Double The Hertz, Double The Pleasure

[tinhead] has opened up a Tekway DST1102B oscilloscope and doubled its bandwidth to 200MHz, sharing his work in the eevblog forum. This is great news to anyone who is looking for a faster sampling rate but can’t afford the high-end models. Mind you, for a lot of us even these Hanteks and Tekways are hard to afford but there are more appropriate options for the ramen-dependent hacker.

In the hacking guide [tinhead] includes comprehensive information on the different scopes he originally considered (a Rigol, Atten and UniTrend) before settling on the Tekway, as well as links to regional distributors for the hackable scope. Good quality benchtop units are invaluable for development and troubleshooting, and it pays off to understand their inner workings. It’s heartwarming to know that even the tools of hacking can be hacked.

41 thoughts on “Double The Hertz, Double The Pleasure

  1. actually it is ARM9 linux 2.6.13 based box,
    so many things are possible … you can
    even still use it as DSO.

    Original idea behind this hack was to get
    some bw for free, but you can do other
    things too.
    There is 16bit bus and user i/o available,
    or even touch screen connector.
    I don’t think that someone will use cam
    interface connected to DSO, but an
    ethernet addon or logic analyzer addon
    could make sense.

  2. I’m still having a hard time believing a software change will turn an oscilloscope with guaranteed specs at 100 200MHz into a scope that with will perform as well as the oscilloscope guaranteed specifications of the 200MHZ. In the case of the Teckway there are differences between the scopes other than bandwidth. Those differences may not matter in some instances,but there will be instance where it may matter. No doubt the pros are aware of it, but it’s going to be some of the electronics hobbyists and hackers that may not be, and will nor understand why their oscilloscope is leasing them in the wrong erection, assuming they know/recognize enough enough everything is not adding up.

  3. @D_

    There are cases where it would work, you just don’t get any certainties.

    Say a company is producing a few similar models. It’s sometimes cheaper to only produce the higher bandwidth version and get the others either in software or maybe by leaving off a daughter board that adds some function. Then when you’re doing your QC run, test at the highest bin first until you get enough that pass spec to meet sales. After that, you take the ones that don’t pass highest-bin spec plus any leftovers that passed or weren’t tested, and test them at the next spec down. And so forth.

    Happens with CPUs and GPUs quite frequently due to things like extra cores and pipelines. If the process only gives you good results at some statistical margin, you can always disable the bad ones and call it a lower spec part. The result still meets the spec it’s sold at, but it’s often way more cost effective than engineering and producing N different models from scratch.

  4. The was a lot of reports on EE forums from people feeling greedy and getting Tekway instead of more trusted brand, a lot of Tekway scopes lose a channel or two after couple weeks of usage

  5. I just got a Rigol DS1052E today, though I haven’t used it yet. It’s a 50MHz scope that can be hacked to 100MHz in software, but there seems to be issues with the new firmware blocking this. It cost $400 (actually $380 from eBay) and is really a nice looking scope for the price (50/100MHz, 1GSamples/sec, dual trace, color LCD, USB storage, etc). I haven’t had a chance to test it but I bought it based on the multitudes of great reviews on eevblog and other forums.

  6. D_

    I designed a product (RFID access control system) at a previous employer which the board of directors / marketing team told me to cripple in firmware . This was so they could sell two variants; a “budget” version and a “full” version. The only difference between the two products was a few blocks of extra C code.

    So I guess this kind of thing happens more often that you would think!

    Cheers,

  7. it’s called price discrimination. if you create several similar products at different prices, every customer pays the maximum price he’s willing to pay. otherwise, you’d have some people paying less than they could and some others not buying at all (much less profit). Of course this strategy requires retarded buyers, not a problem at all on planet earth, we have plenty.

  8. Quote from D_
    “…and will nor understand why their oscilloscope is leasing them in the wrong erection…”

    Got something else on your mind? :p

    As other have said, its extremely common to see company design a single product and then cripple it in software to make several different products. In a lot of cases it would be much more expensive to actually design and build separate devices with different specs than to just make the differences in software.

  9. > In a lot of cases it would be much more expensive to
    > actually design and build separate devices with different
    > specs than to just make the differences in software.

    Even Ford do this with their engines – same engine, different fuel maps means different power outputs and different prices, especially in their Transit van range.

  10. @therian

    you don’t think i will send you money only because someone deep in russia got issues with “flying parts” :P

    But to be very honest, such kind of issues are typical for early devices, in todays world many companys are “testing” devices at user costs.

    Tekway is producing since 11 months these scopes, and Hantek “rebranded” since two. However it seems they already learned a lot (now there are small heatsinks on ADCs – i bet this was the issue in russia, and maybe shippment? who knows), firmware is light years better than 6 months ago, PC software can be installed without these .ini issues. As i got my device, the first thing was to check power supply and air flow inside, and of course i found things which i didn’t liked.

