Atten 200 MHz GA1202 CAL vs Rigol 1054Z: memory depth sucks - fraudulent?

[veek]

The Atten (200MHz 2GSa/s) has a memory depth of 40kPts which halves to 20kPts once I start the second channel. The Rigol 1054Z has 12Mpts default and is a mere 50-100 (hackable) MHz BW.

Price wise, the Atten is about USD 150 lesser than the Rigol and has a larger bandwidth/sample-rate.

My question is this: 200Mhz is the maximum achievable sample rate under optimum conditions on the Atten. Does the Atten make this impossible under real world conditions? Are they, in short, screwing me over by promising me too much by making the marketing look very attractive? In the real world with so low a memory depth, even if I hack it and solder on more memory, will the Atten CPU simply fail in its task?  What good is 200MHz and 2GSa/s if there's not enough memory to store the samples in, for analysis? And since the extra memory is quite cheap, I'm a bit suspicious about this 'deal'. Has anyone on the forum experienced problems with Atten where they find their Ferrari has it's wheels off the ground: like a 5GHz processor that only hits 5GHz once a day? How shitty is the Atten 1202 vs the Rigol 1054Z? What's the wfms/s of the Atten - couldn't find that, Sparkfun mentions everything else except that: https://www.sparkfun.com/products/11766? I don't get it though, the Rigol DS1102E 1GSa/s 100Mhz has 16Kpts/ch so 40kpts ought to be swimming if you are a Rigol fan.. (http://www.docircuits.com/blog/scopewar-tektronix-vs-agilent-vs-rigol/) unless both of them are padding their datasheet with marketing memory? I checked the datasheet for the Rigol 1000Z series and they just have 'Analog Channel 12 Mpts single channel' - so it's not switching between long/normal memory?

(Also no intensity gradient on the Atten - far as I can tell whereas the Rigol has it)

[Fungus]

The Atten (200MHz 2GSa/s) has a memory depth of 40kPts which halves to 20kPts once I start the second channel. The Rigol 1054Z has 12Mpts default and is a mere 50-100 (hackable) MHz BW.

Price wise, the Atten is about USD 150 lesser than the Rigol and has a larger bandwidth/sample-rate.

The 200MHz 2GSa/s version won't be that cheap.

The Rigol has 24Mpts when hacked and the measured bandwidth is well over 100MHz (more like 150). Why would you want anything else?

PS: Why is it "fraudulent"? A manufacturer can put in any amount of RAM they feel like.

[coppice]

Some information says the 200MHz version of the Gratten offers 2Gsps for 1 channel and 1Gsps for 2 channels, which is faster than the lower models in the range. Other information says its 1Gsps for 1 channel and 500Msps for 2 channels, just like the lower models in the range. I wonder which is accurate?

[necessaryevil]

I would only buy a scope if it's reviewed by Dave (or mabye someone else as long as it is a serious review) has approved it. Go for second hand or save up for the Rigol.  That being said, I have no hands-on experience with Atten nor Rigol.

Also check out Dave's oscilloscope tutorials and his reviews.

[Fungus]

Some information says the 200MHz version of the Gratten offers 2Gsps for 1 channel and 1Gsps for 2 channels, which is faster than the lower models in the range.

That seems correct.

Other information says its 1Gsps for 1 channel and 500Msps for 2 channels, just like the lower models in the range.

Looks like a copy/paste error.

http://www.gratten.eu/ga1202cal-200hmz.html

PS: I wonder if the base model is hackable to 300MHz / 2GSps 

[veek]

@Fungus you're kind of missing the point. They can offer 10000 GSa/s, it's just a number in a datasheet. If you hack it to 300MHz, the cpu  and the memory depth are still the same, so your scope will still sample at a lousy sample rate. Yes it can hypothetically sample at 2GSa/s but it won't ever do so without you narrowing the Acquisition window even further and increasing Dead time even further.

I'm just trying to estimate how small the Acquisition window is and how large the dead time is for 40kPts AT 200 MHz at 2 GSa/s. (for the Rigol too) All three values effect performance - he's giving you 2 large and one very small...

