Low Cost PCB's Low Cost Components

Author Topic: Compare apples with pears - Rigol/big mem. depth versus Agilent/small mem. depth  (Read 9131 times)

0 Members and 1 Guest are viewing this topic.

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
We keep hearing the same:

- Rigol has the biggest memory depth (up to date size), but Rigol is slow and software is not stable.

- Agilent has a very low memory depth (outdated size), but Agilent is fast and software is stable.

But isn't it the bigger memory depth, that makes things slower? So how can we compare honestly?

What if we really could compare apples with apples?

If everything about Agilent is great (fast, stable), except the memory depth, how come that Agilent does not just wake up, and make a new scope that keeps everything (all the great stuff) what they have today, but just add more memory?

Then we could finally compare apples with apples!

Then we would not longer have an excuse to remove Agilent from the wishlist because the memory depth is too low.

Then we could finally check if the Agilent is still as fast when the full memory depth is used.

Then we could finally check if the Agilent is still as stable when the full memory depth is used.

Then we could finally check if the Agilent still reaches waveform update rates of 1M waveforms/s.

Now it's Rigol's turn :)

If everything about Rigol is great (cheap, big memory depth, low power), except that the scope is slow, and becomes unstable when loaded too much, how come that Rigol does not just wake up, and implements the heavy CPU load stuff in hardware instead of software (throw in another FPGA - they don't cost a fortune these days, and the DS4000 series isn't necessarily cheap in the first place), and hire some more Chinese software engineers (to fix these bugs once and for all, the DS4000 series costs enough to get a product that actually does what it is supposed to do)?

Things look easy in an ideal world! :)
« Last Edit: August 05, 2015, 08:51:40 AM by pascal_sweden »
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 12572
  • Country: nl
    • NCT Developments
IMHO the problem is that Rigol has overlooked that it takes a lot of processing power to do calculations on a long memory. Agilent scopes do a lot of processing (decoding for example) in hardware to accellerate going through the memory.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline GlowingGhoul

  • Regular Contributor
  • *
  • Posts: 203
Keysight's 2000+ series overall performance far outclasses the Rigol in every substantive aspect irrespective of memory depth.

The memory limitations of Keysight units are directly related to it's high speed architecture, but I'd argue you're much, much more likely to be getting something useful out of the Keysight.

Then again maybe thousands of engineers, scientists, and institutions are just stupid and don't realize they could save 80% on scope costs and go with a Rigol since it's "even better" than a Keysight scope thanks to $4 worth of low grade PC memory.  :-DD

Rigol is great value for the money, but lets not kid ourselves about which is a hobbyist toy and which is professional grade equipment.
« Last Edit: August 05, 2015, 08:35:44 AM by GlowingGhoul »
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 9605
  • Country: nz
    • Taupaki Technologies Ltd.
IMO the quality of the core code and that in FW is likely to be the limiting performance factor, there's every likelyhood Rigol has a faster processor, they're so cheap these days, but that doesn't mean they know how to make it fly.
Anybody done this comparison?
Avid Rabid Hobbyist & NZ Siglent Distributor
 

Offline GlowingGhoul

  • Regular Contributor
  • *
  • Posts: 203
IMO the quality of the core code and that in FW is likely to be the limiting performance factor, there's every likelyhood Rigol has a faster processor, they're so cheap these days, but that doesn't mean they know how to make it fly.
Anybody done this comparison?

And therein lies the problem with even high quality copies of some other company's engineering work. There isn't the fundamental understanding of how to innovate, or even optimize what they build. The most likely scenario under which you'd see a Rigol perform like a Keysight is if general purpose processors become so fast and cheap they can catch up with MegaZoom performance. Of course Keysight won't be standing still either, and I've yet to see a Chinese company continue to improve their software once they've moved on to the next, bug ridden model.
 

Offline LabSpokane

  • Super Contributor
  • ***
  • Posts: 1899
  • Country: us
I don't fully understand this memory depth comparison/complaint, except with taking looooong samples of digital data.  The Agilent system of segmented memory seems to be ideal for analog purposes.  It obviously isn't ideal for logging reams of serial data. 

