At last, Siglent's SDS5054X touchscreen DSO...

[Rerouter]

the AWG they are using is maximum amplitude of 6V pk-pk, with a maximum output of +-4V, e.g. 6V amplitude with a 1V offset,

Above that it triggers the "Output protection" I haven't yet dug into where that is in the margins. and the function can be turned off, But i would like to know what its protecting first.

You can think of it as the second channel of a SDG1025, I am trying to work out if the output voltage can be increased, but not there just yet.

The DAC alone is capable of 125MHz update rate. and it is indeed being updated at that rate, this is obvious because It has 1 sample time worth of jitter when you get close to the maximum frequency (this isn't a design flaw, Its just the limitation of the sampling rate vs output frequency resulting in a poor division)

Equally it has an arb repeat rate of up to 5MHz, and does this by picking intermediate samples when the repeat rate is faster than the update rate can handle. so technically there is no reason this cannot output a 62.5MHz square wave, though it would be much closer to a triangle wave at that point.

Edit: The output has 3 output ranges, 4-63mV pk-pk (+-705mV Offset), 64-631mV pk-pk (+-4V Offset),  and 632-6000mV pk-pk (+-4V Offset) I cannot yet clarify if they have the full 14 bit resolution for all 3 ranges, or if its switching something else.

[tautech]

Why do they call the AWG option "built-in" in the header/title of that section, when it seems to actually be an external device, just being controlled by the scope software (like 2000X-E, which has "module" in the corresponding title)?

Yes, I can see how this could be confusing:
25 MHz function / arbitrary waveform generator, built-in multiple predefined waveforms

Maybe a better wording would be in-built and addition of Optional module, however this is only describing the waveform types available in the optional SAG1021 AWG external HW.

These new SDS5000X, SDS2000X-E and SDS1*04X-E all share the SAG1021 AWG HW.

[Martin72]

Still confusing…
So you must buy the HW (SAG1021) and SW (License) to use the AWG functionality?
Or to use the SAG1021 itself?
So the scopes are only the frontpanel for it, right ?
The option cost 296€ (incl. VAT) plus the SAG1021 for 165€....
Why should I spend about 460€ for a 1-Ch AWG with limited possibilities....

[Rerouter]

Correct Martin, to use it via the scope UI you need to buy both the device, and the licence,

whereas if you have an external siglent AWG, you can still use bode plot mode with it out of the box. Its only the control of the USB AWG from the scope UI that would are paying for with the license.

The other pain I hope they address is the UI controls do not expose everything the device can do, only the "BasicWave" mode, One thing I have put on the table that should make things better is a Waveform Copy - Paste function, So from the UI you could replay waveforms, But I suspect that will take a few firmware updates to become a reality.

This thing can be more than a toy (output voltage limitations aside), you just can't really pull off the fun stuff without using SCPI commands at this point in time, E.g. you can have it setup to repeat a AWG waveform when the frequency counter input gets a trigger event, so you run this to the scope trigger output and can have something that can burst out a AWG waveform on a pattern match of the scope, The functionality is there, I just cant use it without connecting a PC to it.....

To be flat out honest, I bought the device and Licence for my 1000X-E scope, Its useful, but the limitations to the UI hurt when the license is 20% of the purchase price of that scope itself, and feels more like 0.1% of the dev time.

[tautech]

Still confusing…
So you must buy the HW (SAG1021) and SW (License) to use the AWG functionality?

Yes.

Or to use the SAG1021 itself?

Only for Bode plot.........but check the documentation as I believe its not been enabled in SDS5kX firmware just yet.

So the scopes are only the frontpanel for it, right ?

Correct, there's some usage examples in the User manual.

The option cost 296€ (incl. VAT) plus the SAG1021 for 165€....
Why should I spend about 460€ for a 1-Ch AWG with limited possibilities....

Where ? 
US pricing for SDS2kX-E and 1kX-E SAG1021 HW and SW license is $ 288 ...........sure no taxes but 460 Euro ?   
It makes no sense the AWG HW option and licensing for SDS5kX should/could be that much more expensive. 

Edit to add
Just hunted out SDS5kX pricing and shocked !
SAG1021 HW             $ 159 (as expected)
SDS-5000X-FG SW $$$ 289    Seems like an error.................

[2N3055]

Where ? 
US pricing for SDS2kX-E and 1kX-E SAG1021 HW and SW license is $ 288 ...........sure no taxes but 460 Euro ?   
It makes no sense the AWG HW option and licensing for SDS5kX should/could be that much more expensive. 

Batronix lists it as € 296.31 incl. 19% VAT + € 165.41 incl. 19% VAT. Same in Batterfly, few € up or down.

[tautech]

Where ? 
US pricing for SDS2kX-E and 1kX-E SAG1021 HW and SW license is $ 288 ...........sure no taxes but 460 Euro ?   
It makes no sense the AWG HW option and licensing for SDS5kX should/could be that much more expensive. 

Batronix lists it as € 296.31 incl. 19% VAT + € 165.41 incl. 19% VAT. Same in Batterfly, few € up or down.

Thanks but that doesn't explain why just the FG SW option for SDS5kX is the same price as for both the SAG1021 HW and SW license for SDS2kX-E and 1kX-E models.
Seems wrong ......we dealers need ask some questions.

[Performa01]

The RTB2K with History(ext mem) has 160 x 4(320 x 2) Mpts. Not sure about siglent but you get the same memory for logic also but that's not really 1:1. The RTM has something like 200 x 4Mpts.

Okay, I've overlooked the additional memory that comes with the Segmented Memory option. To be honest I didn't expect an A-brand to be so generous, but then again, R&S seems to be different

You can say they're conservative but the datasheet says what it says. 

I have measured the actual channel isolation at 580MHz on my pre-production unit and the worst case scenario was from Ch.3 (FS) to Ch.4, both at 5mV/div, resulting in -53dB crosstalk. In all other cases, the isolation was significantly higher:
Ch.1 (5mV/div – 1V/div, FS) to Ch.4, isolation 79-90dB
Ch.2 (5mV/div – 1V/div, FS) to Ch.4, isolation 64-108dB
Ch.3 (5mV/div – 1V/div, FS) to Ch.4, isolation 53-88dB

Actually thinking about segmented memory reminded me... The Siglent was something like 2us and the RTM is 200ns between acqs, that's pretty significant.

Yes it is. BTW, actually measured lowest re-arm time between acquistions was 1.49µs for the SDS5104X.

Generally, the SDS5000X cannot be both cheaper and better than all the competition and in every regard. While its waveform update rate might not be particularly fast, I’m positive the functionality of the SDS5000X will look rather good once it’s feature-complete.

Why do they call the AWG option "built-in" in the header/title of that section, when it seems to actually be an external device, just being controlled by the scope software (like 2000X-E, which has "module" in the corresponding title)?

That’s just a copy & paste error. There have been several such occurrences in the datasheet, most of them got corrected, but the Header on page 7 has been overlooked…


Those who think the SAG1021 + SW Option is not very attractive, you are not alone. To be honest, I don’t understand the whole idea of a built-in (or external USB powered brick) AWG. On a small portable service oscilloscope okay, where the tech doesn’t want to carry more items than absolutely necessary. But on a lab instrument, especially one that is not intended for the hobbyist market?

If I were in that situation, I would of course get something more versatile and powerful instead, like the SDG1032X (or above, especially for a 1GHz scope), without a second thought.

A complete review of the SAG1021 (well, at least the HW and what’s accessible through the scope UI) can be found in this thread, replies #48/49:

https://www.eevblog.com/forum/testgear/siglent-sds1104x-e-in-depth-review/msg1432665/#msg1432665

[Martin72]

Correct Martin, to use it via the scope UI you need to buy both the device, and the licence,

whereas if you have an external siglent AWG, you can still use bode plot mode with it out of the box. Its only the control of the USB AWG from the scope UI that would are paying for with the license.

Therefore welectron (german siglent distributor) did a suggest to buy a stand-alone AWG.
I bought a SDG1032X, it got two channels and much better specs.
And cost 403 €.....

