General > General Technical Chat
The Rigol DS1052E
flolic:
Ok, new day for a new test ;D
I made sub nanosecond avalanche-mode pulse generator as described here: http://www.i9t.net/fast-pulse/fast-pulse.html
I cheated a little and used 2N708 instead 2N2369 which I didn't have. Good thing is that it works ;D , bad thing is that I don't know exact
pulse rise/fall time.
Anyway, let's see the pictures ;)
These are with high bandwidth resistive 50 ohm terminated probe. You can see excessive ringing, this is most likely because of bad termination (I used old & cheap 10Base2 ethernet T-BNC and terminator)
CH2 unmodified:
CH1 modified:
Next two pictures are with original 150MHz probes:
CH2 unmodified:
CH1 modified:
dimlow:
So were is Dave ? he should see this
Mark_O:
--- Quote from: rossmoffett on March 17, 2010, 06:44:51 pm ---We have 400 MS/s 100 MHz model C Rigol scopes at school, and I know that they pale in comparison to the DS1052E.
--- End quote ---
Hi, Ross. I'm a bit surprised to hear you say that, because I had thought they were pretty close, performance-wise. The specs certainly aren't all that different.
--- Quote ---The low sample rate makes it look like all noise anywhere over 50 MHz or so.
--- End quote ---
But now you've lost me. The maximum (real-time) sample rates on the 1000C are 200 MSa/400 MSa, for 2 and 1 channel respectively. The sample rates on the 1000E are 250 MSa/500 MSa under the same conditions. The oft-quoted 1 GSa/sec sampling rate is limited to single-channel operation, capturing into the smaller ShortMemory only. That's because there's not enough channel bandwidth to transfer acquistions at 1 GSa into the larger ISSI memory. ShortMemory is retained internal to one of the VLSI chips, possibly the Cyclone, though probably the BlackFin, since that's where all the processing is done. I'm sure someone here would know for certain.
So you're saying that a 20% slower (maximum) sampling rate makes a huge difference in performance on the 1000C scopes?
- Mark
Mark_O:
rf-loop wrote:
> It is better if it is more flat, after 100 or more it can drop more fast. Now it continues nearly "endless". I can find components after digitzing even over 500MHz. Some may be ..oh nice... but I think it is more bad. <
I agree with your assessment, but can see why some would like the ability to detect higher-frequency components, and think it was a good thing (More is always Better, right?).
> these over Nyquist frequencies are not allways good to see in digital oscilloscopes, they make many bad effects in my opinion. These scopes need more flatness and then better BW limiting after some reasonable freq. <
And there's the rub. Without proper BW limiting, after digital sampling you wind up with foldover into the desired passband. AKA, aliasing. These alias-imaged components can create some real anomalies, both in measurements, and in interpreting waveshapes. You go, "Hmm, where did that frequency bump/spike come from?" And the answer is that it was a phantom. Chasing phantoms is not a productive use of engineering time, IMO.
- Mark
Mark_O:
--- Quote from: rossmoffett on March 17, 2010, 06:44:51 pm ---...according to Dave, 1/10 of sampling rate is about the best a real-time readout can do.
--- End quote ---
I'm surprised to hear that Dave made that claim. Especially since it's not true. You don't need 10-times oversampling to be able to accurately reconstruct a waveshape. Assuming a gaussian-distribution (the normal case), you can exactly reconstruct any arbitrary waveform, using sin(x)/x with as little as 2.5x oversampling of the highest frequency component. However, to accomplish that requires very good filters (i.e., expensive), and with the quality of filtering on most consumer-grade instruments, you actually need ~4x oversampling to achieve the same results. (To get technical, it varies, depending on the instrument, from 3-5x... but 4x is a good rule-of-thumb.)
But 10x is really overkill. Thus even the 400MSa on the earlier 100MHz 1000C-series was perfectly fine. Based on your claim, they'd only be good to 40 MHz (and the 100 MHz-version of the 1000E would top out at 50 MHz).
- Mark
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version