But, if I needed to take megabytes of digital data, the last thing I would want to do is use my scope to do it - then have to slog through all that crap with the clunky interface.  A PC-based logic analyzer like the Saleae seems much more useful for that purpose.
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
IMO the quality of the core code and that in FW is likely to be the limiting performance factor, there's every likelyhood Rigol has a faster processor, they're so cheap these days, but that doesn't mean they know how to make it fly.
Anybody done this comparison?
That's an interesting point actually. Even if we have to make abstraction about the fact that Agilent does a lot of stuff in hardware, it is a nice homework task to check and compare the processor cores in the Agilent gear with the processor cores in the Rigol gear.

This information is not only interesting for making a comparison study as such. It is generally interesting to learn about what chipsets are used, and which processor architectures are used in modern scopes. The study could be expanded to the system bus, to check the bandwidth, and check the type of interconnect that is used.
« Last Edit: August 05, 2015, 09:33:28 AM by pascal_sweden »
 

Online EEVblog

  • Administrator
  • *****
  • Posts: 24346
  • Country: au
    • EEVblog
If everything about Agilent is great (fast, stable), except the memory depth, how come that Agilent does not just wake up, and make a new scope that keeps everything (all the great stuff) what they have today, but just add more memory?

Because Agilent did what no one else is doing, they spun their own ASIC (MegaZoomIV) that has the 4M sample memory built into the chip. That is why it's so quick. Same with the fixed display waveform window size etc.
An ASIC like that is a huge investment in time and money, and they will want to milk it as long as possible in as many products as possible, hence the 4000 series, the touch etc are all a bit crippled because they had to use this ASIC.
You can bet your bottom dollar they are working in a MegazoomV though.
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
I don't fully understand this memory depth comparison/complaint, except with taking looooong samples of digital data.  The Agilent system of segmented memory seems to be ideal for analog purposes.  It obviously isn't ideal for logging reams of serial data. 

If you don't use the full memory depth in the Rigol scope, it isn't as slow as many Agilent people suggest on this forum.

But many Agilent people (without making any generalization though) on this forum have the tendency to compare the Rigol scope in its slowest configuration with the Agilent scope in its fastest configuration to direct the outcome of the battle towards the Agilent home front :)

Now here is an interesting question: If we would compare the "Pepsi versus Coca-Cola" battle with the "Rigol versus Agilent" battle. Which side (Rigol or Agilent) would be Pepsi, and which side (Rigol or Agilent) would be Coca-Cola? :) Which test gear has the most "bubbles" for money? =)

These kind of comparisons can be found in different fields. Let's take digital cameras for example.
There we have the "Nikon guys/girls" and the "Canon guys/girls" :) Funny actually that there typically are 2 big fronts. I guess it's making the choice easier, and you always have 50% chance to be on the right side :)
« Last Edit: August 05, 2015, 09:50:36 AM by pascal_sweden »
 

Offline dr.diesel

  • Super Contributor
  • ***
  • Posts: 1955
  • Country: us
  • Cramming the magic smoke back in...
Now here is an interesting question: If we would compare the "Pepsi versus Coca-Cola" battle with the "Rigol versus Agilent" battle. Which side (Rigol or Agilent) would be Pepsi, and which side (Rigol or Agilent) would be Coca-Cola? :) Which test gear has the most "bubbles" for money? =)

Agilent would be Mountain Dew, no crappy cola.

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 12572
  • Country: nl
    • NCT Developments
If everything about Agilent is great (fast, stable), except the memory depth, how come that Agilent does not just wake up, and make a new scope that keeps everything (all the great stuff) what they have today, but just add more memory?
Because Agilent did what no one else is doing, they spun their own ASIC (MegaZoomIV) that has the 4M sample memory built into the chip. That is why it's so quick. Same with the fixed display waveform window size etc.
Still you can do the same with an FPGA and DDR memory.But it takes more FPGA development work and a bigger FPGA to do this.

Which test gear has the most "bubbles" for money? =)
I found out for myself that at some point you need a piece of equipment which goes the extra mile and has features which work 100% instead of 75%.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline GlowingGhoul

  • Regular Contributor
  • *
  • Posts: 203
I don't fully understand this memory depth comparison/complaint, except with taking looooong samples of digital data.  The Agilent system of segmented memory seems to be ideal for analog purposes.  It obviously isn't ideal for logging reams of serial data. 

If you don't use the full memory depth in the Rigol scope, it isn't as slow as many Agilent people suggest on this forum.

