Low Cost Digital 'scopes; accuracy ?

[theoldwizard1]

So obviously I am referring to Rigol.  They have put out some pretty good 'scopes at some amazing price points, but a couple of things bothers me.  8 bit AtoD and sample rate.

When I got around to watching Dave's teardown on the DS1054Z i was very surprised to see that it only had an 8 bit AtoD.  The HMCAD1511 is impressive on 1 channel at 1GS/sec but you have to divide by 4 for all channels.  What little bit of signal processing I can remember (from long ago), you really want >10 bits of resolution and >10x sampling rate.  This makes a DS1054Z a bit "sketchy" when in 4 channel mode !

For the price, is is still very impressive, but there is a huge jump in $$$ to get to the DS2000 with higher resolution and (I believe) higher sample rate.  the HMCAD1520 will do 12 bits, but the speed suffers.

Is there "room in the middle" or can the DS4000 be re-engineered to down to about the price point if the DS2000 ?  Or am I making a mountain out of a mole hile ?

[tautech]

What little bit of signal processing I can remember (from long ago), you really want >10 bits of resolution and >10x sampling rate.  This makes a DS1054Z a bit "sketchy" when in 4 channel mode !

And much more so when hacked to 100 MHz.

[Muxr]

When I got around to watching Dave's teardown on the DS1054Z i was very surprised to see that it only had an 8 bit AtoD.

That's more or less the case with most oscilloscopes. If you need high sampling rate the resolution suffers. This is why they have the analog front end to amplify or attenuate the signal in order to get most out of that resolution.

If you look at USB oscilloscopes many rely on more resolution because they don't have sophisticated analog front ends.. but you will notice most of them have fairly low sampling rates.

The accuracy isn't the issue, the dynamic range is.

[heavenfish]

Nearly all the digital oscilloscopes today are using 8 bits A/D including those cost hundreds of k dollars. In general oscilloscope is not a decent measurement tool for voltage, but timing. There're a few real 12 bits (and high sample rate) oscilloscopes offered by Lecroy etc but very expensive. If you want an entry level 12 bits oscilloscope you can check Owon's http://www.owon.com.hk/products.asp?ParentID=57&SortID=87 I don't have it so not sure about its performance, just found the information on the website.

ADC is a big cost driver in oscilloscopes, so more and more manufacturers are lowering sample rate per channel but specing max(interleaving) sample rate in the datasheet. Tektronix used to promote 10x oversample rate per channel as his advantage, but engineers working on digital circuits would rather spend the money on record length and serial decoding today, not much care about signal fidelity.

[theoldwizard1]

The accuracy isn't the issue, the dynamic range is.

I'll agree with that !

[theoldwizard1]

ADC is a big cost driver in oscilloscopes, ...

I understand, that was why I suggested 10 bits.  The HMCAD1520 does 8 or 12 bits.  I'll bet with very little tweaking it could do 10 bit.

[rstofer]

When I got around to watching Dave's teardown on the DS1054Z i was very surprised to see that it only had an 8 bit AtoD.  The HMCAD1511 is impressive on 1 channel at 1GS/sec but you have to divide by 4 for all channels.  S2000 ?  Or am I making a mountain out of a mole hile ?

Well, with all 4 channels running, the sample rate would be 250 MS/s which still meets the Nyquist sample rate and then some if we assume the scope is hacked to 100 MHz.  It is clearly adequate if the scope is run at 50 MHz where it was sold.

Of course the DS4000 series could be sold for a lot less money but why would they?  The market sets the price and scopes of that caliber cost that much money or more.

Clearly your needs are for a much more sophisticated scope than the lowly DS1054Z.  You should probably move up several levels.

[radhaz]

OWON New Model 100MHz XDS3102AV Review

http://blog.goo.ne.jp/electronics_diy_5/e/1be4bdb5f573b00e76701b950b422269?fm=entry_awp

[Someone]

To be accurate the Rigol DS1000Z accuracy is typical of most scopes:
DC Offset: ±0.1 div ± 2 mV ± 1% offset
DC Gain (signal): ±3% full scale

So you won't get a more accurate result with additional bits, but the precision or resolution from additional bits or averaging can be useful in some applications.

[Wuerstchenhund]

When I got around to watching Dave's teardown on the DS1054Z i was very surprised to see that it only had an 8 bit AtoD.

Yes, it has an 8bit ADC, like some 99% of today's digital scopes. There are very few scopes out there with a hardware resolution of more than 8bit (like the LeCroy HDO Series with 12bit ADCs, the Keysight DSO-S with 10bit, or some PicoScopes with I think up to 16bit, and the OWON XDS with 12bit ADC).

Some scopes have a HiRes/ERES mode to get to more than 8bit resolution through oversampling but that comes with various drawbacks, like a reduction in bandwidth.

The HMCAD1511 is impressive on 1 channel at 1GS/sec but you have to divide by 4 for all channels.  What little bit of signal processing I can remember (from long ago), you really want >10 bits of resolution and >10x sampling rate.  This makes a DS1054Z a bit "sketchy" when in 4 channel mode !

No, it doesn't. For the majority of scope applications, 8bit is just fine.

As to sample rate, as long as it satisfies Nyquist-Shannon (f_s>2*f₀) you're fine. In 4ch mode the DS1054z samples at 250MSa/s, which is enough for signals up to at least 100MHz. As long as the sample rate satisfies Nyquist-Shannon for your scope's analog bandwidth then that's fine, and a higher sample rate than that becomes pointless.

Now, for signal processing you very much want more dynamic range than a 8bit scope offers, and then you'd either opt for a scope with a higher resolution ADC (i.e. Keysight DSO-S or LeCroy HDO) or better, a proper Signal Analyzer with 14bit or 16bit ADC instead of a digital scope (which is a Waveform Analyzer).

For the price, is is still very impressive, but there is a huge jump in $$$ to get to the DS2000 with higher resolution and (I believe) higher sample rate. 

Higher sample rate, yes. But like the majority of digital scopes the DS2000 is still an 8bit scope.

Is there "room in the middle" or can the DS4000 be re-engineered to down to about the price point if the DS2000 ?

What for? The DS4000 is just a 4ch variant of the DS2000, and again 'just' an 8bit scope.

Or am I making a mountain out of a mole hile ?

You are, actually. A short look at the specifications of the DS2000, DS4000 and a few other scopes would have shown you that 8bit resolution is, in fact, the standard for a digital scope today.

Also, you haven't said what your actual measurement problem is, which very much decides which instrument will be most useful. If you want to do signal analysis then you can still do that with an 8bit scope if you can live with the lower dynamic range, if not then I hope you have deep pockets 

[joeqsmith]

I've never had much of a need for higher res.  I think we had a scope that was 6-bits at one time.   

I did make an attempt to over sample and use a PC to post process the data.  When I started looking at the non-linearity and trying to compensate for it, I saw another strange behavior with the scope I was using.  For what it is, not bad I guess.

[heavenfish]

ADC is a big cost driver in oscilloscopes, ...

I understand, that was why I suggested 10 bits.  The HMCAD1520 does 8 or 12 bits.  I'll bet with very little tweaking it could do 10 bit.

try Pico's 5000 series, it uses HMCAD1520. https://www.picotech.com/oscilloscope/5000/flexible-resolution-oscilloscope

[bitseeker]

Most of the time, 8 bits works fine. The critical thing is to be sure to use as much of that dynamic range as possible. There have been many posts where someone asks why such-and-such scope has poor accuracy measuring this or that characteristic of a signal. Upon seeing a screenshot of what they're measuring, more often than not, the signal is using half or less of the full range of the ADC.