Update 22 December 2023: I had a repeating problem with the Rigol with a voltage regulator failing on my unit th-cam.com/video/RlWIhsjXf8E/w-d-xo.html. After replacing it about five times, and with decreasing periods before failure, I've thrown in the towel and bought a replacement unit (with a 3 year warranty included this time!). I have no reason to believe this is a common problem, but it did affect my unit. Here's an update after over a year: I am still in two minds between these scopes, here are some thoughts after a year or so... I find that the UI on the RIgol to be easier to work with in that it's more intuitive to me. Almost all functionality is available on the panel knobs, it's rare to need to use the touch screen. After a while you get used to the delay when changing timebase, to such an extent where I found it didn't really impact me that much. There are some frustrations, one of which is the trigger frequency readout which is a floating item rather than being integrated with the instrument's standard readouts. The menu covering up a significant part of the waveform display, and having to continually turn it off is something that could be resolved to some extent with, say, transparent menus or do what the Siglent does, ie squash the waveform display horizontally. The fonts on the Rigol I think are a bit larger than necessary. As for the front end noise on the Rigol I'm not sure whether that will ever get resolved. On the Siglent I am more likely to find my workflow is interrupted while I figure out how to do something. A lot of functionality is only available through the touch screen. The display seems crisper than the Rigol's, and the amount of information available is more comprehensive largely thanks to using smaller fonts and a slightly larger screen (they have the same resolution, 1024x600, Siglent is 10.1", Rigol is 9"). Regarding the LA which I use frequently, on the Siglent is limited to 500MSa/s with the Rigol at 1GSa/s. On the Siglent its memory is limited to 50Mpts with the Rigol at 25Mpts. This is despite the analogue channels being up to 200Mpts. On the Siglent, the Search function isn't available at all on the LA channels, and on the Rigol Search works on the LA but only with basic searches like edge and pulse, you can't do serial trigger search on the LA channels. Note that you can do all the serial decode and triggering on the LA, the limitations are just on the search function on the LA channels. There is also the 1ns trigger delay bug on the Rigol's LA that I mentioned in the video: in practice, you just work around it. I've had two hardware problems with the Rigol, one was a faulty probe that I replaced and the other was rather more difficult: a couple of months ago the unit suddenly died after a few minutes of being switched on. It turned out to be a faulty regulator IC on the main logic board that I replaced myself. It's worked fine ever since. The TL;DR: For a daily driver, I have changed my mind, and I now find I'm slightly more likely to go for the Rigol despite its several shortcomings, and this is mostly because I'm more comfortable with the workflow. The slightly smaller form factor of the Rigol means fits better on my bench, underneath the monitor without obscuring it. YMMV of course!
do you not like using the touchscreen? we find we use the touchscreen for things like cursors, measurements, triggers etc and the buttons and encoders to manipulate the waveform
"As for the front end noise on the Rigol I'm not sure whether that will ever get resolved." Wow. I'm really surprised by this. The trace on the test signal was terrible. Are you sure it wasn't some sort of triggering issue ? I agree that the AWG needs to be 5V, not 2.5V. This is a really good review. Thanks for doing it. Question: is there a better scope in this price range ? The 2.5V AWG is totally useless to me, so I'd be willing to forego that. Any chance that Rigol is going to update this model and fix some of these shortcomings ? Or release a new model to replace it ?
In short, they are both attracting me. My concerns to the MSO5000 is: - No real 2mV and 1mV vertical range. (I don't expect the 500uV range) - If it's really too hot? I don't like it's ventilation design as well. SDS2000+'s concerns: - Hassle UI operation. It really make me remember the experience of Lecroy scope I've been used. - Lack of zoom out feature. I'm accustomed to the Agilent style. I feel without zoom in feature is hassle to use.
@@Dan-oo1rj Good point: I use wireless mice on the two Keysight scopes I have on shelves above the bench: I have under-desk keyboard and mouse shelves. For the Keysight that I have at bench level (my comfy slippers scope!) I use the controls directly.
Thanks for the direct comparison, even 2 years later I was one of the people that still struggled to decide which one to get. Seeing both of them in action has helped me make up my mind. If I hadn't missed Rigols sale last month here in my region I might have gone with Rigol, but now I'm leaning ever so slightly towards siglent, especially since I have a decent offer for one.
Didnt u noticed that rigol is in normal aqusition mode and Siglent isi in peak mode? In peak mode of course line will be very thin.I think u should try both in normal mode
A good spot, however it makes no visible difference on this test whether it's in peak or normal, on either scope. What about this test leads you to conclude that "in peak mode of course line will be very thin"?
In peak detect mode the scope is showing the maximum values found in that sample set. The name says it, de Peaks. So the opposite is true, we would see a thicker line.
I have a rigol scope but i do have several other pieces of Siglent test gear. I will say this about Rigol their support is very good. I have not had to call Siglent so i cant speak to that but product support is another important measure
Just recently bought the Rigol. It was just 400 Euro cheaper. 900 Euro int total, and the digital probes, all protocol, the signal generator was included. I accept that the modern ones are much better but they are also way more expensive.
It's not a bad scope: I'm still a knobs and buttons guy, so I like having the four separate channel controls. It's the best scope I have for remote access as the screen update rate is comparatively fast when accessed using RealVNC (rather than the scope's web interface, which is slow). Some quirks with the LA... Firstly, at high speeds, there's a 1ns offset between trigger point and display. Secondly, the Search function doesn't work with LA channels at all. It's also a bit flaky on more complex serial searches. Thirdly, note that the LA sampling rate (and memory depth) is 1/8th that of the LA channels: this is especially important for serial decodes on the LA channels, as the headline sampling rate shown on the scope's display is that of the analogue channels. You can see the LA sampling rate in the Acquisition -> More menu. Fourthly, the decodes become unreliable on both Analogue and LA channels on busy traces. The sweet spot for maximum reliable decode for a single 100kbps I2C trace seems to be at 5MSa/s analogue sampling rate and 10Mpts memory depth, so that equates to LA sampling rate of 625kSa.s and LA memory depth of 1.25Mpts. That's about 2 seconds' worth. You can dramatically improve this on bursty data by using the Record function: for example, I achieved over 9,000 34 byte packets over a 4.5 minutes duration using the Record function.
With regards to apparent noise on the traces of the MSO5000, I suspect Rigol and Siglent have different "standard" settings for their "normal" operation. I've used the various Display Controls, outlined in Chapters 4 & 16 in the MSO5000 manual, to improve trace fidelity in my "normal' scope activities. Prior to making those adjustments I thought the Rigol and a bad ground, or something more sinister. TRACE VERY FAT, like me! Now my trace is nice and thin with no annoying spurious noise being apparent. HOWEVER, go back to the "factory settings" and all the noise that's really there, is really there. BTW, I remember the days of hair-thin traces on the TEK 485 350Mhz portable scope. NICE! My ancient TEK 475 on my bench, not so much as it has the aluminum mesh dome "lens" to magnify the apparent sweep and help keep the CRT as short as possible for portable instruments. The FAT traces of the early 465 and 475 scopes were an embarrassment to Tek. My old TEK 475 is a rescued scope from the IBM scrap sale... Still meets specs! Love the 2mv vertical sensitivity! The HP's and Philips, of the day, didn't use the mesh lenses, but were still able to keep the CRTs fairly short. The lenses always spread the traces. The mesh lenses were improved after a couple of years into the design. Can't blame those faults on todays LCD displays, can we! I experienced this first-hand as I worked as a Field Sales Engineer/manager for both Tektronix and Philips for most of the 1970's... Yep, I'm OLD and gray but electrons still work the same now as they did then! BTW Dave Jones, on TH-cam EEVBLOG, has a pair of nice videos on the topic of apparent / displayed noise on digital scopes. They are: EEVblog #601 - Why Digital Oscilloscopes Appear Noisy and EEVblog #442 - Analog Vs Digital Oscilloscope Noise. Worth a watch and good information why a more expensive digital scope might appear inferior to a less expensive digital scope or an analog scope. So, my question do any of these phenomenon apply to your comparison. Not having and Siglent SDS2000x on hand, I'll leave it to your capable hands to enlighten the rest of us. I TRULY enjoy learning something new every day! Excerpts from the Rigol MSO5000 manual: CHAPTER 4 - TO SET THE SAMPLE SYSTEM Contents in this chapter: ◼ Timebase Mode (NOT INCLUDED THIS CUT AND PASTE) ◼ ACQUISITION MODE ◼ SAMPLING MODE ◼ SAMPLE RATE ◼ LA Sample Rate (NOT INCLUDED THIS CUT AND PASTE) ◼ Memory Depth (NOT INCLUDED THIS CUT AND PASTE) ◼ LA Memory Depth (NOT INCLUDED THIS CUT AND PASTE) ◼ Anti-Aliasing (NOT INCLUDED THIS CUT AND PASTE) ◼ Horizontal Expansion (NOT INCLUDED THIS CUT AND PASTE) ACQUISTION MODE The acquisition mode is used to control how to generate waveform points from the sample points. MSO5000 supports the following four acquisition modes: Normal, Average, Peak, and High Resolution. By default, the acquisition mode is Normal. Press → Acquisition on the front panel, and then rotate the multifunction knob to select the desired acquisition mode. Then, press down