New Rigol DS1054Z oscilloscope

[Fungus]

Anyone tried these 10x-only probes?
http://www.sefram.com/en/products/accessories/GE2511-250mhz-passive-oscilloscope-probe-x10-300v.html
Around 35€ w/VAT, accessories included.  Seems good match with DS1000Z.

They look good to me!

I am sure they will work fine up to 100 MHz but they would be poor for random 250 MHz oscilloscopes do to lack of high frequency compensation adjustments.  There are lots of 100 MHz probes suitable for the Rigol DS1000Z including switchable x1/x10 probes.

There's been a few threads testing cheapo probes and there's not much difference between them up to about 300mHz.

https://encrypted.google.com/search?q=eevblog+cheap+probes

[McBryce]

Go retro. A decent secondhand high frequency probe from Tektronix, HP or LeCroy is better and cheaper than any new probe you can buy.

McBryce.

[saturation]

I bought and reviewed a few in the archives of eevblog.  Up to 300 MHz and $10 each is good in today's price, at the time I bought them 100MHz wase ~$6 each.  For best results characterize the frequency response as the QC varies from unit to unit, so you know the exact response chart.   I think I bought overall between 4-6 as well as a 500 MHz Pico brand for $70 as a reference probe.  The main drawback of the cheapo probes is less the response but its long term durability with use.  But its still cheap enough to be disposable and easy to replace.

http://www.ebay.com/itm/1X-10X-60-100-200-300MHz-High-Impedance-Oscilloscope-Probe-Alligator-Clip-G0ZU-/171994386891?var=&hash=item280baa11cb:m:mNgrHKQOdqb2js1JFiNnt2g

They are all nameless so choose a best purchase route for you that is local or via International aka China, HK etc.,.

[David Hess]

There's been a few threads testing cheapo probes and there's not much difference between them up to about 300mHz.

I have seen a few at 300MHz and higher that were absolutely horrible.  At least the third party ones which have high frequency adjustments can be calibrated with a suitable signal source to produce a clean response when combined with a specific oscilloscope.

Most users lack a fast enough reference flat pulse generator to make an objective test and are not using the probes in an application were poor performance would be noticed anyway.  There are good reasons to use active probes in place of high bandwidth high impedance probes.

The biggest reason I use 250MHz probes on my 100MHz oscilloscopes is that they have the x10 readout function which my 100MHz probes lack.

[Fungus]

There's been a few threads testing cheapo probes and there's not much difference between them up to about 300mHz.

I have seen a few at 300MHz and higher that were absolutely horrible.

300Mhz (approximately) seems to be the point where you have to start doing things properly.

[uvamosk]

Well guys,
 I just got my RMA replacement scope and it still has jitter but not nearly as bad as the last one. BUT this doesn't look right what do you guys think?
This is with a Probe Attached to the scope but nothing attached to it.

[alsetalokin4017]

Well, it doesn't look like anything I can obtain by matching your _visible_ settings. But we need more information.

Do you have both the probe switch and the channel set to 1x or 10x? Naturally the two settings must match.
Why are you offsetting  the trigger position by 15 microseconds when you are using 100 ns/div for the horizontal timebase?
What Acquire mode are you using? (probably Normal, but really should specify)
When you say the probe isn't connected to anything: do you have the ground clip shorted to the probe tip/spring grabber, or is it open?
Have you warmed up the new scope for half an hour then run the self-calibration routine?

[alsetalokin4017]

This is as close as I could get.

Using 10x probe and channel settings, probe ground lead clipped to grabber tip, Normal Acquire mode. Where the "loop" made by the ground lead and probe tip is located makes a difference in the noise signal, of course. In my case I have it up in the air about 2 feet above the scope, hanging by a string.

[rstofer]

Just shorting the probe and setting 10mV/div,100ns/div I get about 8mV of noise.

[JPortici]

two things i'd try..
1 - shorting the probe but with a veeeeeeeeery short piece of metal, like the spring thingie
2 - terminating the input and see what's the front end noise (in this case what should one use? a 1 Meg termination? still 50 ohm?)
3 - expand the timebase and see what the waveform look like, if it's a bunch of sinusoids, all the same frequency, you may just have have a nasty transmitter not far from you though 5.68 MHz is not a common frequency for me

[TheoB]

In my case I see 1.8mV rms or 8mv rms. That's what you get when your bandwidth is >100MHz 
The tip is shorted to the standard ground lead to form a perfect antenna.
Input coupling set to GND puts the noise at 1.2mV. That's the quantization noise I guess (resolution is 0.4mv at this setting)
First thing, enable the BW limit to 20MHz, unless you are interested in these high frequencies. FFT on internal memory showed me the noise to increase to 8mV because of something around 170MHz. But the FM band (88-108) I can also clearly identify.

[metrologist]

50 dBV over the noise?

[TheoB]

No, it says -42.4dBV. With these settings it is about 40dB above the noisefloor.

Verstuurd vanaf mijn GT-P5110 met Tapatalk

[metrologist]

No, it says -42.4dBV. With these settings it is about 40dB above the noisefloor.

Verstuurd vanaf mijn GT-P5110 met Tapatalk

Wait, I wasn't looking at the numbers, just eyeballing. I'm not that well versed on FFT here, so I took the top line as saying 10 dBV per division. I think the sensitivity is on the next menu? Or what does that 10 dBV mean?

Your cursors show Ay as -42 and By as -113, and the math is -71 dBV (and that does not match the graphic if the scale is 10 dBV/Div, and then shouldn't that delta be in terms of dB rather than dBV?).

[David Hess]

Input coupling set to GND puts the noise at 1.2mV. That's the quantization noise I guess (resolution is 0.4mv at this setting)

The DS1000Z series (and DS2000A series? Others?) lacks the hardware to support ground coupling in the sense that other oscilloscopes do.  Instead it changes the gain of the integrated variable gain amplifier so noise from all of the previous stages is removed.

To measure the actual noise, short out the BNC or attach a 50 ohm coaxial termination or attenuator.

[TheoB]

Wait, I wasn't looking at the numbers, just eyeballing. I'm not that well versed on FFT here, so I took the top line as saying 10 dBV per division. I think the sensitivity is on the next menu? Or what does that 10 dBV mean?

Your cursors show Ay as -42 and By as -113, and the math is -71 dBV (and that does not match the graphic if the scale is 10 dBV/Div, and then shouldn't that delta be in terms of dB rather than dBV?).

I quickly set one of the two cursors on the signal of interest (B). The other cursor (A) is just moved away, so ignore the delta.
-42dBV says I measure a signal of 10^(-42/20)V=7.9mV. The scale is 10dB/division. You just don't see the A cursor as it's hidden behind the measurement.
As a side note, this cheapy is able to detect a signal even 30dB lower. That's a signal of 0.25mV!  But that is with the vertical division set to 10mV. If you put input it to X1 it even drops down to 0.025mV. I don't have a spectrum analyzer, so this is my poor mans substitute 

The DS1000Z series (and DS2000A series? Others?) lacks the hardware to support ground coupling in the sense that other oscilloscopes do.  Instead it changes the gain of the integrated variable gain amplifier so noise from all of the previous stages is removed.

To measure the actual noise, short out the BNC or attach a 50 ohm coaxial termination or attenuator.

Ah, are they cheating? Did not realize that   
With 50 Ohm termination I see 190uV rms (yeap 1mV scale)

[Fungus]

To measure the actual noise, short out the BNC or attach a 50 ohm coaxial termination or attenuator.

Ah, are they cheating? Did not realize that   

Why is that cheating?

On analog scopes you often needed a real GND to help you to:
a) Align the 0V to a particular graticule line on the screen
b ) Find the trace when it was off screen and not visible

Neither of those is necessary on a DSO.

[TheoB]

I was thinking about that they might do this to lower the visible noise (it lowers from 380uV to 180uV). But as an aid for finding the ground level it's fine of course. I won't complain...

[David Hess]

The DS1000Z series (and DS2000A series? Others?) lacks the hardware to support ground coupling in the sense that other oscilloscopes do.  Instead it changes the gain of the integrated variable gain amplifier so noise from all of the previous stages is removed.

To measure the actual noise, short out the BNC or attach a 50 ohm coaxial termination or attenuator.

Ah, are they cheating? Did not realize that   
With 50 Ohm termination I see 190uV rms (yeap 1mV scale)

I do not know if I would call it cheating but I think it is deceptive since Rigol does not document it and users almost always expect ground coupling to reflect the true noise limit of the oscilloscope input.  Rigol is similarly deceptive about other things so this is not an isolated fault.  Their "trigger output" has so much jitter as to make it useless in some applications compared to the real trigger output on an analog oscilloscope and some DSOs.  Their "peak detection" on earlier DSOs is actually envelope detection; peak detection is a premium feature while envelope detection is trivial.  Their "delay" function is just another form of horizontal positioning and only works within the existing acquisition record.  Their "delayed sweep" does the same thing and not not acquire a separate sweep or acquision.

How exactly did you arrive at 190uV RMS of noise?  I assume you used the built in RMS measurement capability.  That *should* work however 190uV RMS of noise seems high to me and I would want to verify that RMS measurements work correctly with wide bandwidth noise.  Maybe something weird is going on there which seems to be a ubiquitous problem with Rigol DSOs.

Just for edification, I accurately measured the noise of a pair of 105 MHz Tektronix 7A13s (1) using the tangential method (2) and it came out to 100uV RMS and 85uV RMS.  A 75 MHz 7A18A came out to 10.5uV RMS but such a low level is difficult to measure; it was absolutely below 15uV RMS.  A 120 MHz 7A12 (vintage 1970!) came out at 11.25uV RMS but again, such a low level is difficult to measure and it was absolutely below 15uV RMS.  The 7A12 should be noisier than a 7A18A because of its more complex input amplifier and slightly higher bandwidth and reassuringly that was the result.  These measurements were all made with a 25 ohm source impedance from a terminated 50 ohm cable so they are directly comparable to a shorted or 50 ohm terminated input or real ground coupling.

(1) The Tektornix 7A13 is my benchmark for oscilloscope input noise because its bootstrapped front end circuits and differential input (noise doubled) increase its noise compared to a less functional and simpler design and its high 1mV/div sensitivity makes its noise easy to measure.  If your 100 MHz oscilloscope has more noise than a 7A13, then it has problems.  The lower noise of the 7A18A and 7A11 plus their maximum sensitivity of 5mV/div make self noise measurement difficult.

(2) Tangential noise measurement(3) allows a very accurate RMS noise measurement to be made on an analog oscilloscope.  The 7A13 being a differential comparator is uniquely suited to make this measurement however I did not take advantage of that; a calibrated graticule measurement was more than good enough.

(3) I do not give much credence to Dave's DSO versus analog noise comparison videos because he did not make any quantitative measurements but to be fair, an accurate qualitative measurement of the 5mV/div 50 MHz Tektronix 2225 he had available would have been difficult; an upper limit could have been easily determined using the tangential method though.

[JossDalVera]

Where is the best place to buy one? I have seen places online but I wanted to know if you can purchase in stores aswell

[TheoB]

Quote

How exactly did you arrive at 190uV RMS of noise?  I assume you used the built in RMS measurement capability.  That *should* work however 190uV RMS of noise seems high to me and I would want to verify that RMS measurements work correctly with wide bandwidth noise.  Maybe something weird is going on there which seems to be a ubiquitous problem with Rigol DSOs.

Yes I used the built-in function AVG. The noise is measure in normal acquisition mode. In High Res mode were you average over a number of over captured values (not sure about what it exactly does), the noise goes down to 85uV. I'm not talking about trace averaging. If I do trace averaging, I still measure around 70uV. That could indicate there is some offset, or a systematic error (crosstalk, quantisation error?)

I notice my screenshots are often corrupted. Can't seem to find why that happens. Is there some USB eject function that I should use?

[TheoB]

Ok, I got carried away. I pulled in all raw samples (at 1mV input 50 Ohm terminated) and post processed them with octave:

The instrument displays 190uVrms. The FFT shows a noise floor around -128dBV at a RBW of 833Hz. If the scope has a 125MHz bandwidth (estimate) it will show noise 52dB higher. -128+52=-76dbV (=154uVrms). I do see spurs at 62.5MHz ,125MHz and 250MHz.

[David Hess]

Quote

How exactly did you arrive at 190uV RMS of noise?  I assume you used the built in RMS measurement capability.  That *should* work however 190uV RMS of noise seems high to me and I would want to verify that RMS measurements work correctly with wide bandwidth noise.  Maybe something weird is going on there which seems to be a ubiquitous problem with Rigol DSOs.

Yes I used the built-in function AVG. The noise is measure in normal acquisition mode. In High Res mode were you average over a number of over captured values (not sure about what it exactly does), the noise goes down to 85uV. I'm not talking about trace averaging. If I do trace averaging, I still measure around 70uV. That could indicate there is some offset, or a systematic error (crosstalk, quantisation error?)

Using averaging or high resolution acquisition mode is note going to return a proper or comparable measurement of noise.  Either will attenuate high frequency noise (and are useful for this reason); my measurements were over the full bandwidth of roughly 100 MHz.  (1) The fact that you measured 85uV RMS and 70uV RMS with high resolution and averaging indicates to me that something is very wrong; the noise should have been much much lower because of limited bandwidth.

