DSO Recomendation

[Gustavo-ar]

Hello! I am new here I am from Argentina so my English is not very good (thanks google translator!), I have been reading a lot in this forum about DSO.

I am looking forward to buying an oscilloscope, in my work I would like to measure noise in PSU, PWM signal etc. and as a hobby I am doing something in Arduino and AVR.

I was watching this (the prices are in US dollars, too expensive because of my country taxes)

Siglent SDS1072CML+      70 MHz     $500
Siglent SDS1072CFL          70 MHz    $600
Rigol DS1054Z       50 MHz    $700

The first siglent has 2Mb of memory, the second only 24k but has 2Gs / sec
I do not know which of the two characteristics may be more important for what I'm going to do.

I think the $500 siglent is ok with good memory and 7 inch screen, i see a review showing the real  bandwith is over 100Mhz, but I was surprised to see a 2Gs/sec DSO for only $100 more

https://www.eevblog.com/forum/testgear/siglent-scopes/msg222002/#msg222002

The third and most complete option is the D1054Z (except for the 2Gs / sec) but I do not know if it's worth the extra $200.

Thanks in advance

[Daruosha]

I understand the prices are quite unreasonable. But the Rigol is more than twice a DSO than DS1072CML+. Save your money and buy something to keep it for a few years. I admit the DS1054 has its own problems, but 4 channel, serial decoders (even on the displayed trace), intensity graded display and so many other options makes Rigol unbeatable in its class.

[Fungus]

as a hobby I am doing something in Arduino and AVR.

Four channels is really valuable for that. It's a big difference.

With the Rigol you also get serial decoders, etc. Also valuable for that sort of work.

[MrW0lf]

that sort of work.

Just word of warning - for analog sort of work where amplitudes and accuracy start to matter this scope is unsuitable. Usable only for general detection of mainly digital electrical phenomenon. Consider this as <10$ multimeter of scopes accuracy wise. I developed deep distrust in mine and it did bite the bullet so to speak... It is tragic you would have to pay so much for one. Maybe source something used and good quality brand locally?

[Fungus]

Just word of warning - for analog sort of work where amplitudes and accuracy start to matter this scope is unsuitable. Usable only for general detection of mainly digital electrical phenomenon. Consider this as <10$ multimeter of scopes accuracy wise.

Oscilloscopes aren't multimeters. The written specification for a DS1054Z's voltage measurements are about 5% accuracy at best. It's an 8-bit DAC, it's never going to measure like a Fluke multimeter, it's silly to expect it to. Nobody but you seems to have a problem with the DS1054Z.

I developed deep distrust in mine and it did bite the bullet so to speak... It is tragic you would have to pay so much for one. Maybe source something used and good quality brand locally?

Maybe you could sell yours to him. Kill two birds with one stone. Win-win.

[MrW0lf]

Oscilloscopes aren't multimeters. The written specification for a DS1054Z's voltage measurements are about 5% accuracy at best. It's an 8-bit DAC, it's never going to measure like a Fluke multimeter, it's silly to expect it to. Nobody but you seems to have a problem with the DS1054Z.

My scopes are comparable to ~4digit DMM in most horizonal scenarios and ~3digit in vertical. I'm used to that and like the fact very much. 8bit is pretty ok if you apply some averaging and scope has on-the-fly DC zeroing capability. Also Sinc ON|OFF should not change anything.

But heres what Rigol does:
3.5Vpp, 80MHz pure sine, 500mV/div, 250MSa/s:
Sinc=ON: 3.7V => 5.4% error
Sinc=OFF: 2.58V => 35.7% error
So total error between two modes about 40% 

Naturally "the thing" is long sold... Expecting arrival of vastly superior tech soon

Why is Rigol "performance" ok with many people? First it look like Tek, then it's often first scope and noone knows any better. Also in simple mostly digital domain work analog performance is not that important I guess? Any line ok as long as it wiggles a bit At the partys also some girl better than no girl

But there are scopes in similar price with vastly better analog performance. Sometimes it makes sense to skip some "features" or "bandwidth" and take a bit more "limited" but excellently implemented device programmed by dudes that visited some decent university not only as tourist 

But psst... you could get banned on this forum if say too loud what Rigol is Maybe soon the case with Keysight also? I could say a thing or two about 3000 X but skip this time because they just sent me free U1282A and it would no be very polite.