the knob to select the mode. You can also press Acquisition continuously to select the mode or enable the touch screen to tap the desired mode and select it. NORMAL In this mode, the oscilloscope samples the signal at a specified fixed time interval to rebuild the waveform. For most of the waveforms, using this mode can produce the optimal display effects. AVERAGE In this mode, the oscilloscope averages the waveforms from multiple samples to reduce the random noise of the input signal and improve the vertical resolution. Greater number of averages can lower the noise and increase the vertical resolution; while at the same time, it will slow the response of the displayed waveform to the waveform changes. When you select "Average" mode, press Averages and rotate the multifunction knob or use the numeric keypad to set the desired number of averages. You can also use the pop-up numeric keypad to input the average count. When the average count value is not in power-of-2 increments, a prompt message "Truncation average error" will be displayed. At this time, a value that is smaller than the one you input and the closest to power-of-2 increments will be input automatically. For example, if you input 9 with the numeric keypad, the average count will be input 8 automatically. The number of averages can be set to 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384, 32768, or 65536. By default, it is 2. PEAK In this mode, the oscilloscope acquires the maximum and minimum values of the signal within the sample interval to get the envelope of the signal or the narrow pulse that might be lost. In this mode, signal aliasing can be prevented, but the noise displayed would be larger. In this mode, the oscilloscope can display all the pulses whose pulse widths are at least the same as the sample period. RIGOL HIGH RESOLUTION This mode uses an over-sample technique to average the neighboring points of the sample waveform. This reduces the random noise on the input signal, generates a much smoother waveform on the screen and improves the vertical resolution. This is generally used when the sample rate of the digital converter is greater than the storage rate of the acquisition memory. Note: • The "Average" and "High Res" modes use different averaging methods. The former uses "Multi-sample Average" and the latter uses "Single-sample Average". • In "HighRes"mode, the signal bandwidth does not exceed 1/32 of the sampling rate. • In "High Res" mode, the highest waveform refresh rate mode is not supported. SAMPLING MODE This oscilloscope only supports the real-time sampling mode. In this mode, the oscilloscope produces the waveform display from samples collected during one trigger event. The highest real-time sample rate on the analog channel of MSO5000 series is 8 GSa/s. The current sample rate is displayed under CH SampleRate, and you can press More → CH SampleRate to view it. TIP: Press RUN/STOP on the front panel to stop sampling, and then the oscilloscope will maintain its last captured graph. You can still expand or zoom the waveforms by using the horizontal and vertical control knobs. SAMPLE RATE Sampling is the process of converting the analog signal into the digital signal at a specified time interval and then restoring them in sequence. The sample rate is the reciprocal of the time interval. The sample rate of the analog channel is related to the current channel mode. The maximum real-time sample rate in the single-channel mode of the oscilloscope is 8 GSa/s. The maximum real-time sample rate in the half-channel mode is 4 GSa/s, and the maximum real-time sample rate in the all-channel mode is 2 GSa/s. SEE ALSO: CHAPTER 16 DISPLAY CONTROL In the display control setting menu, you can set the parameters such as the waveform display type, persistence time, intensity, grid type, grid brightness, and etc. Contents in this chapter: ◼ To Select the Display Type ◼ To Set the PERSISTENCE TIME ◼ To Set the WAVEFORM INTENSITY ◼ To Set the Screen Grid ◼ To Set the Grid Brightness ◼ Scale ◼ COLOR GRADE ◼ Waveform Freeze END of Attachment...
Yes the fat display is the very first thing that jumped out at me. I did several measurements comparing the MSO5000 with a number of other scopes, it has about a 10dB or so worse noise floor: it can barely scrape 5 bits of effective resolution in 8 bit mode. While you can improve things with averaging and oversampling ("High Resolution"), you can do that on other scopes too, but the front end is demonstrably very noisy. See post #1745 here www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg3180030/#msg3180030 The biggest problem for me with the Rigol was a usability issue, in particular the unfathomable delay when changing timebase.
After 11 additional months of use, the noise floor hasn’t been critical for me and I have alternative instruments. Still… where’s the noise source? EMR/poor shielding? Sub-parr devices? Careless layout?? I’m VERY hesitant to crack open the case on my still-warranted, although “hacked” MSO5074 and Dave did a pretty detailed tear-down early on. Anybody out there up to a full necropsy? I’m into a house remodel and don’t have the bandwidth to tackle this question. Anybody out there open to the challenge? A quick call to Rigol might produce a loaner for dissection .
@@Tool-Meister i am very interested. I have Just bought a mso5000 on amazon and i have 2 months to try It and return for free. Now i am in doubt if i should take It or give It back. All these people saying It Is very noisee are making me warried
Re. Siglent’s lack of video out: you can connect to the SDS2000X series using a computer or tablet (or smartphone for that matter) and point a browser to the scope’s IP and see a full-screen image of the scope’s display. Control the scope too.
@@71GAYes it is. However if you're not using the HDMI on the Rigol it also has VNC support that is extremely responsive - much more so than a web interface.
I have the Rigol MSO1104Z and I have lots of problems with it just not triggering correctly, and the logic inputs are flaky sometimes. The lag you mentioned drives me nuts too on the 1104, even though it isn't quite as bad as what you showed on the 5000. My new Siglent SDS2104X+ is on the way to me right now.
I've had an MSO-1104Z-S too for a while, it's my field scope as it's lightweight and compact. While it's a bit laggy, the 5000 is better, although that long delay after changing timebase isn't on the 1104Z. Biggest long term problem on my 1104Z is the LA connections inside the unit on the IDE ribbon connector: I've had to re-crimp it once and may need to do it again.
Nice comparison of the two scopes. There´s one failure in your description though. The Siglent is 2x2 channel oscilloscope (grouped ch1+2 and ch3+4), which means that the sample rate is 1Gs/s at minimum for 3 and 4 channel use and 2Gs/s if 1 or 2 channels are used. Apart from that, we use the Siglent on a daily basis and its a mighty good device.
Yes, it's what I was alluding to at 36:00, but one watching it back I agree it's not clear, not least because I only refer to channels 1 & 2. There are serious antialiasing issues because of it, particularly if it's had its bandwidth upgraded to 500MHz. Nyquist would not be pleased!
@@nezbrun872 Yes, I agree Nyquist would probabely not be amused. ;-) Still though it makes a noteable difference if the samplerate drops to 1/4 because only one ADC is employed or if it drops only to 1/2 because of 2 ADCs. It should also be kept in mind, that a 4x set of probes capable of such high bandwidth typically costs as much as the whole scope. The 500MHz upgrade is rather more for marketing reasons than of actual useability here. But hey, I´d rather receive firmware upgrades even if they´re a bit over the top than no updates ;-) Especially if those updates not only fix existing bugs but also add on functionalities .... as it is the case with Siglent´s updates.
Many thanks for this review, I have been agonising over which of these two scopes to buy. Like you, I have tended to favour Rigol in the past but was starting to move towards Siglent after watching several EEBlog videos. Now, having watched your excellent comparison video, I will definitely be going the Siglent route. I understand that Rigol have designed their own asic front end chip, this might be the reason for the poor noise performance. One possible downside of the Siglent is that it has only one sig. gen. output channel and no modulation modes, whereas the Rigol has two channels with am/fm/fsk modulation.
Pretty much sums it up. That and the performance thing where the Rigol stops triggering for about a second when you change the timebase, that makes if feel more sloggish than it might otherwise be. Considering I'm generally a Rigol fanboy, there's still no doubt which of the two to go for, and the Siglent is undoubtedly a better option overall, something that surprised me.
Only tempting feature of Rigol is its 8GS/s. Otherwise Siglent looks much more a mature product. The scopes one is realisticly / willing to buy for personal non business use slide down regarding the maturity. Siglent seems the most mature of them all. And way down to FNRSI kind of scopes. In my opinion Rigol can take a seat behind the Siglent. Rigol had the front seat before, but has lost it.
All these small annoyances with low and medium scopes drived me cracy, had rigol DS1074Z, then Rigol MSO4054 and always there where tradoffs and limitations bugs and not thought tru solutions when using them fully. decoding was a hit and run, small screen, fideling back and forth with zoom, always woundering if it was the DOT or the scope that played tricks..So when a InfiniiVision MSOX6004T (Incredebly cheap 6000€! fully loaded! A leasing company sold it, had no idea what it was (They had two, bought them both, soldone and had the other for free!)..they mostly sold cars, Trailers, tracktors machines etc I was the only bidder! Since i was in the proces of ordering a brand new one from Keysight) and was earching, it came in view ! Now only look back when having naightmares after a long stint in the lab. Life before and after...All those seconds with becouse of sluggish updates when using the Rigols that gave me gray hair and stomack problems. I recomend the msox6004a!
When are scopes going to move beyond 1990 display tech with OLED 300ppl displays? With these scopes I can't tell the difference between a dac output and the original test signal. Both look stepped instead of pure sine wave.