[tautech]

Those who think the SAG1021 + SW Option is not very attractive, you are not alone. To be honest, I don’t understand the whole idea of a built-in (or external USB powered brick) AWG. On a small portable service oscilloscope okay, where the tech doesn’t want to carry more items than absolutely necessary. But on a lab instrument, especially one that is not intended for the hobbyist market?

If I were in that situation, I would of course get something more versatile and powerful instead, like the SDG1032X (or above, especially for a 1GHz scope), without a second thought.

Agreed, on all counts, however it's not like Siglent hasn't later added AWG HW into a DSO as they did for SDS1000X models where it is an additional internal PCB, first it was called the X-S then later the 1000X+ that offered MSO and internal AWG.
These are still available and the basic SDS1000X MSO model without internal AWG is just some $ 240 less than SDS1000X+ models with the additional internal AWG PCB.

For the SDS2000 and 2000X models AWG HW was incorporated into the mainboard in all models from first release.

[0xdeadbeef]

Talking of weird pricing: Batronix offers the update "SDS-5000X-4BW10" from 4ch 500MHz to 1GHz for 320€ 
Sadly a mistake obviously. I guess they meant 3200€

[maginnovision]

Talking of weird pricing: Batronix offers the update "SDS-5000X-4BW10" from 4ch 500MHz to 1GHz for 320€ 
Sadly a mistake obviously. I guess they meant 3200€

No, I think that's the pricing.

[0xdeadbeef]

Nah...
http://www.saelig.com/product/sds-5000x-4bw10.htm

[tv84]

These romans are crazy!!!  It shouldn't be one or the other but... big/strange marketing strategies in play!

Batronix (VAT included):

Siglent SDS5052X - 5484.71€
Siglent SDS5102X - 6222.51€

Why pay 3.200€ for the upgrade?

I think, we'll soon start seeing this option also:  or Best Offer   

[maginnovision]

Nah...
http://www.saelig.com/product/sds-5000x-4bw10.htm

Yea, that's odd because the upfront cost difference for 500-1000 with same channels is way less.

[0xdeadbeef]

The upgrade option is usually more expensive than the price difference between two models of different bandwidth. Still, this upgrade seems to be especially expensive.
@tv84: you compared the 2ch prices, but this is the 4ch upgrade. The 4ch price difference of the 500MHz and 1GHz model is 1666€. So the upgrade options costs about twice of that.

[tv84]

The upgrade option is usually more expensive than the price difference between two models of different bandwidth. Still, this upgrade seems to be especially expensive.
@tv84: you compared the 2ch prices, but this is the 4ch upgrade. The 4ch price difference of the 500MHz and 1GHz model is 1666€. So the upgrade options costs about twice of that.

I would understand that if it was a hardware upgrade.

Yes, I compared the 2ch but the 4ch was also not 3.2k.

It's a pricing strategy as any other...

[egonotto]

Hello,

Actually thinking about segmented memory reminded me... The Siglent was something like 2us and the RTM is 200ns between acqs, that's pretty significant.

Yes it is. BTW, actually measured lowest re-arm time between acquistions was 1.49µs for the SDS5104X.

In the datasheet of the RTM3000 there is:
 "continuous recording of waveforms in acquisition memory without interruption due to
visualization; blind time between consecutive acquisitions less than 1.5 μs"

I can not find 200ns.

And the total memory (per channel) is "At interleaved mode." in the datasheet of the RTM3000
Is this in the RTB2000 different or forget they the "At interleaved mode." in the datasheet of the RTB2000?

Best regards
egonotto

[Performa01]

Hello,

In the datasheet of the RTM3000 there is:
 "continuous recording of waveforms in acquisition memory without interruption due to
visualization; blind time between consecutive acquisitions less than 1.5 μs"

I can not find 200ns.

And the total memory (per channel) is "At interleaved mode." in the datasheet of the RTM3000
Is this in the RTB2000 different or forget they the "At interleaved mode." in the datasheet of the RTB2000?