[rstofer]

Search this forum for DS1054Z and ignore EVERYTHING prior to November 2016 when the latest firmware was released.  All discussions about this bug or that bug are invalid as of that date EXCEPT for the spelling error of Pluses instead of Pulses.

While searching, pay attention to the trivial process of unlocking 100 MHz, extended memory and decoding options on the DS1054Z.  Hint:  Search for 'riglol' (note the misspelling).

You won't find a scope anywhere near as capable for even twice the price.  But, yes, there are better scopes around.  They just cost a LOT more money.

The DS1054Z is my first DSO so I don't have any frame of reference to compare against other DSOs.  But I have had oscilloscopes for about 60 years and while I concede that my Tektronix 485 has more bandwidth and an interesting set of features, in no way can it do the things I can do with the DSO.  Single shot mode alone is worth the cost of entry.

It's regrettable that your cost is so much higher than here in the US.

[guenthert]

Ok, I bite.

Oscilloscopes aren't multimeters. The written specification for a DS1054Z's voltage measurements are about 5% accuracy at best. It's an 8-bit DAC, it's never going to measure like a Fluke multimeter, it's silly to expect it to. Nobody but you seems to have a problem with the DS1054Z.

My scopes are comparable to ~4digit DMM in most horizonal scenarios

Care to elaborate how DMMs measure time?

and ~3digit in vertical. I'm used to that and like the fact very much. 8bit is pretty ok if you apply some averaging and scope has on-the-fly DC zeroing capability.

Take it or leave it, 8bit is very much industry standard for many years and probably will be for a while longer.  There are exceptions (I believe some picoscopes have fairly fast wider ADCs and the (comparatively cheap) Digilent Analog Discovery has a 14b ADC (albeit only 100MS/s).  Many DSOs (including the 1054z) offer a 'high resoultion' mode, in which they increase the apparent resolution to 14 (or more) bits by reduction.

This isn't generally much a problem as long one understands the limitation and purpose of an oscilloscope.
A 3% error margin is what analogue 'professional' CROs offered too.

Also Sinc ON|OFF should not change anything.

I give you that, it shouldn't.  File a bug report with Rigol.

But heres what Rigol does:
3.5Vpp, 80MHz pure sine, 500mV/div, 250MSa/s:
Sinc=ON: 3.7V => 5.4% error
Sinc=OFF: 2.58V => 35.7% error
So total error between two modes about 40% 

You're measuring a 80MHz signal with a sample rate of 250MS/s?   What on earth do you expect?
And tell me, which DMM measures a 80MHz signal accurately?  My 6.5 digit HP3456a is good only up to about 100kHz, my not quite as ancient HP34401a makes it to 300kHz.  For frequencies in the short wave range you'll need specialized gear.

Naturally "the thing" is long sold... Expecting arrival of vastly superior tech soon

Why is Rigol "performance" ok with many people? First it look like Tek, then it's often first scope and noone knows any better. Also in simple mostly digital domain work analog performance is not that important I guess? Any line ok as long as it wiggles a bit At the partys also some girl better than no girl

But there are scopes in similar price with vastly better analog performance. Sometimes it makes sense to skip some "features" or "bandwidth" and take a bit more "limited" but excellently implemented device programmed by dudes that visited some decent university not only as tourist 

But psst... you could get banned on this forum if say too loud what Rigol is Maybe soon the case with Keysight also? I could say a thing or two about 3000 X but skip this time because they just sent me free U1282A and it would no be very polite.

Nonsense, you wont get banned, just laughed out of the house.