But many Agilent people (without making any generalization though) on this forum have the tendency to compare the Rigol scope in its slowest configuration with the Agilent scope in its fastest configuration to direct the outcome of the battle towards the Agilent home front :)

Now here is an interesting question: If we would compare the "Pepsi versus Coca-Cola" battle with the "Rigol versus Agilent" battle. Which side (Rigol or Agilent) would be Pepsi, and which side (Rigol or Agilent) would be Coca-Cola? :) Which test gear has the most "bubbles" for money? =)

I don't think that's it at all. I can see the difference just doing the most basic things like navigating options on the Rigol. The Keysight is silky-smooth and ultra responsive, the Rigol is not.

Don't get me wrong, I think the Rigol brings otherwise unaffordable equipment to many people that works "well enough". If it weren't for the Rigols of the world I'm pretty sure Keysight wouldn't feel the need to be generous with free app bundles and they'd probably double their prices for good measure.  :-DD
 

Offline hendorog

  • Frequent Contributor
  • **
  • Posts: 821
  • Country: nz
Quote
Don't get me wrong, I think the Rigol brings otherwise unaffordable equipment to many people that works "well enough". If it weren't for the Rigols of the world I'm pretty sure Keysight wouldn't feel the need to be generous with free app bundles and they'd probably double their prices for good measure.  :-DD

Well said. Now all you Agilent guys buy us Rigol guys a beer since we did you a favour  :-DD
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
Yes, the perfect arrangement for these beers is the annual Belgian Beer Weekend in September.

To read all about this wonderful event in Brussels look here! =)
http://www.belgianbrewers.be/en/events/belgian-beer-weekend-171/
 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 3798
  • Country: gb
I use both Rigol and Agilent/Keysight scopes. For day to day work in the lab I use the Agilents pretty much exclusively, but for out in the field I take an MSO1074Z-S because it's so fully featured and its diminutive size and weight. I am out and about at the moment in fact, and am spending quite a bit of time with the Rigol. Each time I use it it grows on me a bit more.

For most things, the Rigol is more than adequate for what I do, but the UI is rather sluggish compared to the Agilents although it's still reasonably usable. Other downsides of the Rigol include muxed vertical controls and fiddly menu navigation.

I was trying to think of some functionality that the Agilents I use have that the Rigol doesn't, and it's hard. CAN and USB triggering for example, but then I hardly ever use those triggers. Being able to easily browse and search a long serial decode from a segmented capture on the Agilent is handy, the Rigol can browse to a degree but can't search. The Rigol _does_ have triggered segmented capability, it's just that it's called Record.

The biggest negative of the Rigol compared to the Agilents for my day to day work is the performance, 100MHz BW with 1GSa/s reducing to 250MSa/s for 4 ch is limiting. But for the money, the Rigol is outstanding value.
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
To make a long story short: Only with Rigol scopes you will get this "most bang for buck" feeling.
The Agilent will never give you that feeling, no matter how great it is :)

Only people who have a Rigol, know what that feeling means!

But of course, maybe some Agilent users are looking for another feeling, such as "I have the most expensive scope in town" feeling, in addition to some of their other feelings, such as "I play tennis on Fridays", "My Cannondale mountain bike has the lightest carbon available for consumer use, and is the second best material next to space craft material from NASA", "I am simply the best" =)
« Last Edit: August 07, 2015, 05:42:34 AM by pascal_sweden »
 

Offline dadler

  • Supporter
  • ****
  • Posts: 842
  • Country: us
Or how about the my scope just works feeling, or my scope is responsive feeling, or the confidence that it's your circuit and not the scope that is the problem feeling, or the well-engineered architecture feeling...

My most recent favorite is the my scope is actually in my possession feeling...since my DS2000 is back at Rigol for a-fixin' ....

That said the DS2000 does have a lot of "bang" for the buck. I think the power supply on mine went "bang" and I got that at no extra charge.

</cranky me>

I still :heart: you, Rigol. flaws and all.
« Last Edit: August 07, 2015, 06:00:00 AM by dadler »
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
Or how about the my scope just works feeling, or my scope is responsive feeling, or the confidence that it's your circuit and not the scope that is the problem feeling, or the well-engineered architecture feeling...

Nice extension of the list! Cool that people play along here in a sporty fashion :)
 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 3798
  • Country: gb
I am willing to be re-educated, but I suspect that some who wouldn't touch a Rigol with a barge pole, and have the privilege of using five or six figure scopes, don't actually buy those scopes with their own money.

Apart from being out on customer sites on a few field visits, every scope I've ever used since I left school I bought myself, whether it be Rigol, Tek or Agilent. It may be that those who buy their own scopes have a different concept of value to those who have their well-equipped labs paid for. That's not meant as a a criticism, just an observation.
 

