Hosted by David L. Jones
And presented in Dave's unique non-scripted overly enthusiastic style!
-
EEVblog #37 – Rigol DS1052E Oscilloscope Teardown
Posted on October 12th, 2009 67 commentsDave opens up his Rigol DS1052E Oscilloscope, and ponders how to convert it into a DS1102E.
And yes, he finally checks out the infamous ADC overclocking issue.Hires images: ADC & FPGA ADC Support Blackfin DSP Front End 1 Front End 2 Front End 3 PSU
Buy the Rigol DS1052E cheap:
67 responses to “EEVblog #37 – Rigol DS1052E Oscilloscope Teardown”
-
I think the 40 pins header are for the digital logic analyzer. They offer it in DS1052D and DS1102D. The D-series are 550USD more expensive than the E-series.
-
You mentioned a good question Ive had. I understand the outer shell is for shielding for RF Noise, What is the interior little CAN for? I found one inside of a satalite reciever I parted out. (It was housing tuning inductors) Is that for shielding of interior RF noise?
-
Correct, it’s shielding the front end from internally generated noise.
Remember that switch mode power supplies are horrible noise generators, and every digital signal is a potential noise source too. So lots of noise floating around inside that case inside the outer shield.
-
-
That’s it, now I’m ready to buy this scope myself
As for the cpld, my guess would be that it is used to program the fpga.
-
Michael Thompson October 13th, 2009 at 00:33
If I remember from the capacitor tutorial, isn’t that voltage regulator going to heat up the space around a couple of those nearby electrolytics as well as be vibration prone?
It’s wedged in there, even though there’s a little space front and back.
It seems like a strip of metal taking the heat up and away from that area and simultaneously securing that component might be a neat little modification that could kill two birds with one stone.
-
The regulator heat is only a problem if it actually heats up! The lack of a heatsink would indicate that it’s hardly dissipating any power, just being used as a low-ish current negative 5V supply, probably for the analog support section. There may be various reasons they used it over a surface mount regulator in this instance.
-
-
Mark-R October 13th, 2009 at 01:47
Looks like they’ve been grinding the numbers of the chips in the front end as well as the ADCs. Seems they are happy to let us see part numbers unless there’s something to hide (like clocking -40 parts at 100MHz!) I wonder if they’re bending the specs on those front end chips too. That might explain the 1052 and 1102 difference: perhaps the 1102 uses full spec parts and the 1052 uses parts out of the reject bin with the numbers ground off.
-
I’m very happy that I followed your advice earlier, and I’m having a great time with my DS1052E. Now I’m even more proud of it! I can’t wait to convert it to a 100MHz one, so do your magic and show us how to do it!!
My father tested it with military grade equipment, and he told me that the signal starts to attenuate at 49,8MHz for ch.1 and at 48,2MHz for ch.2, so this implies that the filter is built into the analog section and not in the DSP. A tolerance this big is a bit strange for the quality of the components that you mention, I would have expected more consistent results between channels if they use capacitors from the same batch… By the way, he was also able to measure a sine wave up to 300MHz!Something to upgrade and dump firmware would be very useful, so we can add Logic Analyzer function (I hope that these 40 pins are just routed to the front panel, and the box they sell is simply for protection of the inputs).
-
That sort of tolerance would be expected between channel bandwidths. The caps are going to be 5% at best for starters.
-
I know, but the 5% should be the maximum tolerance respect to the declared value… I meant that usually (from experience) the tolerance is quite lower from the average value between parts from the same batch, mounted on the same PCB at the same time and tested at the same temperature! Oh, well, who cares?
-
-
-
Mh.. The scratched 16 pin IC in the analog part should be an AD8370
http://www.analog.com/en/amplifiers-and-comparators/variable-gain-amplifiers/ad8370/products/product.html
but I don’t know why they are scratching it! There’s nothing to be ashamed of, it’s a 750MHz part…-
Yes, it is an AD8730, I was still able to read it.
I don’t know why you’d scratch it either, crazy! -
Ling SM October 16th, 2009 at 02:08
Rigol is one of the few companies that do their own R&D in China, and they also suffer from the cloning activities there. But looks like they can’t sand so well.