    Tekway’s answer was:

    “The components we used in this DSO is industrial level, so these conponents can be worked well under the temperature you tested. For example the AD9288-100, work well under 75°. Of course we have reserved the interface for fan supply, but not used in this series. We have considerd about the situation when the DSO running without the fan. It could be worked well. But we will add the fan in the future if our clients need”

    Well, as EE i don’t trust other engineers, so i installed small fan and never got any heating issues (for me Tekway’s answer was like “yes, maybe, but we have to produce cheaper”). I’m glad to see i was right.

    I don’t want to make these scopes better than they are, but they not bad (just read eevblog review for details) – and for “money savers” they hackable.

    P.S. – i said “afaik”, but you know what, i will ask Tekway to send me the 100 bucks, hell i told them months ago that this is potential issue, and now i got busted :)

  11. Where this guy located is irrelevant, he ordered scope from China anyways and it was couple month ago so the model is up to date. I didn’t Google to find someone with broken scope, I was participating on that forum and his experience stick in my memory.

    “you don’t think i will send you money only because someone deep in russia got issues with “flying parts” :P”
    Russians sometimes say it fly instead of it breaks(think of it as soul of dead device fly out)

    Promising something you don’t intend on delivering makes you unworthy

  12. yeah ,i saw your nick on the forum. The translation is a bit crippled and my russian is not good enough anylonger, from what i see he ordered in germany? (pinsonne)
    Whatever, now we can only do some speculations, we both don’t know the serial numbers so hard to say what he got.

    I wish i could follow up on this (I’m in contact with them), but without serial number or at least information where he bought it will be not possible. Don’t get me wrong, i don’t care about this particular single-user issue, it is just because i have a small fight with Hantek/Tekway right now (about GPL Copyright Infringement).

    In case you can get something from “q123” send me a PM on eevblo forum.

  13. looks like “q123” with total of 10 messages only joined forum for advice on purchase and haven’t returned back since. And his photos are too low resolution to read serial from it, but one showing the board looks like have written venison 1.2 or 1.5 perhaps 1.7 on it.
    I dont know what cause his first unit to fail self calibration but second unit losing a channel most likely due to relay failing to stay fully closed so only small capacitive patch left for signal to go.
    Curious what Copyright Infringement you talking about, is it about Hantek/Tekway copping Rigols GUI?

  14. @therian

    regards “q123” device: on the pics he posted there is 1.00.2 version number on power supply – exact the same as on my very early device (SN xxx000911). So let summarize: we have single issue (and not multiple like you said i your first post), unknown user – we don’t know what the user did within “first week of using”, we know he is using breadboards so probably beginner, we don’t know what he measured (broken cable maybe?), etc … so finally didn’t really matter, he send back the DSO and will probably never answer these questions.

    Rigol GUI ??, you cant compare apples and oranges, Rigol have small display with low resolution and single window UI – Hantek/Tekway wide-screen display, with large resolution, two window UI. Just read eevblog, there are enough pictures.

    I said “GPL Copyright”, it is about embedded linux on the DSO, once again just read eevblog.

  15. How is that single issue if a single person went thought 2 broken units ? It was yours idea and you set simple rule : “afaik there is no single incident with Tekway DST1xxxB series DSOs. In case i’m wrong, you will get 100 bucks from me, just give me a link. fair enough?”
    I provided you with 2 incidents, fair enough ?

  16. @therian
    you are real moron… just read what you posted.

    ———————————————-
    The was a lot of reports from people feeling greedy … a lot of Tekway scopes lose a channel or two after couple weeks of usage

    Posted at 6:16 pm on Nov 24th, 2010 by therian
    ———————————————-

    if you don’t know what “a lot of” means, well google for it. Single man, deep in russia got DSO which at deliver time was ok – week later broken. The guy is a beginner and we have no informations what he did. You know him ? I don’t, and probably if i would be an beginner who killed a DSO, i would probably tell the same “i did nothing, it just happens, this my second unit, all bad, i’m innocent, i did disassembled the device but never played with it”.

    Btw, even if he would be EE professional you can always get a bad device, lol, even two. Look here, this is definitely better brand than Rigol:

    http://www.mikrocontroller.net/topic/198807

    For those who don’t speak german – this guy bought Lecroy 44Xi, after 3 days broken, repaired, then again broken and so on …

    So what now, let’s buy Rigol from “therian” because Lecroy is bad brand ?

