Author Topic: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown  (Read 93332 times)

0 Members and 1 Guest are viewing this topic.

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #25 on: September 26, 2012, 12:57:45 pm »
The Rigol website lists only these options:
SD-DS2 (serial decode)
MEM-DS2 (mem upgrade)
AT-DS2 (advanced triggering)

Dave.
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #26 on: September 26, 2012, 01:16:28 pm »
This model must have a lot of the established names CR@PPING themselves. Where do they go, from here?

I think the power supply is custom built for Rigol, by ALPS (see part-covered silk screen printing in bottom left hand corner of the board). No shame in that.
If you don't measure, you don't get.
 

Offline A Hellene

  • Frequent Contributor
  • **
  • Posts: 602
  • Country: gr
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #27 on: September 26, 2012, 01:25:44 pm »
The Rigol website lists only these options:
SD-DS2 (serial decode)
MEM-DS2 (mem upgrade)
AT-DS2 (advanced triggering)
Yes, Dave, I've seen that. These functions above must be already existing in the firmware, since they can be individually unlocked ("installed") by entering the appropriate "serial numbers."

Additionally, the DS2000 front-end does not seem to be much different to the DS1000 front-end, splitting the input signal into its DC and AC components.

The first front-end 16-pin chip might be a specialised oscilloscope front-end amplifier, like the LMH6518, which is a 900 MHz VGA with a programmable bandwidth limit of 20/100/200/350/650/750 MHz that has a suspiciously identical package to the one the DS2000 seems to have!

The following chip must be a standard high bandwidth Differential Amplifier, like the 1.5 GHz LMH6552.


-George
« Last Edit: September 26, 2012, 01:29:05 pm by A Hellene »
Hi! This is George; and I am three and a half years old!
(This was one of my latest realisations, now in my early fifties!...)
 

Offline Eliminateur

  • Regular Contributor
  • *
  • Posts: 179
  • Country: ar
  • Electronic's Technician
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #28 on: September 26, 2012, 07:11:28 pm »
I've seen those through hole metal loops fail on PC motherboards too. There's quite a bit of force on them.
you didn't check the soldering of those in the back Dave!

and yes, i've seen those crappy loops take off clean in a very expensive RACKED HP server (i think it was a DL380), i put racked in caps because the motherboard lays FLAT, INCREDIBLY didn't shorted out ANYTHING in the MB(i think it shorted the NIC and might have smoked one, can't remember).
When we inspected it, it's as if someone had used a solder sucker on it, probably a very cold bad solder...

So we resoldered them with good old full of tasty lead solder, never moved again :D
 

Offline Dread

  • Regular Contributor
  • *
  • Posts: 122
  • Country: us
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #29 on: September 26, 2012, 11:53:29 pm »
This model must have a lot of the established names CR@PPING themselves. Where do they go, from here?

I think the power supply is custom built for Rigol, by ALPS (see part-covered silk screen printing in bottom left hand corner of the board). No shame in that.

Not so sure about that!  Take a look at how the price sky rockets when you go up in features and BW

$839 is for 70 MHz
$1140 for 100 MHz
$1625 for 200 MHz
+$222 for serial decode
+$334 for 56MB memory from the standard 14Mb
+$222 for advance trigger options.

So the full 200 MHz version will cost a cool $2403.  That's way above the price of many other hobby grade 200 MHz scopes.   Would I buy the Rigol base model at $839 probably not because of the 70 MHz limit but that's me, for other people 70MHz is fine and they will certainly get better quality with the Rigol vs most other $800 scopes that can do 200 MHz but I needed  a scope that could go past 140MHz and was under $1000 so the Rigol was out for me.



The Optimist says the glass is half full, the Pessimist says its half empty, an engineer only see's a glass that’s twice as big as it needs to be!
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #30 on: September 27, 2012, 12:38:13 am »
Not so sure about that!  Take a look at how the price sky rockets when you go up in features and BW

$839 is for 70 MHz
$1140 for 100 MHz
$1625 for 200 MHz
+$222 for serial decode
+$334 for 56MB memory from the standard 14Mb
+$222 for advance trigger options.