[bsas]

Hi all,

I am also on the market to buy my first DSO! (very excited about it). I want to restart my electronics hobby

Currently I am playing a lot with analog audio circuits (like guitar pedals for example), so I am very likely yo use a lot of FFT with 100Hz to 25kHz range, but I want to be able to play with digital electronics in the future too.
I am very strongly towards the Rigol DS1054Z, and the price looks just too good to be true.

I am just wondering if this is the best scope you can get for around $500 or less. I am OK to go a little over if it has some killer(s) feature(s).

Thanks all for the recommendations!

[Gustavo-ar]

Mostly what I want to measue are SMPS from chargers, desktop motherboards, laptops motherboards, VGA etc.

Most of the times the swollen capacitors tell me that there is a problem, but other times at first sight they seem to be ok.
That's why I want to measure the noise at the output, to see if it is within normal values.

I could throw the old hardware and put a new one, but in many cases being old models I have to change mother, cpu and ram.
Also often the noise causes very sporadic crashes that cause me to waste a lot of time

Of course I plan to give it more uses, but that's something I could use "out of the box."
I also have some projects as a hobby: a modified atx psu transformed into a variable voltage from 1V to 24V approx., Can also be used as a constant current psu.
A component tester with an Atmega328

[MrW0lf]

Care to elaborate how DMMs measure time?

My main DMM workhorse - Agilent U1272A has frequency, pulse, duty up to 100kHz. Fairly ok to work with mechatronics. Here is test where I compared Rigol & Pico & Agilent DMM at measuring 666.666666Hz square with 66% duty:
https://www.eevblog.com/forum/testgear/testing-dso-auto-measurements-accuracy-across-timebases/msg1096096/#msg1096096

A 3% error margin is what analogue 'professional' CROs offered too.

If get to know instrument, it can usually do much better than spec if does not mess with raw data. Most decent scopes try hard not to.

I give you that, it shouldn't.  File a bug report with Rigol.

I have gotten zero response to bugs filed directly to Rigol (both US & EU). Interesting if at least something will be fixed in next FW...

You're measuring a 80MHz signal with a sample rate of 250MS/s?   What on earth do you expect? And tell me, which DMM measures a 80MHz signal accurately?  My 6.5 digit HP3456a is good only up to about 100kHz, my not quite as ancient HP34401a makes it to 300kHz.  For frequencies in the short wave range you'll need specialized gear.

Sampling rate is irrelevant if averaging long enough (or using ETS), as long as above Nyquist. I did comparable scaled test on normal scope. No problems. Data points pile up and you get true reflection of analog frontend performance. Decent frontent has predictable performance. So no, I do not need some "specialized gear", just need decent scope and some basic dB calculus.

Nonsense, you wont get banned, just laughed out of the house.

I actually did get banned in the middle of quite civilized discussion 
https://www.eevblog.com/forum/testgear/rigol-ds-1054z-or-tektronix-2440/
So... psst

My recommendation is GW Instek if knobs are a must, or Pico otherwise. With GWI you lose decoding and "on paper" bandwith. With Pico only bandwith (load of decoders included). But actually they seem to go much higher bandwith than spec + if you buy Pico MSO then can "hack" up to 16bit accuracy additional analog channel using digital inputs + math formula + Arduino etc.

Edit: Look DC gain accuracy of properly engineered scope:
https://www.eevblog.com/forum/testgear/lecroy-waverunner-lt-224-review-and-(partial)-teardown/msg205402/#msg205402
In fact, this is interesting subject - so I ordered some voltage references and plan to compare all DMMs & scopes I have at hand, thats >10 units so should be interesting.

[rstofer]

I don't know much about scope accuracy, it simply isn't important +- a few percent.

