Siglent SDS2204 + MSO review

[Mark_O]

@Marmad and Mark_O:
I just checked again but I can't select a digital channel for decode. There is something weird with decode and digital channels. When I press digital when decode is active then decode is de-activated. But the other way around is different. If I press decode with digital active both decode and digital are active. Like I wrote before: this area seems to be work in progress.

BTW: I also found out that decode cannot be enabled if there is an automatic measurement active. I don't see why these two functions would interfere.

Thanks for checking that, and clarifying.  I had the impression that it was enable-able, but just didn't work.  What you've described sounds really strange.  Like a section of my code, where I have scaffolding but no implementation. 

[Mark_O]

If I turn on the Digital channels on my unit, I don't see them as possible sources in Decode - but I don't have the MSO cable, so perhaps that makes a difference.

My old Rigol MSO (DS1102CD) has LA head detection, and not only disables certain capabilities when it's not plugged in, but beeps and alerts you just for trying to turn the LA on.

[marmad]

There is something weird with decode and digital channels. When I press digital when decode is active then decode is de-activated. But the other way around is different. If I press decode with digital active both decode and digital are active. Like I wrote before: this area seems to be work in progress.

It works either way for me. But the menu buttons for the extra Vertical set of controls (Ref / Math / Digital / Decode) don't function uniformly. Some of them turn on both the display of the feature plus it's menu (like Digital), but others turn on just the menu (Decode), but you still have to enable the Decode bus display with a menu selection. Plus, the light for the Decode button seems to operate incorrectly:

Start with both things off (and all button lights unlit).
Press Decode
Select Display On from the screen menu
The decode bus will be displayed
Press Digital
The digital channels will be displayed
The button light for Decode will go off - but the decode bus is still enabled and displayed on the screen
Press Decode again
The button lights again and you go back to the Decode menu.
Select Display Off from the screen menu.
The decode bus will disappear from the screen - but the button remains lit - until you turn off Digital (or turn Math on and off).

[nctnico]

MSO probe
The cable itself is a crappy 0.65 pitch ribbon cable which is easy to damage. Worst; most of these ribbon cables have a solid core so it is likely to have a very short life.

Do you know these are solid core?  Or just deprecating them because they might be?  Since they're so "crappy", and you already have much better ones, just chop them up and find out.  Oh, wait.  You might not keep it.

I took the cable apart and inspected one wire at the connector. I was wrong: it is stranded wire and not solid core. That still doesn't mean I'm happy with using a fragile unprotected ribbon cable.

The specifications of the MSO bit are not stellar: 10pf input capacitance is rather high for a logic analyser especially with the input circuitry very close to the inputs.

Hmm.  That's twice in one review you've criticized the same thing.   

No, that is keeping things complete in case people skip ahead instead of reading everything. 

The MSO part
...I think somehow the time cale for the digital channels doesn't follow the timescale for the analog channels. That is a major bug!

Is there a Delay or Skew setting adjustment in the Menu structure?  If not, then it's not a bug, so much as a design fail.  Acquisition for

It is not skew. If you look closely at the screendump I posted earlier you see D0 has the same signal as CH1 but it is delayed by 18ms with a different time scale.

[nctnico]

After having used the oscilloscope a bit longer I really start to miss horizontal cursors. I use these a lot. The tracking cursors don't work if the signal is a thick noisy trace (which is often the case). I want to be able the put a horizontal cursor at two positions on the screen and read the difference between them. Siglent: please add horizontal cursors 

I also found another bug:
If I capture a trace and use the time/div knob to enlarge the waveform every now and then the waveforms are not properly aligned. When the horizontal position is adjusted slightly the signal jumps to the proper position. The picture below shows the effect. All the signals should be in phase but the yellow one is not.

[marmad]

After having used the oscilloscope a bit longer I really start to miss horizontal cursors. I use these a lot. The tracking cursors don't work if the signal is a thick noisy trace (which is often the case). I want to be able the put a horizontal cursor at two positions on the screen and read the difference between them. Siglent: please add horizontal cursors 

Can't you use Manual -> X cursors?  Or maybe Y; I'm not quite sure if you mean measuring horizontally - or measuring vertically.