Yes, it does get pricey.
I can't see too many people buying the 56M upgrade, as 14M is already oodles of memory.
The serial decode might be popular though, and I think reasonably priced.
Advanced trigger is handy, but you can in most cases get by without if needed.
Their real problem is the lack of ability to upgrade the bandwidth after purchase, that sucks.
I guess they didn't want people to be shocked by the $800 it would cost to go from 70-200MHz, and not give them that psychological problem of knowing their scope bandwidth is "crippled" in software. But Agilent decided to go that route.

Dave.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #31 on: September 27, 2012, 01:11:40 am »
I am just waiting a couple more decades (or maybe just a few years) until more Chinese vendors will pop up and sell a 200 MHz scope which is plain 200 MHz but at a 70 MHz price point. Because clearly Rigol can make one at a 70 MHz price point, they just don't want to sell it at that because they wouldn't profit as much. I'm pretty sure they make a decent margin on the 70 MHz version. The most expensive things on that scope are probably the Spartan-6 devices at about $35 each, the CPU is probably $10 and the ADC is probably the same... the rest is R&D, other components, assembly, test, support, accessories, and margin...

To be honest, the argument of recovering dev costs back doesn't mean much here. Sure there are slightly more costs in designing a higher bandwidth scope but nowadays it's things like selecting a higher bandwidth buffer amp/VGA, changing a few passives and making the layout cleaner. Maybe with older analog scopes 200 MHz was very difficult to get but nowadays it is done with so few components it's pretty impressively simple.

Agilent and LeCroy better up their game by improving their lower end offerings... it's no longer funny to just slap your name on a Rigol or Atten and double the price. Tek seems to be struggling in the higher end department too. Got to play around with one of their newer scopes at work (I forget the model but I think it is related to the DPO3034.) Compared to the 3000X we also had on evaluation it was slooowwww (menus took ages to pull up, options hidden behind 20 sub menus etc.) the 3000X was nice and responsive, intuitive to use, and a bit cheaper too.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #32 on: September 27, 2012, 01:23:15 am »
I am just waiting a couple more decades (or maybe just a few years) until more Chinese vendors will pop up and sell a 200 MHz scope which is plain 200 MHz but at a 70 MHz price point. Because clearly Rigol can make one at a 70 MHz price point, they just don't want to sell it at that because they wouldn't profit as much. I'm pretty sure they make a decent margin on the 70 MHz version. The most expensive things on that scope are probably the Spartan-6 devices at about $35 each, the CPU is probably $10 and the ADC is probably the same... the rest is R&D, other components, assembly, test, support, accessories, and margin...

To be honest, the argument of recovering dev costs back doesn't mean much here. Sure there are slightly more costs in designing a higher bandwidth scope but nowadays it's things like selecting a higher bandwidth buffer amp/VGA, changing a few passives and making the layout cleaner. Maybe with older analog scopes 200 MHz was very difficult to get but nowadays it is done with so few components it's pretty impressively simple.

It's an age old argument, but is costs money to run big corporations that do leading edge R&D like this that deliver us our new scopes, and continue to develop and support them.
That system falls apart if it becomes the "race to the bottom" if they all started cutting each others throats and offering the 200MHz scope for the 70MHz price point etc.
And if it was possible, some OneHungLow design house would have done it and would kill them all. And we can kinda see some attempts at it already in the low end of the market. But it basically doesn't happen because very likely it not as easy or cheap to build these things as you might think.

If you want companies to grow and innovate, they simply have to make bigger margins on their higher end products, that's how it always works. Take that away, and the big players collapse into a heap and we are all left with buying single batch run scopes on AliBaba from OneHungLow at the lowest price point, the lowest quality, and the lowest service and support. Innovation, support, stability, and service goes out the window.

