Siglent SDS2000X HD, 12bit, 2GSa/s high definition oscilloscope serie.

[Martin72]

Since a week I got a DMMCheck Plus reference , today I measure the dc and ac voltage output.
First "classic", then using the DVM mode of the scope.
What to say...I´m impressed about the precision.
Deviation from the "reference" ( it´s calibrated with a 8.5 digit meter) in ac mode only 0.2%, in dc mode 0.02%...Nice.
Very nice. 

[teddychn]

About to unlock my SDS2104X HD's bandwidth and functions. But I need someone's kind help.

And I find that my SDS2104X HD has a higher noise level with a 50-ohm termination compared to the 1M-input impedance. It's beyond my expectation. Is it similar to your HDs?

I'll add the screenshots later.

[kladit]

No, here, SDS2504X HD, the noise is slightly higher at 1 MOhm than at 50 Ohm. That is to be expected, the other way round not.

[Martin72]

Hi,

Same here.

[Performa01]

Interesting!

So what's the problem?
Your scope has 50 µV AC-RMS at 50 Ohms compared to 56 µV at 1 MOhm.

The full bandwidth devices of the other users have a bit higher noise, but again it's e.g. 83 µv at 50 Ohms and 88 µV at 1 MOhm, so it's lower in 50 Ohms mode.

All the difference comes from low frequencies anyway, because at higher frequencies the 1 MOhm input becomes low impedance as well (because of its input capacitance), so there is no significant difference to the 50 Ohms mode anymore.

[teddychn]

Interesting!

So what's the problem?
Your scope has 50 µV AC-RMS at 50 Ohms compared to 56 µV at 1 MOhm.

The full bandwidth devices of the other users have a bit higher noise, but again it's e.g. 83 µv at 50 Ohms and 88 µV at 1 MOhm, so it's lower in 50 Ohms mode.

All the difference comes from low frequencies anyway, because at higher frequencies the 1 MOhm input becomes low impedance as well (because of its input capacitance), so there is no significant difference to the 50 Ohms mode anymore.

Hi Performa01,

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm. However, I realized my mistake; my tests were DC coupled, while Martin72 and Kladit's tests were AC coupled.

In fact, I believe the DC offset affects the RMS calculation. I just calibrated my HD and retested it.

There's so much to learn, and I was just curious about the phenomenon. Previous pictures will be taken off preventing misleading.

[2N3055]

Interesting!

So what's the problem?
Your scope has 50 µV AC-RMS at 50 Ohms compared to 56 µV at 1 MOhm.

The full bandwidth devices of the other users have a bit higher noise, but again it's e.g. 83 µv at 50 Ohms and 88 µV at 1 MOhm, so it's lower in 50 Ohms mode.

All the difference comes from low frequencies anyway, because at higher frequencies the 1 MOhm input becomes low impedance as well (because of its input capacitance), so there is no significant difference to the 50 Ohms mode anymore.

Hi Performa01,

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm. However, I realized my mistake; my tests were DC coupled, while Martin72 and Kladit's tests were AC coupled.

In fact, I believe the DC offset affects the RMS calculation. I just calibrated my HD and retested it.

There's so much to learn, and I was just curious about the phenomenon. Previous pictures will be taken off preventing misleading.

Of course DC offset affects RMS. That is because it is true RMS taking into account AC+DC RMS. Relevant measurements here is Stdev which is same as what Keysight calls AC RMS. It is explained in the manual.

Reading the manual is good idea. There are so many functions in these devices..

[Fungus]

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm.

50-Ohm termination is there to avoid reflections in the wires, nothing to do with noise.

[2N3055]

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm.

50-Ohm termination is there to avoid reflections in the wires, nothing to do with noise.

You might want to rethink this statement.... just saying, politely...

[Fungus]

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm.

50-Ohm termination is there to avoid reflections in the wires, nothing to do with noise.

You might want to rethink this statement.... just saying, politely...

I guess it depends on how you define "noise", but the idea behind 50-Ohm termination, it's reason for existing, is to do with signal integrity, not noise per se. Any measured changes above will be incidental, not by design.

[2N3055]

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm.

50-Ohm termination is there to avoid reflections in the wires, nothing to do with noise.

You might want to rethink this statement.... just saying, politely...

I guess it depends on how you define "noise", but the idea behind 50-Ohm termination, it's reason for existing, is to do with signal integrity, not noise per se. Any measured changes above will be incidental, not by design.

Ok so more directly: what is thermal noise (Johnson Nyquist noise) ?

[Fungus]

Ok so more directly: what is thermal noise (Johnson Nyquist noise) ?