@@nezbrun872 Up to now, I've been using a CRT based scope. But the new scopes all have these low resolution VGA displays, which is frustrating, as I shop for a replacement for my 2465B that U800 failed on.
To be fair, the Rigol can decode flex-ray, i2s and mil-std-1553 serial data. This might be an important difference to some people. Rigols last firmware update was in April 2020. I'm hoping that is only because of coronavirus and that future updates will come at some point. I'm looking to replace my old DS 1102C which has served me well having never been updated. Alas, the lcd is starting to exhibit solid color vertical lines that come and go. I've tried reseating the cable, but I fear that it's the actual screen that is failing internally. I'm afraid to pull it completely apart until I have a new scope. If I can fix it, it is destined for portable and possibly automotive service. It's been a great scope, even though it cost a small fortune (>$1000) new. Right now, the MSO 5074 with the free bundle is looking pretty good to me for $1000. I do like that the new Siglent's support wifi, but I'm hesitant to switch from Rigol because I've been so happy with my last purchase from them.
Thanks for the great review! It looks like you have 3 scopes on you bench, Rigol, Siglent and Keysight. Which is your everyday driver? Also any comments on the quietest fan noise of the three?
In this lab, there are five scopes set up permanently on the bench: MSO8104A at the soldering station end, with MSO7104B, MSOX3104T, Tek MDO4104C and an old HP 54121T 20GHz sampling scope on the main bench. My favourite day-to-day scope to use is the MSO7104B at the bottom eight of the video: it has a fast boot time, big screen and is a joy to use with its super snappy & simple, intuitive UI: it's the first one that gets switched on. The 3104T is also nice scope and its UI is very similar to the 7104B. I has more decode and analysis options than the 7104B, but its screen real estate is significantly lower in terms of both resolution and area. Being ably to monitor two serial buses simultaneously can be quite handy. The touch screen I don't use much due to its position on the bench, but I do use a wireless mouse with it in the under-desk keyboard drawer. The MDO4104C is seeing more use these days, but the UI is a right royal bag of spanners. It's inconsistent, and unlike the Infiniivisions (7104B and 3104T) its UI becomes unresponsive when you start adding the digital channels, bus decodes and longer acquisitions. It took me many months to get comfortable with it. I find I'm forever having to adjust the memory depth, something that's simply unnecessary on the Infiniivsions. However the Tek does have some useful additional features over the Infiniivisions that I have, such as USB HS decode, up to three simultaneous buses, 20Mpts memory depth on all channels,
Hi, I'm thinking about picking up that Rigol... and was considering the Siglent also, so I was enjoying your video. You didn't mention the "App" button (lower left on the Rigol). That looked pretty cool to me on some other videos I watched. What are your thoughts on that? I'm also curious as to what the stop, left & right buttons do on the Rigol (in the center of all the controls there).
I don't use these kinds of functions in my day to day work very often, so they rarely hit my radar. For a bode plot, I'm far more likely to use a PC based solution, specifically the Analog Discovery. Bode plots are quite slow on both these scopes.
Regarding channel to channel skew: from the MSO5000 manual: Channel Delay When using an oscilloscope for actual measurement, the transmission delay of the probe cable may bring relatively greater errors (zero offset). MSO5000 allows you to set a delay time for calibrating the zero offset of the corresponding channel. Zero offset is defined as the offset of the crossing point of the waveform and trigger level line relative to the trigger position, as shown in Figure 2-2. Missing Graphic which doesn’t copy. Figure 2-2 Zero Offset Press 1 → More → Ch-Ch Skew, rotate the multifunction knob or use the numeric keypad to set the desired delay calibration time. The available range of the delay calibration time is from -100 ns to 100 ns. Note: This parameter is related to the instrument model and the current horizontal time base setting. The larger the horizontal time base is, the larger the setting step will be.
I think my point was that channel skew isn't available for digital channels. Typically this wouldn't pose a particular problem, except that there is a constant 1ns delay between a digital edge trigger and what's displayed. So if you do an edge trigger on Channel D0, the edge is always displayed a constant 1ns after the trigger. Timestamp 34:00.
Nezbrun My error. First of many each day. Ready, Fire, Aim... Did you try the THRESHOLD ADJUSTMENT and scope auto-calibration? The manual seems imply that it may affect the skew, i.e. “To Set the Threshold and Calibrate Probe”. I’m in the middle of a house sale and associated move so am at a disadvantage regarding getting at my equipment, otherwise I would try to duplicate the problem for you. I received my MSO5074 in July and promptly performed the magic feature and bandwidth spells. I THINK the noise-floor had a noticeable increase but I wasn’t bright enough to make any “before” measurements other than rise-time, which of course decreased. Not to be repetitive, congratulations on a very thorough review. You seem to take the peer review and questions all in stride. Kudos to you!
@@Tool-Meister Regrettably, none of these work. It's the same on any digital channel, there's a constant 1ns delay between the trigger point and the displayed edge.
One last thought. In pure analog scopes, you can’t see the leading pulse edge unless the scope has a “delay-line”. Thankfully, virtually all modern, say post 1975, have a delay-line which allows seeing a pulse’s leading edge. However, it takes special features and techniques to know your exact trigger point. That said, with digital scopes you can know and see the trigger point. As to the 1ns offset, it’s at least known and can be allowed for. I wonder what the tradeoffs are for Rigol’s design approach? WHEN I can get at my OWON scope and my pulse generator, I’ll check to see if it has an offset. To clarify, your observation is there a 1ns offset from the triggering channel’s displayed position compared to all non-triggering channels displayed positions? You are NOT observing that phenomenon on other digital scopes, specifically the Siglent in the comparison. I’m going to reach out to Rigol. I’m only 8 miles from them and can “ mask up” and visit them in person if need be. BTW, where did you terminate the signals? Did you swap channels to verify the signal source’s 2 channels have identical propagation delay and no reflections? I suppect you did but I’m imagining what questions Rigol will ask when we speak. Again, GREAT WORK on your part!
Oooh, have to press a button to select the ch! Nasty! They should be shot for that. (yoking .. really). It also seems that you're a little confused about what the word "DEFAULT" means. No problem. Okay, so this seems to be a bunch of nitpicks with the Siglent edging out the Rigol. Good thing he Rigol is a bit cheaper. (but not as I write this) For me the win goes to the Siglent. I'll agree, neither one is quite as good as a $18,000 Keysight, R&S or Tek, but if I had $18K burning a hole in my pocket I have higher priority outlets for that kind of excess. I'm a hobbyist so either of these scopes are bordering on overkill for what I need. The Siglent (especially considering the present pricing) is the clear winner. One is on the way as I type.
Is the MSO5000 front-end really more noisy? I checked it against a much higher-end 1GHz scope and also a 350MHz analog scope and the noise was exactly the same on all of them when using the freq gen from the MSO. I'll have to check against a better function generator but I don't think it will be much different. ??
I did comparative measurements against a number of other scopes: Keysight MSO8014A, Keysight MS)7104B, Keysight MSOX3104T, Tek MDO4104C, Siglent SDS2104X+, and Rigol MSO1074Z. Effective bits on the Rigol MSO5000 was about 5 bits (varies somewhat depending on input attenuator), while the other scopes were around or just below 7 bits, so you're looking at about 10 to 12dB or so worse noise floor on the Rigol MSO5000X. It's covered here: www.eevblog.com/forum/testgear/benefits-of-going-with-all-siglent-setup/msg3281580/#msg3281580
@@nezbrun872 For the MSO5000 I am getting lower noise numbers than your eevblog post. Have your tried these measurements with the new firmware updates? Using your set parameters, I am measuring the std-dev noise (AC-RMS) as follows: 0.167mV at 2mV/, 1.9mV at 100mV/, 3.8mV at 200mV/, 8.6mV at 500mV/, and 67mV at 5V/. (I am using the available AC-RMS measurement instead of V-RMS, to measure the noise around the mean, thus avoiding any non-zero bias.) Also, I would generally expect to sample/measure and see wider noise peaks on this scope, because of its higher sampling and wfms/s capture rates.
Thanks for the detailed review. Does the high resolution mode on Rigol helps to reduce some of that noise by averaging the additional sample points available?
Iam new to all this but I think on my 1204 x-e siglent oscilloscope. I think incan independently set the x10 as a default on cearten functions. So when I press default it will default to what I set it to on some settings
I can't be 100% certain, but I'd have hoped it could be resolved with a firmware update. It is possible that there may be a hardware related PLL settling time limitation. Rigol had some issues around PLL instability in the DS1000Z series that was corrected with a firmware upgrade. I should note that at the time of writing this, it's been about 18 months since Rigol issued a firmware update to the MSO5000 range. See my new pinned comment for more updated info.
Yes, if you have a sufficiently repetitive signal and you can put up with the reduced update rate. Using averaging isn't a universal panacea, not by a long way.
I have a question, on the Rigol (MSO5104) it says 8GSa/s but the bandwidth is limited to 100MHz so I don't know what the point is. On the Siglent (SDS1204X-E) it is 1GSa/s but the bandwidth goes to 200MHz. Which one is better? My thinking is a higher bandwidth is better and with the Rigol all those extra samples just don't get used because of the limited bandwidth.
It depends: the MSO5000 drops to 2GSa/s as soon as you have more than one channel running, and is designed to support up to 350MHz with software licenses (see eevblog regarding how liberate your scope's options). A high sampling rate makes the anti-alias filtering a lot easier. On the 1204X-E, it drops from 1GSa/s to 500MSa/s depending on how many channels you have on, which doesn't leave much headroom for antialiasing. I'll also add that the MSO option for the 1204X-E is a terrible implementation, the scope becomes almost unusable with it.
Thanks for the video and update. Though I still can't make my mind going for which one. I fully agree with your unbiased opinion. Actually I ordered a MSO5074 and returned it when I have seen the video of "fake" vertical range below 5mV. I can accept it's ADC noise, but I can't accept the digital zoomed vertical range. It dose limit it's ability to measure low level signal like ripple noise. Below are some of my personal points of view: MSO500's strong points: - tangible knob and keys for vertical controls and soft menus MSO500's weak points: - the vertical ranges below 5mV/div are digital zoomed SDS2000X+'s strong points: - more detail in specifications, I feel Rigo's are a bit vague. - fruitful power measurement functions SDS2000X+'s weak points: - the UI seems hassle, not intuitive. - can't acquire waveforms outside the screen although it has got a lot of memory. it should be able to zoom out or shift waveforms which outside the screen. I can't say if the segmented acquisition make sense. But having zoom out ability make things easier. Something I'm not sure: - If MSO5000's full-memory hardware measurement is for all the measurements? I can only find the frequency counter is in hardware. - If MSO5000's full (1M points) memory FFT has high update rate? SDS2000X+'s seems a bit laggy. But it's still much better than those vintage digital oscilloscopes other than HP/Agilents'. Both model don't search events of digital channels. That's a pity.
Yeah rigol is more of a "digital" oriented scope, higher sampling rate higher noise. Price is about 30% more for siglent, worth it if you have the budget. Skew adjustment is in SCPI but doesn't work, reported as a bug, not fixed. But for me 99% of the time logic analysis on the PC is better than a scope.
Whether it's 30% more or not depends I think. You get 350MHz probes on the 70MHz model for example, whereas they're 200MHz probes on the SDS2000X base 100MHz model. The basic serial analysis is thrown in with the Siglent, a considerable additional expense on the Rigol unless you benefit from a marketing push such as the current bundle offer in some markets. Sadly I've seen absolutely no benefit on the Rigol over the Siglent on its 8GSa/s vs 2GSa/s sampling rate. The Siglent will go up to 500MHz analogue bandwidth, but beware of the aliasing issues when using the interleaved channel pairs as mentioned in the video, when the sampling rate drops to 1GSa/s. Possibly where the Rigol benefits in the digital domain is the search facility (only available on analogue channels weirdly, with no serial search at all on the Siglent) and the 1GSa/s LA (500MSa/s on the Siglent's LA). Whether you think logic analysis is better on a PC than a scope is an interesting topic, and I hear it a lot that you're better off doing it on a PC. I think it depends. It's a different work flow, and it's often a personal preference. On a scope it's all about triggers and real time analysis, whereas on the PC it's mostly about bulk capture and post processing. Such PC based LAs tend to have only the most basic triggering facilities. Personally speaking, It's pretty rare I ever do any serial analysis on the PC, it's only when I have an exceptionally long capture to analyse. Budget PC based LAs are hampered by low sample rates so are often useless for many SPI or I2S applications for example. In short I won't deny I was disappointed by the MSO5000, I knew I had to try out the Siglent bearing in mind others' comments on eevblog.
I like the update speed on the siglent better than the rigol and the display is rock solid - the rigol is fluttering about. I'm wondering how much better (if any) a Siglent SDS5034 would be vs SDS2354X-plus - in terms of features and experience... I think the separate vertical controls could be handy but isn't a dealbreaker, but also 5GSa/s (2.5) on the SDS5000 series vs 2GSa/s (1) on 2000. Or whether its just worth getting the base 2104X+ and 'upgrading' :)
Be aware that there are significant aliasing effects when running at 1GSa/s on the SDS2000X+ (i.e., ch 1+2 or 3+4 on) when upgraded to 500MHz bandwidth. I can't comment on the SDS5000, except that yes, I would prefer separate channel controls, but as you say not a deal breaker. There's a zoom function on the Siglent that's far better implemented there than on the Rigol that I didn't show, you can zoom in vertically as well as horizontally, and it shows the 2D area of the captured waveform you're looking at, rather than the ubiquitous 1D zoom overview. You have to see it in the flesh to witness it, but the display on the Siglent is so crisp compared to the Rigol, so much so that initially I was wondering if the timebase was actually running.
@nezbrun872 Hi. Thanks for the review, it really helps a great deal to make a choice on these (for personal use not cheap pieces of equipment). Remark: anyone using equipment has to know about its limitations, as even the most expensive equipment has. Question: as the Siglent does not have the sufficient sample rate to decently cover the highest possible BW of 500MHz with its 2GS/s, does it maybe have any old skool tricks like equivalent or random time sampling? That trick could work for periodic signals. And as mentioned before working on those frequencies would require matching probes or a good coax.
Hi there just watched your video on the yi 4k+ cable, i wonder if you can help me, could I make up a cable just for an external mic connection as I cannot purchase that cable anywhere, Thank you.
@@nezbrun872 I wince every time I hear how Dave Jones pronounces the name ("rygol") The original intention was for the name to sound like "Regal" as in royalty.
@@nezbrun872 No I haven't, but it doesn't really matter as I am in the market for a mid-range scope right now and not a phone, but your attention to detail instills confidence that you know what you are talking about. Like I said in my first comment "I am being petty".
The sample rate is only 8GSa/s on one channel. Using channels 1 and 3, you get 4GSa/s, but you still only get 2GSa/s if you use three or four channels, or two adjacent channels.
What do you actually get with the higher sample rate? The reason I ask is because my scope is twice the bandwidth (200MHz) but has a much lower sample rate. All I see is that with the Siglent I can look at a 200MHz signal and with the Rigol I cannot, so what do you get from the Rigol with the higher sample rate?
@@javacup7579 good refresh rate. almost like live view. i tried Siglent and it have bad screen visual signal live reproduction. also i can’t see distortion in fronts.
@@oleggritsev my 200MHz scope also has "good refresh rate. almost like live view" so i don't know what you get with the higher sample rate other than a lower bandwidth, still wondering what that is.
@@javacup7579 Rigol it is like Apple. Siglent it like Android. Siglent it made for poor people and quality of material is worst! I gave Siglent, Owon. Hantek. Rigol like tectonics
I already made a comment four months ago: 'A good spot, however it makes no visible difference on this test whether it's in peak or normal, on either scope. What about this test leads you to conclude that "in peak mode of course line will be very thin"?'
I had a sheet of paper to the right with all of the points of discussion. I'm sorry you're only able to concentrate for short periods. There were around two dozen points I raised comparing the two, how do you think I could reasonably demonstrate those in a three or even fifteen minute video? May I suggest that you try reading the description and the pinned post, assuming your capable of comprehending a thousand words.
All GUI writers should be locked into a test room for 1,000 days and made to use the scope all that time in order to make these GUI MORRONS write the GUI properly, INCLUDING responsiveness!
Update 22 December 2023: I had a repeating problem with the Rigol with a voltage regulator failing on my unit th-cam.com/video/RlWIhsjXf8E/w-d-xo.html. After replacing it about five times, and with decreasing periods before failure, I've thrown in the towel and bought a replacement unit (with a 3 year warranty included this time!). I have no reason to believe this is a common problem, but it did affect my unit.
Here's an update after over a year:
I am still in two minds between these scopes, here are some thoughts after a year or so...
I find that the UI on the RIgol to be easier to work with in that it's more intuitive to me. Almost all functionality is available on the panel knobs, it's rare to need to use the touch screen. After a while you get used to the delay when changing timebase, to such an extent where I found it didn't really impact me that much. There are some frustrations, one of which is the trigger frequency readout which is a floating item rather than being integrated with the instrument's standard readouts. The menu covering up a significant part of the waveform display, and having to continually turn it off is something that could be resolved to some extent with, say, transparent menus or do what the Siglent does, ie squash the waveform display horizontally. The fonts on the Rigol I think are a bit larger than necessary. As for the front end noise on the Rigol I'm not sure whether that will ever get resolved.
On the Siglent I am more likely to find my workflow is interrupted while I figure out how to do something. A lot of functionality is only available through the touch screen. The display seems crisper than the Rigol's, and the amount of information available is more comprehensive largely thanks to using smaller fonts and a slightly larger screen (they have the same resolution, 1024x600, Siglent is 10.1", Rigol is 9").