Dave.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #33 on: September 27, 2012, 03:10:02 am »
Ah, the 128MB of memory mystery is solved I think.
After a quick play with became obvious why it needs such large memory and how it works.
The waveform "capture and replay" feature (which is neat) actually captures the entire sample memory (up to the 14M record length) and stores it for each sample.
So if the timebase is set to give 14M samples (>=500us), you only get 7 waveform captures. Half those samples for dual channel, but same number of captures.
14M x 7 = 98M which fits in to the 128MB of physical memory. There must be some overhead storage requirement that takes up the rest.
Likewise if the timebase is set to 1.4K samples (50ns) you get 65000 waveform captures. 1.4K x 65000 = 91M
And so on for all the timebase/memory depth settings.

So it's not one 64MB chip per channel, it's shared, and sample memory half is two channels are on.

Dave.
 

Offline northlondonsage

  • Contributor
  • Posts: 11
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #34 on: September 27, 2012, 07:29:08 am »
Great work, Dave.

Just thought I'd let you know you are appreciated. There's so much effort goes into these and you are getting better and better at them.

Michael
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #35 on: September 27, 2012, 04:08:30 pm »
It's an age old argument, but is costs money to run big corporations that do leading edge R&D like this that deliver us our new scopes, and continue to develop and support them.
That system falls apart if it becomes the "race to the bottom" if they all started cutting each others throats and offering the 200MHz scope for the 70MHz price point etc.
And if it was possible, some OneHungLow design house would have done it and would kill them all. And we can kinda see some attempts at it already in the low end of the market. But it basically doesn't happen because very likely it not as easy or cheap to build these things as you might think.

The point I'm making is that the R&D is less significant nowadays for a high MHz scope. That's why most basic digital offerings start at 50-100 MHz.  Once you've designed a scope frontend, you can re-use parts of that or the whole of that design. I don't run these companies, so I don't know how significant the costs are. I imagine Rigol will sell a lot of these scopes especially to the education market. I think Agilent and Tek will win larger contracts for companies still - but I wonder how long that will last. I also imagine it's cheaper to hire engineers in China - though I wonder how much cheaper? The engineering on the scopes and other equipment is obviously very good, so there must be first class engineers working at Rigol.

It will be interesting to watch China over the next decade. Patents and copyrights mean very little in China. It results in a lot of copying between companies. But maybe it will end up better for the small fry and the consumer? In the US, many small start-ups have been quashed by patent law. It's also essentially pointless to attempt to protect a patent - the costs are enormous. Patents are held on the most obvious and pointless things - such as the page up/page down patent by Microsoft. Maybe it will turn out that it is better to simply have MORE copying as it leads to more innovation and cheaper products for the customer...
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #36 on: September 27, 2012, 04:54:17 pm »
Of course, in your mindset 2GS ADCs dont cost anything more than a 1GS ADC? You should know how much it takes to develop a 2GS ADC and a decent high refresh rate display plus the front end.
No no, that doesn't cost anything. Would it? NO!
They used good quality parts here, if you would accept anything from atten i'd say go ahead and waste your money on a crap 200MHz scope if they had one
 

Offline Eliminateur

  • Regular Contributor
  • *
  • Posts: 179
  • Country: ar
  • Electronic's Technician
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #37 on: September 27, 2012, 05:36:09 pm »
i don't see this new Rigol DS2000 as "young player" or "hobbyist" viable, the price is simply too expensive as base alone, with twice the price of the DS1502 for me it's a no-go, with steep import duties, freight charges and middleman's take a DS2000 barebone 70MHz would end up in excess of 1600 USD

 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #38 on: September 27, 2012, 06:00:24 pm »
Of course, in your mindset 2GS ADCs dont cost anything more than a 1GS ADC? You should know how much it takes to develop a 2GS ADC and a decent high refresh rate display plus the front end.
No no, that doesn't cost anything. Would it? NO!
They used good quality parts here, if you would accept anything from atten i'd say go ahead and waste your money on a crap 200MHz scope if they had one

Rigol didn't design the ADC, I'm pretty certain it's just a relabelled Analog Devices part (or maybe NatSemi/TI)... they use the same "RAD" labelling as they do on the the AD9288's in the new DS1052E etc. That ADC probably costs a fair bit... more than the $10 I estimated originally... probably as much as the FPGAs ($35 each.) Still looks like there's only about $150-250 in components there, the rest goes to other stuff.

