Beginner oscilloscope for audio applications, is 8 bits good enough?

[genericpurpleturtle]

Hey everyone, I want to buy my an oscilliscope for hobbyist applications. This guy here is what is what I'm currently considering getting

https://www.tomtop.com/p-e1269eu.html

Seems to be pretty good value for money (even after £30 on shipping), I also like the fact that it comes with a function generator too as I don't have one. I appreciate that it probably isn't the best but it'll be better than nothing. The two things which are currently holding me back is that it's memory could be bigger, 40k points isn't nearly as much as some other slightly pricier scopes which have 1 or 2M points. However I can't really think of too many applications where I will be using it where I would need more, but I would hate to have to buy new scope a year down the line when I start some new project. Is 40k generally enough?

The other thing which gives me pause is I would like to get a bit in audio stuff and I'm considering building my own sound systems; I read an article which said that 8-bit oscilloscopes don't give high enough resolution for audio applications. Do I need to go for a scope with a higher vertical resolution? Would 10-12 bits be enough? Can you actually get anything that is the sub £400 mark with that kind of resolution (even if it a USB oscilloscope)?

[janoc]

Hey everyone, I want to buy my an oscilliscope for hobbyist applications. This guy here is what is what I'm currently considering getting

https://www.tomtop.com/p-e1269eu.html

Seems to be pretty good value for money (even after £30 on shipping), I also like the fact that it comes with a function generator too as I don't have one. I appreciate that it probably isn't the best but it'll be better than nothing. The two things which are currently holding me back is that it's memory could be bigger, 40k points isn't nearly as much as some other slightly pricier scopes which have 1 or 2M points. However I can't really think of too many applications where I will be using it where I would need more, but I would hate to have to buy new scope a year down the line when I start some new project. Is 40k generally enough?

The other thing which gives me pause is I would like to get a bit in audio stuff and I'm considering building my own sound systems; I read an article which said that 8-bit oscilloscopes don't give high enough resolution for audio applications. Do I need to go for a scope with a higher vertical resolution? Would 10-12 bits be enough? Can you actually get anything that is the sub £400 mark with that kind of resolution (even if it a USB oscilloscope)?

There has been another thread on these Hanteks recently. No, it isn't a good value for money. You can get a Rigol DS1054z for the same price, with 4 channels instead of two and full options - 24M of memory and tons of other things. 40k of memory is pretty bad today - deep memory is useful because you can "zoom in" into the signal much more, especially at higher frequencies and also capture a lot more of the signal - useful when dealing with glitches, digital protocol decoding (even audio gear has digital controls!) and lots of other things. If all you are planning to do is audio, then 40k will be enough but are you only ever going to do audio?

The only extra this scope has that the Rigol doesn't is the built-in function generator. If you want that on the Rigols you would have to pay about 200 euro extra for the more expensive variants which have it. However, it is probably better to buy an external generator - it is much easier to use when you have actual dedicated buttons and knobs than having to deal with the multi-function buttons and menus on the scope.

I read an article which said that 8-bit oscilloscopes don't give high enough resolution for audio applications.

Most scopes and especially the cheap ones have only 8bit ADCs. And they have certainly been good enough for audio (and not only). Keep in mind that the final arbiter of what you measure on a scope is your own eyeball in 90% of the cases. Will you be able to see the difference between 8 or (for ex) 12 bits of vertical resolution in a typical audio application?

Typical measurements done with a scope in audio work are checking/adjusting the signal levels and looking for various distortions (not precisely measuring them - a scope is not good enough for that, an experienced tech is able to see as minimum only about 1% or worse  distortion on a sine wave). And neither of those really needs the high resolution ADC.