[tautech]

After having used the oscilloscope a bit longer I really start to miss horizontal cursors. I use these a lot. The tracking cursors don't work if the signal is a thick noisy trace (which is often the case). I want to be able the put a horizontal cursor at two positions on the screen and read the difference between them. Siglent: please add horizontal cursors 

I also found another bug:
If I capture a trace and use the time/div knob to enlarge the waveform every now and then the waveforms are not properly aligned. When the horizontal position is adjusted slightly the signal jumps to the proper position. The picture below shows the effect. All the signals should be in phase but the yellow one is not.

Several edits.

Could it be the noisy waveform is not allowing this.
It can be "tamed" with a lower Mem or Averaging setting.
Can you then get the cursors to track?

You image below shows the Trig Level setting for CH2 above the waveform.
Does a lower setting change results?
Is there a reason why the trigger point is at 640mS delay?
Also Ch2 is AC coupled and all others are DC coupled.
Could that affect result?




Have you done a Self Cal since you got the scope?
Do you have Quick Cal turned on?

My questions in no way imply improper scope use, I like you want to find reasons for your results.

[nctnico]

The problem is not related to triggering. It occurs every now and then when turning the time/div knob quickly to enlarge a captured waveform when the acquisition is stopped (run/stop is red).

Using tracking cursors is very tedious when all I want to know is the voltage difference on a signal. And sometimes I use the horizontal cursors on my TDS744A to set a predefined level on the screen without a signal being present. My TDS744A has vertical, horizontal and tracking cursors with vertical cursors showing time or frequency and horizontal cursors showing the voltage or current.

[tautech]

The problem is not related to triggering. It occurs every now and then when turning the time/div knob quickly to enlarge a captured waveform when the acquisition is stopped (run/stop is red).

OK, the trick will be to duplicate this result in Siglent's lab.

Using tracking cursors is very tedious when all I want to know is the voltage difference on a signal. And sometimes I use the horizontal cursors on my TDS744A to set a predefined level on the screen without a signal being present. My TDS744A has vertical, horizontal and tracking cursors with vertical cursors showing time or frequency and horizontal cursors showing the voltage or current.

When the manual Cursors are first turned on(with no waveform present), they are obscured by the flat line trace, but can individually be adjusted to the position you desire.
When selected and the knob adjusted they then become visible.

EDIT
I believe this answers your previous post:

After having used the oscilloscope a bit longer I really start to miss horizontal cursors. I use these a lot. The tracking cursors don't work if the signal is a thick noisy trace (which is often the case). I want to be able the put a horizontal cursor at two positions on the screen and read the difference between them. Siglent: please add horizontal cursors 

[marmad]

When the manual Cursors are first turned on(with no waveform present), they are obscured by the flat line trace, but can individually be adjusted to the position you desire.

Didn't I just mention these? 

[nctnico]

I have added a screendump of my TDS744A showing the cursor options with horizontal cursors enabled:

The SDS2000 seems only to support Vbars (vertical cursors) and paired (tracking).

[marmad]

The SDS2000 seems only to support Vbars (vertical cursors) and paired (tracking).

The SDS2304 that I've got certainly has them.

[nctnico]

The SDS2000 seems only to support Vbars (vertical cursors) and paired (tracking).

The SDS2304 that I've got certainly has them.

Aha... hidden in some menu  Got them now.

I gave the software another try. I tried to connect to the oscilloscope over the network using the EasyscopeX software but now I'm getting some kind of JIT error because some component from NI cannot be found  The network is OK; I can ping the scope just fine. I guess I'll just have to wait until Siglent implements a webserver in the oscilloscope like other manufacturers.

[tautech]

You have been using your Tek for too long. 
New equipment usually broadens your outlook on how things can be done.
Not saying any variation from any "norm" is right or wrong, its just different to expectations.
But if existing equipment operation is exactly copied, then accusations arise.
As gear gets more complex and featured, the design challenge is to fit it in a package that has required features and function.