-
Yes, Rigol early on had problems with cloners (ironically for China) ripping off their products, so they have become quite protective. This explains why the GM was happy to talk to me about my blog, but gives stone silence when asked about the overclocking… it’s obviously one of their key low price enabler secrets. A standard AD chip is also rubbed off in the analog front end section, so they obviously have some deal with AD they don’t want people to know about…
-
-
-
I think the 16pin TSSOP in the input section is an AD8370 Variable Gain Amp.
The little chip next to it must be some kind of differential amp/buffer/ADC driver. The pinout matches pretty well with ADI chips like AD8139, but it’t not the right kind of package…
-
I’m thinking about buying this unit from some time. And now I’m almost decided.
It would be great if there be possibility to convert it to 100Mhz and add logic analyser.
-
Some quick thoughts about the input stage:
The chip marked “B 7A” down by the input is an AD8510 opamp, and it seems to be applying a DC offset right at the input (to DC-shift the signal?).
The SOT23 marked “J Y” is just an input protection diode.
“K51″ is the input transistor, which must be some fast FET but I can’t seem to find out which one. It is followed by “3EM”, which is a MMBTH10LT1 (bipolar transistor), after which the signal goes to the VGA which probably does all the gain setting (except for the two ranges that are selected by the relay at the input).
We could probably draw up the entire analog frontend schematic pretty easily, but someone may have to be brave enough the desolder som components to figure out all the cap values.
Personally I’d like not only to mod my scope to 100MHz bandwidth but also to get some nice open-source firmware running on it…
-
Open-source firmware is a nice idea but this is not very popular product, I mean compare for example to mp3 players which has a lot of users and they are making for example rockbox firmware.
David. I think that would be nice to make a small forum on this site where we can discuss things related to your video topics.
Then we can start topic about Rigol 1052E and maybe then more peoples will be interested to enhance possibilities of this nice piece of hardware.
-
I agree, Dave should host a forum! I’ts becoming hard to follow all the comments…
-
Dave,
If you would like to host a forum on your server, I can help set it up, if you don’t mind it running on an external server, I could host it here
Regards,
Kat.
-
Hmmm I have to add support for the www in my url. looks like time to do some aliases lol.
Regards,
Kat.
-
Thank for the offer Kathy. The forum is all sorted now and up and running. I’ve got almost unlimited server space and storage
-
-
-
-
Andrew October 14th, 2009 at 04:35
The Wittig / Welec DSO firmware can be had as open source software. Wittig was a German DSO manufacturer that went bankrupt. Remaining stock of Wittig DSOs is sometimes sold on eBay as no-name DSOs, example: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=200393279525
Be warned, this is a DSO for those where the journey (playing with some DSO firmware) is the reward.
-
-
I hope some of the engineers who worked on the this Oscilloscope see your blog. They could teach us all a thing or two.
-
Probably a question for RIGOL, but … I have been thinking about getting one as well. All the pictures show it shipping with a Eurostyle voltage plug. And assume this was designed to run on 220v. If it is a switchmode supply Can I just run it off my USA 110V, If I supply the appropriate plug?
-
David DLC October 13th, 2009 at 08:46
Daniel, I bought one unit and it has a non-USA power cord, but you can use pretty much a standard one, it will work with 110 ~120 VAC
-
I’m not too cluey about electronics yet, but wouldn’t it be easier for the company to just have 2 different firmwares then physically swap components? that might be worth a check as well.
-
You are right, it’s usually easier and cheaper to do that, and I have considered that’s a possibility.
I (and someone else on this forum) have actually measured the front end, and it does show attenuation on the analog channel before the ADC. Although I have not exhastively tested this yet.
So if it is actually being bandwidth limited by some input circuitry, it’s not out the question that it’s controlled via the firmware (e.g. a spare mux switching in an extra cap or something). Perhaps there is a byte in an EEPROM somewhere that tells the firmware the model number and limits accordinly.
That would be real cheeky.
But given Rigol’s paranoia about the hardware, I suspect they might be paranoid about fimrware hacks too. So from that point of view it makes me more inclined to think it’s an actual physical component value difference in the BOM when they assemble the boards.
-
-
Andrew October 14th, 2009 at 05:26
Regarding Rigol DSOs, rumor is Agilent had a helping hand in designing the Rigol DSOs. Agilent DS3000 series are rebranded Rigols.
Regarding Rigol as such, they unfortunately managed to get on my “don’t want” list. End of 2008 they released their new arbitrary function generator series DG10x2. For creating and downloading waveforms they deliver a software called UltraWave. They didn’t mention that two years after the release of Windows Vista UltraWave didn’t run on Vista.
