I've spent a week thinking about a new oscilloscope, help please? (1k-3k budget)

[tggzzz]

I just want to be done with this decision and go back to designing things!

A key part of design is that you specify what and how you test. Therefore you should finish designing before considering what equipment to buy.

The signals I would like to look at are:
- Your typical boring CMOS clock, <= 40 MHz

Irrelevant number. The only thing that matters is the transition time. FFI and concrete examples, see https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/

- 70-140 MHz parallel bus

Once you know you your analogue signal will be correctly interpreted by your digital receiver, flip to debugging in the digital domain by using a logoc analyser.

The other signals I would like to look at but I probably cannot afford are:
- DDR2, 400MHz
- SerDes, too high of a frequency to even comprehend, it's probably best to hope the design just works.

Without defining what you are looking for and why, then using skill and imagination to think of alternative ways of reaping the benefits, you are likely to spend more than necessary.

Good.

Now you will have to think what you need to achieve, and then use skill and imagination. See my .sig, and translate that to electronics! I designed, constructed and tested my first computer (with 128 bytes of RAM) using only an analogue voltmeter, some switches, and some LEDs.

Remember the old aphorism: a week in the lab can save an hours thinking about your circuit.