RESOLVED - Keysight U1272A odd behavior in dual measurement mode - firmware bug?

[ArcticGeek]

Hi All,

This is my first post here, but I've been lurking on this forum for quite some time.  I recently purchased a new Agilent/Keysight U1272A meter for my home hobby use, and I must say that I REALLY like this meter.  Up until now I have been a Fluke bigot and only considered Fluke as a "real" DMM, but this Keysight meter is a mighty nice instrument, and I find myself preferring it over either my Fluke 87iii or 189.  However, I have found some odd behavior, and I'm wondering if others have ever seen this.  The problem seems to occur when measuring an AC voltage with a small DC voltage offset, and having the meter in dual measurement mode.

To investigate this, I used 4 different meters and a function generator:  my trusty old Fluke 189, a brand new Keysight U1252B, and two different U1272A meters (one brand new, and another one that is ~3 yrs old from work).  The Fluke 189 and the older U1272A are both out of calibration by a few years, the U1252B and the new U1272A and brand new and still within the 1yr cal cycle.  In order to observe the problem, the U1272A has to be placed in dual measurement mode, reading ACV on the big display, and DCV on the small display in the upper right.  Here are my findings:

Function gen. set to generate a 2KHz sinewave, with a 1.844V DC offset, with the output connected to all the DMMs in parallel.

Condition                                                                Fluke 189          U1252B (new)          U1272A (new)          U1272A (old)
2kz, 1.84V DC offset                                               318.68mV         0.3244V                  0.3330V                    0.3134V
Press "dual" button once to change U1272A           318.68mV         0.3244V                  0.3221V                    0.3224V
to read freq instead of DC in small display
Press "dual" button 3 times to read DCV in              318.69mV         0.3244V                 0.3330V                    0.3136V
small display again
Hold "dual" button down to exit dual mode              318.69mV         0.3224V                 0.3221V                     0.3223V

What is really odd here is that when your are in dual mode with the DC voltage is being displayed in the small display, the reading is off by ~100 counts.  If you exit dual mode or measure freq rather than DC voltage, the problem does not occur.  The problem does NOT occur at 200Hz, it seems to start at ~1Khz and above.  Also, if you reverse the display and have DCV in the primary display and ACV in the small display, it works fine.  It only occurs when ACV is in the primary display and DCV is the secondary display.

I think I can explain why the old and new U1272A's act differently, the old one has 2.00 firmware and the new one is 2.04, so I assume that is why the difference. I've repeated this experiment numerous times after power cycling all the meters, and the results are the same.   

Anybody else observed this issue?

[electr_peter]

Hi,

I agree Agilent/Keysight U1272A is very good meter indeed - it is accurate and very capable DMM.

I will test U1272A with 2.04 FW over the weekend to see if there is any issue with AC+DC in dual display mode. Can you test DMM with higher voltage levels in both AC and DC? It is not entirely clear what level of AC you are using.

Have you seen this safety note http://literature.cdn.keysight.com/litweb/pdf/U1272A-03-S.pdf ? It can be related - basically it says that in DC measurement mode you should have filter turned on to have accurate measurement.

[ArcticGeek]

In regard to the AC level, I guess I didn't specifically mention it, it was ~320mV AC as indicated by the various meter readings.

Yea, I had seen the warning about the LPF setting, but that only applies in ACV and ACmV ranges, not the AC+DC range that I was measuring with (that is the only range you can use to get both AC and DC readings of the signal).

I'll try a higher level of AC/DC voltages tonight, I just thought I'd throw this out to the community to see if anyone else has observed this....I spent most of the evening last night confirming what I thought I was seeing on these 1272A meters.  I guess I didn't mention it, but the Keysight U1252B meter and the Fluke 189 do NOT have this same issue when in dual measurement mode.

[EVS]

Anybody else observed this issue?

Yes

[ArcticGeek]

Wow,

Very nice data plot!   This definitely illustrates the problem nicely.  What is interesting is that your data shows that your meters are behaving similar to my older meter that has the 2.00 level firmware.  That is, the dual display reading is LOWER than the single display reading.  On my newer meter that has 2.04 firmware (like yours), it exhibited behavior where the reading was HIGHER than the single display.   At any rate, the reading is out of whack compared to what it should be. 

I did some more experiments last night and found that this problem seems to be exclusive to the 3V range.  If I raise the AC voltage to >3V, the meter goes to the 30V range (as expected), and the problem doesn't seem to occur.  With a 3.65V AC sine wave and a 1V DC offset, I did not observe the problem.  Furthermore, if I lower the AC voltage to 270mV and change to the mV range on the meter, I don't see the problem either.  So the problem seems to only exist on the 3V range.

Thanks for taking the data.  I might make a phone call to Keysight and report this problem - we'll see what they say.

[LaurentR]

I have a U1273A with FW 01.95 and haven't been able to see this. With the same parameters, the ACV reading is pretty solid when I change the dual display mode.

What I do see is that it takes about 10 sec to stabilize the first reading, and whenever I press the Dual button, the ACV reading droops a bit (or sometimes glitches to a totally different value) for the duration of exactly one reading, but then goes back to the original value.