This article has a good explanation of benefits and disadvantages of having a higher resolution ADC:

http://www.electronicdesign.com/test-measurement/what-s-really-difference-between-12-bit-and-8-bit-oscilloscope

Also this:
https://www.eevblog.com/forum/testgear/chinese-oscilloscopes-with-12-bit-ad-converter-in-the-analog-front-end/

Lot of scopes claiming to have 10-12 ADC just have the better ADC chip in there but the front end hasn't been upgraded to match - so you can't really use the higher dynamic range because at the bottom end (weak signals) you are going to be limited by noise. If a scope has to really provide the 12 bits of usable resolution it must be designed for it - and that is expensive. And it also eats your sample memory for breakfast because you are filling it much faster when using those wider samples.

The resolution of the ADC is probably the last thing a starting up hobbyist needs to worry about, IMO.

BTW, forget USB scopes - those are either expensive instruments for very special niche applications (e.g. field work) or complete toys. A normal desktop scope will serve you much better - even an used analog one if you can get one cheaply (< 100 euro or so, I wouldn't pay more for a working basic analog scope).

[Andreas]

Hello,

I would go for a 16 bit scope which is specialized for audio purposes:
https://www.picotech.com/oscilloscope/4262/picoscope-4262-overview
Of course to get a low noise level the bandwidth has to be reduced.

And be carefully: some scopes claiming 12 Bit resolution have only a 8 bit converter.
The 12 Bits are then done by oversampling.
So the above scope is a "20 bit scope" if you take the oversampling into account.

with best regards

Andreas

[janoc]

Yeah, but that Picoscope costs more than double of the Hantek/Rigol/entry level category scopes - and is not really useful for anything else but audio because of the low sample rate and bandwidth. So a very specialized piece of kit.

If this was to be my only scope, I certainly wouldn't buy it - you could well end up needing another scope next to it for more mundane testing, especially involving anything digital.

[Fungus]

For "audio work" you can use a PC sound card. They're 16 bits or more and come with signal generator.

https://encrypted.google.com/search?q=sound+card+for+audio+measurement

If you want something fancier there's the Analog Discovery. It's been discussed many times on these forums and can do really good Bode plots, etc.:

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

None of those low end 'scopes mentioned will do an amazing job for Audio analysis.

[janoc]

I think the OP should clarify what kind of jobs he or she wants to actually do with the scope. There is no point in spending a lot of money on specialized audio-specific gear if their "getting into audio stuff" means a beginner learning to build amplifiers, guitar effects or tinkering with filters.

On the other hand if the goal is to measure things like THD, frequency response, phase shifts and what not, then specialized gear (or that Analog Discovery) has its place. But that is not what I would understand as something a beginner would do, given both the equipment and knowledge needed.

[Performa01]

Hey everyone, I want to buy my an oscilliscope for hobbyist applications. This guy here is what is what I'm currently considering getting

https://www.tomtop.com/p-e1269eu.html

Seems to be pretty good value for money (even after £30 on shipping), I also like the fact that it comes with a function generator too as I don't have one. I appreciate that it probably isn't the best but it'll be better than nothing. The two things which are currently holding me back is that it's memory could be bigger, 40k points isn't nearly as much as some other slightly pricier scopes which have 1 or 2M points. However I can't really think of too many applications where I will be using it where I would need more, but I would hate to have to buy new scope a year down the line when I start some new project. Is 40k generally enough?

The other thing which gives me pause is I would like to get a bit in audio stuff and I'm considering building my own sound systems; I read an article which said that 8-bit oscilloscopes don't give high enough resolution for audio applications. Do I need to go for a scope with a higher vertical resolution? Would 10-12 bits be enough? Can you actually get anything that is the sub £400 mark with that kind of resolution (even if it a USB oscilloscope)?

The Hantek DSO4102S is an outdated design, so I would not consider one nowadays. On the other hand it is cheap and comes with a built-in AWG, so I can certainly see why it appears attractive to you. And of course a scope like this would be better than no scope at all.

There are certainly better scopes within your price range up to some €450, but without built-in AWG. A proper external AWG alone would almost eat up your budget.

For the use as just an oscilloscope, 8 bits should certainly do the job. Even analog scopes (equivalent to a near infinite number of bits) can’t give you significantly more information when looking at a signal in the time domain. The limitation only arises when you want to use the DSO as a signal analyzer (which analog scopes cannot do from the outset), because then the 8 bit limit of some 48dB for the dynamic range poses a hard limit at high quality audio applications. The dynamic range can be extended a bit in certain situations by using averaging techniques, but not by nearly as much as some seem to believe, because a low noise floor achieved by digital signal processing does not extend the dynamic range without a dither and the INL (integral nonlinearity) of the ADC sets a hard limit anyway.

As for the memory depth, 40kpts is certainly usable in general – especially at higher frequencies, but a major limitation (or at least inconvenience) for low frequency work. It means that the maximum sample rate can be maintained only up to 40us record length. When working on audio, you are much more likely to use record lengths of e.g. 40ms and that means your sample rate will drop down to 1MSa/s or even just 500kSa/s with both channels in use. Consequently, any signal content above 250kHz will cause signal distortion due to aliasing – that’s quite the opposite of what you want when working on audio. One could claim that this should not be a problem because audio has a well defined frequency spectrum, but this would only be true for a perfect working analog system. Stability problems (ringing or even self oscillation in amplifiers) for instance could cause problems even in old style analog audio applications, not to mention modern class-D equipment and digital audio in general.

When a DSO shall act as a signal analyzer, the FFT function becomes important. With just 1024 FFT points (as in the Hantek DSO4102S) you won’t get very far. Frequency resolution is poor and noise floor will be high with just 8 bits. And with only 40kpts memory, the FFT cannot be fancy, even if the hardware were capable of processing more FFT points within a reasonable time.

So what are your options?

The PicoScope 4262 would certainly be a brilliant choice for low frequency applications, with 16 bits physical resolution, low noise frontend, 5MHz bandwidth, 16Mpts total memory and a powerful FFT up to 2Mpts, which turns this device in a high end dynamic signal analyzer. The built-in AWG produces super clean waveforms with low noise and distortion, but is limited to 20kHz. As a general purpose scope it is obviously limited because of the 5MHz bandwidth, but you can at least utilize the 17 serial decoders (a lot more than in other scopes) for slower serial protocols up to some 1Mb/s.

Here’s a demonstration of its capabilities at analyzing very low frequency signals <1Hz with 725uHz resolution bandwidth and using the multiport capabilities to run two FFT windows at the same time:
https://www.eevblog.com/forum/testgear/low-frequencies-on-spectrum-analyzers/msg1468441/#msg1468441

The main problem with the Pico 4262 is its price. The 12-bit 4224 would be cheaper and has 20MHz bandwidth, but still above your budget. If you can afford it, this would still be my recommendation.

Then there’s the Analog Discovery 2, 14bit, 30MHz, AWG, with pattern generator and logic analyzer, but it’s a soapbox with only 16kpts memory total (8k dedicated for the analog channels by default), which is even worse than the Hantek and might explain why Digilent cannot be bothered to specify the FFT length for the spectrum analyzer. Even more importantly, signal distortions due to aliasing will defeat the purpose of having 14bit vertical resolution in many situations.

That leaves us with 8-bit DSOs.

The Siglent SDS1202X-E would be within your budget; high bandwidth, 14Mpts, 500uV/div full resolution sensitivity low noise frontend, 1Mpts FFT – but no built-in AWG.

So finally your best bet might be a PicoScope 2206B. It stays within your budget quite easily, has 50MHz bandwidth, 32Mpts memory, 1Mpts FFT and a built-in AWG up to 1MHz.  Like any PicoScope, it has lots of mid-range and even high-end features like full HD screen support with multiple views, very advanced math, a bunch of serial decoders and it gets regular updates at no cost. Despite its fairly low price, it comes from a well known reputable brand and is certainly not a toy.

https://www.picotech.com/oscilloscope/2000/picoscope-2000-overview

Mind you, the prices in the Pico web shop are usually higher than the actual street prices, so it pays off to look around a bit.

