You think they are "overly optimistic"?? It's not that hard, my TTi TF930 counter is better than that.
Not really. It has a calibration/correction mechanism and is configured at the factory to be within 0,2ppm but it still has a temperature variation of +/-1ppm and ages 1ppm in the first year. They don't even specify anything beyond that. Now it's easy to implement a calibration in a counter, as it's most probably just a correction factor in the software, i.e. you don't actually need a digitally trimmed TCXO unless there is a timebase output - which isn't the case for the TTi TF930 though.
For a scope, such a cheap correction implementation wouldn't work as long as you have a trigger output which is a basic feature that nearly every scope has. So you'd need a (digitally) trimmable time base and I'm not aware that any scope even in the 20k€ range would have this.
Anyway, the time base spec of the SDS2000X+ and SDS5000X is really top notch compared to most competitors. Actually., most scopes <10k€ have a worse spec.
While technically correct, you are nitpicking a bit...
I'm talking about order of magnitude.
Keysight MSOX 4000 series (a 20000 USD scope) has: Time base accuracy: ± 10 ppm (Denotes warranted specifications, all others are typical. Specifications are valid after a 30-minute warm-up period and ± 10 °C from firmware calibration).
It is much worse than TF930 (from guaranteed spec, mine is still less than 2ppm from ideal after 5 years, compared to GPSDO at my friends labs few months ago). Also TF930/960 is specified ±1 ppm over full temperature range, not ±10 °C from firmware calibration. And it has digitally trimmable TCXO.
Funny thing, cheaper MSOX3000T series is much better (and more detailed specification): Time base accuracy 1 ±1.6 ppm + aging factor (1st year: ±0.5 ppm, 2nd year: ±0.7 ppm, 5 years: ±1.5 ppm, 10 years: ±2.0 ppm) (Denotes warranted specifications, all others are typical. Specifications are valid after a 30-minute warm-up period and ± 10 °C from firmware calibration).
Speaking with some people who should know, it was hinted that is because 4000 has 10 MHz ref in, you can connect outside reference if needed (and you can use it without when not critical), and 3000T doesn't have ref input so they put in tighter oscillator, so they can say, "not ref in input, but hey, it is darn accurate anyways so not really need for it", and also they are marketing built in counter a lot, so they wanted it to be reasonable accuracy...
Reason for this is, like Martin and Nico say, that scopes timebase is not usually important to be that accurate, in low ppm range. Stability is important, but absolute accuracy not so much.
Even those scopes that have External Ref IN, have not so much for absolute accuracy, but for synchronizing separate instruments for time correlation...
Or should I say, for scopes with short memory is not that critical, because of their timing resolution.
With very long memory, it becomes more critical, with sample times that are long enough for problems to show. So for Siglents with 200-250 MPoints of memory, both stability, phase noise and corresponding jitter, and all that crap for timebase, will be more visible than for Keysight with 4 MPoints.
If I was speculating, that is the reason why Siglent put in better oscillator than average.
And on SDS5000X they also wanted to have very good jitter specifications, to enable some measurements and have very good trigger stability.
But really, all of them could (and should) put them in, not because we really need it, but because it is easy and cheap.
Good job Siglent.
Home
Help
Search
Login
Register
