It's your place to connect with experts and peers, get continuous support, and share knowledge.
Contact SchneiderCommunity.Support@se.com if you have any questions.
Collaborate with multiple experts and discuss various topics about Power Meters and Power Quality. From design & implementation to troubleshooting and more, get support from experts and share your experiences by subscribing to the Schneider Electric Exchange forum today.
You can subscribe to this forum after you log in or create your free account..
Posted: 2018-02-20 05:57 AM
I'm writing in to report a pretty strange observation that one of our customers encountered when using the PM8000 and the on-board web server GUI to view the capture voltage and current waveform. To set the basis of discussion, I am in Singapore, our power grid is 400Vac (Line-Line), 230Vac (Line-Neutral), 50Hz and grounding system is TN-S. The meter I am using to replicate the error is connected single phase, and hence only adheres to 230Vac (Line-Neutral), 50Hz and grounding system is TN-S.
The ION Setup version used is,
This is not the latest version which is build 18019 as that build crashes upon attempting to retrieve waveform from the PM8000.
The steps to recreate the problem observed on site is as follows,
1) Firmware update the PM8000 to the latest 1.4.1 firmware and using the 8000_v1.4.0_FAC-PQ_184.108.40.206.0.dcf
2) Set the meter to the following using ION Setup
3) Set the power quality settings to capture 256 samples/cycle for 7 cycles
4) Do a manual waveform capture trigger
Viewing on ION Setup, the waveforms captured on the PM8000 V1 and I1 inputs are as follows,
Nothing out of the norm. The timescale per cycle is exactly 0.020s which is 20ms, corresponding to 50Hz
But now if you view the same waveform (remember, you have flashed the new firmware, everything is erased), this is what you see,
Almost immediately you will realize the time indicated at the lowest point on the first cycle is not as indicated by ION Setup 9:08:18.821 PM but rather 13:08:18:81234 (not quite sure how you get below 1ms accuracy when our timestamp is only accurate down to 1ms). I noted that the web GUI gave the timing in UTC, hence correcting for GMT offset of +8, that gave 9:08:18.81234 PM
To matters worse, looking at the lowest point on the next cycle (that gives 1 full cycle, lowest point to lowest point)
The time indicated on the occurrence is 13:08:18.81806 UTC (9:08:18.81806 PM) where it is indicated in ION Setup as 9:08:18.841 PM
Taking the difference between the 2 adjacent lowest points on the waveform obtained from the web GUI, the total time elapsed is (9:08:18.81806 - 9:08:18:81234) = 0.00572s = 5.72ms. Since when has the frequency changed from 50Hz to well over 170Hz?
I repeated steps (1) to (4) with the exception that I modified step (3) to capture 32 samples/cycle for 54 cycles,
The capture was indicated as started at 13:29:58.07553 and the end time is indicated as 13:29:58.11548, that gives 0.03995s = 39.95ms and 54 cycles elapsed? That gives over 1350Hz in the voltage waveform.
Credit for the above goes to Shanelle Chan who was corresponding with the customer and detected this error in the web GUI. Without her persistence, I doubt this error would have been surfaced. In summary, the errors found as below,
- Web GUI reports timing in UTC while the device is set to 12 hour clock
- Time reported by ION Setup difference significantly from that reported on the web GUI waveform viewer
- Waveform timescale on the web waveform viewer is not correct
- Timescale error changes from bad to worse when more cycles of waveform is captured
Appreciate the engineering / development teams look into this matter with urgency as it is impacting the credibility of our products.
Tan Kuan Khoon
Senior Specialist, PAE
Schneider Electric Singapore
Hi Tan Kuan Khoon,
Have you opened a technical support case for this issue? I would like to have somebody looking into it from our expert team.
Thanks for bringing this to our attention.
The issue with the ION Setup waveform viewer is an issue with the IONTKN.DLL which is not properly decoding the ION timestamp with a high enough resolution. The latest version of ION Setup now has a temporary fix to address this until we can figure out if we can easily fix the DLL in the future or some other means to address this issue.
To fix the crash issue, I have =S= only version as I'm still running some verification test on some fixes for this version:
As for the web viewer discrepancies, the Web GUI I believe uses the ComTrade files on the device and at the moment does not adjust for the programmed time zone. There may be other issues as well but I will let someone with more knowledge of those comment on that.
Hi Kuan Khoon TAN
I have logged this issue for further investigation with PME.
Thanks for reporting it.
Hi Tan Kuan Khoon,
I have created one case in BFO, case number is 46488333.
I will investigate it and contact you through the bfo system.
Thanks for the follow up on this. Looking forward to getting some concrete replies from our engineering departments.
Simply update the web pages (as described above) to the newest.
The best method to follow for latest firmware and web pages are these steps: