I think that is the most beautiful use of the 2 cylinder CS.  Lovely boat, great sound!  Thanks for the video.
Bruce

An update:
I built an insulated tower for air intake at 6 foot for cooler air (10F hotter than ambient here near the groundin still air from sun heating of the earth), and an evaporative pre-cooler.  The precooler is just a single aspen pad, with a standard 230VAC evap cooler pump (70W).  I upgraded to a new (surplus on ebay) Panasonic 1350W compressor after finding that the used LG compressor had starting problems, and lousy performance at higher current draw. 

I added an Arduino nano, power monitor via I2C, 6 Dallas one wire temperature sensors, and a 4x20 character LCD display, and a stepper motor driver board for a Chinese EEV.  I also got a larger brazed plate heat exchanger, and increased my water flow rate from 2 GPM to 3.2GPM with a Lang D5 12V pump.

It now cranks out 15-17K BTU (10 degrees of cooling at 3.2GPM) on 1400W to total power (320 of those watts for fan and water pump), the compressor is loafing and runs at only 115F on the top of its case. 

The one area that needs work still is that even going to a larger EEV I have found that my filter dryer seems to be restricting flow enough that there is no change from opening beyond half way.  That's an easy fix problem but today I just wanted to take advantage of the heat and give it a long shake down run.

As I write I am now (ever so slowly) cooling my house while its about 95F outside, on my solar PV/inverter system.  Its been a very educational project.

Evaporative pre-cooling makes for a near doubling of performance at the same power level on hot days in the high desert.

PS-  The line of temperatures on the display are:  Incoming air, cooled air,  HP refrigerant line,  Suction refrigerant line, Water into BPHE,  Water out of BPHE.  The dT and KB (K Btus) are full accuracy.  EEV position is 300.  Ill try to find and add one more photo.  The difference between the displayed watts and volts times amps is due to power factor.  An inverter must provide the full VA, not just the "real" watts so I should change my display to that.

PPS- The Control box with Arduino Nano is squeezed in next to my propane (backup) water heater.  I used all standard drivers for the power and temperature monitoring. The Dallas One wire software is very slow, about 1.5 seconds for updating all 6 sensors plus a bit of floating point math for that and the power monitor board but that's OK.  Right now I have manual switches as inputs to the Arduino for controlling the EEV and power relays. By next summer, it will be automated.  Te perf board left of the Arduino is just a 5V linear regulator and an analog circuit that senses water flow rate and shuts down the compressor power immediately if flow rate drops below 2 GPM; that assures no water freezing and the resulting wrecking of the heat exchanger.

He's not just very bad at welding.  He's seriously technically impaired.
Nothing about the project makes any sense technically or  financially.
First he throws away the built in, fan cooled 3 phase rectifier (6 diodes), then he adds an external 3 phase rectifier on a heatsink.

His power source is a 1 HP motor ?!  Obviously grossly inadequate.  3/32 rod needs 24V at 90 amps absolute minimum.
How do you get 2160W from a 1 HP motor?  You don't. So he's limited to 1/16 inch rods?

I also note that he is using an AC powered DC supply to provide rotor excitation.  Good grief, what a frig. Welders can be had for about $100.  He's spent way more than that already.

The roughly 17% conversion losses are well worth it to have a real welder that will maintain your set welding amps and is designed to start and maintain arc stability. 

I agree that the Lister CS is a poor choice for a portable welder as others have noted.  Portable screamers are designed for just that- light, compact.  Earplugs work.  Propane portables have a slightly lesser bark and have the advantage of being able to sit unused with no fuel issues, unlike gasoline.  For shop power, with occasional welding, I'd like the quiet of a well muffled, stationary CS, but I'd put in an aluminum piston and find some suitable flywheels to make it an 8/1.  Then you've got the power for your larger motors and also any welder.

 I didn't realize you were in the UK.

You may find it useful to add a step up transformer to get 230V. You didn't mention the frequency of the generator AC output.
That's an issue as well as far as compatibility, though not for things like most power tools.

Your GFCIs are more commonly called an RCD outlet or residual current detectors.  Our GFCI outlets are dual socket, and can protect more outlets, with current limited to the GFCI unit capacity, often 15amps of 115/120V (60 Hz).

