No Script, No Fear, All Opinion
RSS icon Home icon
  • EEVblog #19 – Rigol caught with their pants down!

    Posted on July 19th, 2009 EEVblog 21 comments

    Rigol caught with their pants down! An issue with the Rigol DS1052E Oscilloscope.
    Or perhaps it’s Analog Devices caught out?
    (correction – there are only 2 PLL’s in the model FPGA used in the Rigol)

    Be Sociable, Share!
    • John Mcvirgo

      Do you have a a 1Ghz source to test it?

    • Katie Wasserman

      Very interesting! Do you think that the over clocking explains their need for the extra air flow and hence the loud fan noise? I understand that Rigol’s ‘B’ series scopes which are 2GS/s have a much quieter fan. These are the ones they make for Agilent — Rigols’s DS1102B is the same as Agilent’s DSO1012A, for example.

      • http://www.alternatezone.com/eevblog admin

        That did cross my mind as a contributor to the decision to include a fan. But my unit barely gets warm on the fan output, so it’s hard to see why it’s actually needed in the first place. Overclocked parts almost certainly won’t use any more power than the proper 100MHz versions anyway.

        The Agilent 3000 series are also old rebadged Rigol 5000 series scopes.

    • Y.Man

      Good Job! ^_^

      Waiting for Rigol GM reply.

    • Tracy

      I’m not sure if I think this is a shady trick on their part or darn clever engineering. If the scope works as advertised, and is as sturdy and well built as it seems from your review, I have to give them credit for being able to make a good product from less expensive hardware. I’m waiting to see what Rigol’s response is to this, I’m still keeping them on my list for a new scope. If I were them I wouldn’t worry so much about their customers finding this out as much as the competition. “Tektronix: We cost more, but we use REAL 100 Mhz chips!”:-)

    • Tomi

      Hello
      I found some more inside pictures
      original
      translated

      • http://www.alternatezone.com/eevblog admin

        Awesome!, thanks for the link.
        I wonder if it’s possible the owner has any more photos?

    • Luke

      We do this all the time at my job with the large chip manufactures. We just call up analog devices (or whatever company we are dealing with) app engineers, sales rep, or sometimes talk directly to the designer of the chip. We ask if the same die is used in all the chip versions, what QC is involved with each chip, the calibration process, if different die jumpers are used on each chip, etc. We then get a written statement from the manufacture saying the lower grade chip is the same die, QC, cal procedure, etc of the higher grade part to cover ourselves if anything should come up.

      We buy in very high volume and have a mutual NDA in place with most large chip manufactures so we have access to all this information. Which I am guessing the case with your scope.

      A lot of parts use the same die and QC process so we always buy the lowest speed grade part on when that is the case. Speed grades are commonly implemented when the part is first released and they have not fined tuned the silicon or cal process so they end up with some parts that do not meet the high grade specs. Instead of trashing these parts they just rate then at a lower grade until they fine tune there process. After a few years most parts only have one die and the tolerances are tuned in so tight there is no difference between parts.

      To summarize, this is no big deal and done all the time in large, high volume hardware design companies.

      Hope this makes sense,
      Luke

      • http://www.alternatezone.com/eevblog admin

        Thanks for the heads-up Luke.
        I’m sure Rigol haven’t used the lower spec part indiscriminately. If they have done it right then I don’t understand why they don’t respond and simply say so, and clear up the issue with users. Perhaps it is because of an NDA in place…

    • Luke

      I forgot to also mention the price listed on the analog devices web page is not the cost high volume companies buy parts at. We negotiate with different companies before we choose a part in our designs and get different manufactures to fight for the lowest cost since most chips are pretty similar performance. Then we design around the lowest cost chips selected that meet our requirements.

    • Luke

      That’s correct, they are not allowed to officially tell you that they are the same part because that information was disclosed to them under the NDA with AD. This is extremely common practice, most engineers who work for companies with large volumes will tell you about it off the records but they are not allowed to disclose it in say an email from their work email address, etc.

