Tablet Oscilloscope Claims 100 MHz, But Is It?

[LearnElectronics] grabbed a FNIRSI tablet oscilloscope from a vendor from China. The device has a seven-inch touchscreen and claims to be a two-channel 100 MHz scope. But is it? Watch the video below and you’ll see.

Spoiler alert: [LearnElectronics] was skeptical of the 100 MHz claim and it looks like it is more like a 30 MHz analog bandwidth. Despite that, it does seem like a pretty capable 30 MHz scope in a very handy form factor and a very cheap price: as little as $120 or so, depending on where you shop.

The test setup was a bunch of can oscillators and up to about 30 MHz, the scope did OK. After that, the results were less than stellar. However, we aren’t sure that the test setup — on a solderless breadboard — wasn’t part of the problem and we’d have liked to see the test done with some known good quality probes as that can also contribute to bad readings even if the scope’s circuitry is up to par. We’ve seen reviews of these cheap probes that suggest they aren’t bad, but they aren’t actually all they claim, either.

The verdict? [LearnElectronics] likes it for a 30 MHz cheap scope. Of course, cheap is a relative term here. You can get a much better scope, but it will probably cost more than this one. For most of what you are probably doing with a scope, this seems like it would be adequate. On the other hand, throw in a few hundred extra dollars and you could have more channels, a more likely chance of measuring high-frequency signals, and probably enhanced measurement capabilities, too.

We will admit, though, having a portable battery-operated scope can be super handy sometimes, and you don’t always need the highest speed. You probably use your cheap multimeter more than your six and a half digit bench scope, too, right?

While the screen isn’t large, it is more convenient than a tube full of flames. And while it may not quite make 100 MHz, it is cheaper than one that will do 100 GHz.

 

35 thoughts on “Tablet Oscilloscope Claims 100 MHz, But Is It?

  1. Yeah this test is worthless when he’s using a solderless breadboard lol. He’s just measuring the stray capacitance in there, who knows if the tablet is the problem? Although I also doubt it measures up to 100MHz, this test doesn’t tell people anything and is close to just being libelous… he’s basically accusing them of false advertising without any real evidence

    1. dude theres people out there doing stuff in gigahertz range on breadboards… i think this guys bumbled together rig is probably valid for this piece of crap

      1. I figured the 1Gsps was from interleaving on repetitive signals. Op-amps and ADC’s at that speed are hard to come by. Is there an “average” setting as you will find on a Tek scope? I bet it is like that.

        1. My 200MHz scope only have 200MS/s (real time) sampling rate, but it can sample at 25GS/s (repetitive). All of that are labelled on the front. The repetitive rate is limited by the internal sampling clock jitters, accuracy of the triggering circuits etc.

          The interleave sampling is a bit on the slow side and assumes that your waveform is repetitive, but it is better than nothing. 1Gs/s (repetitive) for a 100MHz signal would only get you to 10 points/cycle.

  2. I wrote a bigger comment, but hackaday’s dog ate it, so here’s a short:
    That setup, not exactly 100% as said Al and the video maker himself. No comparison with a “real” scope. He jumps from 28.somethingsomethingMhz to 100Mhz with nothing in between. Would be nice to see a better test.

    1. @Jii said: “hackaday’s dog ate it”

      I feel your pain Jii. There are probably thousands if not million HaD comments buried in the Big-Bit-Bucket in the sky since HaD changed to this horrible comment system.

      Hey HaD, bring back the days of all lower-case; not that I like all LC, but at least back then your posts didn’t randomly VANISH into thin air.

  3. is he really testing above 10Mhz with a bread board source there without coupling capacitors? And no known-good second oszi cross-check of the setup either? hmm…k…

    1. If by “same conclusion” you mean… totally opposite? This video literally has a 5Vpp 100 MHz sine wave going into it, and the 10X probe records around 1.5Vpp, which is completely consistent with a 3 dB 100 MHz bandwidth. It looks like the trigger’s goes kinda wonky at 100 MHz exactly, but that could just be a trigger running at 200 Msps internally.