I would certainly not buy something from Atten I'm just saying that it will not be long before you start seeing budget 200 MHz scopes because the extra R&D cost (which nowadays is only a little extra work) is easily outweighed by the extra marketability of the scope. I have a DS1102E myself, it's a fine scope and fairs well compared to its competitors.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #39 on: September 27, 2012, 06:15:36 pm »
Dude seriously, the AD9286 costs 75 in 250 quantities what about a 2GS ADC?
TI ADC08D1020 477$ in 60 off quantities
What about the higher speed FPGA now? Blackfin DSP? Good LCD? Good PSU? PCB price? Encoders? Design time? Salaries? Shipping? RF Metal shielding? The rising price of DDRII?
No one's paying for the R&D costs. It already looks like rock bottom margins
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #40 on: September 27, 2012, 06:25:49 pm »
Great tear down.  Compared to another similar one posted there is no comparison, Dave's is a league apart.   Side by side when comparing the Rigol 1052e to its clones, e.g. Atten, clearly there is a edge in Rigol quality and again it shows in this scope.

BTW; in regards to the icon on secretary bird, its the classic pose:

 

It seems so long ago, back in 2009, hi res photos:

http://www.eevblog.com/2009/10/12/eevblog-37-rigol-ds1052e-oscilloscope-teardown/

and hi res DS2000 photos:

http://www.flickr.com/photos/eevblog/sets/72157631618295437/

Looks like the heat sink issue is again QC, another reason to favor someone local to you to be able to return it for exchange.  Although you could fix it DIY there is no need too given the warranty. 

Now, the next thing to see, how well its functions.

Best Wishes,

 Saturation
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #41 on: September 27, 2012, 06:41:15 pm »
Dude seriously, the AD9286 costs 75 in 250 quantities what about a 2GS ADC?

It's not a 2GS ADC. It's a dual 1GS ADC, and it does 2GS for one channel, 1GS for two channels. That's what I thought at least. Still a pricey part! No kidding.

Rigol have shown it is possible to make a 1GS ADC for a scope using 5 dual ADC chips with a 1ku cost of $3.44 each. I would not be surprised if they are using a much cheaper part than the 9286. Noise, INL, DNL and all those other expensive specifications are less important in a scope.

I thought the argument/discussion/commentary was over bandwidth anyway... you don't unlock sample rate on this thing...
« Last Edit: September 27, 2012, 06:49:42 pm by tom66 »
 

Offline RFguy

  • Newbie
  • Posts: 5
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #42 on: October 01, 2012, 01:10:16 am »
So Dave, when will you make us the the honor to watch a review of this babe ?
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #43 on: October 01, 2012, 02:21:50 am »
It's not a 2GS ADC. It's a dual 1GS ADC, and it does 2GS for one channel, 1GS for two channels. That's what I thought at least. Still a pricey part! No kidding.

Yes, it drops to 1GS/s with both channels, so the ADC's would be 1GS/s at best.
Wouldn't surprise me it is was a 4 channel 500MS/s and they simply interleaved in the FPGA.
Actually they have to do some interleaving in the FPGA, as the ADC is clearly 1GS/s tops.

Dave.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #44 on: October 01, 2012, 11:35:50 am »
Dude seriously, the AD9286 costs 75 in 250 quantities what about a 2GS ADC?

It's not a 2GS ADC. It's a dual 1GS ADC, and it does 2GS for one channel, 1GS for two channels. That's what I thought at least. Still a pricey part! No kidding.

Rigol have shown it is possible to make a 1GS ADC for a scope using 5 dual ADC chips with a 1ku cost of $3.44 each. I would not be surprised if they are using a much cheaper part than the 9286. Noise, INL, DNL and all those other expensive specifications are less important in a scope.

I thought the argument/discussion/commentary was over bandwidth anyway... you don't unlock sample rate on this thing...
I know that, that's why i filtered for a "Dual 1GS" ADC
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #45 on: October 01, 2012, 12:04:07 pm »
The point I'm making is that the R&D is less significant nowadays for a high MHz scope. That's why most basic digital offerings start at 50-100 MHz. 

