I moved the BMO from the workbench to the lab, which is 5 meters away. Powered on it showed "temperature warm up" - as the unit was off for thirty seconds or so this seems to depend only on the uptime, not on the temperature. No big thing as I guess the extended calibration will not be done often.
The Magnova continuously monitors the front-end temperature to determine when the thermal end state has been reached. Once there is no significant temperature change for several minutes, the calibration dialog changes from "warming up" to "warm."
The measurement is quite precise (to a tenth of a degree), and the system correctly identifies a reboot as a significant temperature change, meaning it recognizes that the final thermal state hasn’t yet been reached anymore. In this case, however, it should not have taken too long for the warm state to be reached again.
The DMM used should have at least 0.1% accuracy on 10KHz AC measurements. One might think, that 14 month old Keithly DMM is way better than that and it is still in it's first calibrated intervall. Maybe a mistake. Accuracy of the DMM6500 in the needed frequency and voltage ranges is 0.05% of reading + 0.03% of range. At near full range readings we will have therefore 0.08% accuracy; not so much better than 0.1% and the temperature coefficient is still not considered. Looking at the datasheet I was surprised that the calibration cycle for the 1VAC and 10VAC ranges is 90 days. The 2 year cycle I had in mind is valid for the 750VAC range.
While it would be possible to use a less precise digital multimeter (DMM), the 0.1% requirement ensures that the DMM's measurement error has a negligible effect on the oscilloscope calibration. A 6.5-digit DMM, such as the Siglent SDM3065X, should work well for this as well (0.06% + 0.03%, within ±5 °C for ACV/2 V/10 kHz for a 1 year period).
Maybe another error could result from the needed cabling for the calibration. It takes 4 T connectors to connect all input channels and the DMM to the function generator. As the hp 3325B only has a 50 Ohms output, I also added a feedthrough terminator. To be sure I checked each connection for the scope input channels with the DMM and of course I got small differences. For a 1Vrms / 10KHz sine wave the deviations were <0.1mV.
The absolute voltage error was fine tuned on the hp but never vanished.
The manual (and I think the text on the screen as well) told me that the difference from the ideal voltage value and the measured voltage should be entered at some point of the procedure but I missed that. Maybe some extra colored hint would help at this point.
That's an important point. Since the generator output can’t be set exactly right, the calibration lets you enter the actual value measured by your DMM manually. You can tap either the amplitude or rms voltage field. Entering a value in one field automatically updates the other. The calibration will then be based precisely on the measured and entered DMM value, so there’s no need for tedious fine-tuning of the generator.
Thank you for pointing out that we should make the input fields more visible.
At the last step I noticed that the frequency had changed from 10KHz to 10Hz - or was it some steps before? I am not sure but at the moment I do not want to repeat the time consuming procedure.
Thanks for pointing that out! The frequency note could certainly be highlighted. It's easy to overlook the change, especially since every step except the last one uses 10 kHz.
At the first sight, the result of the extended calibration was good - I did not see any level differences.
Excellent! That's how it should be.
As I got the AWG some weeks ago, I wanted to know if it has a good voltage accuracy. Connected to the DMM in Hi-Z mode it measured 1.0003 VAC +/- 0.07mV for a 1Vrms sine wave.
Then I put it in 50 Ohm mode and used the same feedthrough terminator I used before and the measured values worsened ... up to a 5 mV error.
So doing the calibration again without the terminator and half the voltage setting on the generator would maybe result in less need for voltage fine tuning. All the other terminators I have are non feedthrough and would require another T connector.
The point is: With the AWG installed you can think of an automated extended calibration procedure. Maybe in a first step the accuracy of the AWG should be checked manually but then the rest could be run by a program.
Yes, more advanced automated calibration with a built-in generator would be possible. However, an accurate external DMM and some manual steps would still be necessary.
The accuracy you observed in Hi-Z mode sounds very good (as expected).
When using a 50-ohm termination, the voltage is subject to the tolerances of the internal output resistance and the external termination.
The Magnova has a special feature that sets it apart: With the Magnova generator, not only can you specify Hi-Z and 50 ohms as the load, but you can also adjust the resistance within a wide range with milliohm precision. This allows you to compensate for the tolerance of the output resistance, and the load (termination) resistance, as well as perform a quick calibration for a specific load case.
Home
Help
Search
Login
Register
