Siglent SDS1000X and (MSO) SDS1000X+ series oscilloscopes

[rf-loop]

LF frequency response is not perfect flat. As we know all is only total sum of errors + some unknown amount of truth and nothing more. Ideal perfect shapes can see only in kids school books.

Here is typical SDS1kX  LF frequency response. From under 1Hz to 300Hz.
Scale in image is 25Hz / div (and linear)

In this individual unit difference between 25 and 125Hz is under 0.6dB.

Perhaps it can tweak better but not without opening. (Front end circuit is dual pathway type and perhaps dc-lf / hf balance can tweak more perfect but also error is not high )

[nctnico]

This is more like an adjustment error in the front end. There is no technical reason why a modern oscilloscope shouldn't have a better than 0.1dB (1%) flat response in 0Hz to 100kHz.

[cjm]

LF frequency response is not perfect flat. As we know all is only total sum of errors + some unknown amount of truth and nothing more. Ideal perfect shapes can see only in kids school books.

I wouldn't expect perfection, but I'd have thought a DSO should show a 100Hz square wave as being square (within a few screen pixels anyway)? It looks dead square on my DS1052E - how about other scopes?

(BTW, both channels on my SDS1202X show the same distortion.)

[rf-loop]

LF frequency response is not perfect flat. As we know all is only total sum of errors + some unknown amount of truth and nothing more. Ideal perfect shapes can see only in kids school books.

I wouldn't expect perfection, but I'd have thought a DSO should show a 100Hz square wave as being square (within a few screen pixels anyway)? It looks dead square on my DS1052E - how about other scopes?

(BTW, both channels on my SDS1202X show the same distortion.)

It looks like it need tweak better.

Here identical shots using X+  and  X-E

As can see 1202X-E  LF response is better adjusted. (DC-LF pathway  and HF pathway better balanced)

Signal is exactly same around 500mVpp Sine and both scope have all settings same.
freq sweep ~0 - 300Hz (25Hz/div linear)

Siglent need adjust  better least SDS1000X/X+ models analog front ends. This can do better and  it do not cost "anything".

Things like this can do right and wrong with same money -  but it need desire to do better.
Since professional "I want do better" do not affect, then money should use power - if it would help it better.


SDS1102X+       0 - 300Hz  (Horizontal 25Hz/div, Vertical signal 500mVp-p. 50mV/div)




SDS1202X-E       0 - 300Hz  (Horizontal 25Hz/div, Vertical signal 500mVp-p. 50mV/div)

[cjm]

Thanks for doing the comparison between the X+ and X-E rf-loop. I'd hoped to use my 1202X for audio work but seems it mightn't be the most suitable for that, oh well...

[vpv]

This is my result on SDS1102X.
Signal is around 500mVpp Sine, freq sweep ~0 - 300Hz (25Hz/div linear).




This is a 100Hz square wave. My generator is the MHS-5200A.

[cjm]

Thanks for that vpv - definitely much better than my SDS1202X - Must be significant variation from unit to unit...

[nctnico]

LF frequency response is not perfect flat. As we know all is only total sum of errors + some unknown amount of truth and nothing more. Ideal perfect shapes can see only in kids school books.

I wouldn't expect perfection, but I'd have thought a DSO should show a 100Hz square wave as being square (within a few screen pixels anyway)? It looks dead square on my DS1052E - how about other scopes?

See for yourself...
edit: updated screendump with 100Hz (what happened there    )

[skench]

cjm,

Could it be that your input is set to 50ohm be causing your problem?

[cjm]

cjm,

Could it be that your input is set to 50ohm be causing your problem?

No, afraid not. I see the same result with 1MOhm input, and with a different generator.

[StillTrying]

E.g. below is a screen capture of distortion I see in a 100Hz squarewave (the same is seen using a X10 probe or using a different generator). I assume this isn't normal?

Your SDSX's LF can't be that bad, there must be something wrong somewhere, are you sure there's no digital filters set. A simple CMOS 555 would be fine for test square waves at ~100Hz.

I don't trust a square wave's top or bottom levels from an opamp or SG, so the 85Hz square below is the output of one gate of a 74AC14, probe is X10, so 0.5Vpp at the scope, it looks fine to me.

[cjm]

Your SDSX's LF can't be that bad, there must be something wrong somewhere, are you sure there's no digital filters set. A simple CMOS 555 would be fine for test square waves at ~100Hz.

I didn't believe it myself at first, so I did also build an LMC555 circuit - but same distorted result. I also split the signal simultaneously to another scope (DS1052E) - it's square on the DS1052E, distorted as shown on the SDS1202X. SDS1202X doesn't have digital filters.

My frequency response is very similar to the one shown by rf-loop too - around +0.5dB rise at low frequencies. Maybe rf-loop could post a 100Hz square wave from his SDS1kX for comparison?

[2N3055]

ntnico

See for yourself...

Ntnico, if you'd be so  kind to show squarewave of 100Hz, not 100kHz please?
Dual path switching in vertical amplifier happens at very low frequencies.. It would not show distortion on squarewave even for 1kHz..