This: https://en.wikipedia.org/wiki/Johnson%E2%80%93Nyquist_noise

But whether the measured "noise" at the input of an oscilloscope goes up or down when you connect a 50-Ohm resistor across it isn't determined by that, it could go either way depending on the layout of the circuit, any RF in the vicinity, and the characteristics of the resistor itself (eg. physical size).

Two people could probably measure it and get different results depending on their surroundings.

Try putting some tinfoil over your BNC connector when you take the measurement. See if that helps.

[rf-loop]

I thought the noise level at 50 Ohm termination would always be lower than at 1M-Ohm.

50-Ohm termination is there to avoid reflections in the wires, nothing to do with noise.

You might want to rethink this statement.... just saying, politely...

I guess it depends on how you define "noise", but the idea behind 50-Ohm termination, it's reason for existing, is to do with signal integrity, not noise per se. Any measured changes above will be incidental, not by design.

Ok so more directly: what is thermal noise (Johnson Nyquist noise) ?

Some oscilloscopes have separate true 50ohm pathway from input connector to ADC but most have 1Mohm pathway with only 50ohm input termination when 50 ohm is selected.

Other words: In many oscilloscopes there is always 1Mohm normal "noisy" front end independent of if user have selected 1Mohm or 50ohm. When user select 50ohm input it only add 50ohm termination before this 1Mohm "noisy" pathway and ADC still see this 1Mohm pathway noise. Only positive thing in this solution is lower price. With 5-8 bit oscilloscopes this noise is not big deal but when we start rise resolution we need also care more about this noise. J-N noise (thermal noise)  but also other way generated noises. Example we can see many times nearly poor 1/f noises aka flicker noise. But yes, we want cheap... aka compromise between money and performance.  Also we have there other noise generating components than just resistances.

[Martin72]

It will be worth the wait...
Firmware update:
Every current model has already received one this year, I'm curious when "we" are on it.

 

[Performa01]

With the high input capacitance of the 1 meg path it is rather doubtful that a simple 50 ohms resistor across the input would take us very far. The input matching is very poor and essentially the same as with an external through terminator. The last scope from Siglent where I've encountered such a concept was the old SDS2000X (without plus!). Look at the attached plot that shows the input VSWR over frequency:

SDS2304X_VSWR_100mV_c

Even at moderate 100 MHz we already exceed the usual 1.5:1 limit - but this is a 300 MHz scope - and we have a peak VSWR of 3.44:1 at only 241 MHz!

Hint: even just 10 pF input capacitance are XC = 50 ohm at ony 318 MHz.

SDS2000X Plus and all higher class oscilloscopes from Siglent have a proper 50 ohms path, hence maintain an input VSWR <1.5:1 up to the full bandwidth.

[JimKnopf]

Now I am also on board with a SDS2104X HD. I have replaced my old DS1054Z with it. Ordered yesterday, arrived today.

Seems i have to design a new wall mount because the one i created for the DS1054Z doesn't fit.

BTW: No Mouse included.

[Martin72]

For 3 years, since I once bought the SDS2104X+, I kept sneaking around it.
Could not decide, sometimes yes, sometimes no and then the high price...
But now the time has come, I did it because I need more than 4 channels for one thing:
I have ordered the SPL2016....

[JimKnopf]

I ordered the clone version of the SPL2016.

[kripton2035]

I also bought the clone version of the digital probes. came from ebay, shipped from the usa took less than 2 weeks to arrive.
works fine for around 50 EUR.

[Martin72]

Arrived today, looks and feels very valuable.
That should also be the case for the price.

[rf-loop]

Arrived today, looks and feels very valuable.
That should also be the case for the price.

Also least I give some extra value for these coaxial ribbon cables.

[tv84]

Arrived today, looks and feels very valuable.
That should also be the case for the price.

Now you must do a battle with the clone version. 

[Andreas]

Same here.

Hmm,
what are you doing guys?
50 Ohms + AC?

The 50 Ohms are shurely directly at the connector.
Then the 100nF capacitor for AC in series.
Then the 1 Meg input resistance to ground.

so the 50 Ohms are mostly isolated by the 100 nF capacitor.

With best regards

Andreas

[INLET1011]

When I went to a hardware exhibition, a salesperson of a certain company said "Even though the products of Chinese makers and ours are the same 12-bit, they cannot be compared, especially in terms of noise floor."
However, looking at the catalog, the noise floor of Siglent (SDS2000X HD) does not seem to be higher than that of Tektronix or LeCroy.
Should I trust the catalog?

[Martin72]

The salesman wouldn´t say take what you want, they´re all the same..
At work I have access to a 12-Bit lecroy, can measure the noisefloor tomorrow.