But I have a calibrated DMMCheck device so I thought I would check my DS1054Z:

DC Volts    Calibrated = 5.0000V        Measured On Scope = 4.99Vavg
AC Volts    Calibrated = 4.9989V        Measured On Scope = 5.00V at 100 Hz
Frequency   Calibrated = 100.004 Hz,    Measured On Scope = 100.005 Hz
Frequency   Calibrated = 10.0004 kHz,   Measured On Scope = 10.0005 kHz

I suppose I could use my DP832 power supply to generate other voltages, verify them against my most accurate, but not currently calibrated, Fluke 189 and see how the scope performs over a range but I won't bother.  If I had a signal generator that went high enough, I suppose I could try the frequency response of the scope but I don't have the signal generator and I certainly don't have a DMM that would measure AC accurately (like to 4 decimal digits).

All I need is wiggly lines plus or minus a few percent.  If I need bandwidth, I'll use my Tek 485 and have no measurements of any kind.

[MrFox]

Hi all,

I am also on the market to buy my first DSO! (very excited about it). I want to restart my electronics hobby

Currently I am playing a lot with analog audio circuits (like guitar pedals for example), so I am very likely yo use a lot of FFT with 100Hz to 25kHz range, but I want to be able to play with digital electronics in the future too.
I am very strongly towards the Rigol DS1054Z, and the price looks just too good to be true.

I am just wondering if this is the best scope you can get for around $500 or less. I am OK to go a little over if it has some killer(s) feature(s).

Thanks all for the recommendations!

If FFT is going to be among your most useful tool, make sure you ask DS1054Z owners opinions about this specific feature. I love my DS2072 on most aspects, and the DS1054Z is newer so maybe they improved on this, but IMO Rigol's FFT implementation is just plainly horrible. (to express myself mildly).

[Fungus]

Currently I am playing a lot with analog audio circuits (like guitar pedals for example), so I am very likely yo use a lot of FFT with 100Hz to 25kHz range, but I want to be able to play with digital electronics in the future too.
I am very strongly towards the Rigol DS1054Z, and the price looks just too good to be true.

The Rigol's FFT is a weak point.

There's a 'scope by GW-Instek that actually costs a little bit less than the Rigol and has a better FFT - the GWS 1054B

http://www.gwinstek.com/en-global/root/Oscilloscopes/Digital_Storage_Oscilloscopes/GDS-1000B

The (unlocked) Rigol is better for just about everything else. Take your pick!

[rstofer]

Care to elaborate how DMMs measure time?

Fluke 189 will measure frequency up to 1 MHz, Duty Cycle between 10 and 90% and pulse width to 1 Sec with a resultion of 0.1 mS or to 0.5 Sec with a resolution of 0.01 mS.  Not stellar performance as a counter but adequate in many cases.

http://assets.fluke.com/manuals/187_189_umeng0200.pdf

[rstofer]

Currently I am playing a lot with analog audio circuits (like guitar pedals for example), so I am very likely yo use a lot of FFT with 100Hz to 25kHz range, but I want to be able to play with digital electronics in the future too.
I am very strongly towards the Rigol DS1054Z, and the price looks just too good to be true.

The Rigol's FFT is a weak point.

There's a 'scope by GW-Instek that actually costs a little bit less than the Rigol and has a better FFT - the GWS 1054B

http://www.gwinstek.com/en-global/root/Oscilloscopes/Digital_Storage_Oscilloscopes/GDS-1000B

The (unlocked) Rigol is better for just about everything else. Take your pick!

Ahmen on the Rigol FFT - it's much improved but still not stellar.  The good news, for me, is that I have done Fourier Analysis exactly ONCE since I graduated in '73.  It just isn't important - to me...

If I were just starting out and my frequencies (and voltages) were moderate, I would skip all the scope stuff and buy the Digilent Analog Discovery.  It has a bunch more capability at a far lower cost.  Two channel scope, two channel arbitrary waveform generator, dual adjustable power supplies (won't replace a good bench power supply but adequate for lower voltage projects), 16 bits of digital IO that can also be used as a logic analyzer, a network analyzer (freq and phase response of two terminal networks) and it does a decent job of FFT.  Besides, it is running on the BIG screen - 27" displays are way better than 5" displays.