I would run self-cal, just in case... Self cal should equalize amplification between low frequency and high frequency gain...

[tautech]

I can only suspect the generator.
For reference cjm's screenshot:



SDS1102X+
Source SDG1032X
No additional attenuation, cheap BNC cable.

[cjm]

I can only suspect the generator.

hmm.. two different generators show me the same result and splitting the same signal simultaneously to another scope shows undistorted square wave on the second scope. Here's another example, same signal on the two scopes, this time from a LMC555, for variation... (I also tried running self cal)

would be interesting to see rf-loop's 100Hz square wave, as his frequency response looks similar to mine...

[nctnico]

Time to send it back to the dealer for repair!

[rf-loop]

would be interesting to see rf-loop's 100Hz square wave, as his frequency response looks similar to mine...

 0.01Hz - 300Hz sweep show  what is also low freq square wave response as long as we believe math.

[boggis the cat]

Did you reverse the connections when you split the signal to the two 'scopes?  Sometimes you can get weird issues with BNC connectors.

Otherwise, perhaps the front end is damaged.  Is this consistent between channels?

[rf-loop]

I can only suspect the generator.
For reference cjm's screenshot:



SDS1102X+
Source SDG1032X
No additional attenuation, cheap BNC cable.

I can suspect there is differencies between individual units. Perhaps due to internal fine adjust or component tolerances. Problem is or looks like it is between signal two pathways inside front end. There it is splitted for DC pathway and AC pathway. (nothing to do with input AC coupling)  This error can see in sweep from nearly DC to low freg and it can see in square shape (mirrored because sweep f was from down to up)


Here in last image blue is  ref A, stored using 1ms/div and 100Hz square. Yellow is 1kHz square. As can see its frequency freq components are inside  quite flat range in front end frequency band. As can see 100Hz square shape follow same shape what can see in image where was low freq sweep. In sweep of course shape is "reversed" because sweep is from low to high but in square shape top and bottom shape from left to right is from high to low if think freq)





Here first this previous low freq sweep (~0 - 300Hz)   (note vertical scale)




In this image signal from Hewlett-Packard FG (this signal shape I trust and it is also confirmed using Tek 2465 and HP54522A)  And when someone talk about signal transfer lines quality,  matching etc... it is good to realize that we are now talking about nearly DC. 

[cjm]

Time to send it back to the dealer for repair!

Yes indeed, or at least calibration, if this can be calibrated out.

Did you reverse the connections when you split the signal to the two 'scopes?  Sometimes you can get weird issues with BNC connectors.
Otherwise, perhaps the front end is damaged.  Is this consistent between channels?

I have tried different cables, but at such low frequencies I don't think this should matter.
Both channels show almost identical distortion - so probably not damage.

I can suspect there is differencies between individual units.

Yes, my guess too. So, we now have two SDS1kX samples showing similar distorted LF reponse and two other samples showing good (flat) responses. Not much data to draw a firm conclusion, but some indication that any given SDS1kX might have this issue. We also have a number of other captures from other scopes showing no LF problems, which I think shows that we should expect to see an undistorted 100Hz sqarewave on any scope.

Here in last image blue is  ref A, stored using 1ms/div and 100Hz square. Yellow is 1kHz square. As can see its frequency freq components are inside  quite flat range in front end frequency band. As can see 100Hz square shape follow same shape what can see in image where was low freq sweep. In sweep of course shape is "reversed" because sweep is from low to high but in square shape top and bottom shape from left to right is from high to low if think freq)

Many thanks rf-loop for doing this additional capture. It's good to confirm we both see similar result in time domain too. (For comparison, below is my result with cursors added). At higher frequencies (>~200Hz), I don't see any problems on my unit either.

So, I've contacted my dealer (a German supplier) and will report back if there is an outcome there. Many thanks to everyone for taking the time to do screen captures and respond - this really is a very nice and super helpful forum! 

One last thought is that perhaps this is a firmware issue, a problem with the self cal routine? If it's possible to do, I'll try to revert to an older firmware and see if that makes a difference (I'm using the latest version at the moment). (no, not the firmware).

Thanks all!

[cjm]

Ok, I've received a reply from the dealer :

...
we already had a few cases where the SDS1000X devices have a weakness at low frequencies (<300Hz) AND low voltages (<500 mV).
If the deviation is greater then 3% (this is what is specified in the datasheets) we can send the device to Siglent for adjustment.
Unfortunately there is no other way to solve this.
...

So my SDS1202X will be going on a little trip! 

[StillTrying]

Did cjm get the SDS1202X's  LF response fixed/improved ?

[cjm]

Did cjm get the SDS1202X's  LF response fixed/improved ?

I did indeed receive my adjusted/repaired unit back from Siglent (via the distributor) but haven't been at my home lab for a while to run a check. Will report back as soon as I can..

[cjm]

So... my SDS1202X was returned from repair and now shows a much improved low frequency response (100Hz square wave response attached). However, my dealer advised that Siglent service centre just replaced the entire main board - so unfortunately I'm none the wiser as to the actual problem on the original board.

[StillTrying]

Looks fine for 100Hz,  I assume CH2 is fine as well.