Same Spec'ed Rigol Probes but Not The Same?

[Mechatrommer]

While probing 100MHz signal using both Original Rigol's Probes set to x10, both showing different phase and magnitude (x10_probe.jpg). The scope is left warmed up 30 minutes and calibrated and both probes are compensated. But when using direct DIY x1 pcb/bnc connection, both showing okay the same signal phase and magnitude (x1_direct.jpg). If the x10 difference is due to ground loop stray inductance, they have both using ground loops and the inductance should be modeled the same between both probes i think. Is this suggesting both probes have different capacitance? and since there is slight attenuation in channel 2 throughout all frequency in x10 setup, Is this suggesting both probes have different resistance? as well?

I read that to better measure high frequency signal, its better to use x10 probe (higher impedance, lower loading , and compensated internal dso capacitance), but from case above, i'm not sure whats going on. i hope to hear some explanation/assumption. thanx.

[Kiriakos-GR]

Question : What is the max P-P voltage input level for this Rigol ? 


 

[Mechatrommer]

Question : What is the max P-P voltage input level for this Rigol ? 

300Vrms, somewhere ±424V or 848Vpp, theoritically, thats what the front panel said.

[bilko]

From the manual ....

Measurement Category
The DS1000E, DS1000D series Digital Oscilloscope is intended to be used for
measurements in Measurement Category I.

Measurement Category Definitions
Measurement Category I is for measurements performed on circuits not directly
connected to MAINS. Examples are measurements on circuits not derived from
MAINS, and specially protected (internal) MAINS derived circuits. In the latter case,
transient stresses are variable; for that reason, the transient withstand capability of
the equipment is made known to the user.

WARNING
IEC Measurement Category I. The input terminals may be connected to circuit
terminal in IEC Category I installations for voltages up to 300 VAC. To avoid the
danger of electric shock, do not connect the inputs to circuit’s voltages above 300
VAC. Transient overvoltage is also present on circuits that are isolated from mains.
The DS1000E, DS1000D series Digital Oscilloscopes is designed to safely withstand
occasional transient overvoltage up to 1000Vpk. Do not use this equipment to
measure circuits where transient overvoltage could exceed this level.

[Kiriakos-GR]

Ok , I have one last question .  
The 100MHz it is RF , and so logically the oscilloscope will need an more powerful source than the 1V
that you have as test source.

Can you build an small buffer to amplify the signal that you have , to 100Vpp ?
Or amplify it in to a minimum of 5% from the total Max range, and this translates to 42Vpp.
And retest ?    

[Mechatrommer]

it got nothing to do with mains, why its popped up in your mind about the mains concern? and why on hell should i amplify the original DUT from 1Vpp (2Vpp actually) to 42Vpp. i know you work in HV-LF, but we work in LV-HF, and amplifying it is not as simple as putting HP step up transformer in the middle. i've not shown you the mV range yet. just explain to me why there is ~45 degree phase shift? on the very same signal probed with 2 probes/terminal?

[vk6zgo]

Your test seems to rule out the DSO,but just in case,swap the probes over (change the BNC connectors over).
If the lead waveform swaps over,it is a problem with the probes,if it doesn't,it is a DSO problem.

If,as seems likely,it is the probes,one suggestion is to adjust one or both probes for minimum phase difference,consistent with a non-distorted calibration waveform.

VK6ZGO

[Mechatrommer]

Your test seems to rule out the DSO,but just in case,swap the probes over (change the BNC connectors over).

how come i didnt think about this. ok, trying it now...

[Mechatrommer]

yes. the probes are the culprit, i've swapped the probe and the display got reversed in respect to probe position (1) and (2). they are not similar. changing the variable capacitance of probe channel 2, i only manage to make either one, make the magnitude closest as possible, or make the phase as close as possible, (3) is decompensated for closest phase match, but worst magnitude difference, but worst, the rigol calibration output is out of compensation (4). so what should i do? take it apart? mod the probe?
note: rigol is just switched ON cold, so smaller phase difference maybe due to rigol is out of calibration. as it gets warmer, the phase difference is got bigger, visibly.