Another year later UltraWave still doesn’t run on Vista. And I didn’t manage to get a clear statement when they will have a Vista version.
Their English web site is pretty useless, not even mentioning the DG10x2 series, only the older DG10x1 series. Their Chinese web site, well, is in Chinese. E-Mail support is not worth to mention.
I just hope people have more luck with Rigol DSO PC software and Rigol DSO support.
-
Ling SM October 16th, 2009 at 02:04
Have you contacted them recently? Jane was helping me, and I am very pleased with her service and her English skill. They would want you to contact the shop/person that you bought from,
-
Andrew October 16th, 2009 at 07:04
Yes, I did. First problem, find the right e-mail address. Note: It is not support@rigol.com.
Second problem, after finding and using the right e-mail address service@… they responded with some questionnaire, so they “can help me to chose the proper instrument”. WTF? What about reading help requests, instead of sending out questionnaires?
After insisting on getting an answer, and not a questionaire, I got the reply that maybe sometime in the future they might have Vista software for it.
Hello? Three years after Vista entered the market, one year after they started selling the generator they don’t have Vista software, and can’t make their mind up when they will have anything? For a US$700 instrument. An instrument where you need the software so you can reasonably load arbitrary waveforms into the instrument.
I can get throw-away instruments from other manufacturers for 1/5th of that price with dodgy Vista software, that will break with the next Microsoft service pack or security fix. But at least these other Manufacturers have something, unlike Rigol. And that’s why Rigol ended up on my “don’t want” list.
-
-
-
Hasn’t anyone noticed the horrible noise on these scopes ? at all V/div settings there is a 100-120 KHz signal that is very small on the screen but a real nuisance, it grows in voltage as the V/div is increased so stays the same size on the screen so its an internal thing that is bypassing the V/div circuitry, I beleive it to be from the SMPS which is not screened ! not very clever ! anyone else have this ? I will attempt to screen this and see if it makes any difference, I can eliminate it with the scopes own digital filters but that limits its use, the scope makes heavy default use of filtering to avoid this “noise” ever wonder why they make a hooha about filtering out AM/FM broadcasting ? yea really !!! I thing a screened scope with no probes attached will not be picking up A/FM broadcasts
Simon
-
It’s not horrible noise, it’s typical for this class of scope.
I did a quick comparison between the 1052E and the Tektronix TDS210 at 2mV, 200mV and 2V inputs with nothing connected and no digital filtering, all an arbitrary 10us timebase, and no bandwidth limit:
http://www.eevblog.com/images/RigolNoise1.JPG
http://www.eevblog.com/images/RigolNoise2.JPG
http://www.eevblog.com/images/RigolNoise3.JPG
http://www.eevblog.com/images/TDS210Noise1.JPG
http://www.eevblog.com/images/TDS210Noise2.JPG
http://www.eevblog.com/images/TDS210Noise3.JPGAs you can see, the Rigol performs similar, if not better than the Tektronix, esp on the lowest 2mV scale.
Adding a 50ohm terminator makes little practical difference.Dave.
-
-
Ian Banks October 16th, 2009 at 01:24
Anyone have any experience finding one in Australia? Like many suppliers here the ones I’ve found seem to be both mysterious and more expensive than a simple currency conversion would otherwise suggest.
-
Everyone I know in Oz just gets their Rigol’s direct from China or Hong Kong via eBay, that’s how I got mine. I think the current price is around the AU$580 mark inc delivery.
There is also an Ebay guy in Sydney who has local stock you csan inspect before you buy, but you pay a bit more.
Emona are one of the official reps in Oz, but the price is around $800! Standard rep markup for Oz unfortunately. It’s almost never just the conversion rate
-
It’s interesting to look at the cost of the DS1052E in China. If you search for it on http://www.taobao.com (the china ebay equivalent) they’re going for 1750 RMB ($256 US) I wonder if these are fake or whether the price is increasing that much when it gets sold outside of China.
-
-
-
Well .. here in UAE the official distributor of Rigol sells 1102E 240% of its original price (795$)! It seems I’m gonna get through ebay after all ..
-
MisS_E July 1st, 2012 at 00:13
I’m living in UAE and also looking for one! Did you manage to find one with reasonable price?
-
Recently I purchase Rigol Oscilliscope from Anaum International Electronics, they keep most of Rigol products in stock and prices are very reasonable.