[D3f1ant]

Sorry this post doesn't really add anything new other than more statistics.
Mine goes up exactly 40 counts on the main display when in dual display mode with dc in 2nd display. Oddly enough it takes few seconds to go back to the presumably correct single display value when cycling through other 2nd display readings. And it is does seem to only affect the 3V range.
FW 2.04. Test parameters where 1.8V offset Sine 2Khz 1VP-P

[ArcticGeek]

Sorry this post doesn't really add anything new other than more statistics.
Mine goes up exactly 40 counts on the main display when in dual display mode with dc in 2nd display. Oddly enough it takes few seconds to go back to the presumably correct single display value when cycling through other 2nd display readings. And it is does seem to only affect the 3V range.
FW 2.04. Test parameters where 1.8V offset Sine 2Khz 1VP-P

Actually this is useful information, because this tells me that this is much more widespread than just a few units.  I too noticed the ~2 second delay to go back to the correct display value when cycling thorugh the other 2nd display readings.  I do plan to contact Keysight about this, I'll post what I hear from them as soon as I get a chance to talk to their technical support group.  I find this problem very annoying, I dob't like it when my test equipment lies to me about the value the reading I'm taking....I don't like it one bit.

[ArcticGeek]

I don't know if very many people were following this thread, but if so I can report that Keysight has fixed this bug in their latest firmware release 3.03.  This version is available on their website, and if you have the USB/IR cable the download is pretty straightforward.  I downloaded the new code and verified that it fixes this issue.

Kudos to Keysight for listening to me and fixing the problem - I just wish they would have told me about the fix sooner!   The release date on this code was 9/21, just a few days after I reported it to them, but I didn't find out about it until now - grrrr!   Oh well, at least Keysight took the initiative to fix the problem.

[LaurentR]

I don't know if very many people were following this thread, but if so I can report that Keysight has fixed this bug in their latest firmware release 3.03.  This version is available on their website, and if you have the USB/IR cable the download is pretty straightforward.  I downloaded the new code and verified that it fixes this issue.

Kudos to Keysight for listening to me and fixing the problem - I just wish they would have told me about the fix sooner!   The release date on this code was 9/21, just a few days after I reported it to them, but I didn't find out about it until now - grrrr!   Oh well, at least Keysight took the initiative to fix the problem.

I am also surprised nobody noticed the new firmware. I didn't even notice new U1273A firmware from 2014. Strangely, their support page shows the 2013 firmware as the latest, but when clicking on the link, it actually goes to the newer version...
http://www.keysight.com/main/techSupport.jspx?pid=1896657&pageMode=OV&cc=US&lc=eng
"Firmware" tab

[HighVoltage]

New FW version again from keysight:

Version:    3.1.51130.01
Release Date:  2015-12-03

I am just installing it now.

[electr_peter]

I plan to test my U1272A with 2.0.4 FW first on AC/DC source, then do FW update and test it again. Just to be sure issue existed and was fixed.

[ArcticGeek]

HighVoltage,

Where did you find the new 3.15 version of firmware?  I checked the Keysight website and didn't find it.

[electr_peter]

I see "3.03 Firmware update for U1271A & U1272A" here. Firmware changes are listed here - it mentions only 3.0.3 on 2015-09-21.

2015-09-21
Version 3.03
-       Change to Keysight identity
-    Change shunt calibration limit from 10% to 20%
-    Remove AC offset calibration limit
-   Bug fixed for Spider LOG issue
-    Bug fixed for voltage alert error when setting count > 65535
-    Bug fixed for APO icon not dim in setup mode when APO = OFF
-   Bug fixed for Bar0 missing when exit view mode
-    Bug fixed for incorrect BAR polarity in dBm/dBV mode
-   Bug fixed for Event log index delay one record
-    Bug fixed for AUTO HOLD feature not clear when press dual button
-    Bug fixed for ACV + dual DCV

[ArcticGeek]

The 3.03 is the newest firmware I have found that is actually downloadable and usable.  It is interesting though that in the Technical Support section of Keysight for the U1272A, the current firmware is listed as:

Current Firmware/Software
Version:    3.1.51130.01
Release Date:    2015-12-03

Even though 3.1.5 is the latest firmware level, from what I can see only 3.03 is available for download and updating your meter.  I have tried the 3.03 on my meter, and it fixes the AC+DC problem that I originally noticed.

[electr_peter]

I have tested U1272A with 2.04 firmware version with [0.6; 5] AC + [-2; 2] V DC offset. AC frequency tested was 200-2500 Hz. While pressing buttons to get dual display, I only got ~10 counts change in AC voltage reading (0.6180 +- 0.0010V). I consider this to be very small difference. Various changes to setup did not result in any significant reading differences.

After installing 3.03 firmware this small issue seems to be solved (or issue changed status from almost undetectable to completely undetectable).

Note on FW update. I used Win7 PC with Agilent USB/IR lead. PC saw USB cable as Prolific serial port - this setup worked with Agilent handheld meter logger SW, but it did not work with FW update SW. I had to reinstall Prolific 2303 drivers (provided with FW update) to update DMM FW.
DMM needs to be turned on during FW update. DMM screen shows "Pro" with flashing "u" during update.