I find the character of the sound is more important than sound level.  The Listeroid chug, chug, at 650 rpm I find quite pleasant.  I don't care for 1800 rpm engine noise either, though it is certainly better than 3600 rpm.  Admittedly I have chronic brain inflammation from MS and resulting epilepsy.  So finding a bearable generator was one of my first efforts when moving way off grid.  The Listeroid has served me well.  It's inaudble  near my home and shop site.  My setup uses a large automotive muffler, as the exhaust is used for induced draft through the radiator.

Other factors for those intent on putting the Lister CS type engines close to the home are the clatter of the valve train is quite loud; I found this even for the DES 8/1 clone I converted to propane for my neighbor.  There is no diesel knock, but the valve train clatter is still there.  Louvered ventilation vents down low and not facing the home site  help, as does the insulated engine shed.  The valve train clatter is  audible for about 50 feet. 

The exhaust note on my neighbor's propane 8/1 is completely gone via a leach field muffler.  It is barely detectable with your ear right against the outlet pipe.  It also cleans up the exhaust-  no odor but earthy air.  Any of the various earth mufflers seem to perform exceptionally well thanks to the very large expansion area.

The other issue for those planning very close proximity is the vibration of the earth close to the Lister CS types.  This is best mitigated by proper balancing using the 38AC method. A slightly overbalanced engine will have a bit of rocking but the least vertical vibration.  You could adjust to suit the nearby structure. Still, I would avoid putting it very close to the home slab or foundation and would try to isolate the engine concrete from the nearby structure via foam or sand if trying to have it very close (less than 10 feet).

Best Wishes,
Bruce

If its putting out 115VAC then you could certainly wire in a AC receptacle for your angle grinder or other tools. 

Hey Mike,
You got it right, regarding diagonals.  It does not matter which AC terminal is neutral.
Four small  AC voltage rated, 0.1 uF capacitors, between each of the terminals, one on each side, will greatly clean up the EMI on the AC power  for very little cost. It stops much of the diode noise at starting and stopping conduction right at the source.

Here's one rated for 275VAC:
https://www.digikey.com/en/products/detail/kemet/R46KI310040H1M/5730927

I usually solder them to the quick connect tabs near the base, so the connectors still work fine.

Best Wishes,
Bruce

One of the issues that is swept under the rug is the power consumption of an inverter, even when there is no load. It's not unusual to have 50W or or more for just running the inverter, all night long.  That affects you DOD which affects you lead acid battery life, and increases your charging time the next day.  Straight 120VDC (or pulsed DC) gets around that inverter power loss.

One issue of modified sine and even worse, square wave inverters when used for induction motors is the alleged 20% loss of efficiency due to heating (something I have not confirmed personally). I wanted to get around that so when designing my own inverter for 120VDC input, I opted for a 5 step sine (originally 7 step but later settled on 5 after much testing).  It has a measured THD of 12%, though I could do better with better software for waveform timing.  I only used the "that looks about right" method, using a small Basic program to generate my timing tables.  The 5 step keeps even my washer's timer motor happy (it gets warm and makes noise on ST-3 power, measured at  15% THD.   My inverter is small, designed for 1500W continuous, but since it uses two 1000W toroidal trransformers with secondaries in series, it can start a 1.5 hp air compressor or my well pump (3500W starting surge) as if they were on grid power.  The inverter idle current is 15W, but I only turn it on when needed.  A single twisted pair allows for start/stop control of the inverter at multiple points of use. 

I did not invent the concept of adding a step; there was a single research paper proposing before I did it which I had found showing a dramatic reduction in THD by adding just one more step to the typical MSW inverter (a 3 step waveform).  I used the now expired patent, Trace SW series method of multiple transformers with secondaries in series as a means of doing a true, all low frequency, single conversion design in order to use very slow, soft switching and thus dramatically reduce the EMI on both AC and DC.  I used high efficiency toroidal transformers and got a measured 92% efficiency.

I'm attaching a Picoscope image of what my 5 step sine looks like, using a couple surplus transformers for 120VAC ouput.  The final version uses Antek transformers, with rewound secondaries for 230VAC for my well pump and electric air compressor.  I step it down and have both 230 and 120VAC at my shop/laundry.