-
-
I got dizzy from the shaky camerawork !
I love the subject matter of your podcasts, I keep coming back for the videos of the internals of devices, but the presentation is awful !
-
Kyriakos October 19th, 2009 at 16:09
Some more pictures of the DS1052E guts are here:
http://www.ourdev.cn/bbs/bbs_content.jsp?bbs_sn=3428440&bbs_page_no=1&bbs_id=3043
A nice review with screen captures is there:
http://www.science.smith.edu/departments/NeuroSci/courses/bio330/labs/L1cro.html
and an extensive (in depth) discussion on Rigols scopes (with problems and solutions, sceen captures, internal pictures and software pieces) is over there:
http://www.rcgroups.com/forums/showthread.php?t=663958&page=47
-
vespas October 22nd, 2009 at 22:25
Hi Dave,
speaking of cheap scopes, what do you think of the Owon ones?
http://www.owon.com.cn/eng/pdsSeries.asp
around US $350 in the UK for a 25MHz unit, $680 for the 100MHz (from a store, not ebay). Can also be fitted with battery for portable use.Also, what do you think is the bandwidth normally needed for hobbyists? i.e. what can and can’t you do with the different models? What common applications are there for e.g. the 60MHz that cannot be done with a 25MHz? It seems to me that either you need a cheap low bandwidth one for audio and similar analogue stuff, or a very high b/w one for RF. Am I missing something here?
vespas
-
Im pretty sure MachXO CPLD is used only for LA. Firmware is universal across all of the 1000 line so there must be a serial eprom with config stored somewhere (8 pin chip above blackfin? or the one next to it above ram/flash?)
-
I suspect the test pads visible immediately between it and the FPGA indicate it’s actually bringing up the FPGA too. I don’t see any special Altera flash chip, so a CPLD would certainly be necessary to serialise the FPGA’s bitstream.
That FPGA has enough PLLs on board to provide the 5 shifted clocks necessary for the ADCs and have some spare. I’m a bit suprised they picked an FPGA as large as that one. Maybe they squeezed some of the signal processing in there too to make ripping off the DSP firmware that much more difficult.
-
-
Troy Lavigne November 6th, 2009 at 07:22
Can anyone point me to decent driver install instructions for ultrascope? I’ve gone to Rigolusa.com and downloaded/installed the app, but haven’t had any luck resolving the USB driver issues. Thanks!
-
Hi Troy,
I’ve come to learn myself that the 1052E/1102E have had different Ultrascope software sets. One of which is the older 1000C/CD style software/driver, and the other utilizes the NI-VISA driver that it looks Rigol has been using in newer products. It probably depends on your firmware which one to use.
NI-VISA version:
http://www.rigol.com/upload/accessory/20098/20098191453126177769.rar
Non NI-VISA version (older style USB driver):
http://www.rigol.com/upload/accessory/20097/20097251310266496387.rar
Good luck!
Alan
-
-
signal7 November 10th, 2009 at 04:41
Where are you buying this for under $500USD? I’ve looked at various places online and I can’t find it for that price. $600 is more like it from what I’ve seen… Is it possible there’s a currency conversion error somewhere in here?
Also, I don’t think you’ll be able to mod this the way you would like without also rewriting the firmware. It’s fairly common to design a high end version of a product and then to disable various features in software to sell the product at various price points. The company making the product saves money on components via volume pricing (all of the products use the same parts), and the ultra low end products become loss leaders that entice new users into trying out the product.
I don’t doubt that there are probably a few component differences in the analog section. That certainly would be one way to help offset the difference in cost between models, especially if the 1052E is their loss leader product.
-
@signal7
Most people are buying their Rigol direct from China on Ebay (I did).
Current prices from most sellers are under US$400 including delivery!
Or if you don’t like ebay for some reason, here is an online dealer for US$412 including delivery:
http://www.dealextreme.com/details.dx/sku.30573-
ericwertz February 13th, 2010 at 10:48
This is no longer the case. In December, Rigol went to EBay and told them to pull all of the Chinese sellers’ listings that were sub-$500. You will now still find that this is the case. This was likely done because their official distributor channel complained (see Alan below) about non-o.d.’s in China selling for ~$385USD.
The other issue is that if you do get one from China (DealExtreme included) you get a China power cord and China-only warranty. So if you have a warranty claim, it has to be sent to Beijing. Rigol North America (for example) won’t honor your warranty, which makes sense. If you don’t buy it through their distributor channel, then you don’t get their service. I know this first-hand.
