Siglent SDS2000X Plus

[tautech]

One question I have yet to find an answer to is whether the x axis can be in log vs linear scale. The y axis can. This would allow many more harmonic markers to be shown and hence much more useful.

Not available in FFT, only in Bode plot mode.

[mawyatt]

Of course, you can't average a waveform any faster than you can complete a sweep.

It does not follow this thinking, the average is much much slower than sweep time.

Best,

[ci11]

One question I have yet to find an answer to is whether the x axis can be in log vs linear scale. The y axis can. This would allow many more harmonic markers to be shown and hence much more useful.

Not available in FFT, only in Bode plot mode.

Thanks. But what a shame. Would have been so nice.

[mawyatt]

One question I have yet to find an answer to is whether the x axis can be in log vs linear scale. The y axis can. This would allow many more harmonic markers to be shown and hence much more useful.

Not available in FFT, only in Bode plot mode.

Thanks. But what a shame. Would have been so nice.

Maybe a firmware update for the future

Best,

[jemangedeslolos]

Hello 

I had asked here I believe how the SDS2000X+ did well for triggering on a video pal signal and the answer was..... no.
What were your settings ?
I haven't yet been able to test on a noisy signal but on a clean signal, it trigs without problem on a line and it is rock stable 

[mawyatt]

Hello 

I had asked here I believe how the SDS2000X+ did well for triggering on a video pal signal and the answer was..... no.
What were your settings ?
I haven't yet been able to test on a noisy signal but on a clean signal, it trigs without problem on a line and it is rock stable 

You can get a "noisy signal" from the built-in AWG. Just set the AWG to a low amplitude and use a 10X scope probe. For example set AWG to 10mvpp, and the input to the scope will be 1mvpp, or set the AWG to 8mvpp and the input will be 800uvpp.

On my DSO I need about 800uvpp for a reliable trigger with a 1MHz sinewave and a little more for the counter. Here's examples with a noisy 800uvpp signal and using the average function with 16 count to nicely "clean up" the waveform. One is sine wave and the other is a Chirp Waveform, the sine wave trigger is very steady, the Chirp is not as stable but still acceptable, both at 800uvpp.

Added another using the AWG SNR function, which is a sine wave with a PN noise pattern. The frequency was reduced to 1KHz because the "noise" is not random but periodic and at 1MHz this could easily be seen when averaging.

I think this is impressive for a DSO at this modest cost point

Best,

[Howardlong]

Hello 

I had asked here I believe how the SDS2000X+ did well for triggering on a video pal signal and the answer was..... no.
What were your settings ?
I haven't yet been able to test on a noisy signal but on a clean signal, it trigs without problem on a line and it is rock stable 

I saw that, and I did originally achieve it reasonably easily on an NTSC signal: I did a pulse trigger on the frame, and added holdoff so it triggered on alternate frames, but other than that it wasn't difficult, it was solid, no problem seeing the colour burst for example thanks to the deep memory.

I've since found a PAL source, I can trigger rock solid on a PAL signal using a negative pulse trigger, it's a little more fiddly, so I'll provide a couple of screen shots. Trick is, as well as getting the pulse width right, to get the trigger level and holdoff right too.

Edit: This was done using the internal 50 ohm load on the 75 ohm signal so levels are a little lower than one might otherwise expect.

[rf-loop]

Hello 

I had asked here I believe how the SDS2000X+ did well for triggering on a video pal signal and the answer was..... no.
What were your settings ?
I haven't yet been able to test on a noisy signal but on a clean signal, it trigs without problem on a line and it is rock stable 

You can get a "noisy signal" from the built-in AWG. Just set the AWG to a low amplitude and use a 10X scope probe. For example set AWG to 10mvpp, and the input to the scope will be 1mvpp, or set the AWG to 8mvpp and the input will be 800uvpp.

On my DSO I need about 800uvpp for a reliable trigger with a 1MHz sinewave and a little more for the counter. Here's examples with a noisy 800uvpp signal and using the average function with 16 count to nicely "clean up" the waveform. One is sine wave and the other is a Chirp Waveform, the sine wave trigger is very steady, the Chirp is not as stable but still acceptable, both at 800uvpp.

Added another using the AWG SNR function, which is a sine wave with a PN noise pattern. The frequency was reduced to 1KHz because the "noise" is not random but periodic and at 1MHz this could easily be seen when averaging.