Essentially there should be no "norm" but a state of constant improvement.

 

[nctnico]

I agree but you have to admit that the way the Tektronix scope shows all the available options on the screen is quite clever  O0

[David Hess]

Touch screens, integrated batteries, and single button mice are all improvements as well.  Maybe you are holding the Siglent wrong.

[nctnico]

Today I have tried to use the FFT function to look for distortion (clipping) in a 1kHz signal. The frequency scale seems to be driven by 2 factors:
- memory length
- time base setting

With the memory length set to 7kpts (shortest) and the timebase set to 20ms I could get a frequency scale low enough to see the 1kHz signal and distinctive harmonics (2kHz and 3kHz). Unfortunately setting the timebase this slow makes the signals invisible.

IMHO the FFT function would be better if the FFT uses more points if you zoom in (more detail) and when averaging can be applied to the FFT results. I wouldn't mind slower FFT updates.

[David Hess]

It sounds like it is decimating the display record before performing the FFT so the FFT resolution is not increases by having longer record lengths.  That makes sense if the FFT size is limited which is often the case.

There are algorithms for computing just part of an FFT but if a large fraction is desired, it is better to compute the whole thing.

[Siglent]

After having used the oscilloscope a bit longer I really start to miss horizontal cursors. I use these a lot. The tracking cursors don't work if the signal is a thick noisy trace (which is often the case). I want to be able the put a horizontal cursor at two positions on the screen and read the difference between them. Siglent: please add horizontal cursors 

I also found another bug:
If I capture a trace and use the time/div knob to enlarge the waveform every now and then the waveforms are not properly aligned. When the horizontal position is adjusted slightly the signal jumps to the proper position. The picture below shows the effect. All the signals should be in phase but the yellow one is not.

Dear nctnico,
SDS2000 have horizon cursors.
I have a question that your waveform is not triggered (the trigger level is not at right position ), how can you confirm this is the oscilloscope problem in stop mode?
If the waveform is normally triggered, it will be OK. see below.

[nctnico]

@Siglent:
My signals are perfectly triggered and in phase because they are the same signal on different channels. What I did was:
- setup the oscilloscope to trigger
- press the run/stop button to put the oscilloscope in stop mode
- change the time/div knob quickly

After a couple of tries you should be able to reproduce the same problem (signals out of phase).

[Siglent]

 
all right,
I have tried many times and finally reproduced it.

[tautech]

all right,
I have tried many times and finally reproduced it.

Well done Jade, now you have marmad's list to work through. 
https://www.eevblog.com/forum/testgear/review-siglent-sds2304-a-comparison-of-features-with-rigol-ds2000-series/msg505048/#msg505048

[nctnico]

all right,
I have tried many times and finally reproduced it.

That is good to read. If you need more information/clarification on reproducing the other bugs I think I have found let me know on the forum or through a private message.

[nctnico]

Today I received the MSO probe cable replacement: A Dell 0H924 SCSI cable. The Dell SCSI cable is a very flexible round cable. I had to remove the strain reliefs and pulling tabs to make the connector fit in the shell. I also found out that the connector on the oscilloscope is recessed too deep. I had to mill a little bit from one of the shells to make the connector lock properly.

The end result is a cable which is much more convenient to use. There are still wires exposed although much less than before.

[tautech]

Today I received the MSO probe cable replacement: A Dell 0H924 SCSI cable. The Dell SCSI cable is a very flexible round cable. I had to remove the strain reliefs and pulling tabs to make the connector fit in the shell. I also found out that the connector on the oscilloscope is recessed too deep. I had to mill a little bit from one of the shells to make the connector lock properly.

The end result is a cable which is much more convenient to use. There are still wires exposed although much less than before.

Much better  although I'm sure we all think it could be better still.
I wonder if we can convince Siglent to change their sockets to receive a similar SCSI cable in a future HW revision?