[bilko]

Are the Rigol probes rated for 100MHz ?

[Mechatrommer]

here's from the manual:

The passive probes have a 6MHz bandwidth with a rating of 150V CAT II when the switch is in the 1X position, and a Full oscilloscope bandwidth with a rating of 300 V CAT II when the switch is in the 10X position.

[Kiriakos-GR]

here's from the manual:

The passive probes have a 6MHz bandwidth with a rating of 150V CAT II when the switch is in the 1X position, and a Full oscilloscope bandwidth with a rating of 300 V CAT II when the switch is in the 10X position.

Logically no one would use an 6MHz probe , to measure 100MHz.
Not to say that the oscilloscope are hacked to operate at 100Mhz.  

And the point is that I do not see any point in all this tests.
And as long this oscilloscope it was intended to operate at 50 MHz,
why it should came with 100MHz probe ?


Oh sorry its not your mistake, probably no one told you,
that if you hack your oscilloscope to operate at 100MHz,
that you are forced to spend more cash about getting an true 100MHz probe !!    

The 6MHz bandwidth at 10X becomes 60MHz !! 

[Mechatrommer]

i'll use my intelligence then, instead of spending more money

[saturation]

Although probes are rated the same, they are not identical; that's why you need to 'tune' each probe by compensating them ...  recall there is a series resistor in there and the variable capacitor.  In addition there are LCR parasitics built into the probe.  There is a slight phase change, but its very small.

If you bypass the series resistor by using x1 and lose the phase change, chances are its the resistor causing the issue. Note, x1 is not bad, its just it its frequency response is less, and will roll off at 6 MHz or so.

A 60 MHz Rigol scope needs a 300-500 MHz probe, because if you measure a square wave, just the 5th harmonics will roll the edges at 300 MHz using a 100 MHz probe.  The Rigol I have shipped with RP2200 passive probes, rated at 200 MHz.   300 MHz passive probes really start to cost a lot, about $40-100 each, so I can live with the 100-200 MHz probes which are $5-10 each.



More on probes.

https://www.eevblog.com/forum/index.php?topic=2289.msg32036#msg32036

While probing 100MHz signal using both Original Rigol's Probes set to x10, both showing different phase and magnitude (x10_probe.jpg). The scope is left warmed up 30 minutes and calibrated and both probes are compensated. But when using direct DIY x1 pcb/bnc connection, both showing okay the same signal phase and magnitude (x1_direct.jpg). If the x10 difference is due to ground loop stray inductance, they have both using ground loops and the inductance should be modeled the same between both probes i think. Is this suggesting both probes have different capacitance? and since there is slight attenuation in channel 2 throughout all frequency in x10 setup, Is this suggesting both probes have different resistance? as well?

I read that to better measure high frequency signal, its better to use x10 probe (higher impedance, lower loading , and compensated internal dso capacitance), but from case above, i'm not sure whats going on. i hope to hear some explanation/assumption. thanx.

[Mechatrommer]

mine are also RP2200 written on em.

[saturation]

Yes, as I recall, 200 MHz ships with 60 MHz scope, I don't know what ships with the real 100 MHz 1102e.

http://www.aidetek.com/New_products_info/Photo/RIGOL/RP2200_Manual.pdf

mine are also RP2200 written on em.

[jahonen]

Many scopes (but sadly I think that Rigol does not belong into that class) have adjustment called probe skew adjustment, to compensate variations of the flight time between different probes. That is important if one measures something like digital timing margins.

But it might be the scope itself (sampling not done exactly at same time on each channel), try swapping probes and see if the time difference is reversed between channels.

Regards,
Janne

[A Hellene]

This is Rigol's probes characteristics table:


According to Rigol's RP2200 specifications, the DS1052E's probes are rated to 150MHz and 300VAC when set x10 and to 7MHz and 150VAC when set x1.

The oscilloscope probes are extremely complicated animals, despite the textbooks that describe them as simple passive devices with resistive and capacitive dividers. For example, no x1 passive probe can go further beyond the 20..50 MHz barrier. Why?
Please, read an excellent article by Doug Ford on the oscilloscope probes, that was published at the Silicon Chip in 2009:
The secret world of Oscilloscope Probes.


I barely use Rigol's probes, since I have better results using much better ones (x100/250MHz). Just watch the table below, to see what a difference can a better probe make. By the way, I use the fragile RP2200's only when the signal is very low in amplitude to be measured by the x100 probes.

Now, this is an eloquent example of probe loading, while reading the crystal oscillator output pin ("TOSC2") of an Atmel ATmega8 μC running at 16MHz:

ATmega8: Vcc = 5.04V, DDS disconnected, CKOPT fuse enabled:
TOSC2: 4.76Vpp, loaded with a x100 probe (100MΩ//6.5pF)
TOSC2: 4.48Vpp, loaded with a x10 probe (10MΩ//17pF)
TOSC2: 3.56Vpp, loaded with two x10 probes

ATmega8: Vcc = 5.04V, DDS disconnected, CKOPT fuse disabled:
TOSC2: 582mVpp, loaded with a x100 probe
TOSC2: 312mVpp, loaded with a x10 probe
TOSC2: 184mVpp, loaded with two x10 probes

The "CKOPT fuse" controls the xtal oscillator output swing amplitude and, in consequence, the oscillator's power consumption and fan-out.
P.S. I did not even bother to make any measurements of TOSC2 with the probes set x1 (1MΩ//100pF)...


By the way, "No probes are made equal"!


-George

[Leo Bodnar]

The probes with a x1 / x10 switch are amongst the worst.  I'd much rather prefer plain x10 then a switched one.  I have never seen passive Agilent, HP or LeCroy probe with a switch.

[saturation]

@hellene, great links. The original treatise on passive probes in the 1960s circa Tek manual describes all. However the article linked in siliconechip does have a new lean to it and worth keeping.

See here as was linked on the previous thread.
http://www.slack.com/TE/TekConcepts/TekProbeCircuits.pdf

The RP2200 probe specs is interesting, the link I posted is nearly identical in content, except the US Rigol link has the bandwidth and rise time derated, for the same probe.

@Bodnar, yes, I'd see the switch introduce finite switch resistance and some residual capacitance that can change as the switch is simply used, ages or wears.  Seems like an unnecessary variable when most of the time you'd like the probe to support the highest bandwidth possible with lowest loading.  I'd think in practice 1x is rarely used vs the 10x setting and one can also accidentally set it to 1x by error.

[A Hellene]

Ah, Joe Weber's classic! Thank you for the link, even though I have the document very well hidden, somewhere in my HD drives!

Doug Ford's article above is not linked to Silicone Chip's magazine; it is kindly provided by Doug Ford himself!


Yes, it is interesting that Rigol keeps and distributes two different versions of the same document, even if both of them state to have been printed in 2007; the difference in bandwidth is not due to some sort of some newer edition reprint. Well, either the products are promoted differently or they are different actually, despite the fact that they share the same product name (RP2200), shape and --probably-- manufacturer. Maybe that is because of a possible difference in their target groups. But, who knows these things better than Rigol itself...

By the way, I just spotted another very important difference in those two documents, in the color of two of the probes' marker rings!
This probably indicates that the two different sets of the RP2200 probes might be supplied with two different oscilloscopes.


-George

[Kiriakos-GR]

A new question just born ...

Does those probes haves any markings, serial numbers on them ? something printed on the cable ?
This sort of info it would help the many so to stop scratching their heads about the probe type that they hold in their hands.

 

[A Hellene]

The set of probes that came with my 1052 included a printout of the 150MHz version of the data-sheets, above.

[Mechatrommer]

But it might be the scope itself (sampling not done exactly at same time on each channel), try swapping probes and see if the time difference is reversed between channels.

did that in Reply #8. yes time difference is reversed, so its the probes (capacitance), but attenuation is not reversed suggesting there is difference between dso internal input impedance. picture replayed here.


edit: thanx saturation and a hellene for the link and description. i never saw the RP2200 before due to my ignorance.

[Kiriakos-GR]

Hey what are your doing in here ?
Your local time is 4:33am , go to get some rest.