Once you've designed a scope frontend, you can re-use parts of that or the whole of that design. I don't run these companies, so I don't know how significant the costs are.

There is a lot more to a good scope than just the analog front end, which yes, it's not overly hard for a few hundred MHz if you know what you are doing. That is why there are lot of cheap 200MHz class scopes around.
Notice those two huge FPGA's in the new Rigol scopes? Care to guess how much engineering effort has gone into those?
And I can also say that the UI contains an impressive amount of work in it's own right, perhaps the most work of the entire design?
All this stuff costs a crap load of time and money to produce and innovate, and only the big companies can afford to do it. How? By making those big margins on the higher end units that people like to complain about.
The lesser companies often don't have those same expenses because they are much simpler and more primitive designs in terms of (fast update) systems engineering and UI firmware etc. They are not innovating in terms of R&D. That is why you don't see Tekway or Owon producing a 50K waveform update scope for example.

Quote
The engineering on the scopes and other equipment is obviously very good, so there must be first class engineers working at Rigol.

That's because Rigol have been working hard at it for about 12-13 years now.
Rumour has it that Rigol have hired a bunch of (disillusioned?) Tek guys.

Dave.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #46 on: October 01, 2012, 12:05:50 pm »
i don't see this new Rigol DS2000 as "young player" or "hobbyist" viable, the price is simply too expensive as base alone, with twice the price of the DS1502 for me it's a no-go, with steep import duties, freight charges and middleman's take a DS2000 barebone 70MHz would end up in excess of 1600 USD

Really?
Even in Australia I can buy the official 70MHz Rigol DS2000 for around $900 (<US$1000)

Dave.
 

Offline Tepe

  • Frequent Contributor
  • **
  • Posts: 572
  • Country: dk
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #47 on: October 01, 2012, 01:53:13 pm »
Really?
Even in Australia I can buy the official 70MHz Rigol DS2000 for around $900 (<US$1000)

Dave.
Batronix in Germany has the DS2072 for €710 with free shipping within the EU. That's around 880 Dave dollars.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6678
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #48 on: October 01, 2012, 03:16:44 pm »
It's not a 2GS ADC. It's a dual 1GS ADC, and it does 2GS for one channel, 1GS for two channels. That's what I thought at least. Still a pricey part! No kidding.

Yes, it drops to 1GS/s with both channels, so the ADC's would be 1GS/s at best.
Wouldn't surprise me it is was a 4 channel 500MS/s and they simply interleaved in the FPGA.
Actually they have to do some interleaving in the FPGA, as the ADC is clearly 1GS/s tops.

Dave.

And it may end up cheaper or easier to use an FPGA capable of "only" 500 MHz clock rates, instead of 1 GHz, speed grades and that...

I'm interested that they went for an FPGA for the display processor. It's an obvious choice of course for 50k waveforms per second. Except, does the LCD actually do that many (best case LCD refresh rate is 3-4ms right now), or is it a DPO effect? Also, the FPGA costs $35 in 100 units, I'll bet much less in 1k to 10k - I wonder if Agilent considered it but found ASIC design to be cheaper (thinking they'd get the chips for $5-10, but invest several million in developing it.)

If they hired former Tek engineers, won't they be expecting a similar salary? I was under the impression that salaries would be lower in China. After all, everything else is.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: EEVblog #360 - Rigol DS2000 Oscilloscope Teardown
« Reply #49 on: October 01, 2012, 03:36:55 pm »
Nah, the ASICs agilent makes are not cheap.
At least that's the path agilent took since the first megazoom ASIC, and the latest brethen goes from 100k (X2000) to 1M (for the X3000) so yeah
The RIGOLs are not really much cheaper, but by a few hundred bucks
I'm interested in the 4k series guts, it's apparently 110K instead of 50k of the 2k series might be a ASIC maybe?
I bet most of the cost went into designing it, as it looks like quite a lot of time spent
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf