Planetary Magnetic Fields and Metallic Hydrogen
      The difference between metals and non-metals has to do with how hard or easy it is to move electrons around in the outer part of an atom, or from one atom to another. Big atoms with lots of electrons like iron can easily have the outermost electron moved around. There are lots of other electrons between it and the nucleus, all pushing it away from the nucleus, and the attraction of the nucleus is reduced by the fact that the outermost electron is far from it, so it is easy to move it around or tear it away. So easy that in a liquid or solid form of such a material, where the atoms are ‘in contact’, the outermost electron can move from atom to atom to atom almost as though there is a free ‘gas’ of electrons moving through the material. It is this freely moving electron cloud that causes metals to have the heat and electrical conductivity that makes us call them metals.
     In normal hydrogen and helium this does not happen. They have only one or two electrons, very close to the nucleus, held onto very tightly by the nucleus, and no other electrons in between them and the nucleus trying to push them away. As a result, it is hundreds of times harder to move electrons around inside these atoms than in an iron atom. And in a liquid or solid state, the electrons stay with the atoms they belong to, and do not wander from atom to atom, so there is no heat or electrical conductivity associated with such motion, and these materials are non-metals.
     However, deep inside Jupiter the weight compressing the gas (mostly hydrogen) is tens of times the weight compressing the core of the Earth, and the liquid hydrogen is compressed to ten, twenty, and perhaps even thirty or forty times the normal density of liquid hydrogen. That’s why Jupiter is twice as dense as Saturn, even though it almost certainly has a greater proportion of hydrogen. It is so much heavier and so much more compressed under its greater weight that the hydrogen in the core is two or three times denser than inside Saturn.
     Now if 30 or 40 hydrogen atoms occupy the space normally occupied by one hydrogen atom, each of the electrons in each of those atoms must be around 3 times closer to the nuclei of other hydrogen atoms than they normally are to their own nucleus. So it might not be surprising if they got ‘confused’ as to which atom they belong to, and move from atom to atom in a way similar to the movement of electron gases in metals, thereby turning the hydrogen into metallic hydrogen. (Note that physicists do not use such anthropomorphic reasoning, but instead rely on what the average person would consider horrific math, but the results are the same -- electrons move from atom to atom, causing the hydrogen to behave like a liquid metal.)
     Because Jupiter has such a strong magnetic field, if that field is produced, as we feel certain it is, by a liquid dynamo (motions in a liquid metallic substance), then there must be hundreds of Earth masses of liquid metal inside Jupiter. But because of its low density the ONLY thing that there can be hundreds of Earth masses inside Jupiter is hydrogen. So it must be metallic hydrogen.
     For Saturn the field isn’t nearly as strong or as big, although is as strong as the Earth’s field and much larger, so you still need quite a few Earth masses of electrically conducting fluid (again, almost certainly metallic hydrogen) to explain its magnetic field. And because it is a very big planet and considerably compressed under its weight, although not nearly so much as Jupiter, so that it is only half as dense, it should indeed have some metallic hydrogen.
     For Uranus and Neptune, on the other hand, the weight compressing the bottommost part of the hydrogen layer isn’t nearly as large -- there’s not as much hydrogen, so you don’t go down through as much of it before you reach the core of the planet, and there’s not as much weight. For Neptune, in particular, where most of the planet must be heavy liquids like water, it is essentially impossible for any metallic hydrogen to exist. The pressures at the bottom of the hydrogen layer on Neptune are almost certainly less than laboratory experiments which fail to create metallic hydrogen. And for Uranus, although the hydrogen should go down twice as far because there should be about twice as much of it, the presence of metallic hydrogen is considered very doubtful.
     So for Jupiter and Saturn, their powerful magnetic fields REQUIRE the existence of metallic hydrogen -- hundreds of Earth masses for Jupiter, a few tens of Earth masses for Saturn. But for Uranus and Neptune, the magnetic fields require another explanation: "salt"-water (that is, water mixed with other materials that make it electrically conducting). Water (and similar materials) in the outer core of those planets is probably partially liquid, and electrically conducting.
     SUMMARY:
     In Terrestrial planets, convective motions in a molten iron core produces the magnetic field. If there is no field, there is either no molten iron core or the molten region is not large enough or hot enough to produce significant convective motion, and a magnetic field. If there is a magnetic field, there is certainly a molten iron core and substantial convective currents within the core.
     For Jupiter and Saturn, convective motions with central regions filled with metallic hydrogen must produce their magnetic fields.
     For Uranus and Neptune (and large moons of the outer planets that happen to have magnetic fields), electrically conducting fluids such as seawater must produce the magnetic field.