DS2000 vs. DS1000Z record, playback, navigate, analyze, mask, etc. ??

[marmad]

If you need a Function/AWG generator and have limited bench space (or want one built-in for portable use, with fewer boxes to carry), it's an available option on the 1000Z-S, but not on the 2000.

I suspect the DS2000A-S models will be released in the West fairly soon (I think they've been sold in China for awhile already).

[Mark_O]

I suspect the DS2000A-S models will be released in the West fairly soon (I think they've been sold in China for awhile already).

That's definitely worth pointing out, but I prefer to discuss things that exist, as "in available for purchase".  Also, its accuracy is dependent on your definition of "fairly soon".  I'd guess realistically (maybe) 3 months?, and waiting that long may not be desirable.

Also, it should be mentioned that he could buy the cheaper 1074Z and get the same BW as the more expensive 1104Z he was comparing at price parity against the 2072.  In fact, for that current parity, he could get the 1074Z-S.  Which is certain to be significantly LESS expensive than a forthcoming 2072A-S.  (likely 25-30% less, just as the 1074Z is 30% cheaper than the 2072A.)

[Mark_O]

BTW, the scope I'd like to see is the DS2074.  Mainly because I can't afford/justify the DS4014 I'd like (or even better, an MSO4014).   

But I suspect that's not highly probable, any time soon.  (The DS2000 Rev2 hardware  has a space for a future LA, as well as the funcGen.  But no extra channels.  )  So Rigol has holes in its product line.

[marmad]

yes, 4 channels is probably fine, but what about the 1 vs 2G Samplingrate?, is there other things that makes the choice easier?

It is perhaps worth pointing out that the maximum sample rate while using all 4 channels is 250MHz (and I assume that rate is only while using AUTO/3MB up until 1ms/div) - so the 70MHz (or 100MHz) bandwidth specification while running all 4 channels is rather pushing the envelope.

Another byproduct of the reduced sample rate, which might be meaningful to a prospective buyer, is the maximum Peak Detect, which is only 4ns on the DS1000Z (down from 500ps on the DS2000).

[Mark_O]

It is perhaps worth pointing out that the maximum sample rate while using all 4 channels is 250MHz (and I assume that rate is only while using AUTO/3MB up until 1ms/div) - so the 70MHz (or 100MHz) bandwidth specification while running all 4 channels is rather pushing the envelope.

Worth noting, yes.  However sampling at 2.5x is sufficient to achieve a 100% perfect sinx reconstruction.  So "pushing the envelope", but not exceeding it.

Another byproduct of the reduced sample rate, which might be meaningful to a prospective buyer, is the maximum Peak Detect, which is only 4ns on the DS1000Z (down from 500ps on the DS2000).

Definitely meaningful!     Especially when dealing with high-speed digital signals.  I assume the 500ps is in single-channel mode, and it would be 1ns in dual (since I know the 4000 is 250/500ps).  But the 1000Z is 4ns, regardless of # of active channels.  So there are potentially glitches (in very fast systems) that would fall through the cracks on the 1000Z.  Good to know.

[marmad]

Worth noting, yes.  However sampling at 2.5x is sufficient to achieve a 100% perfect sinx reconstruction.  So "pushing the envelope", but not exceeding it.

Well, yes, mathematically speaking, at least 2.5x higher. But unfortunately, real world limitations of the DSO can interfere with this. Linear interpolation should have, ideally, a sample rate 10x higher.

I don't know exactly how the DS1000Z series works (although I'd be curious to know), but the DS2000 series follows these rules:

If the sample rate <= 500MSa/s it automatically uses linear interpolation.
If the sample rate >= 1GSa/s it automatically uses sin(x)/x interpolation.

Here is a sine wave sampled at 2.5x it's frequency and reproduced using linear interpolation:

[Mark_O]

Here is a sine wave sampled at 2.5x it's frequency and reproduced using linear interpolation:

  Well... that's just butt ugly. 

I don't know exactly how the DS1000Z series works, but the DS2000 series follows these rules:

If the sample rate <= 500MSa/s it automatically uses linear interpolation.
If the sample rate >= 1GSa/s it automatically uses sin(x)/x interpolation.

I was unaware of that.  Thanks!   

[marmad]

I don't know exactly how the DS1000Z series works, but the DS2000 series follows these rules:

If the sample rate <= 500MSa/s it automatically uses linear interpolation.
If the sample rate >= 1GSa/s it automatically uses sin(x)/x interpolation.

I was unaware of that.  Thanks!   

