Agilent DSO-X 2004A DC precision way off?

[jrgandara]

Hello!


I was measuring a 7805 regulator yesterday and instead of 5.00V DC I got 5.31V! Well, I know scopes are not the best for DC voltages, but is not too much off? I decided to test every one of those 4 inputs and every input gave me different results for the 5V regulator:

IN1 = 5.31V
IN2 = 5.04V
IN3 = 4.82V
IN4 = 5.01V (the only one which match my calibrated Fluke 187!)


Is this normal or my Agilent scope is bad? Have you tested your inputs?

Ps: Does not matter which probe I use, the difference is in each input.

[w2aew]

Does the scope have a self-calibration routine?  If so, let the scope warm up, and run it!

[Kibi]

Your scope has a user calibration routine which you should run. The scope needs to be warmed up too, I tried it once when it was cold and it failed.
You will need four 50? cables all the same length which need to run from the back of the instrument to each channel input.

[jrgandara]

Yes, it has User Cal, but my unit is only two months old and there is a warning in the manual which says if I perform a User Cal the calibration certificate will void. Is it normal a 5% error in DC with a new unit which specs says 1%?

I´m not sure if worth loose the factory calibration for everything to try to fix this DC error.

[echen1024]

Yes, it has User Cal, but my unit is only two months old and there is a warning in the manual which says if I perform a User Cal the calibration certificate will void. Is it normal a 5% error in DC with a new unit which specs says 1%?

I´m not sure if worth loose the factory calibration for everything to try to fix this DC error.

Not user calibration. Self calibration. There is no way in hell you will void your warranty by running SELF cal

[Mark_O]

I was measuring a 7805 regulator yesterday and instead of 5.00V DC I got 5.31V! Well, I know scopes are not the best for DC voltages, but is not too much off? I decided to test every one of those 4 inputs and every input gave me different results for the 5V regulator:

IN1 = 5.31V
IN2 = 5.04V
IN3 = 4.82V
IN4 = 5.01V (the only one which match my calibrated Fluke 187!)

Is this normal or my Agilent scope is bad?

First off, performing a Self-Cal as others have suggested will probably improve things a fair bit on your most divergent channels.

But to answer your questions directly:

a)  no, that's not too far off, and
b)  yes, it's normal, and
c)  it's within spec (a question you didn't explicitly ask, but was implied)

Is it normal a 5% error in DC with a new unit which specs says 1%?

I'm curious where you got the impression it has a 1% error spec?  From the datasheets I have, it says 3% of FS, over 10mV/div.

I don't know what settings you were using, but if we keep it simple and say 2V/div, with 0 offset, that's 16V FS.  3% of that is 0.48V.  All 4 of your channels are within that spec.  (Actually, there's an additional 0.25% FS error term, which would make it 0.52V, max rated error).  So Agilent only claims it will be within ~10% of reading at 5V, and they all are well within that range.

Your average error is 2.7%, dominated by your worst channel (6.2% of reading, 1.9% of FS).  A Self-Cal will improve that, but still is unlikely to get you anywhere close to 1% of reading.  2% of reading here (+/-0.1V) would be very good indeed.  Expecting it to "match my calibrated Fluke 187" is completely unrealistic.

[Mark_O]

I wrote:

> I don't know what settings you were using, but if we keep it simple and say 2V/div, with 0 offset, that's 16V FS.

To extend the analysis a little more, let's say you decided to zoom in a bit, to examine things closer, and minimize that nasty FS error term.  So you switch to 1V/div, and slide it down 2V, so you have some headroom above 5V to see excursions. 

You've cut your FS error in half (260 mV), and introduced a new Offset error term into the equation.  That's +/-0.1div +/-2mV +/-1% offset setting.  So another 122 mV, for a total potential error of 382 mV.  That's better than 520 mV, but not by a lot.  7.6% of reading, vs. 10%.

People often don't like to take the time to read all the "fiddly bits" in the specs, but it's important for engineers to do so.  Understanding how they impact the readings you get from your instrumentation is the only way you can make effective use of your tools, and derive results you (and others relying on your expertise) can have confidence in.  The very fact that you're concerned about the 3rd digit (which is completely meaningless here) betrays that lack of understanding.

[ot1]

I read a post on this forum recently that Agilent sells a 4 channel bnc accessory to calibrate your scope  I looked it up and it was $35 from one of the online distributors