      1. I should clarify that that video also comes to a “I wouldn’t use this for over 30 MHz” conclusion, but it’s for a totally different reason than what’s in this video: the analog bandwidth of the scope is definitely 100 MHz (and c’mon, it’s not like that’s *difficult*).

        The reason *this* video says “I wouldn’t use it for over 30 MHz” is because it’s *actually* 200 Msps, not 1 Gsps, which I don’t think is mentioned in this video.

      2. and, of course, I have to amend myself because I didn’t actually read the numbers right at first (and then I copied them correctly here, wtf me) so obviously yeah, it’s more like 30 MHz bandwidth. But Dave Jones’s video shows that quite a bit more clearly than this video, which really seems clearly degraded by the setup.

  4. While this is Hackaday, and we applaud people doing their best with minimal equipment, there are very well established techniques for testing oscilloscopes.

    They always start with ‘Using a calibrated signal source’. And never start with ‘Grab a handful of can oscillators that have unmeasured output characteristics’. Dave’s video where he actually measures the device with an actual signal generator that he trusts is what should have been linked in the article.

    At the risk of attracting flack, there is much opportunity for the HaD authors to help people learn. Lets do that.

    1. Sure. And if you really want to measure front-end bandwidth, you do it with a super-quick edge, rather than a generic signal generator.

      For the practical:
      https://hackaday.com/2016/10/03/the-fastest-rise-time-in-the-west-making-a-truly-quick-pulse-edge/
      https://hackaday.com/2016/10/04/a-quickly-hacked-together-avalanche-pulse-generator/

      And for the theory:
      https://hackaday.com/2019/07/16/say-it-with-me-bandwidth/

      Point is, though, with an analog bandwidth of 30 MHz or so, displaying the shortcomings should be easy.

      1. The worst part of that “oscilloscope” isn’t even the less-than-advertised/wildly optimistic/fraudulent (make your pick) bandwidth claim but that the frontend response is not flat. Dave’s video shows it clearly even though her doesn’t really talk much about it.

        The consequence is that the scope is pretty useless for making amplitude measurements because you can’t trust that the amplitude at the given frequency is what the gizmo actually tells you it is.

      2. You would also eliminate the scope probe by connecting the high speed pulse to the BNC (or other high speed connector) and using 50 ohms termination. Quick often the “complementary” scope probe isn’t all that great.

        That and the dangling scope ground cable attachment is a great sign of a noob making a high speed measurement.

  5. The hell happened to my comment?
    Shorter short, because Hackaday’s dog keeps eating my comments.
    Not ideal test setup, jumps from 28,somethingsomething to 100Mhz without anything in between, no comparison to a “real” scope.

    And i’ll add this, EEVblogs video is much more thorough. Thanks Robert for the link (he was the first).

  6. So here’s my question. If I was looking for a scope to play with, and the $30 2″ square single channel 20Mhz sucker from wherever isn’t cutting it, but I’m not at the point of dropping $400 on a Rigol DS1052z (or whatever the 2 channel model is), would this be a valid first scope for basic learning, hacking, etc?

    How is the UI? I’m not wild about touch screen… knobs are nice and simple and intuitive to use, the more the better I suspect. Overloading the UI with too many functions on one knob/four way controller, etc just leads to frustration.

    And yes, I keep a vague eye on ebay and other sites for deals/steals.

    1. There are quite a few options, it just depends upon where you want to compromise – do you want 2/4 channels? MSO? sig-gen? analogue performance specs you can trust? wedded to standalone or happy to go for USB based?

      1. I don’t think I want USB based because I use linux exclusively for my developement work. Standalone should be just fine, and actually preferred. Though support for Linux to pull data from it easily would be nice. MSO isn’t need since I already have a small digital logic analyzer and sigrok.

        I’m a *beginner* with ‘scope work, but I would have to get too limited too quickly. I have one of those kit KT15001 (and a multitude of other names) suckers, with just one channel and a suspect quality probe.

        I suspect a cheap 50Mhz two channel from Rigol or other name brand like that would be more than good enough for my needs for a long long time.

    1. Audio watts have always been somewhat specious… to the extent that if you made something 20% efficient at sound to electrical power conversion, you’d theoretically get free energy from any consumer stereo looped with it.