No problem. But I really would be curious as to how they implement interpolation on the DS1000Z series. On the DS2000, you have a menu switch for 'Type:' 'Vectors' or 'Dots' - on the DS1000Z series they've changed that to 'Sin(x)/x:' 'On' or 'Off' - so perhaps they've dropped using linear interpolation.

[Mark_O]

marmad, just for clarification, you said, "automatically uses", but can you override that behavior?  I.e., select sin(x)/x in some menu?

[marmad]

marmad, just for clarification, you said, "automatically uses", but can you override that behavior?  I.e., select sin(x)/x in some menu?

Nope - it's automatic.

[marmad]

marmad, just for clarification, you said, "automatically uses", but can you override that behavior?  I.e., select sin(x)/x in some menu?

One other point: I believe the Agilent X-Series also uses linear interpolation at slower sample rates (although unlike the DS2000, you can't turn off interpolation and just see the real sample points).

From what I understand, the reason for this is that sin(x)/x attenuation will give you aliasing if the original signal contains components above the Nyquist frequency (the higher components will fold back into the pass band). Linear interpolation is cruder but will be less wrong in that it won't introduce false peaks at slower sample rates - although it will introduce discontinuities in the gradient.

[Mark_O]

If the sample rate >= 1GSa/s it automatically uses sin(x)/x interpolation.

It's just starting  to dawn on me that the 2000 never uses sin(x)/x smoothing, unless you have the sample rate cranked all the way to Max (or next to Max with a single channel).

I guess that's "good" in the sense that you don't get a mis-impression from the interpolated points, and therefore increase the sampling (as you say) to 10x to get better fill.  I'm not sure I've ever run into a scope before that enforces that policy (and doesn't let you override it). 

Perhaps they decided to do so based on the extreme memory depths these scopes offer.  You'd sure never get away with it on scopes with shallow sample memories.

I really would be curious as to how they implement interpolation on the DS1000Z series

Same here.  It wouldn't surprise me, based on the above, if they defaulted sinx to On only at 250 MSa/s rates and up.  But let you override that in either direction.  (I.e., turn it Off at 500 MSa, or On at 6 kSa, if you wanted.)

[Mark_O]

From what I understand, the reason for this is that sin(x)/x attenuation will give you aliasing if the original signal contains components above the Nyquist frequency (the higher components will fold back into the pass band). Linear interpolation is cruder but will be less wrong in that it won't introduce false peaks at slower sample rates - although it will introduce discontinuities in the gradient.

Yes.  That's certainly true.  However, that's where the Anti-aliasing option comes in (that apparently doesn't do anything at all on the 2000).

[Mark_O]

Perhaps they decided to do so based on the extreme memory depths these scopes offer.  You'd sure never get away with it on scopes with shallow sample memories.

I'd check what my Tek TDS460A does on this, with only 110k samples/chan.  And my LeCroy 9314L, with 1M samples/chan.  And my Tek TLA7e2, which samples to 5 GHz, but has only 15k samples/chan.  But they're still boxed up from my recent move.

[marmad]

It's just starting  to dawn on me that the 2000 never uses sin(x)/x smoothing, unless you have the sample rate cranked all the way to Max (or next to Max with a single channel).

Well, as you surmised later in your post, it likely has to do with the large memory on the DSO (i.e. higher sample rates @ slower time base settings). Using AUTO/14MB, sin(x)/x interpolation is active down to 1ms/div (down to 500us/div with two channels) - and with the 56MB option enabled, down to 2ms/div for one or two channels.

I'm not sure I've ever run into a scope before that enforces that policy (and doesn't let you override it).

Well, as mentioned already, the Agilent X-Series are worse (in terms of what options can't be overridden): it has anti-aliasing and interpolation always on - and it determines itself whether to use sin(x)/x or linear.

Same here.  It wouldn't surprise me, based on the above, if they defaulted sinx to On only at 250 MSa/s rates and up.  But let you override that in either direction.  (I.e., turn it Off at 500 MSa, or On at 6 kSa, if you wanted.)

Umm... I'd be rather surprised if the DS1000Z gave you MORE control over certain functions than the DS2000. From what I've seen, it's always in the direction of 'less'.  Actually, I was wrong. I thought the 'Sin(x)/x' switch replaced the 'Vector' switch on the DS2000, but reading through the manual it appears it has both. So in this regard, the DS1000Z has more functionality than the DS2000, allowing you to manually switch between sin(x)/x interpolation, linear interpolation, or none.

