New Rigol DS1054Z oscilloscope

[Mark_O]

Before the "upgrade" I measured the bandwidth as almost exactly 50 MHz with a sharp rolloff above 50 MHz, consistent with bandwidth limiting in software.

Potentially, that could be a concern.  I.e., software BWL does nothing to eliminate aliasing, because the damage has already been done by then.  (Higher frequency components have already folded back into the pass-band.)  That's why there is a programmable LPF in the gain-stage of the front end, in the first place. 

However, this isn't unexpected, since the chip they're using has no 50 MHz LPF setting.  I'd guess they're using the programmed 70 MHz setting in hardware, which in practice should be fine.  But for those paying attention to small details, the 70 MHz models are only down 3dB at ~90 MHz.  And therefore the more "conservative" 50 MHz BW (with respect to a 4-channel 250 MHz sample rate), doesn't buy you any more immunity from aliasing, because it's imposed in software. 

Assuming of course that swperk is correct (and I believe he is).

[ragerino]

I'm wondering if the waveform generator from the DS1000Z-S versions can be added by adding the two bnc connectors.
Here's a teardown video of a DS1000Z -> http://youtu.be/xMPPuAOoD8c?t=1m15s

Unlikely. Fitting DACs to all units would not be very cost efficient. Also, there is a cutout on the board in place where gen BNCs would normally go so they are either on separate board in ds1000z-s or the main board is different.

Are you sure, that the board woudl need additonal DACs. What I've seen on the video only the chip for the "not yet available" Logic-Analyzer option and the connectors for the waveform generator are missing. I could imagine that a replacement of the firmware would add the waveform generator.

[alank2]

However, this isn't unexpected, since the chip they're using has no 50 MHz LPF setting.  I'd guess they're using the programmed 70 MHz setting in hardware, which in practice should be fine.  But for those paying attention to small details, the 70 MHz models are only down 3dB at ~90 MHz.  And therefore the more "conservative" 50 MHz BW (with respect to a 4-channel 250 MHz sample rate), doesn't buy you any more immunity from aliasing, because it's imposed in software. 

Does the 1000Z series use an amplifier chip like the 2000 series?  I was watching the teardown on youtube, but I didn't see anything like that mentioned....

[Mark_O]

Does the 1000Z series use an amplifier chip like the 2000 series?  I was watching the teardown on youtube, but I didn't see anything like that mentioned....

Hey, Alan.  I thought so, but I could be wrong.  I know the DS2000/4000/6000 all do.  But the 1000Z could be an exception.

[But even if not, it still has to have at least a switchable LPF, for 70/100MHz, so my comments above would still apply.  (Unless it had two of them, to enable a hardware 50MHz mode.  Then I'd be FOS.  )]

[olewales]

I'm wondering if the waveform generator from the DS1000Z-S versions can be added by adding the two bnc connectors.
Here's a teardown video of a DS1000Z -> http://youtu.be/xMPPuAOoD8c?t=1m15s

Unlikely. Fitting DACs to all units would not be very cost efficient. Also, there is a cutout on the board in place where gen BNCs would normally go so they are either on separate board in ds1000z-s or the main board is different.

Are you sure, that the board woudl need additonal DACs. What I've seen on the video only the chip for the "not yet available" Logic-Analyzer option and the connectors for the waveform generator are missing. I could imagine that a replacement of the firmware would add the waveform generator.

Not sure until we'll see some teardown photos of ds1000z-s but there seems to be no circuitry handling analog outputs on the main board and it would be a significant waste of money for rigol to include ICs for features that are not supposed to be present.

Anyway, just upgraded my BW. Attached are some before and after screenshots. Signal generated by IO toggle on msp430. I cannot measure effective bandwidth but trace definitely gained some high frequency details.

[alank2]

There is a teardown video on youtube.

[godFather89]

Has anyone tried just pressing the source button with the front cover off? Maybe it just works with no modification. 

[netdudeuk]

Searching for 'Rigol DS1054Z' or 'DS1054Z ' doesn't show it.   Got the URL for the teardown video please ?