Regarding the LA which I use frequently, on the Siglent is limited to 500MSa/s with the Rigol at 1GSa/s. On the Siglent its memory is limited to 50Mpts with the Rigol at 25Mpts. This is despite the analogue channels being up to 200Mpts. On the Siglent, the Search function isn't available at all on the LA channels, and on the Rigol Search works on the LA but only with basic searches like edge and pulse, you can't do serial trigger search on the LA channels. Note that you can do all the serial decode and triggering on the LA, the limitations are just on the search function on the LA channels. There is also the 1ns trigger delay bug on the Rigol's LA that I mentioned in the video: in practice, you just work around it.
I've had two hardware problems with the Rigol, one was a faulty probe that I replaced and the other was rather more difficult: a couple of months ago the unit suddenly died after a few minutes of being switched on. It turned out to be a faulty regulator IC on the main logic board that I replaced myself. It's worked fine ever since.
The TL;DR: For a daily driver, I have changed my mind, and I now find I'm slightly more likely to go for the Rigol despite its several shortcomings, and this is mostly because I'm more comfortable with the workflow. The slightly smaller form factor of the Rigol means fits better on my bench, underneath the monitor without obscuring it. YMMV of course!
do you not like using the touchscreen? we find we use the touchscreen for things like cursors, measurements, triggers etc and the buttons and encoders to manipulate the waveform
"As for the front end noise on the Rigol I'm not sure whether that will ever get resolved." Wow. I'm really surprised by this. The trace on the test signal was terrible. Are you sure it wasn't some sort of triggering issue ?
I agree that the AWG needs to be 5V, not 2.5V.
This is a really good review. Thanks for doing it.
Question: is there a better scope in this price range ? The 2.5V AWG is totally useless to me, so I'd be willing to forego that. Any chance that Rigol is going to update this model and fix some of these shortcomings ? Or release a new model to replace it ?
@@ikbendusan I prefer to use USB mouse. Screen is clean and touch point is more precise :)
In short, they are both attracting me.
My concerns to the MSO5000 is:
- No real 2mV and 1mV vertical range. (I don't expect the 500uV range)
- If it's really too hot? I don't like it's ventilation design as well.
SDS2000+'s concerns:
- Hassle UI operation. It really make me remember the experience of Lecroy scope I've been used.
- Lack of zoom out feature. I'm accustomed to the Agilent style. I feel without zoom in feature is hassle to use.
@@Dan-oo1rj Good point: I use wireless mice on the two Keysight scopes I have on shelves above the bench: I have under-desk keyboard and mouse shelves. For the Keysight that I have at bench level (my comfy slippers scope!) I use the controls directly.
Thanks for the direct comparison, even 2 years later I was one of the people that still struggled to decide which one to get. Seeing both of them in action has helped me make up my mind. If I hadn't missed Rigols sale last month here in my region I might have gone with Rigol, but now I'm leaning ever so slightly towards siglent, especially since I have a decent offer for one.
The rigol can be unlocked up to 350MHz and it works up to 400MHz! The instructions are on google.
Didnt u noticed that rigol is in normal aqusition mode and Siglent isi in peak mode? In peak mode of course line will be very thin.I think u should try both in normal mode
A good spot, however it makes no visible difference on this test whether it's in peak or normal, on either scope. What about this test leads you to conclude that "in peak mode of course line will be very thin"?
In peak detect mode the scope is showing the maximum values found in that sample set. The name says it, de Peaks. So the opposite is true, we would see a thicker line.
I was thinking first about Sigilent, and now about Rigol and than i found your video :)
I have a rigol scope but i do have several other pieces of Siglent test gear. I will say this about Rigol their support is very good. I have not had to call Siglent so i cant speak to that but product support is another important measure
Rigol definitely has some of the best customer service I've had
Just recently bought the Rigol. It was just 400 Euro cheaper. 900 Euro int total, and the digital probes, all protocol, the signal generator was included. I accept that the modern ones are much better but they are also way more expensive.
It's not a bad scope: I'm still a knobs and buttons guy, so I like having the four separate channel controls. It's the best scope I have for remote access as the screen update rate is comparatively fast when accessed using RealVNC (rather than the scope's web interface, which is slow).
Some quirks with the LA...
Firstly, at high speeds, there's a 1ns offset between trigger point and display.
Secondly, the Search function doesn't work with LA channels at all. It's also a bit flaky on more complex serial searches.
Thirdly, note that the LA sampling rate (and memory depth) is 1/8th that of the LA channels: this is especially important for serial decodes on the LA channels, as the headline sampling rate shown on the scope's display is that of the analogue channels. You can see the LA sampling rate in the Acquisition -> More menu.
Fourthly, the decodes become unreliable on both Analogue and LA channels on busy traces. The sweet spot for maximum reliable decode for a single 100kbps I2C trace seems to be at 5MSa/s analogue sampling rate and 10Mpts memory depth, so that equates to LA sampling rate of 625kSa.s and LA memory depth of 1.25Mpts. That's about 2 seconds' worth. You can dramatically improve this on bursty data by using the Record function: for example, I achieved over 9,000 34 byte packets over a 4.5 minutes duration using the Record function.
@@nezbrun872 Thx.
With regards to apparent noise on the traces of the MSO5000, I suspect Rigol and Siglent have different "standard" settings for their "normal" operation. I've used the various Display Controls, outlined in Chapters 4 & 16 in the MSO5000 manual, to improve trace fidelity in my "normal' scope activities. Prior to making those adjustments I thought the Rigol and a bad ground, or something more sinister. TRACE VERY FAT, like me!
Now my trace is nice and thin with no annoying spurious noise being apparent. HOWEVER, go back to the "factory settings" and all the noise that's really there, is really there. BTW, I remember the days of hair-thin traces on the TEK 485 350Mhz portable scope. NICE! My ancient TEK 475 on my bench, not so much as it has the aluminum mesh dome "lens" to magnify the apparent sweep and help keep the CRT as short as possible for portable instruments. The FAT traces of the early 465 and 475 scopes were an embarrassment to Tek. My old TEK 475 is a rescued scope from the IBM scrap sale... Still meets specs! Love the 2mv vertical sensitivity! The HP's and Philips, of the day, didn't use the mesh lenses, but were still able to keep the CRTs fairly short. The lenses always spread the traces. The mesh lenses were improved after a couple of years into the design. Can't blame those faults on todays LCD displays, can we! I experienced this first-hand as I worked as a Field Sales Engineer/manager for both Tektronix and Philips for most of the 1970's... Yep, I'm OLD and gray but electrons still work the same now as they did then!
BTW Dave Jones, on TH-cam EEVBLOG, has a pair of nice videos on the topic of apparent / displayed noise on digital scopes. They are: EEVblog #601 - Why Digital Oscilloscopes Appear Noisy and EEVblog #442 - Analog Vs Digital Oscilloscope Noise. Worth a watch and good information why a more expensive digital scope might appear inferior to a less expensive digital scope or an analog scope. So, my question do any of these phenomenon apply to your comparison. Not having and Siglent SDS2000x on hand, I'll leave it to your capable hands to enlighten the rest of us. I TRULY enjoy learning something new every day!
Excerpts from the Rigol MSO5000 manual:
CHAPTER 4 - TO SET THE SAMPLE SYSTEM
Contents in this chapter:
◼ Timebase Mode (NOT INCLUDED THIS CUT AND PASTE)
◼ ACQUISITION MODE
◼ SAMPLING MODE
◼ SAMPLE RATE
◼ LA Sample Rate (NOT INCLUDED THIS CUT AND PASTE)
◼ Memory Depth (NOT INCLUDED THIS CUT AND PASTE)
◼ LA Memory Depth (NOT INCLUDED THIS CUT AND PASTE)
◼ Anti-Aliasing (NOT INCLUDED THIS CUT AND PASTE)
◼ Horizontal Expansion (NOT INCLUDED THIS CUT AND PASTE)
ACQUISTION MODE
The acquisition mode is used to control how to generate waveform points from the sample points.
MSO5000 supports the following four acquisition modes: Normal, Average, Peak, and High Resolution. By default, the acquisition mode is Normal.
Press → Acquisition on the front panel, and then rotate the multifunction knob to select the desired acquisition mode. Then, press down the knob to
select the mode. You can also press Acquisition continuously to select the mode or enable the touch screen to tap the desired mode and select it.
NORMAL
In this mode, the oscilloscope samples the signal at a specified fixed time interval to rebuild the waveform. For most of the waveforms, using this mode can produce the optimal display effects.
AVERAGE
In this mode, the oscilloscope averages the waveforms from multiple samples to reduce the random noise of the input signal and improve the vertical resolution. Greater number of averages can lower the noise and increase the vertical resolution; while at the same time, it will slow the response of the displayed waveform to the waveform changes.
When you select "Average" mode, press Averages and rotate the multifunction knob or use the numeric keypad to set the desired number of averages. You can also use the pop-up numeric keypad to input the average count. When the average count value is not in power-of-2 increments, a prompt message "Truncation average error" will be displayed. At this time, a value that is smaller than the one you input and the closest to power-of-2 increments will be input automatically. For example, if you input 9 with the numeric keypad, the average count will be input 8 automatically.
The number of averages can be set to 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384, 32768, or 65536. By default, it is 2.
PEAK
In this mode, the oscilloscope acquires the maximum and minimum values of the signal within the sample interval to get the envelope of the signal or the narrow pulse that might be lost. In this mode, signal aliasing can be prevented, but the noise displayed would be larger.
In this mode, the oscilloscope can display all the pulses whose pulse widths are at least the same as the sample period.
RIGOL HIGH RESOLUTION
This mode uses an over-sample technique to average the neighboring points of the sample waveform. This reduces the random noise on the input signal, generates a much smoother waveform on the screen and improves the vertical resolution. This is generally used when the sample rate of the digital converter is greater than the storage rate of the acquisition memory.
Note:
• The "Average" and "High Res" modes use different averaging methods. The former uses "Multi-sample Average" and the latter uses "Single-sample Average".
• In "HighRes"mode, the signal bandwidth does not exceed 1/32 of the sampling rate.
• In "High Res" mode, the highest waveform refresh rate mode is not supported.
SAMPLING MODE
This oscilloscope only supports the real-time sampling mode. In this mode, the oscilloscope produces the waveform display from samples collected during one trigger event. The highest real-time sample rate on the analog channel of MSO5000 series is 8 GSa/s. The current sample rate is displayed under CH SampleRate, and you can press More → CH SampleRate to view it.
TIP: Press RUN/STOP on the front panel to stop sampling, and then the oscilloscope will maintain its last captured graph. You can still expand or zoom the waveforms by using the horizontal and vertical control knobs.
SAMPLE RATE
Sampling is the process of converting the analog signal into the digital signal at a specified time interval and then restoring them in sequence. The sample rate is the reciprocal of the time interval. The sample rate of the analog channel is related to the current channel mode. The maximum real-time sample rate in the single-channel mode of the oscilloscope is 8 GSa/s. The maximum real-time sample rate in the half-channel mode is 4 GSa/s, and the maximum real-time sample rate in the all-channel mode is 2 GSa/s.
SEE ALSO:
CHAPTER 16 DISPLAY CONTROL
In the display control setting menu, you can set the parameters such as the waveform display type, persistence time, intensity, grid type, grid brightness, and etc.
Contents in this chapter:
◼ To Select the Display Type
◼ To Set the PERSISTENCE TIME
◼ To Set the WAVEFORM INTENSITY
◼ To Set the Screen Grid
◼ To Set the Grid Brightness
◼ Scale
◼ COLOR GRADE
◼ Waveform Freeze
END of Attachment...
Yes the fat display is the very first thing that jumped out at me.
I did several measurements comparing the MSO5000 with a number of other scopes, it has about a 10dB or so worse noise floor: it can barely scrape 5 bits of effective resolution in 8 bit mode. While you can improve things with averaging and oversampling ("High Resolution"), you can do that on other scopes too, but the front end is demonstrably very noisy. See post #1745 here www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg3180030/#msg3180030
The biggest problem for me with the Rigol was a usability issue, in particular the unfathomable delay when changing timebase.
THNKX U MASTER !!
After 11 additional months of use, the noise floor hasn’t been critical for me and I have alternative instruments. Still… where’s the noise source? EMR/poor shielding? Sub-parr devices? Careless layout?? I’m VERY hesitant to crack open the case on my still-warranted, although “hacked” MSO5074 and Dave did a pretty detailed tear-down early on. Anybody out there up to a full necropsy? I’m into a house remodel and don’t have the bandwidth to tackle this question. Anybody out there open to the challenge? A quick call to Rigol might produce a loaner for dissection .
@@Tool-Meister i am very interested. I have Just bought a mso5000 on amazon and i have 2 months to try It and return for free. Now i am in doubt if i should take It or give It back. All these people saying It Is very noisee are making me warried
Very good video showing everything working ... Thanks for sharing
Re. Siglent’s lack of video out: you can connect to the SDS2000X series using a computer or tablet (or smartphone for that matter) and point a browser to the scope’s IP and see a full-screen image of the scope’s display. Control the scope too.
Is this also possible for Rigol?
Hdmi would still be useful
@@71GAYes it is. However if you're not using the HDMI on the Rigol it also has VNC support that is extremely responsive - much more so than a web interface.
I have the Rigol MSO1104Z and I have lots of problems with it just not triggering correctly, and the logic inputs are flaky sometimes. The lag you mentioned drives me nuts too on the 1104, even though it isn't quite as bad as what you showed on the 5000. My new Siglent SDS2104X+ is on the way to me right now.
I've had an MSO-1104Z-S too for a while, it's my field scope as it's lightweight and compact. While it's a bit laggy, the 5000 is better, although that long delay after changing timebase isn't on the 1104Z.
Biggest long term problem on my 1104Z is the LA connections inside the unit on the IDE ribbon connector: I've had to re-crimp it once and may need to do it again.
Nice comparison of the two scopes. There´s one failure in your description though. The Siglent is 2x2 channel oscilloscope (grouped ch1+2 and ch3+4), which means that the sample rate is 1Gs/s at minimum for 3 and 4 channel use and 2Gs/s if 1 or 2 channels are used. Apart from that, we use the Siglent on a daily basis and its a mighty good device.
Yes, it's what I was alluding to at 36:00, but one watching it back I agree it's not clear, not least because I only refer to channels 1 & 2. There are serious antialiasing issues because of it, particularly if it's had its bandwidth upgraded to 500MHz. Nyquist would not be pleased!
@@nezbrun872 Yes, I agree Nyquist would probabely not be amused. ;-)
Still though it makes a noteable difference if the samplerate drops to 1/4 because only one ADC is employed or if it drops only to 1/2 because of 2 ADCs.
It should also be kept in mind, that a 4x set of probes capable of such high bandwidth typically costs as much as the whole scope.
The 500MHz upgrade is rather more for marketing reasons than of actual useability here.
But hey, I´d rather receive firmware upgrades even if they´re a bit over the top than no updates ;-)
Especially if those updates not only fix existing bugs but also add on functionalities .... as it is the case with Siglent´s updates.
Many thanks for this review, I have been agonising over which of these two scopes to buy. Like you, I have tended to favour Rigol in the past but was starting to move towards Siglent after watching several EEBlog videos. Now, having watched your excellent comparison video, I will definitely be going the Siglent route. I understand that Rigol have designed their own asic front end chip, this might be the reason for the poor noise performance. One possible downside of the Siglent is that it has only one sig. gen. output channel and no modulation modes, whereas the Rigol has two channels with am/fm/fsk modulation.
Pretty much sums it up. That and the performance thing where the Rigol stops triggering for about a second when you change the timebase, that makes if feel more sloggish than it might otherwise be.
Considering I'm generally a Rigol fanboy, there's still no doubt which of the two to go for, and the Siglent is undoubtedly a better option overall, something that surprised me.
Only tempting feature of Rigol is its 8GS/s. Otherwise Siglent looks much more a mature product. The scopes one is realisticly / willing to buy for personal non business use slide down regarding the maturity. Siglent seems the most mature of them all. And way down to FNRSI kind of scopes. In my opinion Rigol can take a seat behind the Siglent. Rigol had the front seat before, but has lost it.
All these small annoyances with low and medium scopes drived me cracy, had rigol DS1074Z, then Rigol MSO4054 and always there where tradoffs and limitations bugs and not thought tru solutions when using them fully. decoding was a hit and run, small screen, fideling back and forth with zoom, always woundering if it was the DOT or the scope that played tricks..So when a InfiniiVision MSOX6004T (Incredebly cheap 6000€! fully loaded! A leasing company sold it, had no idea what it was (They had two, bought them both, soldone and had the other for free!)..they mostly sold cars, Trailers, tracktors machines etc I was the only bidder! Since i was in the proces of ordering a brand new one from Keysight) and was earching, it came in view ! Now only look back when having naightmares after a long stint in the lab. Life before and after...All those seconds with becouse of sluggish updates when using the Rigols that gave me gray hair and stomack problems. I recomend the msox6004a!
you can set the horizontal expansion point on the siglent to be anywhere.
Indeed, see the corrections and clarifications in the notes above.
What do you think of the new siglent 2000 HD? Would be awesome if you do a review on it. Or the new 12bit Rigol dho1204
I second this comment.
When are scopes going to move beyond 1990 display tech with OLED 300ppl displays? With these scopes I can't tell the difference between a dac output and the original test signal. Both look stepped instead of pure sine wave.
You're probably using the wrong tool for the job at hand. FWIW CRTs were still very much king into the early 2000s for higher resolutions.
@@nezbrun872 Up to now, I've been using a CRT based scope. But the new scopes all have these low resolution VGA displays, which is frustrating, as I shop for a replacement for my 2465B that U800 failed on.
amen
@basspig RIP Tek 2465. Best scope for nigh on 3 decades
To be fair, the Rigol can decode flex-ray, i2s and mil-std-1553 serial data. This might be an important difference to some people. Rigols last firmware update was in April 2020. I'm hoping that is only because of coronavirus and that future updates will come at some point. I'm looking to replace my old DS 1102C which has served me well having never been updated. Alas, the lcd is starting to exhibit solid color vertical lines that come and go. I've tried reseating the cable, but I fear that it's the actual screen that is failing internally. I'm afraid to pull it completely apart until I have a new scope. If I can fix it, it is destined for portable and possibly automotive service. It's been a great scope, even though it cost a small fortune (>$1000) new. Right now, the MSO 5074 with the free bundle is looking pretty good to me for $1000. I do like that the new Siglent's support wifi, but I'm hesitant to switch from Rigol because I've been so happy with my last purchase from them.
> To be fair, the Rigol can decode flex-ray, i2s and mil-std-1553 serial data
To be fair, so can the Siglent.😀.
If you call Rigol and talk to a human they’ll often give you a 10% discount coupon number.
@@Tool-Meister I called, they only gave me a 5% off code. :(
Siglent can decode: i2c, SPI, UART, CAN, LIN, FlexRay, CAN FD, i2S, 1553B, SENT, Manchester.
Thanks for the great review! It looks like you have 3 scopes on you bench, Rigol, Siglent and Keysight. Which is your everyday driver? Also any comments on the quietest fan noise of the three?
In this lab, there are five scopes set up permanently on the bench: MSO8104A at the soldering station end, with MSO7104B, MSOX3104T, Tek MDO4104C and an old HP 54121T 20GHz sampling scope on the main bench.
My favourite day-to-day scope to use is the MSO7104B at the bottom eight of the video: it has a fast boot time, big screen and is a joy to use with its super snappy & simple, intuitive UI: it's the first one that gets switched on.
The 3104T is also nice scope and its UI is very similar to the 7104B. I has more decode and analysis options than the 7104B, but its screen real estate is significantly lower in terms of both resolution and area. Being ably to monitor two serial buses simultaneously can be quite handy. The touch screen I don't use much due to its position on the bench, but I do use a wireless mouse with it in the under-desk keyboard drawer.
The MDO4104C is seeing more use these days, but the UI is a right royal bag of spanners. It's inconsistent, and unlike the Infiniivisions (7104B and 3104T) its UI becomes unresponsive when you start adding the digital channels, bus decodes and longer acquisitions. It took me many months to get comfortable with it. I find I'm forever having to adjust the memory depth, something that's simply unnecessary on the Infiniivsions. However the Tek does have some useful additional features over the Infiniivisions that I have, such as USB HS decode, up to three simultaneous buses, 20Mpts memory depth on all channels,
Hi, I'm thinking about picking up that Rigol... and was considering the Siglent also, so I was enjoying your video. You didn't mention the "App" button (lower left on the Rigol). That looked pretty cool to me on some other videos I watched. What are your thoughts on that? I'm also curious as to what the stop, left & right buttons do on the Rigol (in the center of all the controls there).
I don't use these kinds of functions in my day to day work very often, so they rarely hit my radar. For a bode plot, I'm far more likely to use a PC based solution, specifically the Analog Discovery. Bode plots are quite slow on both these scopes.
perfect, thanks, I was looking at these 2, I hate everything about rigols UI on my ds1054
Regarding channel to channel skew: from the MSO5000 manual: Channel Delay
When using an oscilloscope for actual measurement, the transmission delay of the probe cable may bring relatively greater errors (zero offset). MSO5000 allows you to set a delay time for calibrating the zero offset of the corresponding channel. Zero offset is defined as the offset of the crossing point of the waveform and trigger level line relative to the trigger position, as shown in Figure 2-2.
Missing Graphic which doesn’t copy.
Figure 2-2 Zero Offset
Press 1 → More → Ch-Ch Skew, rotate the multifunction knob or use the numeric keypad to set the desired delay calibration time. The available range of the delay calibration time is from -100 ns to 100 ns.
Note: This parameter is related to the instrument model and the current horizontal time base setting. The larger the horizontal time base is, the larger the setting step will be.
I think my point was that channel skew isn't available for digital channels. Typically this wouldn't pose a particular problem, except that there is a constant 1ns delay between a digital edge trigger and what's displayed. So if you do an edge trigger on Channel D0, the edge is always displayed a constant 1ns after the trigger. Timestamp 34:00.
Nezbrun My error. First of many each day. Ready, Fire, Aim... Did you try the
THRESHOLD ADJUSTMENT and scope auto-calibration? The manual seems imply that it may affect the skew, i.e. “To Set the Threshold and Calibrate Probe”. I’m in the middle of a house sale and associated move so am at a disadvantage regarding getting at my equipment, otherwise I would try to duplicate the problem for you. I received my MSO5074 in July and promptly performed the magic feature and bandwidth spells. I THINK the noise-floor had a noticeable increase but I wasn’t bright enough to make any “before” measurements other than rise-time, which of course decreased. Not to be repetitive, congratulations on a very thorough review. You seem to take the peer review and questions all in stride. Kudos to you!
@@Tool-Meister All done, the 1ns delay seems to be a software feature, on any digital channel used as an edge trigger.
@@Tool-Meister Regrettably, none of these work. It's the same on any digital channel, there's a constant 1ns delay between the trigger point and the displayed edge.
One last thought. In pure analog scopes, you can’t see the leading pulse edge unless the scope has a “delay-line”. Thankfully, virtually all modern, say post 1975, have a delay-line which allows seeing a pulse’s leading edge. However, it takes special features and techniques to know your exact trigger point.
That said, with digital scopes you can know and see the trigger point. As to the 1ns offset, it’s at least known and can be allowed for. I wonder what the tradeoffs are for Rigol’s design approach?
WHEN I can get at my OWON scope and my pulse generator, I’ll check to see if it has an offset. To clarify, your observation is there a 1ns offset from the triggering channel’s displayed position compared to all non-triggering channels displayed positions? You are NOT observing that phenomenon on other digital scopes, specifically the Siglent in the comparison.
I’m going to reach out to Rigol. I’m only 8 miles from them and can “ mask up” and visit them in person if need be. BTW, where did you terminate the signals? Did you swap channels to verify the signal source’s 2 channels have identical propagation delay and no reflections? I suppect you did but I’m imagining what questions Rigol will ask when we speak.
Again, GREAT WORK on your part!
Oooh, have to press a button to select the ch! Nasty! They should be shot for that. (yoking .. really).
It also seems that you're a little confused about what the word "DEFAULT" means. No problem.
Okay, so this seems to be a bunch of nitpicks with the Siglent edging out the Rigol. Good thing he Rigol is a bit cheaper. (but not as I write this) For me the win goes to the Siglent. I'll agree, neither one is quite as good as a $18,000 Keysight, R&S or Tek, but if I had $18K burning a hole in my pocket I have higher priority outlets for that kind of excess. I'm a hobbyist so either of these scopes are bordering on overkill for what I need. The Siglent (especially considering the present pricing) is the clear winner. One is on the way as I type.
that noise in the rigol might as well be the fuckton of waveforms/s overlapping
Okey - which one to choose for a home lab?
Hello. Have you already screwed the screw on the shelf into place? 😅
Nah, she'll be right ;-)
Is the MSO5000 front-end really more noisy? I checked it against a much higher-end 1GHz scope and also a 350MHz analog scope and the noise was exactly the same on all of them when using the freq gen from the MSO. I'll have to check against a better function generator but I don't think it will be much different. ??
I did comparative measurements against a number of other scopes: Keysight MSO8014A, Keysight MS)7104B, Keysight MSOX3104T, Tek MDO4104C, Siglent SDS2104X+, and Rigol MSO1074Z. Effective bits on the Rigol MSO5000 was about 5 bits (varies somewhat depending on input attenuator), while the other scopes were around or just below 7 bits, so you're looking at about 10 to 12dB or so worse noise floor on the Rigol MSO5000X.
It's covered here:
www.eevblog.com/forum/testgear/benefits-of-going-with-all-siglent-setup/msg3281580/#msg3281580
@@nezbrun872 Thanks, I have some more testing to do.
@@nezbrun872 For the MSO5000 I am getting lower noise numbers than your eevblog post. Have your tried these measurements with the new firmware updates? Using your set parameters, I am measuring the std-dev noise (AC-RMS) as follows: 0.167mV at 2mV/, 1.9mV at 100mV/, 3.8mV at 200mV/, 8.6mV at 500mV/, and 67mV at 5V/. (I am using the available AC-RMS measurement instead of V-RMS, to measure the noise around the mean, thus avoiding any non-zero bias.) Also, I would generally expect to sample/measure and see wider noise peaks on this scope, because of its higher sampling and wfms/s capture rates.
Have you tested the new Rigol HDO4000?
Thanks for the detailed review. Does the high resolution mode on Rigol helps to reduce some of that noise by averaging the additional sample points available?
It does to some degree, but not as much as one would hope.
Iam new to all this but I think on my 1204 x-e siglent oscilloscope. I think incan independently set the x10 as a default on cearten functions. So when I press default it will default to what I set it to on some settings
Could the sluggish rigol Behavior be adressed by firmware updates in future , or is that hardware related?
I can't be 100% certain, but I'd have hoped it could be resolved with a firmware update. It is possible that there may be a hardware related PLL settling time limitation. Rigol had some issues around PLL instability in the DS1000Z series that was corrected with a firmware upgrade. I should note that at the time of writing this, it's been about 18 months since Rigol issued a firmware update to the MSO5000 range.
See my new pinned comment for more updated info.
4:42 I dont thinks so "disappointment" you can change the thickness with measure>averega> such as 128.
Yes, if you have a sufficiently repetitive signal and you can put up with the reduced update rate. Using averaging isn't a universal panacea, not by a long way.
I have a question, on the Rigol (MSO5104) it says 8GSa/s but the bandwidth is limited to 100MHz so I don't know what the point is. On the Siglent (SDS1204X-E) it is 1GSa/s but the bandwidth goes to 200MHz. Which one is better? My thinking is a higher bandwidth is better and with the Rigol all those extra samples just don't get used because of the limited bandwidth.
It depends: the MSO5000 drops to 2GSa/s as soon as you have more than one channel running, and is designed to support up to 350MHz with software licenses (see eevblog regarding how liberate your scope's options). A high sampling rate makes the anti-alias filtering a lot easier.
On the 1204X-E, it drops from 1GSa/s to 500MSa/s depending on how many channels you have on, which doesn't leave much headroom for antialiasing. I'll also add that the MSO option for the 1204X-E is a terrible implementation, the scope becomes almost unusable with it.
@@nezbrun872 do you mean because you can't use history or the rolling display or why? what happens with MSO that is different?
@@nezbrun872 You can get 4GSa/s on two channels if you use channels 1 and 3.
Dude, connect USB mice to Siglent - and you don't have to touch the screen! :)
Thanks for the video and update. Though I still can't make my mind going for which one. I fully agree with your unbiased opinion.
Actually I ordered a MSO5074 and returned it when I have seen the video of "fake" vertical range below 5mV. I can accept it's ADC noise, but I can't accept the digital zoomed vertical range. It dose limit it's ability to measure low level signal like ripple noise.
Below are some of my personal points of view:
MSO500's strong points:
- tangible knob and keys for vertical controls and soft menus
MSO500's weak points:
- the vertical ranges below 5mV/div are digital zoomed
SDS2000X+'s strong points:
- more detail in specifications, I feel Rigo's are a bit vague.
- fruitful power measurement functions
SDS2000X+'s weak points:
- the UI seems hassle, not intuitive.
- can't acquire waveforms outside the screen although it has got a lot of memory. it should be able to zoom out or shift waveforms which outside the screen. I can't say if the segmented acquisition make sense. But having zoom out ability make things easier.
Something I'm not sure:
- If MSO5000's full-memory hardware measurement is for all the measurements? I can only find the frequency counter is in hardware.
- If MSO5000's full (1M points) memory FFT has high update rate? SDS2000X+'s seems a bit laggy. But it's still much better than those vintage digital oscilloscopes other than HP/Agilents'.
Both model don't search events of digital channels. That's a pity.
Yeah rigol is more of a "digital" oriented scope, higher sampling rate higher noise. Price is about 30% more for siglent, worth it if you have the budget. Skew adjustment is in SCPI but doesn't work, reported as a bug, not fixed. But for me 99% of the time logic analysis on the PC is better than a scope.
Whether it's 30% more or not depends I think. You get 350MHz probes on the 70MHz model for example, whereas they're 200MHz probes on the SDS2000X base 100MHz model. The basic serial analysis is thrown in with the Siglent, a considerable additional expense on the Rigol unless you benefit from a marketing push such as the current bundle offer in some markets. Sadly I've seen absolutely no benefit on the Rigol over the Siglent on its 8GSa/s vs 2GSa/s sampling rate. The Siglent will go up to 500MHz analogue bandwidth, but beware of the aliasing issues when using the interleaved channel pairs as mentioned in the video, when the sampling rate drops to 1GSa/s.
Possibly where the Rigol benefits in the digital domain is the search facility (only available on analogue channels weirdly, with no serial search at all on the Siglent) and the 1GSa/s LA (500MSa/s on the Siglent's LA).
Whether you think logic analysis is better on a PC than a scope is an interesting topic, and I hear it a lot that you're better off doing it on a PC. I think it depends. It's a different work flow, and it's often a personal preference. On a scope it's all about triggers and real time analysis, whereas on the PC it's mostly about bulk capture and post processing. Such PC based LAs tend to have only the most basic triggering facilities. Personally speaking, It's pretty rare I ever do any serial analysis on the PC, it's only when I have an exceptionally long capture to analyse. Budget PC based LAs are hampered by low sample rates so are often useless for many SPI or I2S applications for example.
In short I won't deny I was disappointed by the MSO5000, I knew I had to try out the Siglent bearing in mind others' comments on eevblog.
Do you use logic analysers for time logic analysis?
@@71GA Yes, dslogic pro is great.
I like the update speed on the siglent better than the rigol and the display is rock solid - the rigol is fluttering about. I'm wondering how much better (if any) a Siglent SDS5034 would be vs SDS2354X-plus - in terms of features and experience... I think the separate vertical controls could be handy but isn't a dealbreaker, but also 5GSa/s (2.5) on the SDS5000 series vs 2GSa/s (1) on 2000. Or whether its just worth getting the base 2104X+ and 'upgrading' :)
Be aware that there are significant aliasing effects when running at 1GSa/s on the SDS2000X+ (i.e., ch 1+2 or 3+4 on) when upgraded to 500MHz bandwidth. I can't comment on the SDS5000, except that yes, I would prefer separate channel controls, but as you say not a deal breaker. There's a zoom function on the Siglent that's far better implemented there than on the Rigol that I didn't show, you can zoom in vertically as well as horizontally, and it shows the 2D area of the captured waveform you're looking at, rather than the ubiquitous 1D zoom overview.
You have to see it in the flesh to witness it, but the display on the Siglent is so crisp compared to the Rigol, so much so that initially I was wondering if the timebase was actually running.
@nezbrun872 Hi. Thanks for the review, it really helps a great deal to make a choice on these (for personal use not cheap pieces of equipment). Remark: anyone using equipment has to know about its limitations, as even the most expensive equipment has. Question: as the Siglent does not have the sufficient sample rate to decently cover the highest possible BW of 500MHz with its 2GS/s, does it maybe have any old skool tricks like equivalent or random time sampling? That trick could work for periodic signals. And as mentioned before working on those frequencies would require matching probes or a good coax.
Hi there just watched your video on the yi 4k+ cable, i wonder if you can help me, could I make up a cable just for an external mic connection as I cannot purchase that cable anywhere, Thank you.
I am being petty here, but thanks for pronouncing "Rigol" correctly.
Wriggle. ;-)
@@nezbrun872 I wince every time I hear how Dave Jones pronounces the name ("rygol") The original intention was for the name to sound like "Regal" as in royalty.
@@chrisbenton3365 I guess you haven't seen my video on Huawei then ;-)
@@nezbrun872 No I haven't, but it doesn't really matter as I am in the market for a mid-range scope right now and not a phone, but your attention to detail instills confidence that you know what you are talking about. Like I said in my first comment "I am being petty".
Rigol Is great machine...
The SDS2104X Plus can be upgraded to the SDS2354X Plus?
It appears so. Search here:: www.eevblog.com/forum/testgear
Dead URL link..
@@Tool-Meister GOOGLE IT
www.eevblog.com/forum/testgear/
Rigol have better support and after firmware update works better. Also it have many free bonuses and options. Sample rate 5GS instead Siglent 2GS.
The sample rate is only 8GSa/s on one channel. Using channels 1 and 3, you get 4GSa/s, but you still only get 2GSa/s if you use three or four channels, or two adjacent channels.
What do you actually get with the higher sample rate? The reason I ask is because my scope is twice the bandwidth (200MHz) but has a much lower sample rate. All I see is that with the Siglent I can look at a 200MHz signal and with the Rigol I cannot, so what do you get from the Rigol with the higher sample rate?
@@javacup7579 good refresh rate. almost like live view. i tried Siglent and it have bad screen visual signal live reproduction. also i can’t see distortion in fronts.
@@oleggritsev my 200MHz scope also has "good refresh rate. almost like live view" so i don't know what you get with the higher sample rate other than a lower bandwidth, still wondering what that is.
@@javacup7579 Rigol it is like Apple. Siglent it like Android. Siglent it made for poor people and quality of material is worst! I gave Siglent, Owon. Hantek. Rigol like tectonics
thanx.... rigol useless for us
if you can't comment on Tank's msg this review cannot be trusted
I already made a comment four months ago:
'A good spot, however it makes no visible difference on this test whether it's in peak or normal, on either scope. What about this test leads you to conclude that "in peak mode of course line will be very thin"?'
aimless waffling. Maybe try organising a structured review next time.
I had a sheet of paper to the right with all of the points of discussion. I'm sorry you're only able to concentrate for short periods. There were around two dozen points I raised comparing the two, how do you think I could reasonably demonstrate those in a three or even fifteen minute video?
May I suggest that you try reading the description and the pinned post, assuming your capable of comprehending a thousand words.
All GUI writers should be locked into a test room for 1,000 days and made to use the scope all that time in order to make these GUI MORRONS write the GUI properly, INCLUDING responsiveness!
Maybe the terrible ADC is because the Rigol has a terrible ADC? :) ...a compromise for the 8GS/s ADC!