  7. Hey peeps – Is it just me or at first glance does that “FNIRSI” logo on the device look a lot like the “ANRITSU” logo? If yes, I wonder if that was done on purpose…

  8. So I watched this and some other reviews of the Fnirsi 1013D.

    Memory depth 240kbit instead of bytes or samples, which both Davey Jones and this review missed to mention.
    Most sensitive range 50mv/division, while any “normal” scope at least goes down to 5mv/div, and the trend is to more sensitive (limited by noise). With a normal 1:10 probe this would be 500m/div which is a serious limitation. (You do not want to use any oscilloscope without the 1:10 probes)
    Often you see the waveforms slowly turning into new waveforms.
    This is caused by some averaging, for example @08:08 and @ 10:32.
    The “nice signal there” @08:43 is completely bogus. If I could see the grid and text properly you could do a risetime & bandwith calculation from it, but the rounded tops are still too suspicious, you also see the averaging here.

    Also, I despise manufacturers bluntly lying about the specifications. Such behaviour should not be tolerated.

    From another reviewer I saw pretty decent X-Y graphs on this scope, but it’s hard to trust them because of the averaging.

    For someone with experience with oscilloscopes it’s relatively easy to recognize when this scope thing goes wrong, but for beginners this can easily lead to loads of frustration.

    I have not used this thing myself, and it’s easy to get some stable screen when you hook this up to a function generator, but is this thing capable to capture the more elusive stuff when doing real measurements? I have also not seen a “real” review of this thing. None of the reviewers for example did anything with the single capture mode.

    So the limitations of this thing are:
    * Seriously limited frontend (50mV/div).
    * Very small memory depth, misleading specification. (bits!!!)
    * Serious aliasing (See Dave’s Review for that).
    * Bluntly lying about bandwidth.
    * Bluntly lying about sample rate.
    * Weird averaging (Can this be turned off?)

    For me, reviewing what a gadget can do is just half of the review. The absence of serious bugs and limitations is at least just as important, and this thing has just too many flaws.

    You can get regular decent oscilloscope for around EUR300 nowaday’s. Sure, it’s twice the money of this toy, but my guess is that this thing will cause so much frustration that you want to buy a “real” scope within a year or two, while such a EUR300 scope does not have the glaring pitfalls of this thing and may well give you good service for years to come, and it’s also more likely to be sold second hand for a decent price. Take for example the old Rigol DS1052E, second hand prices seem to be quite near the new price ???

    If you really can’t afford to spend EUR300 for a new scope, then you will probably be better of to buy a second hand scope instead of this.

  9. I picked one up. Seems to be going really fast on Amazon, as I had a hard time finding one. I could not find it at the $120 price he kept quoting (not on amazon). More like $160 – even on ebay. It has a really nice, responsive UI. Simple and easy to use. Great for home projects. Nothing I measure gets close to 100Mhz, so its great for me and the display is much larger than the other scope I have. I think its worth is as long as you dont need the bandwidth.

  10. Not commenting specifically on the merits of this oscilloscope but there seems to be a general misconception of what a 100 MHz spec means. This is the frequency at which the measured value of the amplitude of a sine wave will be 3db lower than its actual value. This represents about a 30% error in measurement so you should not expect to be able to accurately capture 100 MHz frequency signals. At about 30 MHz, the expected error rate would be around 3%.

    Check out this post from Siglent about this: https://siglentna.com/operating-tip/determine-bandwidth-scope-require-application/

    Also, this Tektronix technical brief on the subject (hosted by the Electrical and Computer Eng. dept. at UBC): https://people.ece.ubc.ca/robertor/Links_files/Files/TEK-Understanding-Scope-BW-tr-Fidelity.pdf

    Searching Google for “oscilloscope frequency 3db” should also bring up similar results.

  11. From my understanding and echoed in the Siglent link from @hzc, to measure a square wave reasonably (i.e. at least the first 3 odd harmonics) you need a scope with 5x the bandwidth of the frequency. Hmmmm that seems to come out to about 20 MHz for 1/5 of 100 MHz. Seems to me to be right on target. Just sayin’!

Leave a Reply to RW ver 0.0.1Cancel 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.