The kVAR value of PM5560 is unstable

[Tomohiro-Okada]

Hi, 

I did measurement test of PM5560 using three phase AC generator.

According to the test circuit diagram in the attached pdf data,

when inputting current and voltage with a phase angle of 0 degrees,

the measurement of var (reactive power) should be "0 kvar" as a correct value.

However, the display value is unstable,

fluctuating between -100 and 100 kvar without stabilizing around 0 kvar.

As W and VA show correct readings,

it is unlikely that there is a wiring mistake or a connection error with the testing instrument.

Do you have any idea?

I have one more question,

when the real value of "VAR" is zero, How do I think  acceptable error range?

For example, the real value of "VAR" is zero and PM5560 displays 0.01kVAR then, is it acceptable?

Thanks,

Okada

[N4th4n13L]

Hi @Tomohiro-Okada ,

That is an interesting observation! What is the test kit you're using?

Have you tried testing at near unity? Like instead of 0 degrees difference, at like 2 degrees leading and lagging? My calculation show that roughly 2 degrees off is causing the error your seeing. And, that is outside the accuracy class of the device. 

It seems to hold more steady at other test values though so you're not seeing the same error across the board, just at unity PF. Have you also looked at the phasor diagram of the PM5XXX during the test to see how it is measuring the phase angle? Some screenshots of the phasor diagram might help narrow down the issue. If we're seeing the phase angle moving a lot on the meter, and then narrow down if it's in the meter or the test kit.

[Miyai]

Hi @N4th4n13L ,

I am conducting the exam with Okada.

Our tester uses RX4763.

https://www.nfcorp.co.jp/english/pro/pp/p_tes/index.html

We added an iON9000 to the test circuit and tested it; the soundness of the test circuit and tester was confirmed since all iON9000s have the same theoretical and measured values.

Therefore, we believe PM5560 is responsible.

We checked the phase angle of PM5560 and found a slight error.

The active power is within the error range, but the reactive power values wobble.

When V2 and VN were connected, the reactive power was within the error range and the condition improved.

If the neutral terminal is involved, I would expect the active power to be affected as well. Therefore, I do not know why shorting V2 and VN would stabilize the reactive power.

There is no mention of shorting V2 and VN in the PM5560 installation guide.

Do you think the measure to connect V2 and VN is correct?

[Charles_Murison]

@Ramasamy_N Are you able to comment on wiring suggestion to connect V2 and VN when meter is using PTs and Delta wiring?

Regards,

Charles

[N4th4n13L]

Hello @Miyai ,

Thanks for the additional information. Could it have to do with how the meter determines it's reference voltage internally. I had assumed that the Vn/Vref terminal was used to determine reference. Delta knows to use Vab for reference, but is the Vref terminal needed to improve that measurement? If the reference voltage is left "floating" it would cause the phase angle to adjust slightly.

When I had installed ION 7X50s at power plants in the past I had always connected V2 to Vref when wiring the meters in Delta circuits. I know we had found issues when testing and only connecting Vb to the Vb or Vref terminal only. This doesn't provide the cause, but I have found connecting V2 to Vref to improve measurements on some devices. 

[Ramasamy_N]

Hello @Tomohiro-Okada @Charles_Murison 

Yes, The V2 and Vn terminals need to be grounded as shown in @Miyai comment. 

However the PM5560 measure and update the kVAR value if the phase angle is not 0. It does not round off the kVAR value to 0 like ION9000. 

[N4th4n13L]

Hello @Ramasamy_N ,

The ION9000 does have a use fund kVAR framework, that filters out noise and uses the power harmonics module for the kVAR values. It’s an external Boolean module that controls the logic. 

I would think the test kit signal should be clean though, but something to check that could be causing a difference in measurements. 

[Ramasamy_N]

Hello @N4th4n13L 

Thanks for the information, 

I requested the Power factor & phase angle values from PM5560 to verify the measured kVAR. 

[Tomohiro-Okada]

Hello @Ramasamy_N , @N4th4n13L,

Thank you for your information,

I understand that V2 and Vn terminals need to be grounded when using PM5560.

In this case,  is it OK if V2 and Vn are not grounded, as long as they are at the same potential?

and if the above wirings are needed, do you know the reason why above wirings are not mentioned in the latest user manual of PM5560?

And according to @Ramasamy_N  's request, I attached the result of additional verification test in my environment.

In this test, the measurement value of PM5560(VAR, VA and PF) was unstable when phase difference between voltage and current was 30 or -30 degrees, and there were measurement errors in entire test.

However, the measurement value of ION9000 was stable and there were no measurement errors

(completely same as the theoretical value).

It means the supplied voltage and current from three phase generator was correct. 

In this test, the values of PM5560's active power were stable,

do you have any idea about it (only active power of PM5560 was stable)?

Regards,

Okada

[Ramasamy_N]

Hello @Tomohiro-Okada 

Thanks for the details, 

The difference in active power measurement between the ION9K and PM5560 might be due to the power calculation method that followed in the meters. 

PM5560 follows the Arithmetic method (fixed calculation method) and the calculated PF in PM5560 (0.99261 LD) is not matching with ION9K reading (1.0000LG). 

Can you please check this calculation method configured in ION9K ? 

Regards, 

Ramasamy N

Can you please check the 

[DanL]

Hi @Ramasamy_N 

I believe you misspoke regarding "The difference in active power measurement" as active power (kW) shows very good agreement, while both reactive power (kVAR) and apparent power (kVA) show poor agreement.

I'm fairly certain that both devices are using the 2-element watt-meter method for computing active power (with DELTA volts mode) and the total active power is the summation of the 2 elements.