      It is also important noting that this does not apply to all parts. Many parts have different grades that are sorted out during the factory QA and calibration process. Cal involves laser trimming resistors, inductors, and caps on the metal layer. There is a small specified trim range. If the cal process runs out of trim range the part will just get marked as a lower grade instead of scraping it. Also some parts have jumpers set on the metal layer that are used to create many different parts from one die. In some cases the higher speed grade parts are the same dies with different jumper configurations to and are less input noise tolerant, etc.

      Many parts go through redesign over time to increase yields, this is normally when the lower grade becomes irrelevant if there is no jumper difference between grade. Also parts are often labeled at a much later time then production since parts are made in high volume runs and stored. If i remember correctly they can store the silicon for a year without packaging it into a standard IC footprint and then store the packaged part for two years.

      – Luke

    • David DLC

      What do you think about low cost USB oscilloscopes ? What will be better a Rigol or $200 dlls USB oscilloscope ?

      Thanks.

      David

      • http://www.alternatezone.com/eevblog admin

        I think they are toys, and generally poor value for money.
        Watch blog #13 Part 2 of 2

    • David DLC

      Do you think with this issue, is still worth to get one ?

      I found them cheaper on E-Bay (new) but the vendor is in China and I’m affraid of this.

      Were did you buy yours ?

      • KTP

        I took the risk and bought one off ebay from China…at least it was from a reputable almost 100% positive seller who has sold over 39 of them already on ebay. Price was under $400 US including shipping! I will respond to this post when/if it arrives safely.

        I have a TDS 210 and have always hated the 2.5k memory and lack of PC transfer (I never bought the upgrade modules). I decided to take the gamble and see if this Rigol can outperform the monochrome TDS210.

        • KTP

          My ebay Rigol 50mhz scope arrived today from China. It took about 8 days from clicking the buy it now to getting it on my doorstep in great condition.

          A few quirky things, like the power cord is definately not a USA type and the unit comes with everything in Chinese, but a few button presses and presto English!

          I can’t believe this scope was $340.

    • David DLC

      Sorry: Where did you buy yours ?

    • JoeF

      Very good discussion! After reading all this, it occurs to me that parts that have been speed binned to a given Frequency (eg 40MHz) typically run well at that speed with a degree of margin (usually chosen to be fairly conservative). Two things can be done to improve the ability to run a higher speeds. One is to raise the voltage of the logic and the other is to lower the temperature of the die. If this scope was designed to use 40MHz parts to run at 2.5x there bin speed, then I would expect more aggressive provisions to prevent failure. A bigger fan with more air flow is one. Heatsinks would be an easy, cheap one. Putting parts on the bottom would not be something good since the cooling there is limited (but they are there, most likely, to get around routing and trace length limitations). Having said all this, I occurs to me that simple test to determine if these parts are indeed running with little margin would be to run the scope at elevated temperature. If these ADC’s are riding the hairy edge, they should fold over after some thermal testing….? I don’t have one of these scopes, but has anyone tried this?

      Joe

    • David DLC

      A question came to my mind, if you have this unit:

      Can you replace the 40 MHz A/Ds with the 100 MHz version ?

      What would be the implications other than recalibrate the unit ?

      I hope somebody can help me to answer this question.

    • Toli

      Hi Guys,

      1. I can’t find the 1052E at the prices you’ve stated. Lowest price on eBay is about 550$ now. I see that dealextreme.com has them for 404USD shipped, but I’m now sure how good the warranty is with them.
      Any thoughts as to where I should buy it?

      2. What would you say is better, the 1052E or the tek 2235 (used obviously) for about the same price (400USD locally).

      3. About the frequency issue, an easy test would be to hook it up to a waveform generator and test the maximum frequency it goes up to, no? Can someone who has the required equipment do this?

    • http://www.lynx-india.com Amarbir[Lynx-India]

      Hi ,
      I wish someone could measure the voltage going to those chips .I am damn sure it has to be higher then rated voltage as overclocking is impossibly impossible .Its just like overclocking computer systems ? .