http://store.digilentinc.com/analog-discovery-2-100msps-usb-oscilloscope-logic-analyzer-and-variable-power-supply/

I have attached a sample FFT from the Analog Discovery and the DS1054Z.  Remember, I don't really know much about these displays, I never use them.  The settings are probably less than optimal. The Analog Discovery is generating the 1 MHz square wave

[MrW0lf]

But I have a calibrated DMMCheck device so I thought I would check my DS1054Z:

So as you can see, scopes can be pretty decent with DC. For many scopes this carries to HF AC also, because of "by the book" implementation of analog frontend and software processing, which does not decimate raw data. If you use scope with accurate HF AC for long enough, DMMs gonna stay in the drawer for most of the time because one can change practices a bit. Maybe not even need to buy them so much and invest money in better scope instead.

But if scope software has tricks in it you gonna be in for surprises when frequency does not divide well with decimation ratio or Sinc implementation is programmed so it causes large overshoot even with pure sine. In short - scope response is not simply predictable and depends heavily on input signal and simple dB corrections are not going to help - accuracy is permanently lost.

So in general I do not see why fight the progress - it was time when single frame of video was blurry mess. Now at 4k it's way crisper than old film photos. Same with scopes. Highly accurate AC is going to be norm. Not coincidence that top of the line DMMs are starting to implement scope features and vice versa.

[nctnico]

Ahmen on the Rigol FFT - it's much improved but still not stellar.  The good news, for me, is that I have done Fourier Analysis exactly ONCE since I graduated in '73.  It just isn't important - to me...

Maybe not yet. On my previous scopes I have not used FFT much but since I have a scope on which FFT works well (quick & high frequency resolution) I find myself using FFT more and more. Though I have to admit I do a reasonable amount of projects involving (digital) signal processing.

[bsas]

Thanks guys!

But let me rephrase my request
I don't care about the FFT being super precise, since the audio applications I am thinking about are not super precise at all, I just need a visual aid about the harmonics that the circuit is amplifying, for example.

My questions for DS1054Z owners is:

1) If I need, for some specific problems, a very high quality FFT, is it possible to connect it via USB to a more complex data analysis software? Because if so, I can use the FFT in the DSO just for quick analysis and deep analysis I can dump the data to the computer and do it there.

2) For everything else, do you recommend the DS1054Z or anyone here have a unit in mind at this price range that is better?

Again, thanks all!

[rstofer]

Thanks guys!

But let me rephrase my request
I don't care about the FFT being super precise, since the audio applications I am thinking about are not super precise at all, I just need a visual aid about the harmonics that the circuit is amplifying, for example.

My questions for DS1054Z owners is:

1) If I need, for some specific problems, a very high quality FFT, is it possible to connect it via USB to a more complex data analysis software? Because if so, I can use the FFT in the DSO just for quick analysis and deep analysis I can dump the data to the computer and do it there.

I have done offline FFT from data collected from external power measurement devices.  It worked well but I had to practically go back to school to work it out.  I don't have any idea how to sample with the DS1054Z.  I know you need periodic samples and there are constraints on the sampling interval.  Has only been an interest once, a long time ago.

2) For everything  else, do you recommend the DS1054Z or anyone here have a unit in mind at this price range that is better?

Those of us who like the DS1054Z don't believe there is anything less than twice as costly that even comes close.  Those who hate the DS1054Z freely admit that their preferred device is at least $1000 and feel we should just pony up the bucks.

I don't speak for anyone other than myself - I am entirely satisfied.   When that is no longer true, I'll just buy something else.  Remember, technology isn't standing still.  No matter how good the DS1054Z is compared to the rest of the market, something better or cheaper will come along soon.

If you want to spend a lot more money, there are better scopes on the market.