  17. you called me a moron but you yourself assumed there is perfect manufacturing process exists and set yourself a bet. All you asked is a link to broken device, you did not specify where person should live what version of device it was and what experience of user is, so all BS you saying now is irrelevant to original bet:
    “there is no single incident with Tekway DST1xxxB series DSOs. In case i’m wrong, you will get 100 bucks from me, just give me a link. fair enough?”
    You own words…
    And I see no point of discussing which brand is good or bad, why scopes fail, bla bla bla since it all irreverent to bet

  18. By the way due to customs they pay almost twice in Russian for same electronics we have in US, and average income is much lower too so I will highly doubt that beginner would buy new digital scope instead of cheap second hand soviet analog oscilloscope.

  19. The thing that would concern me is that the modded ‘scope might not be as accurate as it was. Since you use a ‘scope to debug and find problems the last thing you need is errors creeping in due to the ‘scope itself.

    I have a 100MHz GW Instek ‘scope and am very happy with it.

  20. @MoJo
    Accuracy depends on clock domain and ADCs. All my test cacdidates have AD9288, except UNI-T with almost “no-name” ADCs. All are using low cost XO and ADC clock over FPGA PLL – same for GW Instek.
    There is GW Instek review, you can compare it

    http://welecw2000a.sourceforge.net/docs/Hardware/GW_Instek_GDS-1152A.pdf

    What i not like on GW-Instek is analog signal routing (over board-to-board connectors) but it this class of DSOs didn’t really matter that much.

  21. @tinhead
    The Russian report of over heating, and causing channel 1 failure was strange as reported. Showing the two pics of of 2 channel square waves. The second pic showed channel 1 distortion, caused by overheating? If you look at the timestamps on the pics, the distortion is 21:57, and it clears up at 21:59, how is overheating the cause? And the difference is 2 minutes not 10 minutes. Doesn’t make sense. Looks like a probe problem to me.

  22. well, actually it can be everything, from “monday” device , over QC issues up to user problem.

    For those who not following the thread on eevblog:

    In the meantime the hack has been improved, actually we on 250Mhz -3db (with proper HF active probes – don’t expect too much from cheap ones).
    I found a way to manipulate the digital bw filter, allowing the user to setup the max bw, however max. usefull from ADC point of view is 450Mhz allowing measurments up to 400Mhz in single chan mode.

    I developed (and published proj. details) Ethernet adapter card, only 10mbit but that’s suficient for most things.

    Hantek/Tekway did changed hardware revision, adding some nice features , like SDCARD slot on board, phone/mic connectors and even spare place for I2S audio codec, extended i/o port giving chance for even more hardware hacks.
    It is now more Samsung S3C2440 dev board with DSO functionality :)

    Another user is developing custom SDK, currently he is able to access display/memory data, manipulate menu structure (to add own functions into the fw). He haven’t posted yet, but is close to do it.

    So yeah, this platform (DSO) makes still enough fun to play with it – and of course use as DSO.

  23. @tinhead
    How do we know before we buy the scope that we will get the latest hardware revision? Is there a question to ask the distributor so I can be sure? I know he won’t open the scope to check.

    I am very glad I didn’t jump onto the ADS, I was looking at the ATTEN specs and the LCD resolution that was holding me up so I was still researching and found the hack threads. Great Thank you.

  24. @Ot

    actually all official dealers in EMEA/NA will have the latest hardware revision, you can find the old HW revision only if you buy from chinese dealers.

    The dealer have to answer the question, in principle easy because system information is showing 0Xxxxx for old hw and 1005Xxxx for new hw.

  25. Picked up the 5102 from Morton which was on sale ths holiday. The high res display is very nice and the menu doesn’t eat up the screen because there is plenty of width. One thing that is irratating though is that the built in Help menu doesn’t have word wrap so its quite awkward to read since the line breaks occur anywhere within the word. This would have been a rather easy function to include in the firmware. Maybe Chinese characters are words so it doesn’t matter to them but Western languages need word wrap to display properly. The “a” charater isn’t kerned properly either and the index has a dozen typo errors. But overall I’m very happy with this scope. @Tinhead, is this part of the firmware accessable to fix these issues or do we have to wait for Hantek to jump on it some day?

  26. @tinhead,
    I haven’t opened up the scope but regarding the Help.db, you are suggesting that I insert the CRLFs where necessary based on how many characters fit on a line? My thoughts on the approach was to call a function in the LCD write routine and keep track of a character counter, looking at if the next word would exceed the max characters allow for the screen width.
    “,”
    “.”
    ” ”
    CRLF
    Is there another way you were thinking of?
    It would be a pain to manually count the characters in each line to see where a CRLF need to be inserted.

    I don’t have any experience in SQLite, but I looked at the manual, I didn’t see anything that would make things simpler to implement.

Leave a Reply to MoJoCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.