The power cord is a non-issue because it takes the standard PC power supply cord (absent any shielding sauce that could theoretically be present in the factory cord).
Still a good buy at either $385 or $515. I’m guessing/knowing that you can still get them at the former price by going around EBay and directly communicating with the sellers, but the aforementioned caveats would still apply.
If you want local support, you have to pay for it. I couldn’t, and didn’t. Your mileage may vary.
-
-
-
I think we have the best Rigol and Owon prices as official USA distributors for both of these manufacturers – and we ship worldwide!
-
Saelig Company Inc’s website didn’t show up on the last entry (best Wown/Rigol prices) – http://www.saelig com
-
Troy Lavigne November 14th, 2009 at 06:33
You might also want to ask for a quote from T-equipment.net for better than list pricing.
-
I never thought that a video about an oscilloscope could be so captivating (and my company builds them!). Love your blog and your style!
Suzanne-
@Suzanne
Nice to hear from someone from Tek, thanks for the support. Sadly though I haven’t reviewed any Tek gear yet…
-
-
Has anyone managed to get the API working?
Ultrascope works fine but the example code does not seem to use the right dev drivers.
Mike -
vasile March 9th, 2010 at 04:24
I don’t know why, but I have the feeling somehow Dave’s right hand is close related to this product. I’m thinking to the old MK3 design.
However, there are a few design problems which someone noticed: The switching power supply is opened, there is no shield. Imagine an electrolytic blowing inside, spreading on the CPLD, FPGA and what ever Blackfin is inside…
However $400 sound incredible. Still have limited trust in chinese products. Even my tektronis MSO4104 since is manufactured in China looks worse than it’s real predecesors manufactured in the US back in 1980.
-
I guess the Chinese Ebay sellers of this scope now simply add ‘best offer’ to their listings and accept anything over $400.
-
Apparently Rigol is putting the DS1052D on
sale for $399. (5/12/10) And maybe the
sale will go through December.I recommend Saelig.com. I ordered the scope
and they phoned me to tell me they were giving
me the lower price! Saelig is an authorized
distributor in the US.Dave, do keep telling us more about this scope.
-
MarkH May 17th, 2010 at 11:40
Also got mine from Saelig.com, one week before the $399 sale started. They called me and gave me the lower price as well. Plus they send it with a nice padded carrying case.
-
-
Howdy all!!!
Just got my DS1052 (now a 100 Mhz!!).
Was wondering if anyone knows the pinout of the power supply and the voltages? Would like to determine if building a battery supply for it would be feasable, or just use a UPS.
Any ideas??? -
Debraj July 23rd, 2010 at 14:06
Hi David,
I have a query related to most “low end” digital oscilloscope — Tek1204, Rigol-DS1052 etc.. The query is explained as below: -
1. Connect the oscilloscope channel to the 1KHz square wave provided on oscilloscope (for probe compensation). You can see the nice and clean square wave on scope screen.
2. Change the time base to 100mS and you can see a band (individual square wave merged into a band).. Not “STOP” the oscilloscope.
3. Now zoom the recorded waveform such that time base changes to 500uSec. And we can see a nice triangular wave. (Square –> Triangle) why??
I can feel the issues might be due to aliasing, but is that the right reason and is this problem common to all oscilloscopes?
-
That’s just the sample rate at play. You see a triangle wave because the sample rate was not fast enough.
Deep memory scopes will perform better in this respect than shallow memory ones like the Tek 1000 series.
Dave.
-
-
Hello, relatively new to EEVBlog, I came over after listening to The Amp Hour
Dave mentions prime-spec. components a bunch of times and surprise at the selling price.
I wonder why nobody has mentioned the possibility of the components inside the Rigol being counterfeit. The practice of ripping off a design, matching some nominal specs and faking a brand name onto something is extremely common in China.
I understand that some of the technology cannot be duplicated but I still think it is possible not everything is as it seems.
My company just this week caught a totally obvious fake bearing coming from a reseller in China. The first clue was how quickly the thing failed and the evidence came when we disassembled the bearing. I do understand that mechanical is different from Electronics but still…
Leave a reply
-
Follow the EEVblog
Use coupon code “EEVblogViewer” for 25% off!
Timothy Tee October 12th, 2009 at 11:44