The "Arithmetic" vs "Vector" is configurable on the ION9000 for total apparent power (default is Vector).

The ION9000 will always report the Arithmetic Sum for active (kW) and reactive (kVAR) totals.

If I remember correctly, the per-element reactive power algorithm is very different on the PM5500 (fixed to using "Vector" for reactive) vs. ION9000 using either fundamental (default) or 1/4 cycle time shift.

[DanL]

When using PTs (connected to a primary side voltage exceeding the meter's safety limits), the secondary must be ground referenced. Normally the PT secondary common is tied to ground.

For your scenario, this grounded PT secondary common should be connected to the meter's V2 and VN measurement terminals.

[Tomohiro-Okada]

Hi @Ramasamy_N, @DanL ,

Thank you for your information.

By the way, I tried Direct VT connection wiring, too.(please refer to attached wiring method)

The setting of PM5560's "CT on Terminal" was "I1 I2 I3" and

the other test conditions and settings were same as previous verification tests.

In this case, the fluctuation and measuring error of VAR(KVA, power factor) happened again.

If I use the attached wiring diagrams of PM5560,

How could I avoid the problem?

(Could I connect V2 and VN same as the "2VT, 2CT" wiring method?)

Regards,

Okada

[Ramasamy_N]

Yes, @DanL I suppose to mention "The difference in reactive power measurement". 

We need to repeat this same test after configuring the VA calculation in Arithmetic in ION9000. Please swap the ION9000 and PM5560 meter (Sequence: Source - PM5560 - ION9000) 

Hello @Tomohiro-Okada 

Please repeat the above test condition after connecting the V2 and VN same as the "2VT, 2CT" wiring method. 

Also share the same results by connecting the PM5560 and ION9000 individually. 

Regards, 

Ramasamy N

[Tomohiro-Okada]

Hi @Ramasamy_N,

Thank you for your reply.

First, I did test about PM5560 and I attach the result of the test.

I connected PM5560 only (not with ION9000) as you said.

The CT Mode of PM5560 was "I1 I2 I3"

The Volts mode of PM5560 was "3Ph 3Wire Corner Grounded Delta"

Test Condition 1: Normal wiring (Same as latest user manual of PM5560.).

Test Condition 2: Normal wiring + VN wiring (Connected 3 phase ac generator's neutral bar to VN of PM5560)

Test Condition 3: Normal wiring + connecting the V2 and VN same as the "2VT, 2CT" wiring method. 

The result of Test Condition1 was not good.(the fluctuation and measuring error of VAR(KVA, power factor) happened)

The result of Test Condition2 was good, but I think when I use this wiring method, It is necessary that

V2 is not grounded, because if V2 is grounded, the potentials of the V2 and VN is same level and it means

the result will be same as test condition 3.

The result of Test Condition3 was not good, 

Although the result looked good, it was not different from 2CT mode.

The wiring method can't detect change of I2.

Do you have any idea about how to avoid the fluctuation and measuring error of VAR(KVA, power factor)?

Regards,

Okada

[Ramasamy_N]

Hello Okada, 

Thanks for sharing the test results.

The kVAR values in test diagram 1 and 3 is expected since Vn is open. 

Can you repeat the test diagram 2 with actual VT connections (as attached - short VT output and connect with V2 and Vn of meter voltage input) and share the results? Also share the harmonics applied from NF Rx4763 and meter. 

Please capture the PF (true PF and displacement PF (Modbus register-3092)) for the same time and share. 

Regards, 

Ramasamy N

[Tomohiro-Okada]

Hi @Ramasamy_N 

Thank you for your prompt reply.

I did test of the diagram 2 with actual VT connections and when I used VN phase Wiring, the result was OK.

(as attached - short VT output and connect with V2 and Vn of meter voltage input) .

The above result is for "2VT, 2CT" and I imagine that it is no problem if I connect V2 and VN when the wiring system is "2VT, 2CT" or "2VT, 3CT",
but I would like to know how to avoid the problem if the wiring system is
"Direct connect (not using VT), 3CT" or "Direct connect (not using VT), 2CT".

In this case, PM5560's CT Mode setting is "I1 I2 I3"(not "I1 I3") and I imagine that if the setting is "I1 I2 I3", PM5560 doesn't use 2-element watt-meter method for computing active power and using 3-element watt-meter method instead.

Do you have any ideas?

Thanks,
Okada

[Ramasamy_N]

Hello @Tomohiro-Okada 

The I2 current is calculated value when the CT on terminal is selected as I1, I3. 

If the CT on terminal configured as 3CT (I1, I2, I3 - generally used for unbalanced delta system), the meter will use actual I2 measured current for the calculation. 

Regards, 

Ramasamy N

[Tomohiro-Okada]

Hi @Ramasamy_N 

Thank you for your reply! and I apologize for the late reply.

I understand difference between calculation method of "I1 I3" and "I1 I2 I3".
if you don't have any idea about how to avoid the problem if the wiring system is "Direct connect (not using VT), 3CT" or "Direct connect (not using VT), 2CT", I'll ask it of schneider's "BFO".

Do you have any idea?

Thanks,
Okada

[Ramasamy_N · Accepted solution]

Hello @Tomohiro-Okada 

The CTs on terminal (2CT or 3CT) can be configured by the customer based on the load current balance %. 

It is general practice to select any 2 CTs in 3PH3W systems since the load current will be balanced.  

Customer can select 3CT even for 3PH3W systems, if they want to measure current in all 3 phase and calculate the total power rather than calculated value from 2CTs. 

[Tomohiro-Okada]

Hi @Ramasamy_N 

Thank you for your reply.

I understand your information.

Thanks,

Okada