[JPortici]

1) If I need, for some specific problems, a very high quality FFT, is it possible to connect it via USB to a more complex data analysis software? Because if so, I can use the FFT in the DSO just for quick analysis and deep analysis I can dump the data to the computer and do it there.

yes, you can. keep in mind that you still have 8 bit samples! which means that using more points won't give you better results unless tricks are used.. kinda..
see: https://www.eevblog.com/forum/testgear/fft-spectrum-analysis-reviewed/msg920781/#msg920781

2) For everything else, do you recommend the DS1054Z

i don't.
i probably fall into the "hate it" category, i have been very invested in the holy war in the past but now i simply don't care anymore.
given that what those who recommend it say is true i don't recommend it because:
1) measurements are done with data from screen memory, not from acquisition memory.. and screen memory data apperas to be just a decimation, which leaded to measurements inaccuracies (there are countless topics about it) one could argue that the numbers are still inside the accuracy spec but i have lost trust in them
2) decoding is done by analyzing screen memory, which is bullshit. 24 6 MPts (because we are using 4 channels) can contain a lot of packets! have to zoom in so that the decoder can do its job, have to manually scroll very slow or you will find you to be at the other end of the track in a single turn which leads to
3) I find the UI to be painfully slow and unacceptable to be so, even when stopped. i'll give you that it has improved in recent updates to the point it's actually usable, when you don't have decoding and measurements on.

there are other quirks but if i mentioned them i would really be nagging, they haven't been a real problem to me yet.
if when stopped 1 and 2 weren't true because you're not acquiring new data, you can process the full memory.. it wouldn't be a showstopper.

There are people who still use it despite this problems but i am not one of those. The scope doesn't work the way i want it to do (like what i use at work) so i'll change the scope.
I am waiting for mid march to see how the new keysight is. then i'll unload the rigol to a friend and my bedroom lab will have either the new keysight or a siglent 1000x
i can work with 2 (actually 3) channels, i do it all the time! not always straightforward but it can be done for what the op asks

to the OP: you need more memory than sample rate/bandwidth

at this price range that is better?

it depends on what you actually need, a bit of everything or something specific, and if you are willing to spend 1-2 hundred more for it

[djnz]

Afaik, Tektronix does measurements from acquisition memory, but Agilent / Keysight use screen data as well. I am sure the scope makers who don't use acq data but instead use screen data had their good reasons, I just sometimes wonder what they were and what sort of trade-offs went into that decision. 

[JPortici]

yes i remember that thread (i should mention that i use picoscope/tektronix at work)
i don't have direct experience with keysight but i am willing to try and see if its way of work will be okay for me

i am sure that there is a way to use screen data without compromizing measurements too much. still, if i wanted to overlook the measurement point there is all the rest, decoding and UI, that are done differently on other brands.

[MrW0lf]

Agilent / Keysight use screen data as well

That's not entirely correct. They seem to use reduced size secondary buffer, but it is way larger than screen pixel count. Siglents follow similar logic (~70k on SDS1000X). Look this video:



MSO-X 3000T vs Tektronix MDO3000

Note that author arrives at somewhat incorrect conclusion that Agilent needs "higher sample rate" to get risetime measurement right.  With 1GHz Agilent scope you need to be at about 1us/div, 5GSa/s to get ~12ns risetime correctly. Tek does similar at 200us/div in given video. According to my calculation MSO-X secondary buffer used for calculus is about 20k in size. Suppose it is done to mask lacking processing power. Rigol took same concept to the max with 0.3...0.6k buffer. So pick your poison. If you need measurements then full mem / large buffer scopes are in different class to Rigol/Agilent. If no need, other factors come to play.

Edit: Looked up actual tests here:
https://www.eevblog.com/forum/testgear/testing-dso-auto-measurements-accuracy-across-timebases/msg1091323/#msg1091323
Indeed, Siglent SDS1102X+ gets numbers right at about 5us/div for similar risetime. So better than 1GHz Agilent in this regard  Confusing stuff, gotta really understand your actual requirements before buying something...

[Fungus]

All I need is wiggly lines plus or minus a few percent. 

This. Wiggly lines and enough accuracy to know the power supply is OK.

Absolute voltage measurement accuracy isn't very important. Oscilloscopes are for seeing the wiggles.