[genericpurpleturtle]

Thank you everyone for your comments and suggestions!

You can get a Rigol DS1054z for the same price

Janoc, the Rigol DS1054z is about £380 brand new. I know because I had to pick a cheap scope to buy for our new automated test set up at work! It's still within my budget, but I wouldn't say it's "the same price". Although being a known brand and model, I presume it wouldn't depreciate as quickly and would be easier to sell in the future if I ever did need to upgrade.

Also the DS1054z actually has hardware capable of a 100MHz bandwidth if I'm not mistaken. I've heard that its possible to hack the software to allow it to go up to this, do you know if that's the case?

The Hantek one is already 100MHz and cheaper. I guess really the main issue I see with it if 8 bits is good enough is the memory.

Keep in mind that the final arbiter of what you measure on a scope is your own eyeball in 90% of the cases.

Isn't the final arbiter of of what one can measure, the measurement features such as the cursors?

The Hantek DSO4102S is an outdated design, so I would not consider one nowadays.

Performa01 What makes the Hantek DSO4102S outdated?

The one thing I worry about a USB scope is it being a bit of a hassle to use compared to a bench scope. Do people have experience actually using them?

[Gandalf_Sr]

My vote is for the Rigol DS1054Z which can magically be transformed into a DS1104Z (so I'm told).

[Andreas]

Seems to be pretty good value for money (even after £30 on shipping),

Hello,

Perhaps on the first view. Its not only the GPB 30 for shipping but also the VAT (20%) that you will have to pay if you import the device on your own.

Do people have experience actually using them?

We have several devices of 12 Bit PicoScopes (mainly automotive) at work  (PS4423, PS4425, PS4824) since years.
At home I use a PS5444A.
For best performance you should have a battery supplied laptop to avoid ground loops.

Performa01 What makes the Hantek DSO4102S outdated?

How can you ask with a 1kPts FFT and 40 kPts memory?

1MPts FFT and >1MPts memory is standard.

On better scopes with enough (segmented) memory you can even average the results of FFTs to go lower in noise floor.

With best regards

Andreas

[nctnico]

@genericpurpleturtle: can you do a step back and tell us what you want to do with an oscilloscope? Audio is a rather broad term and not every measurement needs a lot of bits and for some type of measurements an oscilloscope is the wrong tool (measuring signal levels accurately for example).

[genericpurpleturtle]

@genericpurpleturtle: can you do a step back and tell us what you want to do with an oscilloscope? Audio is a rather broad term and not every measurement needs a lot of bits and for some type of measurements an oscilloscope is the wrong tool (measuring signal levels accurately for example).

Well I want a scope for general electronics tinkering. And I would also quite like to build an audio soundsystem of reasonable proportions, that I can perhaps run of a 12V leisure battery. Although I don't know how much of it that is going to "building" actually building the speakers and amplifier circuits vs "assembling" by buying already manufactured speakers, crossovers, amps and the like as I'm very much a newb when it comes to all things audio. I can't really give much more detail as I don't know have much of an idea

When reading an article on the picoscope website about how to choose an oscilloscope it mentioned needing higher resolution ADC and so not knowing much about the subject I thought I would ask for everyone's opinions before buying something.

https://www.picotech.com/library/application-note/oscilloscope-tutorial

[Fungus]

In that case: Rigol DS1054Z

[nctnico]

@genericpurpleturtle: can you do a step back and tell us what you want to do with an oscilloscope? Audio is a rather broad term and not every measurement needs a lot of bits and for some type of measurements an oscilloscope is the wrong tool (measuring signal levels accurately for example).