[godFather89]

What URL? For the source button? The source button is present only on DS1000Z-S. Still, I'm sure it's on the PCB of DS1054Z. From the 1074Z teardown, I don't see any connectors for the signal generator. Maybe it's just a software feature.

[netdudeuk]

I clicked on Reply to the wrong post.

The URL for the teardown video is what I was looking for thanks.

[Fungus]

What did it have before?

What does if have with different numbers of channels enabled?

Before the "upgrade" I measured the bandwidth as almost exactly 50 MHz with a sharp rolloff above 50 MHz, consistent with bandwidth limiting in software.

There is no real difference in the bandwidth with multiple channels enabled

Thanks! That's exactly what I (and probably lots of others) wanted to know.

[alank2]

1000Z teardown:

[DanielS]

Before the "upgrade" I measured the bandwidth as almost exactly 50 MHz with a sharp rolloff above 50 MHz, consistent with bandwidth limiting in software.

Potentially, that could be a concern.  I.e., software BWL does nothing to eliminate aliasing, because the damage has already been done by then.  (Higher frequency components have already folded back into the pass-band.)  That's why there is a programmable LPF in the gain-stage of the front end, in the first place.

The simplest way to avoid that would be to apply the programmable filter at the full 200+MSPS regardless of time base and then down-sample to whatever output sample rate from there. You need little more than an FIR filter for that, which takes a trivial amount of space in modern FPGAs.

[netdudeuk]

1000Z teardown:

I've seen that one thanks.  He was spot on when he commented that it looked like a mixed signal scope could be on the cards.

[David Hess]

Potentially, that could be a concern.  I.e., software BWL does nothing to eliminate aliasing, because the damage has already been done by then.  (Higher frequency components have already folded back into the pass-band.)  That's why there is a programmable LPF in the gain-stage of the front end, in the first place.

The simplest way to avoid that would be to apply the programmable filter at the full 200+MSPS regardless of time base and then down-sample to whatever output sample rate from there. You need little more than an FIR filter for that, which takes a trivial amount of space in modern FPGAs.

Some DSOs work this way but it has drawbacks and less expensive oscilloscopes usually have different maximum sample rates depending on how many channels are active.  It would be really annoying if the character of each vertical input changed more than necessary depending on how many channels are active.  For a similar reason, a Nyquist filter which scales with sample rate is a bad idea in an oscilloscope.

DSP bandwidth filtering also tends to screw up the transient response.  I ran across this on a Tektronix MSO5204 that I evaluated a couple years ago; the transient response changed significantly between the analog and digital filters of the same cutoff frequency.  I suspect now this had to do with the Gibbs phenomenon.

[DanielS]

Some DSOs work this way but it has drawbacks and less expensive oscilloscopes usually have different maximum sample rates depending on how many channels are active.

That's why I used 200+MSPS in my hypothetical scenario instead of 1GSPS: the DS1000z has 250MSPS with all four channels enabled.

A properly designed FIR filter will not 'screw' the transient response any worse than extra orders of analog filtering do assuming the analog front-end does a good job of eliminating signal beyond Nyquist. To accommodate different input sampling rates when turning channels on/off, you simply need a set of matching FIR filter coefficients for each input rate.

[swperk]

The actual -3 dB point is about 133 MHz using a leveled signal generator.

Thanks for that. 

Did you happen to notice how far out the -10dB or -20db points were?  I'm curious about how fast the falloff is, since in 4-channel mode, the extended BW makes things pretty dicey (aka, probable aliasing).  That would not be an issue, with it's original bandwidth.

In other words, if I had one, I think I'd be more likely to bump it to a 1074z model, instead of 1104z.

ADDED:  that's assuming I was going to be using it in 4-channel mode a lot.  Which I probably would be.