What should work is an automated RMS measurement of the trace without averaging or high resolution mode enabled.

The measurement should *not* change significantly at different sample rates or record lengths (although it may look visually different); if it does, then something is wrong and this is easy to demonstrate.  The RMS value is equal to the standard deviation which is what the tangential measurement I used returned. (2)(3) If half of the sample points are removed by using a lower sampling rate, the standard deviation remains the same.  If half of the sample points are removed by halving the record length, then the standard deviation also remains the same.  This applies *even if* the noise bandwidth is undersampled so aliasing is not important (!) which is how RF sampling voltmeters can make RMS measurements into the microwave RF bands. (4)

The measurement should *also not* change significantly when operating in real time or with a single shot acquisition.  I only bring this up because Rigol may be doing some DSP voodoo on its real time display to produce index grading which could interfere with an accurate RMS measurement so this should be checked.  If the results differ, then I would trust the single shot measurement more than the real time measurement.

(1) I am ignoring shape factor of the bandpass.  All of my measurements were on instruments with a single pole Gaussian rolloff which increases the measured noise by 1.6 times because noise above the 3dB bandwidth is also included.  Do not sweat the small stuff until the big stuff is taken care of. 

(2) A couple years ago I verified that tangential RMS noise measurement on an analog oscilloscope agreed with RMS measurement using sampling RMS voltmeter and automated RMS measurement on a good DSO.  I mean it should, right?  So does it?  Actually, tangential measurement agreed so well while some DSO RMS measurements did not that I now trust analog tangential RMS measurement more than random DSO measurements.

(3) I do not think the Rigol can do it but some higher end DSOs can produce histograms and make standard deviation measurements.  The standard deviation measurement should return the same value as the RMS measurement.

(4) This suggests another test which can be done besides changing the time/div or record length.  The Rigol has to reduce sample rate when more channels (1 GS/s for 1 channel, 500 MS/s for 2 channels, 250 MS/s for 3 or 4 channels) are used but this should *not* affect the RMS noise measurement made on one channel.  If the measurement changes, then something is broken.

[TheoB]

Using averaging or high resolution acquisition mode is note going to return a proper or comparable measurement of noise.  Either will attenuate high frequency noise (and are useful for this reason); my measurements were over the full bandwidth of roughly 100 MHz.  (1) The fact that you measured 85uV RMS and 70uV RMS with high resolution and averaging indicates to me that something is very wrong; the noise should have been much much lower because of limited bandwidth.

I averaged to see if it is actually random noise. And it is not. Default display:

1024 traces averaged were all the noise is uncorrelated should bring the trace back to the vertical quantization resolution of 80uV (about 40uV rms) (1mV/20). I can clearly see a 125MHz signal. That's what I also found in the FFT result from the raw samples:

If I use the raw sample data I have a bandwidth up to sample rate/2. That's 5 times oversampling for a scope bandwidth of 100MHz. I think I have shown clearly that the rigol has a lot of noise and a spur. I don't consider the scope to be defective though, it's just not the high quality that other scopes might offer.

The measurement should *not* change significantly at different sample rates or record lengths (although it may look visually different); if it does, then something is wrong and this is easy to demonstrate.

I checked, it does not change with memory depth settings.
It does go up (a lot) when I change the time base. The same test done at 50us/div displays a much wider noise band and it reports 500uV RMS noise. That should not happen. Hi RES solves that and lowers the noise back to 70uV (also single shot). And that's not averaging. With averaging the noise drops to 0-40uV (and the display is a nice noiseless line). Sample rate is 1Gs, mem depth 1.2M

[Sredni]

The same test done at 50us/div displays a much wider noise band and it reports 500uV RMS noise. That should not happen. Hi RES solves that and lowers the noise back to 70uV (also single shot). And that's not averaging. With averaging the noise drops to 0-40uV (and the display is a nice noiseless line). Sample rate is 1Gs, mem depth 1.2M

Perhaps I've had too much wine, but... I cannot find these values in the pictures you attached.
You say with NORMAL acquisition at 50ns/div you get 500 uV RMS noise, but the first attached picture shows 164 uV RMS measure.
You say in averaging the noise drops to 0-40 uV and yet the picture with AVERAGE acquisition mode, 1024 averages at 1GS/s and 1.2 point of memory shows 72.3 uV of rms noise.

What am I not understanding?