Well I want a scope for general electronics tinkering. And I would also quite like to build an audio soundsystem of reasonable proportions, that I can perhaps run of a 12V leisure battery. Although I don't know how much of it that is going to "building" actually building the speakers and amplifier circuits vs "assembling" by buying already manufactured speakers, crossovers, amps and the like as I'm very much a newb when it comes to all things audio. I can't really give much more detail as I don't know have much of an idea

In that case just start with a generic scope and don't worry about the number of bits too much. You can always buy more specialised gear later on.

[nanofrog]

The Rigol DS1054Z can be unlocked which pushes the BW to 100Mhz, doubles the memory to 24Mpts, as well as permanently activate other software options such as serial decoding.

I can confirm this works as I did it with mine (very easy to do BTW).

In regard to audio, a 12bit or better ADC does help with some measurements (increased vertical resolution), but it's also expensive. If you find you're in need of this, you may actually be better off obtaining an analog scope for these measurements as they can be had much cheaper (you lose the niceties of a DSO, but gain the full vertical signal being tested).

I wouldn't stress over the vertical resolution as you can do a lot with a DSO with an 8bit ADC. 

[SaabFAN]

Don't know if it has been said already, but if you're looking purely for audio-applications, your best bet is most likely a modern PCIe Sound Card with 24 bit / 192 kHz Sampling. Software to turn the soundcard into an oscilloscope should be available as Open Source.
You might need to build some simple protection-circuits, as these soundcards generally don't like anything more than Line-In voltages.

Also the Output of the PC Soundcard is probably the best function generator per Euro. If you want higher output-voltages, put a simple amplifier-circuit between your DUT and the Soundcard.

To debug other things, the Rigol DS1000Z-Series or the equivalent series by Siglent are most likely your best bet.

[tautech]

Well I want a scope for general electronics tinkering. And I would also quite like to build an audio soundsystem of reasonable proportions, that I can perhaps run of a 12V leisure battery. Although I don't know how much of it that is going to "building" actually building the speakers and amplifier circuits vs "assembling" by buying already manufactured speakers, crossovers, amps and the like as I'm very much a newb when it comes to all things audio. I can't really give much more detail as I don't know have much of an idea

When I post this most will say you must have a 4ch scope but to learn on and for general tinkering 2 are often enough. SDS1202X-E are very popular in this market segment and it's what I'd recommend for you as its FFT is good in this class of instrument.
https://www.siglenteu.com/digital-oscilloscopes/sds1000x-e-series-super-phosphor-oscilloscopes/

[Gandalf_Sr]

My Rigol DS1054Z DS1104Z has FFT too

[rhb]

I have a DS1102E which has a completely useless FFT.  The Instek 2000E line has a good FFT.  Can't say about the others. I considered a DS1000Z, but the FFT was not in the same class as the Instek.  Of course, now I have an 8560A spectrum analyzer, so it's less important than it was. The Instek UI is annoying.  The Rigol has much better ergonomics.  I got lucky and got a GDS-2072E for $244 delivered from Amazon when they were clearing out old stock.  But it's not competitive with the DS1054Z at regular prices.

For audio work, a good sound card and Octave beats anything you can afford.  So if you want to build a super clean audio oscillator (THD too low to measure)  a la "Max Wien, Mr. Hewlett and a Rainy Sunday Afternoon" by Jim Williams, the sound card will let you measure the distortion to even lower levels than Jim could. In fact, with a pair of sound cards you can assemble an audio vector network analyzer with very high performance for testing speakers, crossover networks, etc.

In the end it depends on where your main interests are.  If it's mostly audio, the sound cards are the best choice.  If it's more general the Rigol is a good choice and has the advantage of being easy to sell if you want something different.

[xani]

You might want to look into Analog Discovery 2. For the price you get decent (14 bit, differential, 100MS/s) input, signal gen and Network Analyzer which is invaluable for testing bandwidth.

You can also easily measure THD directly, altho in that case the 14 bit input is a limitation.

