For the purpose of getting it to work, there’s a good chance getting two capacitors the wrong way round isn’t a problem. Short answer is in bold below, the following explanation can be ignored by anyone with TL;DR!
The reason is static converters are imperfect even when set up optimally. True 3-phase power consists of 3 lines each alternating 120° out of phase with each other. Static converters use single-phase L and N to produce a 3-phase system in which two phases are 180° apart, and then create a third phase about 90° from L/N by adding capacitors. A capacitor only circuit gives 0°, 180°, and about 270° rather than 0°, 120°, 240°. The ‘about 270°’ is improved somewhat by adding an auto-transformer.
In electronics, the time taken by AC to pass through a capacitor by charging and discharging it, is equivalent to a -90° phase shift. An inductor adds a +90° phase shift. Approximately! Both depend on the frequency and the impedance of the load. Impedance is a complex mix of inductance, capacitance and resistance and in a motor, inductance varies with current drawn. Start current is usually much bigger than Run current, and Run current varies with the load. It’s impossible for a static converter circuit to get this right automatically so they come with a bunch of switches typically labelled START, LOW and HIGH HP. These allow the owner to find the combination of capacitors that start and run his motor. He has to experiment. If Trevor gets C1 and C3 the wrong way round, all that will happen is the HIGH, LOW switches will be set differently for motor best running. A guess: C1 (6uF) is connected to the LOW switch, and C3 (10uF) to the HIGH switch.
As static converters are a bit of a bodge, no surprise that not all motors work with them! Or that the power output and smoothness of the motor are both somewhat inferior compared with accurate 3-phase.
Rotary converters do better by putting an idler motor in the circuit. The idler motor responds to the current drawn by the working motor, so the false phase doesn’t jump about so much. Better technical answer, but noisy, physically bigger, more expensive, and relatively inefficient. They do support having a number of different machines plugged into them, which is important in some workshops.
These days Variable Frequency Drives are usually preferred for fixed installations. Unlike static and rotary converters, VFDs output correctly phased 0°, 120°, 240° power. Quietly and with high efficiency. Plus speed control and a long, long list of other goodies. Disadvantages include: understanding the manual, not plug-and-play; affordable units only support one motor, no sharing; EMC issues; and a spiky synthesised waveform likely to poke holes through the insulation of an electric motor made before 1970.
In the UK at least, Electricity suppliers are extremely reluctant to provide 3-phase to domestic users. I guess this is partly economics – small users are unprofitable! In the past single-phase installation costs were partly amortised over a 40 to 50 year supply life, usually in a scheme doing a whole village or new housing estate. In contrast today, singleton domestic customers seem be quoted the full initial installation cost. Not too bad if a 3-phase connection point is adjacent, but horrifyingly expensive if a deep trench is dug across a busy main road, poles put up, or. god forbid, a new transformer is needed. Counter-intuitively, the per customer cost of converting a whole street to 3-phase is cheaper than running 3-phase to one small workshop; it’s because the installation expense is shared by many customers. Electric cars may be the best bet for getting Model Engineering 3-phase. With luck, the government will encourage us to install 3-phase chargers en-mass by running an efficient roll-out scheme, where they carry the initial cost. Unlike a small workshop using tiny quantities of 3-phase, car chargers will pay for themselves by being used! And once 3-phase is on the premises for a car charger, then running a cable to the workshop will be much cheaper.
Dave