New scope - LeCroy HDO9000

[jjoonathan]

It's not bullshit, it's a fact of advanced FFT design that anyone who has made a FFT with "spectrum-analyzer-like controls" will be able to tell you.

Averaging M FFTs of N points is is mathematically equivalent to calculating the FFT of the M*N points and taking every Mth point of the spectrum, but faster -- O(n) instead of O(n*ln(n) in time and O(1) instead of O(n) in memory. Multiply the M*N time-domain samples by a window function before taking the FFT and you convolve/smooth the frequency response that you would get *before* downsampling -- that is, you mix information from the bins that the fast technique discards into the bins it doesn't discard, resulting in the same process gain / noise floor and scalloping improvements as the M*N FFT while maintaining O(n) time cost and O(1) memory cost. Obviously with O(1) memory cost you can keep averaging until the cows come home, so it doesn't really make much sense to say it has a certain size to it.

If you actually want M*N frequency bins, you're out of luck and you have to pay full t=O(n*ln(n)) s=O(n) price. If you just want the noise floor & scalloping benefits of a M*N FFT, you only have to pay t=O(n) s=O(1).


> charging extra for something like 16bit HiRes

Now *that* is bullshit.

[madmax96]

Hdo9k vs S-Series is not easy not make a choice. You need to consider that the S-Series is a true 10bit ADC just at max sample rate (10Gs/s on 4 ch or 20Gs/s on 2 ch).
It means that at 5-2-1Gs and 500Ms and below is just an 8 bit...
So the LeCroy seems better.
Then I see that LeCroy put the Enob on their datasheet and this is the key point...the effective number of bits,

[Wuerstchenhund]

It's not bullshit, it's a fact of advanced FFT design that anyone who has made a FFT with "spectrum-analyzer-like controls" will be able to tell you.

Averaging M FFTs of N points is is mathematically equivalent to calculating the FFT of the M*N points and taking every Mth point of the spectrum, but faster -- O(n) instead of O(n*ln(n) in time and O(1) instead of O(n) in memory. Multiply the M*N time-domain samples by a window function before taking the FFT and you convolve/smooth the frequency response that you would get *before* downsampling -- that is, you mix information from the bins that the fast technique discards into the bins it doesn't discard, resulting in the same process gain / noise floor and scalloping improvements as the M*N FFT while maintaining O(n) time cost and O(1) memory cost. Obviously with O(1) memory cost you can keep averaging until the cows come home, so it doesn't really make much sense to say it has a certain size to it.

If you actually want M*N frequency bins, you're out of luck and you have to pay full t=O(n*ln(n)) s=O(n) price. If you just want the noise floor & scalloping benefits of a M*N FFT, you only have to pay t=O(n) s=O(1).

You're right that the process you described isn't BS, but I'm not sure that this is the way the RTO does FFT (there are reasons why scopes do FFT the way they do).

Anyways, the more worrying issue is the complete lack of information, which has pretty much been replaced by marketing speak.

[Wuerstchenhund]

Hdo9k vs S-Series is not easy not make a choice. You need to consider that the S-Series is a true 10bit ADC just at max sample rate (10Gs/s on 4 ch or 20Gs/s on 2 ch).
It means that at 5-2-1Gs and 500Ms and below is just an 8 bit...
So the LeCroy seems better.

Yes, but you'd still be stuck with the tiny memory.

Then I see that LeCroy put the Enob on their datasheet and this is the key point...the effective number of bits,

Listing ENOBs is nice but the figure can be misleading, as it depends on how the number was derived. And without that information it's just a number with little meaning.

[madmax96]

Tiny memory...it depends on application. But typically when you are looking at signal with bw>=1Ghz (and so sample rate >=2Gs/s) , why do you need a memory higher that 100M points?
I see needing of bigger memory for power application , not for fast signal

About the Enob you are right, this is the point for every scope supplier.
There is an IEEE recommendation that explain how to measure the Enob...this could be a good question to do to LeCroy technician

[Wuerstchenhund]

Tiny memory...it depends on application. But typically when you are looking at signal with bw>=1Ghz (and so sample rate >=2Gs/s) , why do you need a memory higher that 100M points?
I see needing of bigger memory for power application , not for fast signal

I regularly capture long sequences for further analysis and more memory is always helping.

Also, we're talking about high BW/high samplerate scopes here, and at 40GSa/s 128Mpts memory is only enough to capture a 3.2ms sequence (or 6.4ms on 20GSa/s). I agree that this isn't exactly bad but on the other side with the DSO-S I can capture up to 40ms at full sample rate (20GSa/s). The memory also affects FFT so the HDO9k can do 128Mpts vs up to 800Mpts on the DSO-S.

I'd have expected a new scope to offer more memory than the predecessors (which for the HDO9000 is probably the WavePro 7zi-A with up to 256Mpts), not less. Plus the loss of some unique features which were really nice, like the swivelable/pivotable display on the WR6zi or the detachable frontpanel (which can be replaced with one with individual vertical controls) on the WP7zi(-A)/WM8zi(-A).