In this question I can see the advantages of using AC electric power over DC in large aircraft. However, Embraer decided to produce a "DC aircraft": the ERJ 135/145.
What are the advantages of such a DC power system over the more commonly used AC one?
Now, DC is easier to produce than AC and easily to inverted to AC
To produce AC of a specific frequency, you will need one of below
Method #1 is used by most American GE diesel locomotives. As a result, when the locomotive is providing AC power to the train cars, the engine needs to run at a fixed rpm, hence can not run at full rpm and full output. For this reason, a lot of other countries have switched to DC bus for train cars, essentially switched to method #3 (e.g. China uses DC600V for the train bus and it is inverted to 220V/380V 3-phase within each car).
Method #2 I believe is used by most airplanes, because fixing the engine at a certain RPM is simply not possible. The variable transmission is certainly not zero weight and is pretty complex and is certainly a point of failure. It seems like some airplane designs are switching to method #3 as well. They didn't switch to #3 earlier simply because aviation industry relatively is slow on new technology (so is the space industry).
Method #3 is basically what everyone else chooses since the beginning of this century, when semiconductor finally makes it cheap and simple and efficient to invert DC to AC. From small gasoline generators and home air conditioning, to electric cars and trains and ships, whenever the frequency of the power generator doesn't match the consumer, using DC and semiconductor power electronics to bridge the gap is the simple choice.
Most avionics are electronics, and most electronics runs on low voltage DC
Except the gyro, the CRT display, and environment control, everything else runs on DC internally. If you use an AC bus, then every device needs its own converter for it's internal working DC requirement. If you need a converter for each device anyway, then why not a DC-DC one and it's lighter and smaller and possibly cheaper?
Changing voltage used to be easier for AC than DC, now it's sort of the other way.
Before semiconductor, the only efficient way of converting voltage is via a transformer, hence AC. But transformers are heavy, which is why aviation uses 400Hz instead of 50/60Hz: transformers for higher frequencies are lighter and smaller. With modern semiconductor power electronics, you can basically convert DC voltages with near-zero weight, which is certainly desirable for airplanes. Also, semiconductor are made from sand, a lot cheaper than copper that makes a transformer.
DC-DC voltage conversion is more robust than AC-AC
An AC power transformer converts the voltage at a fixed ratio, so when the input side's voltage or frequency fluctuates, it propagates to the output, potentially affecting the downstream devices. DC-DC converters and DC-AC converters outputs at a fixed spec, so the input fluctuations are isolated from the output. This additional protection is certainly desirable for the aviation industry where safety, reliability and fault tolerance are the top of concerns.
With DC generators (or AC alternators immediately rectified to DC)... each generator simply has its own voltage regulator set to output the agreed DC voltage. The regulator, which is simple, well-established tech, simply increases/decreases the generator excitation field to get the correct voltage to come out. It will do this at any engine or generator speed. Your car already does this with its alternator.
The only issue is on multi-engine vehicles. Starting an engine draws a tremendous amount of current. The generator on the already running engine might overload itself trying to help start the other engine. However this is easily solved by adding a current-limiting circuit to the regulator, and that can be done entirely with passive components.
When you have AC distribution, and you want to feed that distribution from 2 generators, it gets byzantine fast.
First, the generators must be synchronized to each other before the second one connects -- both in frequency and phasing. If not, it's a really bad idea to bring them online.
If the frequency is not matched, extreme current will flow, as each generator tries to "force" the others to run at its own speed. Imagine 2 engines connected with a clutch: they're at different speeds and then you abruptly "pop the clutch". The engines will be forced into sync, and the forces on the clutch will be tremendous. Well, the same thing happens, but with electric force which will be tremendous. Likely, this will simply trip their circuit breakers.
If frequency is matched but phase is not, the generators will still fight each other, as each tries to maul the other into sync. It's like if the engines were running at the same speed, but the clutch had gears and the teeth weren't aligned. There will be a big shock, and again that will occur with electric force onto the AC buses.
Synchronizing is straightforward as you saw in the linked video. But since you don't have a flight engineer to watch the 3 lights and turn on the generator at the right instant, this all needs to be automated. Not so simple - lots of complex things to go wrong.
Once synchronization has been achieved, there is a reactive force flowing between all generators to keep each generator in sync and phase-lock. If a generator starts to overspeed, its share of total load will increase, and this load will make its engine bog down. If a generator underspeeds, load will come off it entirely and it will speed up. At least that's how it works with engines *whose primary load is the generator. The tail wags the dog: the grid wags the engine.
You caught the part above where the 2+ generators will "fight" to force each other into sync. If the generators are hard-coupled, this means they are trying to force the entire engines into sync also, like an electric version of the cross-shaft on a V-22 Osprey.
That "tail wagging the dog" is fine when the engine's main output is the generator. But the plan falls apart if the generator is an auxiliary load.
So that means thrust-generating aviation engines can't be hard-coupled to their generators. There must be a variable-speed mechanical coupling that frees the generator to run at its ideal/target speed. This is more complexity still, and more stuff to break.
This concept of a 'mechanical coupling' also exists in the DC generator, but this happens electrically -- the voltage regulator adjusts field excitation to suit engine speed and electrical loads. That's what voltage regulators already do. Easy as pie!