I think this is impressive for a DSO at this modest cost point

Best,

For this kind of chirp we can also try use, instead of rising edge and level threshold, example slope trigger (depending if signal is so that there is one detectable individual slope, as here example first rising slope) or just individual time interval inside one chirp what exist only once in one chirp.
Yes there is some learning curve before can reach level when can easy select and setup best trigger mode and parameters for different situations. We must learn out from  "rising edge and level + holdoff time"  thinking. I know, I have born analog scope in hand at 50's and it was nearly only way, if can not ext trig.

One example, still many of peoples may use edge- threshold + holdoff time when they trig for AM modulated signal. Because they are not familiar with lot of better methods. Yes it works whit quite narrow changes in modulating frequency but... in practice mostly signal are more complex that just simple fixed freg AM from signal generator. Same for this kind of signal in images.

As can see now there is visible some extra trig from other position in chirp. Just find unique detail inside one chirp and set trig trig to just it it. It can be example unique slope inside whole repetitive signal, like example in one chirp. In this example first falling slope. Or if think interval trig, last rising edge in end of chirp  to first rising edge in start of chirp happen only once in period.

Adjusting these may first feel some frustrating but after short learning time ... and so on. Modern scopes can do lot of...

[mawyatt]

Agree, looking for a unique waveform attribute then attempting to trigger is the best approach for complex waveform triggering.

The very old Tektronix analog scopes had some the best triggering circuits available then, they were almost "magical" in their triggering ability. They used some very clever circuits which employed ECL gates as very fast analog limiters/comparitors.

Another way to trigger on complex waveforms like these low level noise corrupted "Chirp" waveforms is to employ the Zone Triggering feature. Takes a few minutes to setup and the touch screen implementation isn't very easy, but this does work well. Here's a couple examples using the Zone Triggering, note the waveform levels, impressive ability to reliability trigger IMO.

Best,

[ci11]

Agree, looking for a unique waveform attribute then attempting to trigger is the best approach for complex waveform triggering.

The very old Tektronix analog scopes had some the best triggering circuits available then, they were almost "magical" in their triggering ability. They used some very clever circuits which employed ECL gates as very fast analog limiters/comparitors.

Another way to trigger on complex waveforms like these low level noise corrupted "Chirp" waveforms is to employ the Zone Triggering feature. Takes a few minutes to setup and the touch screen implementation isn't very easy, but this does work well. Here's a couple examples using the Zone Triggering, note the waveform levels, impressive ability to reliability trigger IMO.

Best,

Just for fun - would you mind running a couple of noise floor tests with the following settings?

DC Coupling
50 Ohm
10-bit
20MHz Bandwidth Limited

One without trace average
One with 50 trace averages

Attached please see results from my LeCroy 12-bit HDO for reference: without average: 56µV, with averages: 6.2µV

Thanks!

[mawyatt]

Just for fun - would you mind running a couple of noise floor tests with the following settings?

DC Coupling
50 Ohm
10-bit
20MHz Bandwidth Limited

One without trace average
One with 50 trace averages

Attached please see results from my LeCroy 12-bit HDO for reference: without average: 56µV, with averages: 6.2µV

Thanks!

I did a couple plots. One has the vertical scale at 1mv/div and shows the 64 average (it's 64 since the Avg Function only scales by powers of 2).
Next is with vertical scale 500uv/div, then using the Avg Function and Eres Function (3.0), with both scaled by 50X to 10uv/div.

As you can see the noise floor is very low in the 1mv/div & 500uv/div, and with the amplified vertical scales at 10uv/div.

Best,

[Martin72]

Hi,

Source for measuring could be also the function channels.
Did you use the formula editor for the 50 times averaging of eres (C1) ?

[2N3055]

Just for fun - would you mind running a couple of noise floor tests with the following settings?

DC Coupling
50 Ohm
10-bit
20MHz Bandwidth Limited

One without trace average
One with 50 trace averages

Attached please see results from my LeCroy 12-bit HDO for reference: without average: 56µV, with averages: 6.2µV

Thanks!

I did a couple plots. One has the vertical scale at 1mv/div and shows the 64 average (it's 64 since the Avg Function only scales by powers of 2).
Next is with vertical scale 500uv/div, then using the Avg Function and Eres Function (3.0), with both scaled by 50X to 10uv/div.

As you can see the noise floor is very low in the 1mv/div & 500uv/div, and with the amplified vertical scales at 10uv/div.

Best,

Mawyat,
I would suggest to use Stdev measurement instead RMS. RMS is AC+DC measurement and will include channel offset.
STDEV is AC RMS without DC component.
Regards,
Sinisa

[ci11]

Just for fun - would you mind running a couple of noise floor tests with the following settings?

DC Coupling
50 Ohm
10-bit
20MHz Bandwidth Limited

One without trace average
One with 50 trace averages

Attached please see results from my LeCroy 12-bit HDO for reference: without average: 56µV, with averages: 6.2µV

Thanks!

I did a couple plots. One has the vertical scale at 1mv/div and shows the 64 average (it's 64 since the Avg Function only scales by powers of 2).
Next is with vertical scale 500uv/div, then using the Avg Function and Eres Function (3.0), with both scaled by 50X to 10uv/div.

As you can see the noise floor is very low in the 1mv/div & 500uv/div, and with the amplified vertical scales at 10uv/div.

Best,

Thank you - it does look quiet. It does appear that your plots are in AC Coupling so here's one on the LeCroy HDO with AC Coupling, same filters, for a comparison. Also, to 2N3055's point, StDev may be a good measurement. It was a little confusing to me at first that Siglent called their AC RMS without DC "StDev" because "StDev" is also the variation across samples in cumulative measurements, and in Keysights, what Siglent refers to as "StDev" is known as "AC RMS", which is without the DC component.

[Martin72]

The fact, that the sds2k+ will compare seriously to a true 12 bit lecroy...And doesn´t suck....Wow, just wow.

[tautech]

The fact, that the sds2k+ will compare seriously to a true 12 bit lecroy...And doesn´t suck....Wow, just wow.

13 bit mode. 

[2N3055]

Thank you - it does look quiet. It does appear that your plots are in AC Coupling so here's one on the LeCroy HDO with AC Coupling, same filters, for a comparison. Also, to 2N3055's point, StDev may be a good measurement. It was a little confusing to me at first that Siglent called their AC RMS without DC "StDev" because "StDev" is also the variation across samples in cumulative measurements, and in Keysights, what Siglent refers to as "StDev" is known as "AC RMS", which is without the DC component.

AC RMS is Stdev ...

[mawyatt]

Hi,

Source for measuring could be also the function channels.
Did you use the formula editor for the 50 times averaging of eres (C1) ?

Yes.

Best,

[ci11]

AC RMS is Stdev ...

Please see attached red-circled measurements from my KS - the "RMS without DC" is called "AC RMS" and "Std Dev" is a statistical standard deviation of measurement values. This "RMS without DC" measurement on a Siglent (and some others) are called "Stdev". Threw me for a loop a few times until I got it straight.

[mawyatt]

The fact, that the sds2k+ will compare seriously to a true 12 bit lecroy...And doesn´t suck....Wow, just wow.

Agree, impressive DSO

This DSO has exceeded my expectations in every way, including the build quality (had an "incident" that caused me to completely disassemble the entire scope)

Want to thank Tautech for "steering" me in the right direction

Now I just need to get couple more channels (I wanted the SDA2104X Plus, but they were backordered (still are I believe), so got the 2 channel version that was in stock).

Best,

[Elasia]

The fact, that the sds2k+ will compare seriously to a true 12 bit lecroy...And doesn´t suck....Wow, just wow.

Agree, impressive DSO

This DSO has exceeded my expectations in every way, including the build quality (had an "incident" that caused me to completely disassemble the entire scope)

Want to thank Tautech for "steering" me in the right direction

Now I just need to get couple more channels (I wanted the SDA2104X Plus, but they were backordered (still are I believe), so got the 2 channel version that was in stock).