What version of the datasheet do you look at?

Version 06.00, January 2019 (on page 24):

Fast-segmented mode continuous recording of waveforms in acquisition memory without interruption due to visualization; blind time between consecutive acquisitions less than 200 ns
(up to 2 000 000 waveforms/s)

The waveform update rate in fast segmented mode is really impressive - less so during regular operation, where it's just 64k wfm/s.

Yes, the total memory is specified for interleaved mode (= per channel pair) and there it can reach a maximum of 428Mpts, but only in segmented mode and only with the K15 Option (which is some €1000,- without tax). For regular acquisition it is just 80Mpts (interleaved).

The RTB2000 has 2.5µs re-arm time (300k wfm/s) in fast segmented mode and up to 320Mpts memory (interleaved).

[egonotto]

Hello,

thanks for the answer.

I used the datasheet from Batronix it is Version 05.00,  December 2017.

It is an impressive improvement.

Best regard
egonotto

[0xdeadbeef]

Actually thinking about segmented memory reminded me... The Siglent was something like 2us and the RTM is 200ns between acqs, that's pretty significant.

Yes it is. BTW, actually measured lowest re-arm time between acquistions was 1.49µs for the SDS5104X.

Interesting though that the SDSX5000 datasheet states a minimum (!) dead time of 2µs.
In contrast to this, the 200ns given for the RTM3K is the maximum value. So actually comparing both value doesn't make much sense anyway.
Still, while the 200ns dead/blind time sure is impressive, I don't think 2µs or so would be much of an issue in 99.9% of all the cases where you use segmented memory.
The only thing that concerns me is that Siglent doesn't specify a maximum value at all which makes it impossible to judge if this is a major issue or not.

[2N3055]

I'm confused:
That RTM3000 /RTA4000 spec says :

continuous recording of waveforms in acquisition memory without interruption due to
visualization; blind time between consecutive acquisitions less than 200 ns
(up to 2 000 000 waveforms/s)

They should decide whether is:
200 ns and (up to 5 000 000 waveforms/s)
or
500 ns and (up to 2 000 000 waveforms/s)

 

[nctnico]

I'm confused:
That RTM3000 /RTA4000 spec says :

continuous recording of waveforms in acquisition memory without interruption due to
visualization; blind time between consecutive acquisitions less than 200 ns
(up to 2 000 000 waveforms/s)

They should decide whether is:
200 ns and (up to 5 000 000 waveforms/s)
or
500 ns and (up to 2 000 000 waveforms/s)

There will be a minimum amount of time needed to do the acquisitions as well (time per sample * number of samples). So the total time per frame equals to the acquisition time plus the dead time.

Interesting though that the SDSX5000 datasheet states a minimum (!) dead time of 2µs.
In contrast to this, the 200ns given for the RTM3K is the maximum value. So actually comparing both value doesn't make much sense anyway. or not.

IMHO Siglent probably got minimum and maximum confused. It doesn't make sense to specify minimum dead time. Maybe they reasoned the smallest guaranteed dead time is 2us and thought they should put that under minimum because it is the smallest (which is wrong ofcourse).

[maginnovision]

I'm confused:
That RTM3000 /RTA4000 spec says :

continuous recording of waveforms in acquisition memory without interruption due to
visualization; blind time between consecutive acquisitions less than 200 ns
(up to 2 000 000 waveforms/s)

They should decide whether is:
200 ns and (up to 5 000 000 waveforms/s)
or
500 ns and (up to 2 000 000 waveforms/s)

Presumably they assume it takes time to capture data, it's not 100% re arm time.

[2N3055]

There will be a minimum amount of time needed to do the acquisitions as well (time per sample * number of samples). So the total time per frame equals to the acquisition time plus the dead time.

Of course, but that still means that blind time + sweep cannot be less than 500 ns.
And if you set it to 10ns /div that is 120 ns +200 ns = 320 ns.
Blind time is not only rearm time, but rearm + processing overhead time.
Max. 2 Mwaveforms means a trigger every 500 ns.

It's nitpicking, I know but specifications are specification.