I just remeasured my "upgraded" DS1054Z bandwidth points. I used the amplitude of a 10 MHz sinewave input as my reference level, as I've observed that the amplitude drops by about 1.5 dB at 100 MHz as compared to the level at 10 MHz. I set my Agilent E4436B to deliver a signal that measured 0 dBm at the 50 ohm terminated scope input. I had only Channel 1 active so the scope would sample at 1 GSa/s to ensure the best waveform fidelity. Here are my results:

Frequency      Amplitude
  10 MHz          0.0 dBm
100 MHz         -1.5 dBm
150 MHz         -3.0 dBm
393 MHz        -10.0 dBm
447 MHz        -20.0 dBm

As would be expected, the waveform above 400 MHz became quite distorted (it looked like an amplitude modulated carrier) because of undersampling, but surprisingly, the triggering was rock-solid all the way up. The frequency measurement by the scope was reasonably accurate as well, except that at frequencies above 400 MHz the displayed value jumped around quite a bit. At 500 MHz, the amplitude dropped to 0 (Nyquist frequency).

[olewales]

Solid piece of valuable info. Thank you.

[Fungus]

I think I hate the multifunction knob.

When I open up a menu with a lot of selections, scroll down to the one I want... it quite often jumps the the next selection when I press the knob to select it. So annoying.

It would be so much nicer if there was a separate "Enter" button, or even if I could just press the same button that opened the menu to close the menu again (selecting the new value, obviously).

Pressing a menu button multiple times should just repeatedly open/close the menu without changing the value.

If you want to go up/down to a new value in the menu by pressing buttons instead of using the multifunction knob you could use the blue up/down arrows, not the button that opened the menu.

[marmad]

I think I hate the multifunction knob.

Yes, selecting things with the multifunction knobs on Rigols (and some Siglents also) can be a pain in the ass.

If you want to experience very few skips when selecting, it helps if you can develop a specific way of holding and turning it when selecting things: for me, my thumb is on the bottom edge of the knob front, while my index finger is on the top edge at the very back of the knob (snug against the case), with my hand supported by my 3 free fingers. The knob is turned precisely with opposing motions of the thumb and index finger (their positions don't change), and when the desired selection is reached, my grip tightens slightly with my index finger pushing slightly down and back against the case - while the thumb pushes slightly up and inward - essentially locking the knob from rotating during the click. It works pretty well most of the time.

[Fungus]

It works pretty well most of the time.

"Most"?

I wonder if we can get Dave to rant about it in his video. Maybe they'll do something....It just seems unnecessary to have to push that fiddly little knob with all those other buttons lying around.

[true]

I got used to it, but it took a long time. I agree, it's awful. Also awful is the lack of any kind of acceleration / speed control, and sometimes the knob is just somewhat unresponsive, or overresponsive. I don't get it.

[marmad]

"Most"?

Yes, I'd say 98% of the time. Nothing is foolproof.

I wonder if we can get Dave to rant about it in his video. Maybe they'll do something....It just seems unnecessary to have to push that fiddly little knob with all those other buttons lying around.

Again, this is a problem that has been around for YEARS already: the DS1000E has it (and probably most other Rigol products) - plus it stretches across Chinese brands: e.g. the new Siglent SDS2000 series has it. The Chinese (at least the brands that I have tested) don't seem to be able to properly write the code for selecting from a non-indented rotary encoder.

Whether they add more code for selecting a different way other than using the multifunction knob (which wouldn't be a bad idea), I don't see them solving the selection problem using the encoder anytime soon (since it seems to have eluded them for years).

[alank2]

Is the issue that it moves when pressing the button, or that it has a delay before reaching its final location.  If I wait for it to stop on a selection, then carefully press the button straight on with no rotation, it usually works fine.

Acceleration would be great for trying to get those large value changes though.

[marmad]

Acceleration would be great for trying to get those large value changes though.

That's one of the reasons I think the DS2000 is so nice = acceleration via the navigation knob.  When I was reviewing the Siglent SDS2000 recently, it was driving me crazy every time I had to get through a large selection of numbers via the multifunction knob.... sooo slow... but on the DS2000, just zip right to the end