Offline GlowingGhoul

  • Regular Contributor
  • *
  • Posts: 203
I am willing to be re-educated, but I suspect that some who wouldn't touch a Rigol with a barge pole, and have the privilege of using five or six figure scopes, don't actually buy those scopes with their own money.

Apart from being out on customer sites on a few field visits, every scope I've ever used since I left school I bought myself, whether it be Rigol, Tek or Agilent. It may be that those who buy their own scopes have a different concept of value to those who have their well-equipped labs paid for. That's not meant as a a criticism, just an observation.

Most of the personally purchased Keysight scopes mentioned here are well within the FOUR digit price range.

You can buy a brand new four channel, 100MHZ Keysight DSOX2000 series with the full package of Keysight options (except MSO) for around $2000 USD. That's well within the range of what many hobbyists pay for performance bicycles, photography equipment, or gaming PCs.

Or you can save a little money that will cost you a lot of time.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 12572
  • Country: nl
    • NCT Developments
I am willing to be re-educated, but I suspect that some who wouldn't touch a Rigol with a barge pole, and have the privilege of using five or six figure scopes, don't actually buy those scopes with their own money.
There is always the second hand market which makes once expensive equipment affordable. Depending on your requirements a piece of second hand equipment can be a better buy especially if a 'Rigol' costs several $k as well.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline hendorog

  • Frequent Contributor
  • **
  • Posts: 821
  • Country: nz
Quote
You can buy a brand new four channel, 100MHZ Keysight DSOX2000 series with the full package of Keysight options (except MSO) for around $2000 USD.

Thats pretty cool regarding the options, although its a shame though that Agilent haven't figured out that to play in the low end of the market they need to suck it up and publish their damn prices on their pretty website. It's not rocket salad, geeze.

Looks like they won't sell to NZ unless its through a distributor, changing the country to US shows a _base_ price of $2313
Any brand name stuff here is traditionally seriously overpriced, anyone on a budget imports it themselves.

The LAN connectivity is an option (which is a bit lame in this day and age, maybe in the 1990's that was OK)

To get any further with the pricing I have to give them my details, which will no doubt result in regular badgering from the local agent.
 

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1373
  • Country: no
You can buy a brand new four channel, 100MHZ Keysight DSOX2000 series with the full package of Keysight options (except MSO) for around $2000 USD. That's well within the range of what many hobbyists pay for performance bicycles, photography equipment, or gaming PCs.

Modern tachometers on bikes, or even some sports watches today, have a higher memory depth than the DSOX2000 series :)

Why do they call it InfiniiVision? Maybe they should call it FiniiVision, SlimiiVision, or TunnelVision =)

Disclaimer: FiniiVision (R), SlimiiVision (R) and TunnelVision (R) are Registered Trademarks by Wan Hung Lo Factories. All Rights Reserved :)
 

Online wraper

  • Supporter
  • ****
  • Posts: 6012
  • Country: lv
You can buy a brand new four channel, 100MHZ Keysight DSOX2000 series with the full package of Keysight options (except MSO) for around $2000 USD. That's well within the range of what many hobbyists pay for performance bicycles, photography equipment, or gaming PCs.

Modern tachometers on bikes, or even some sports watches today, have a higher memory depth than the DSOX2000 series :)

Why do they call it InfiniiVision? Maybe they should call it FiniiVision, SlimiiVision, or TunnelVision =)

Disclaimer: FiniiVision (R), SlimiiVision (R) and TunnelVision (R) are Registered Trademarks by Wan Hung Lo Factories. All Rights Reserved :)
And what the point of all of that memory depth if it is cheap slow useless crap? How much money costs 1 gbit ram nowadays? On Rigol, using any significant amount of memory depth cripples the performance drastically. When you try to use the same amount of memory as Keysight have, it will have much slower waveform update rate already, not to say if you try to use more. At the end of the day you'll find out than mostly use Rigol scope on the lowest memory settings, only rarely using more. However with the keysight scope, you use max memory always, no matter what.
« Last Edit: August 07, 2015, 11:39:15 AM by wraper »
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 9605
  • Country: nz
    • Taupaki Technologies Ltd.
At the end of the day you'll find out than mostly use Rigol scope on the lowest memory settings, only rarely using more.
Really?  :o

I've never experienced that.  :phew:
Avid Rabid Hobbyist & NZ Siglent Distributor
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf