How extensive is digital oscilloscope self calibration?

[Ecklar]

Watching Dave or Wolke perform an analog tektronix scope calibration exercise is really interesting but looks somewhere between time consuming to a little difficult to OMG I don't have the proper tools.

With that in mind, just how extensive are the self calibration features on the new Rigol DS2000 and the Siglent SDS2000 series scopes?  Will these scopes essentially be within acceptable tolerance forever if I execute the self cal mode from time to time?   Or, will I still need to crack open the case once a year and fiddle with settings or mail it off to some test equipment calibration specialists?  In addition, are digital scopes much more inherently stable and less prone to drift than the older analog scopes ie Tek 465B?

Thank You

Eck

[MartyD]

Calibrating an analog scope is a rather easy task, IF you have the right tools to do it. Basically you need a good signal generator that goes as high as your scopes bandwidth, a timemark generator with fast rise time, a reference voltage source and, last but not least, a precise multimeter. At least thats what I need for my Tek 2467B and also my combi-scope Tek 2230, they can both be adjusted with rather simple gear.

However, things get WAY more complicated, expensive and time consuming when you try to adjust an older digitizing scope like e.g. a Tek TDS series. Have a look at this video here: .

Regarding modern digital scopes, I think one cannot calibrate them at home at all. There are no pots to twiddle, it's all done in software. And I betcha, the manufacturers take good care of the calibration software and procedures not becoming available publicly. However, it can be expected that their long time stability is way better than on the old scopes.

Marty

[T3sl4co1l]

Simple answer is...

If your organization's quality manual requires that equipment be calibrated by a third party...

...Ya gotta calibrate it by a third party...

If you mean practical expectations rather than written assurance, it's unlikely to drift much over time (how much should be in the manual), and I would at least expect a digital scope to account for gain and offset errors in an automatic procedure, but probably not bandwidth or high tolerance voltage reference stuff.

Tim

[HighVoltage]

My first digital scope was a combiscope from Fluke, the PM3394B.
It had one designated button for "Self Calibration"
Once the scope was warmed up, this self calibration took about 3 min and it was important to
do this self calibration at least once a month.
A full calibration also could be performed by the user with the service manual.

Everything newer than that, I just have given to the calibration lab, because it required software and special equipment.
 

[MartyD]

Siglent e.g. has service manuals in their download section: http://www.siglent.com/ENs/down.aspx?id=11.
So, according to that manual, all I need to calibrate my $400 SDS1102CML is a $20k Fluke 9500B with an active probe, a SCPI-Card and a computer script which is not available for download.

[tautech]

Watching Dave or Wolke perform an analog tektronix scope calibration exercise is really interesting but looks somewhere between time consuming to a little difficult to OMG I don't have the proper tools.

With that in mind, just how extensive are the self calibration features on the new Rigol DS2000 and the Siglent SDS2000 series scopes?  Will these scopes essentially be within acceptable tolerance forever if I execute the self cal mode from time to time?   Or, will I still need to crack open the case once a year and fiddle with settings or mail it off to some test equipment calibration specialists?  In addition, are digital scopes much more inherently stable and less prone to drift than the older analog scopes ie Tek 465B?

A new DSO with a current calibration sheet is only intended to remain within cal limits for up to 1 Year.
This is the normal period for calibration for most test equipment.
Old CRO's could drift a little in that time but modern DSO's with their "self cal" feature are much more likely to remain very close to spec.
But all precision equipment has a factory recomended time of warm up before accuracy is guaranteed while in calibration.
This also applies to pre-"self cal" prcedure. For the SDS2000, Siglent specifies 30 minutes warm up before self cal. They also recommend self cal be performed if the ambient temp has altered more than 5 degrees C.

It requires a serious set of gear to properly calibrate a CRO scope, however a quick check can be done with a good multimeter, power source and a timemark generator as MartyD has described.
Often another scope of a specified BW is also required.

Calibrating the modern DSO is usually done with the front panel from within Service menus and still requires the correct gear for official calibration. Remember also this gear must be in calibration.
All the knobs and buttons take on different roles when in the Service mode and the manuals outline the procedures. But it is stil quite complex, for example an old Tek TDS210(60 MHz monochrome) has 55 steps.

[Howardlong]

IMHO these days the annual cal isn't much more than money for old rope most of the time, a hangover from the valve/tube era and dirty pots.

A half decent engineer will know when to question the results of their test equipment, and will frequently have secondary ways of checking.

Of course, I may be firebranded for such a suggestion, but experience tells me different!

[HighVoltage]

IMHO these days the annual cal isn't much more than money for old rope most of the time, a hangover from the valve/tube era and dirty pots.

There is a deep truth to this.
I have to get my gear calibrated once a year, because I have clients with ISO and TS certification.
Not only that... there is more to it these days.
I have to have prove calibration and "Measurement Capability"
That is a complete new game and takes a lot of work.
Essentially it shows that the gear stays precisely stable, even after several thousands measurements right after each other.
This also needs to be certified and some times cost more than the calibration.

In the good old days, it came down to the engineer and his common sense and a quick check for accuracy.
These days, it is mostly about making extra money for cal lab and we have to prove the last digit in a 6 1/2 digit multimeter
although we only measure with 3 precision.



 

[saturation]

For personal use, you never need to send a modern DSO for calibration. 

If you are in professional setting, calibration is very likely part of your instrument maintenance schedule you can do nothing about, as many certifying bodies require it. 

When making measurements with a DSO, what matter is relative change in real time, from the device itself as a 'standard' versus other components in the DUT.  True calibration and adjustment is correct absolute measurement, correct for all time against the national standard, which is harmonized with the world standard.

DSO are not accurate measurement devices, relative to DMMs.  If you check the manuals, typically they specify 1-5% of reading, plus various offsets.

What DSO calibration buttons do is compares the internal references against the measurement algorithm and adds correction factors into the flash memory tables.  It just realigns the algorithms to correct for component drift. 

In the lowly Rigol 1052e, the measurements are against the screen image, rather than internal memories, so in a way it calibrates the V and H axes of the scope.

If a DSO can be calibrated, those internal references are compared against a national standard and adjusted to conform to it.  Then the automatic DSO calibration algorithm runs against the internal reference which is now sync'd to the national standard.