Yeah, no doubt I've learned a whole lot in the process of doing this work.
The Buzz is only right when the unit kicks on. No more than 1 second. It was pretty loud a few weeks ago just before the motor went out. It's definitely quieter now, but it started off quiet originally and progressively got worse. So I'm thinking there's still a problem somewhere which caused my compressor cap to go bad and a blower motor. I haven't sat outside to listen for sounds at the unit when it kicks on but after it's already running I go out and listen and don't hear any noises. The contactor is in the unit. The time delay relay is in the unit as well but installed on a circuit board.
It sort of makes sense that an older motor would cause more noise, since it may have increased friction from wear, breakdown of lubricants, contamination, etc... But, I would've assumed a new motor and breaker would've corrected the problem. Have you ensured all wire connections, where current is pulled from (including mains), are tightened properly? Another long shot... what about moisture getting into the panel (breaker especially)?
It really sounds like you're near the trip point of the breaker. Kind of hard to believe on a 50A breaker. Have you disconnected your heat strips and tested this system? I realize the heat strips shouldn't be pulling current during warm weather, but I think it's good to subtract possibilities (such as miswired). I'm also assuming this breaker buzzing is only a problem running your AC? Are all the motors turning on at the same time or is there a confirmed delay between each? The total sum FLA of your motors is about 21A FLA (1/4hp = 2.6A, 1/3hp = 3.9A) and an approximate 119A LRA (using the LRA/FLA ratio established from the compressor for worst case).
Have you been able to measure operating current? If you measure it using an ammeter with a current clamp, make sure it's an RMS meter with it's maximum range just over that of the current drawn. "AmpClamps" are generally not very accurate anyway, but a 300A clamp will be worse since the significant digits (rounding effect) will be greater near the lower end of a range.
Another issue I'm going to throw out is CEMF (aka... Back EMF or Back Voltage). CEMF, perhaps from another motor requiring much more current (such as a compressor motor) can create high enough negative voltage swings to breakdown transformer wire insulation and reduce the effectiveness of breakers (even destroy electronics). The insulation breakdown is not noticeable right away, but eventually a motor will become less efficient and draw more current, eventually leading to operational failure. You might be able to see this by placing a good analog meter (600V rated) across your 240VAC. You would need to watch the meter as the blower kicks in and see how high the meter goes or if it pegs the needle. It will be very fast, such that a <$500 digital meter will not catch (they'd need a fast transient integrating RMS IC). CEMF is present as soon as current passes through a winding and opposes EMF, giving a voltage drop (impedance) across the winding. V=L*di/dt, if you're into calculus. When current suddenly drops to zero, because the delta in di/dt is so rapid, huge voltage swings into thousands of volts can be generated. Such a concept is used on ignition coils, where 12V In = 30kV out.
There are MOV suppressors you can by that are designed for these applications. These suppressors run about $15.00 + S&H. The instructions will normally tell you to connect into your panel, but you would be better placing these as close to the potential source or sources as possible. The reason is you have other things on the same circuit, so they'll see the elevated voltage levels before the MOV does (yes, there really is a delay). Any surge suppressors you buy should be in compliance to ANSI/UL 1449 preferably Third Edition. This is a newer update and much more rigorous than previous editions. But, even an old edition will most likely serve you long enough.
Here's a link for an ICM MOV. If you get these, you must make sure the ground wire is connected to ground for full protection. Even without ground, you will still have clamping from L1 to L2. Again, I am just throwing the CEMF out there and it may not be the cause, but low-cost protection doesn't hurt.
http://www.supplyhouse.com/ICM-Cont...qdsE8jBAfD17gn2rfA_e1LabG0hah0HdPXkTNdl_D_BwE
A motor shouldn't simply die in 2 years. These motors are designed for their environment and should last until the mechanics fail, probably about 10 or so years.