I guess the reason that they HAD to do this was the very reason we mentioned above: forcing linear interpolation at a maximum sample rate of 250MHz was going to cause some bandwidth issues.

Or, as Rigol puts it:

"Enable the dynamic sine interpolation can restore the original waveform more real." 

[Mark_O]

marmad, thanks for the update.  We all want "the original waveform more real".   Less real would be bad. 

[hobbes]

Actually, I was wrong. I thought the 'Sin(x)/x' switch replaced the 'Vector' switch on the DS2000, but reading through the manual it appears it has both. So in this regard, the DS1000Z has more functionality than the DS2000, allowing you to manually switch between sin(x)/x interpolation, linear interpolation, or none.

Speaking of that, how do I disable sin(x)/x on the 1000Z? I see the option under 'acquire' but it's always grayed out and set to "ON".

[seronday]

The sin(x)/x ON / OFF only becomes available when 3 or more channels are turned ON.
Do not know if this is intentional or a bug.

[marmad]

The sin(x)/x ON / OFF only becomes available when 3 or more channels are turned ON.
Do not know if this is intentional or a bug.

This makes sense based on what was mentioned previously in this thread:

When 3+ channels are turned on in the DS1000Z, the sampling rate is quartered - with 250MSas/s the maximum rate. At 250MS/s, the Nyquist frequency is only 125MHz, and according to your posted bandwidth chart, will only be attenuated by a bit less than 4dB. So using sin(x)/x interpolation risks giving wrong results if the original signal contains components above that frequency.

[neslekkim]

yes, 4 channels is probably fine, but what about the 1 vs 2G Samplingrate?, is there other things that makes the choice easier?

It is perhaps worth pointing out that the maximum sample rate while using all 4 channels is 250MHz (and I assume that rate is only while using AUTO/3MB up until 1ms/div) - so the 70MHz (or 100MHz) bandwidth specification while running all 4 channels is rather pushing the envelope.

Another byproduct of the reduced sample rate, which might be meaningful to a prospective buyer, is the maximum Peak Detect, which is only 4ns on the DS1000Z (down from 500ps on the DS2000).

I'm starting to lean towards the DS2072(A), I understand that the A is not hacked yet, but I guess it will eventually be.
I guess that 70 should be ok, for some reason, the price difference up to toe DS2102 is a bit steep, about $340 here.

[marmad]

I'm starting to lean towards the DS2072(A), I understand that the A is not hacked yet, but I guess it will eventually be.
I guess that 70 should be ok, for some reason, the price difference up to toe DS2102 is a bit steep, about $340 here.

No one has ever found any difference between the DS2072 or DS2102 - except the front sticker (and price tag)  The un-hacked BW of the DS2072 is at least ~110MHz, so don't waste your money on the DS2102.

[JDubU]

No one has ever found any difference between the DS2072 or DS2102 - except the front sticker (and price tag)  The un-hacked BW of the DS2072 is at least ~110MHz, so don't waste your money on the DS2102.

Assuming that the vertical input amplifier is, in fact, a TI LMH6518, the only software controlled bandwidth settings available for it are:  20MHz, 100MHz, 200MHz, 350MHz, 650MHz, 750MHz and "Full" (900MHz).
Unless some other software controllable parts are involved in bandwidth selection, there would be no way to differentiate a 70MHz version from a 100MHz version.

[neslekkim]

ah, great!

Things I want to use it for, is spi and i2c etc, for development with mcu's, not for  decoding the signals, but for verifcations etc, to see that the signals are good on the other end.
Looking for an saleae or something for decoding and logicwork, someday.

It seems that in other places, the A series is replacing the normal ds2000 series?, here in Norway they gave me one price for the DS2072, and another for the DS2072A..

[Teneyes]

To DS1000z owners:

Marmad is working on a new version of my Rigol Ultravision Utilities to handle all 4 channels of the DS1000Z . Since he does not own the DSO, He'd appreciate any owners willing to be alpha/beta testers. The link for the current alpha (as well as some explanatory info) is here.

[David Hess]

One standard post-processing technique (and super-easy to implement) that Rigol failed to include for segments is "Display All" (showing all acquired segments simultaneously to look for deviations). My RUU software can do this, but you have to get all the segments out of the DSO, which takes a fair bit of time (each segment is considered a unique acquire by the DSO - you can't just export the entire segmented memory in one transfer).

That is about the only reason that I would want segmented memory.  I feed vindicated having not bought a Rigol after deciding they were being deceptive in their marketing.