[tautech]

Some information says the 200MHz version of the Gratten offers 2Gsps for 1 channel and 1Gsps for 2 channels, which is faster than the lower models in the range.

That seems correct.

Other information says its 1Gsps for 1 channel and 500Msps for 2 channels, just like the lower models in the range.

Looks like a copy/paste error.

http://www.gratten.eu/ga1202cal-200hmz.html

PS: I wonder if the base model is hackable to 300MHz / 2GSps 

Of course it's hackable, just that nobody has yet AFAIK.

These Attens and Siglents are all quite similar, Historically the PCB's have the Siglent logo silkscreened on them.
This Atten differs from the current Siglents as it has a little more memory depth, the similar Siglent model having only 24 kpts
http://www.siglentamerica.com/pdxx.aspx?id=58&T=2&tid=1

Siglents offering with 40 kpts is the SDS1000CNL range, however it only offers 1Gs/s but no 300 MHz model.
http://www.siglentamerica.com/pdxx.aspx?id=50&T=2&tid=1

@Fungus you're kind of missing the point. They can offer 10000 GSa/s, it's just a number in a datasheet. If you hack it to 300MHz, the cpu  and the memory depth are still the same, so your scope will still sample at a lousy sample rate. Yes it can hypothetically sample at 2GSa/s but it won't ever do so without you narrowing the Acquisition window even further and increasing Dead time even further.

I'm just trying to estimate how small the Acquisition window is and how large the dead time is for 40kPts AT 200 MHz at 2 GSa/s. (for the Rigol too) All three values effect performance - he's giving you 2 large and one very small...

Discussion and links on DSO blind/dead time:
https://www.eevblog.com/forum/testgear/siglent's-new-product-msosds2000-series/msg464842/#msg464842

[Fungus]

@Fungus you're kind of missing the point. They can offer 10000 GSa/s, it's just a number in a datasheet.

No, I didn't miss that. This guy asked a question, I answered it.

If you hack it to 300MHz, the cpu  and the memory depth are still the same, so your scope will still sample at a lousy sample rate. Yes it can hypothetically sample at 2GSa/s but it won't ever do so without you narrowing the Acquisition window even further and increasing Dead time even further.

Yes, it's all horribly unbalanced and senseless. It's not a DS1054Z killer. You'd regret buying one.

[Fungus]

I wonder if the base model is hackable to 300MHz / 2GSps 

Of course it's hackable, just that nobody has yet AFAIK.

And with that memory size... would you really want to? It still wouldn't be a good scope to use. The '300MHz" is just a number.

[tautech]

If you hack it to 300MHz, the cpu  and the memory depth are still the same, so your scope will still sample at a lousy sample rate. Yes it can hypothetically sample at 2GSa/s but it won't ever do so without you narrowing the Acquisition window even further and increasing Dead time even further.

Yes, it's all horribly unbalanced and senseless. It's not a DS1054Z killer. You'd regret buying one.

If you "needed" the BW there's no way the 1054Z could compete as their -3dB rolloff is just over 100MHz.

I wonder if the base model is hackable to 300MHz / 2GSps 

Of course it's hackable, just that nobody has yet AFAIK.

And with that memory size... would you really want to? It still wouldn't be a good scope to use. The '300MHz" is just a number.

Some have also upgraded their Siglent CNL(40 kpts) models to 2 Mpts with just a chip replacement, if the base model's price is right this becomes an option to improve specs. 
Thread on it somewhere in the forum.

[ebastler]

@Fungus you're kind of missing the point. They can offer 10000 GSa/s, it's just a number in a datasheet. If you hack it to 300MHz, the cpu  and the memory depth are still the same, so your scope will still sample at a lousy sample rate. Yes it can hypothetically sample at 2GSa/s but it won't ever do so without you narrowing the Acquisition window even further and increasing Dead time even further.

I think you don't have your terminology straight here. I assume that you are not doubting the sample rate of 2 GHz, but are concerned about a low number of waveforms per second, or a low duty cycle -- right?