[genericpurpleturtle]

Okay I suppose I am leaning most towards the Rigol DS1054Z (although not completely set just yet). Is it worth getting a second one hand if I can? How much do people think a second hand one is worth?

[bd139]

I sold mine for £280 cash if that's a useful metric for second hand value.

[Alex Nikitin]

For audio work nothing really can replace a plain old analogue CRT scope (just try to observe a mild HF oscillation near clipping in a power amplifier on a DSO). On top of it a computer with a half-decent sound card (plus a sensible front-end) can help a lot.

Cheers

Alex

[genericpurpleturtle]

(just try to observe a mild HF oscillation near clipping in a power amplifier on a DSO).

What about a DSO makes that tricky? I was initially considering getting an analogue one jsut becaues of how cheap they are, but everyone at work advised me to steer clear of them for an assortment of reasons. Big, need calibration, screens are difficult to read.

On top of it a computer with a half-decent sound card (plus a sensible front-end) can help a lot.

What is a sensible front end? I've found this https://www.zeitnitz.eu/scope_en? I literally had no idea you could use a sound card as a scope until this thread! I think I should have a spare one lying around somewhere so might try it out for fun.

I sold mine for £280 cash if that's a useful metric for second hand value.

That is a useful metric. £280 is to be honest more than I thought one would go for. Is a plus for getting one though, as it'll probably keeps its value more so than if I got a less known model.

[bd139]

There are very few problems with the analogue scopes for audio class applications. In fact in some cases they are superior, mainly because the noise floor is a lot lower than the cheap DSOs are. Most of the signals you see are repeating ones. DSOs only come into real usability for rare/single events, automated measurements etc.

There are two things you need to think about i.e. two classes of signals:

1. Time domain, as viewed on an oscilloscope. From that, as another person said, you can see HF oscillation (blurry trace), clipping (odd shaped sine at top/bottom), minor distortion (odd shaped sine), amplitude (peak-peak voltage) etc.
2. Frequency domain, as viewed on a spectrum analyser. This is the important one. You can see distortion via harmonic content etc. The sound card has some remarkable capabilities this space for both injecting and monitoring signals.

The analogue ones are big but that means they're pretty easy to repair and use a host of standard bits inside them. If you can get a relatively decent one (personally Tek, Hameg, Philips) then you're sorted. As for calibration; everything needs calibration on and off. The DSOs are not immune from this although they have self-calibration features built in some of them. However they do drift around a bit and require regular self-cal processes run (from my experience with Rigol and HP/agilent DSOs).

The "easier to read" thing is true but that's really not a problem in practice. You spend more time working out how to get the right measurement on the screen on a lot of DSOs and pissing around with the trigger than you do counting the divs on an analogue scope. Perhaps 90% of the time you don't even need to actually do a real measurement but just get a feel that the signal is about right before cracking out the other tools to do the measurement. That's where the scope comes in most of the time.

With respect to DSO signal visibility, DSOs tend to have a "fake phosphor" display of some sort which is an emulation of analogue scopes. This is quite a poor emulation in some cases and masks some problems which are obvious on an analogue scope. As someone mentioned HF oscillation, if your amplifier is oscillating then you can usually see this as a function of the definition (blurryness) of the trace on an analogue scope. The digital ones, not so good because of the mathematical averaging/interpolation can hide this away.

Honestly one of the reasons I dumped my Rigol is (a) I don't need it and (b) I can get good enough indication on a £30 Tek I repaired and (c) I can use other methods which are far more accurate for actual measurements. Point C: frequency counter, sound card, power meter etc.

Just remember the oscilloscope is the jack of all trades, but a master of none. It's better to invest in something which can measure distortion if you are analysing audio amplifiers and the sound card is pretty damn good at that.

Edit: forgot the software link I was going to add. Look at this for software: http://audio.rightmark.org/products/rmaa.shtml