Not sure about the advice not getting a loan. If a piece of equipment generates revenue to cover the costs of a loan then it is not a bad investment. But it should be noted that test equipment devaluates faster than a car. Buying something cheap now and upgrading later may end up costing more compared to a loan. OTOH debt is debt and it causes a long term outgoing cash flow; the business should be able to support that.
You are right that test equipment depreciates pretty fast, but the worst depreciation is with test equipment from the big brands. It's as simple as that, unless it's some exceptionally specialist or rare piece of kit. Big brand scopes are really poor when it comes to depreciation, and funny enough the B-brands tend to lose a lot less over the years (at least what younger test equipment, i.e 5 years or so) is concerned. Not because they hold their value better but because they already start from a much lower price point when new. So based on the purchase price to resale factor alone, big brand scopes fare pretty bad. So there's that.
Considering this is for electronics consulting in (as it seems) a wide and flexible range of areas, it's unlikely to assume that any instrument bought today will satisfy all requirements which come up with the next 10 years, not only because the accelerated progress of technology. With this in mind, it's futile to buy excessively now just to hopefully be "on the safe side" later on. And even if the original assumption turned out to be correct, the feature will very likely have been bought at a premium compared with what it could cost later if it was acquired when it's actually required. So it makes sense to buy something which satisfies the current needs, and when the time comes the instrument no longer suffices then move to something new.
As to getting a loan, there are situations where it certainly can make sense to get a loan to buy better equipment, but frankly based on the requirement (a 350MHz scope for general purpose work) I can't see any justification for that. Getting a gold-plated solution with no need for it is nothing more than a waste of money.
Because of the end of the day, this thing has to amortize over set period (which also depends on the specific country;'s tax laws). Which for a more expensive scope means you either have to use a longer amortization period (if you can, but which also means you might have to stick with it longer than you want to), or you have to charge your customer more (or live with a lower profit).
To the OP: Regarding the Siglent: one thing to watch out for is that it uses a different memory management compared to the Tektronix you are used to; Siglent typically cuts the memory short to have just enough samples to fit the screen. This has to match your usage. For me this kind of memory management is a hard fail. All in all it still is a good idea to compare several scopes yourself.
The only fly in the ointment here is that most scopes work that way (and I think that others have tried to make you understand this in a long-winded discussion not too long ago). For the simple reason that people normally capture long sequences on a scope and then zoom *in* on the details, while you for some reason seem to insist on capturing short sequences and then zoom out(?) somehow?
This is true for the majority (all?) scopes out there, maybe aside from the Keysight DSO-X (InfiniVision) Series, which allow no manual memory management whatsoever.
Now as to what Tek does: the MDO3000, like other Tek scopes, have a function called Auto-Magnify, where on very short timebases the scope continues to use the full record length (or appears to, as in reality it doesn't use that short time base). In Auto-Magnify, when you reduce the timebase, the actual time base is also reduced but only until you reach a point where the scope can now achieve the full sample rate. Any further reduction in timebase will then not further reduce the actual time base, but instead the scope shows a zoomed-in screen while capturing at the unchanged timebase setting. Further reducing the timebase just zooms further into the signal.
That's it.
Also, Auto-Magnify doesn't work when zoom mode is enabled, for obvious reasons.
While this is an automatic mode, the *exact* same effect can be achieved with any other scope using the zoom function (turn the timebase up right until the sample rate reaches max, then zoom in).
Which is what most people do anyways.
I think the R&S RTB2004 also fits the budget without needing a loan.
The RTB is a very basic scope, and the real only thing it has going for it is the 10bit resolution. Aside from the basic functionality, it also lacks an active probe interface or even simple 1M/50ohms switchable inputs.
I can't see how such a basic instrument would fit the requirements.
If you really need a lot of bandwidth then getting a used higher end oscilloscope is a cheap solution.
As I understand the 350MHz BW is sufficient. But yes, if more BW (a lot more, i.e. >1GHz) is required, the 2nd hand market becomes a great option.