Best,

Importing is the best way to get one, i got mine from https://www.batronix.com/shop/oscilloscopes/Siglent-SDS2104X-plus.html

If you ask they will give you a discount if you mention the forum sent you.. also it gets coded to a tariff code that does not get taxed, least mine didnt

[mawyatt]

Mawyat,
I would suggest to use Stdev measurement instead RMS. RMS is AC+DC measurement and will include channel offset.
STDEV is AC RMS without DC component.
Regards,
Sinisa

Here's a plot with the STDEV for Input C1, C2 and the average(C1) multiplied by 10. C1 is the input with a 10X probe, C2 is the direct input with no attenuation @ 50 ohms (set to 1mv peak to peak at generator), F1 is the average of C1 times 10. So F1 is the attenuated (10X) input multiplied by 10, thus the same as C2 with direct input except with the additional C1 noise due to the 10X gain.

The reason C1 is showing such a high peak to peak reading is there is some narrow high amplitude "glitches" likely from other electronics or my computer nearby causing this to be high. With a 1mv peak to peak (354uv rms) on C2 and C1 with 10X probe should be ~ 100uvpp (35uv rms), note that F1 (10 times average C1) is showing better results due to the filtering effect on the random "glitches". I've also added a couple more plots showing F2 as 10*average(eres(C1)) with eres at 3.0, and included the statistics for F2.

Best,

[2N3055]

AC RMS is Stdev ...

Please see attached red-circled measurements from my KS - the "RMS without DC" is called "AC RMS" and "Std Dev" is a statistical standard deviation of measurement values. This "RMS without DC" measurement on a Siglent (and some others) are called "Stdev". Threw me for a loop a few times until I got it straight.

I know it can be confusing, and Keysight chose to call it with same terminology as in multimeter world, probably to avoid confusion too..
But Stdev IS AC RMS when you apply it to signal. When you apply it to a number of measurements you want to look at statistically, then you are looking at statistical measure.

Stdev is simply a formula that calculates mean of incoming data series (that would be DC component in a signal), subtracts that from data (to make AC only, only variable part of signal) and then calculates Root Mean Square only of that variable data (AC component) of that data, hence AC RMS...

So mathematically it is correct, but can be a bit confusing for those expecting engineering terminology instead of math terminology...

Regards,

Sinisa

[rf-loop]

AC RMS is Stdev ...

Please see attached red-circled measurements from my KS - the "RMS without DC" is called "AC RMS" and "Std Dev" is a statistical standard deviation of measurement values. This "RMS without DC" measurement on a Siglent (and some others) are called "Stdev". Threw me for a loop a few times until I got it straight.

I know it can be confusing, and Keysight chose to call it with same terminology as in multimeter world, probably to avoid confusion too..
But Stdev IS AC RMS when you apply it to signal. When you apply it to a number of measurements you want to look at statistically, then you are looking at statistical measure.

Stdev is simply a formula that calculates mean of incoming data series (that would be DC component in a signal), subtracts that from data (to make AC only, only variable part of signal) and then calculates Root Mean Square only of that variable data (AC component) of that data, hence AC RMS...

So mathematically it is correct, but can be a bit confusing for those expecting engineering terminology instead of math terminology...

Regards,

Sinisa

It need perhaps also note that example many multimeters really can not do Stdev for get "AC RMS". They just can not because there is not sampled true data for this. So there is other methods for do this estimate.

Now with oscilloscopes it is bit more possible but... then we need ask: Stdev math from what data.
Siglent calculate Stdev from all samples in acquisition memory, user know amount and samples interval and can also estimate how accurate true AC RMS (and true total RMS) it can calculate including possible higher frequency components in signal under test. So we know data used for Stdev math and can estimate its limits also for possible higher freq components in signal under test.

Keysight some models are bit more tricky and perhaps it is best they name it as in multimeters AC RMS. 
Who knows what data Keysight exactly use for estimate this AC RMS in models what use tiny intermediate buffer for measurements. How much signal is decimated. How it compute AC RMS. Where is definition how it is done.

[mawyatt]

AC RMS is Stdev ...

Please see attached red-circled measurements from my KS - the "RMS without DC" is called "AC RMS" and "Std Dev" is a statistical standard deviation of measurement values. This "RMS without DC" measurement on a Siglent (and some others) are called "Stdev". Threw me for a loop a few times until I got it straight.

Never thought about the AC RMS (DMM) and Std Dev relationship, thanks for pointing that out 2N3055.

In the KS 34465A (TrueVolt) the RMS value is achieved in a different manner than with the older 34401A which used a "Analog RMS (AD636) circuit". Think this is a sampled data type RMS computation, rather than an Analog type RMS computation.

Best,