes. Because these species have completely different root depths and growth rates, the pasture is highly resilient. During extreme droughts or heavy rains, there is always a species in the mix that survives and keeps the soil covered. Growing different crops together simultaneously increases below-ground microbial diversity, optimises nutrient availability, and allows the harvesting of a diverse, home-grown forage silage for winter feed, eliminating the need for imported feed or synthetic fertilisers.
I found myself walking around increasingly wondering why all farms aren’t doing this. The buzz of the insects and the variety of life visible made this seem like common sense.
The work we are doing is to introduce methods for measuring the biodiversity and natural capital that is so obviously present in these “old” ways of farming. The Yeo Valley “measuring to prove” approach is key to avoiding abstract claims by anchoring their biodiversity and carbon tracking in hard data. We are contributing through our bioacoustics technologies to automatically monitor, quantify, and verify changes in bird, bat, and insect activity over time. We are also looking towards how to scale this bioacoustic tech to their 100+ supply farms.
More photos on Flickr Yeo Valley Album.
Photos from the week:
Front Garden blooming
Wet Leg at Ally Pally
Was a roadie at Somerstock
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
A week of dissertation project support and working on solar powered bioacoustic sensors. Visited House Mill to re-install Satria’s sonar sensor (dissertation project from 2024) - the facade looks really impressive after its paint job. The picture above shows the lora sensor and in the distance the top of the Orbit in QEOP that the sensor is connecting to (about 1.3km). Enjoyed the visit to Houses of Parliament for the Can AI save the planet event on the terrace.
Below a photo from Houses of Parliament - can AI save the planet? event. I liked Henry’s response in the discussion, “I don’t think we do have the solutions since we are not implementing them” therefore there must be some additional barriers. Thanks to Kate for the invite and the great job chairing the discussion.
Also, don’t think I have parked this close to HoP before…
Allotment update - white onions are looking good this year, red ones less so. Not quite the bumper crop of previous years but green house is smelling lovely.
Looking at links.duncanwilson.com, these parts of the internet crossed my path this week:
