[W600][BL0937]Smart plug: power measurement 1 w without any load

I find that the power measurement report "1.xx Watts" consumption also with no LOAD
Also if the plug is set to OFF

I find a FLAG for not sent via MQTT the values when the plug is OFF (I think it is a workaround),
but I would like to understand if there is:
- a problem with the BL0937
- a problem with soldering of W600 chip (TW-02)
- a problem with some kind of pin configuration

I have two smart plugs and One is more stable than the other,
so I presume it can be a hardware problem?!?

Someone else have this "phantom power" ?

P.S.
I am on the last firmware version OpenW600 1.17.540

First of all, have you calibrated your device? With VoltageSet/CurrentSet/PowerSet commands, etc? Or via Web App (tools section)?

p.kaczmarek2 wrote:

First of all, have you calibrated your device? With VoltageSet/CurrentSet/PowerSet commands, etc? Or via Web App (tools section)?

yes, all calibrate with 60W lamp

the phantom current is 0.011A
the phantom power is 1.43W


for now, i found only a "workaround" to not send this "phantom power" to MQTT with the command:
VCPPublishThreshold 0.5 0.015 1.45 0.01
(i set this in startup command)

but, i would like to know if there is some problem with the W600/TW-02

Quote:

Someone else have this "phantom power" ?

Kind of:


It is comparatively rare, about +/- 0.02 W, "OBK-Smartplug1" and "OBK-Smartplug2" are not synchronized and the following situation occurs very rarely:


Both were plugged into the same strip and were therefore energized at the same time. Nothing connected - empty sockets.

It's Action LSC Smart Power Plug 320 2087
Build on Apr 8 2024 09:55:01 version 1.17.540
Short name: OBKSP1, Chipset BK7231N (CB2S)

P.S.: Both were calibrated as a compromise with a 60W and a 100W bulb, a 4-digit true RMS DMM (nnn.n V) and a not true RMS DMM (n.nnn A)

Added after 55 [minutes]:

The situation is now as follows (Energy last hour/today)



(The screenshots are two browser-windows side by side at the same time)

Before we debug any futher, can you tell us whether you're using PowerSave, the so called dynamic sleep?

We already had reports saying that PowerSave breaks BL0937 precision because BL0937 works by sending pulses which are counted by GPIO interrupt and dynamic sleep is messing with that...

Still, I don't think it's PowerSave, judging from the description of the problem, but it's still worth to check.

>>21039272
exactly like that:






the one in left have less "phantom power" than the one on the right... but they are the same Chipset W600 smart Plug with the same firmware version 1.17.540

The question is:
it depend on firmware?
or on hardware?

Added after 34 [minutes]:

p.kaczmarek2 wrote:

Before we debug any further, can you tell us whether you're using PowerSave, the so-called dynamic sleep?

We already had reports saying that PowerSave breaks BL0937 precision because BL0937 works by sending pulses which are counted by GPIO interrupt and dynamic sleep is messing with that...

Still, I don't think it's PowerSave, judging from the description of the problem, but it's still worth to check.

normally I have "PowerSave 1" in startup command,
but I send the command "Powersave 0" to test...
Nothing change

Does everything else works reasonably good and accurate with your current calibration?

BL0937 reports the voltage/current/power by sending a pulses of certain frequency to the WiFi module and WiFi module counts those pulses.

I don't think that pulses can be counted incorrectly if the measurements for higher loads are correct.

It may be possible that we just need to ignore very low measurements like 0.5W, etc...

That's the current state (Energy sums)


My startup command is:
backlog startDriver ntp; ntp_setServer 192.168.0.1; ntp_timeZoneOfs 1; ntp_setLatlong 50.012345 7.012345; waitFor NTPState 1; setupenergystats 1 60 180;

Higher loads I'll try in the evening... The (apparent and reactive) "powers" are only sometimes and only briefly non-zero, the real power is very rarely, but it adds up.

Maybe, as a workaround, I could introduce a configurable command like:

which will effectively cull all power values less than given value?

p.kaczmarek2 wrote:

Does everything else work reasonably well and accurately with your current calibration?

BL0937 reports the voltage/current/power by sending pulses of a certain frequency to the WiFi module, and the WiFi module counts those pulses.

I don't think that pulses can be counted incorrectly if the measurements for higher loads are correct.

It may be possible that we just need to ignore very low measurements like 0.5W, etc...

- the "phantom power" in my case is 1.43W
- this "phantom power" in my opinion is also "added" with all the power measurements

ES:
initial 0W -> sometimes measurement 1.43W
with load 5W -> sometimes measurement 6.43W
with load 10W -> sometimes measurement 11.43W
etc...

I think so because the timing when the measurement is 0W or 1.43W
is the same that changes the measurement "onload" change... from NORMAL LOAD (W) and NORMAL LOAD + 1.43W
(I don't know if I can explain better)

P.S.
The MinPower 1.45 can eventually help with false measurement
Thanks

BUT I think that does not resolve the issue
Maybe the low quality of BL0937, the +/- 2W is present on the datasheet (?)

My last state is


Apparent and reactive powers IMO should not add anywhere. They are certain aspects of loads, that shouldn't be counted by the electricity meters, but they place a (higher/additional) load on the power lines due to the (higher/additional) current they are associated with. This is the reason why a PFC* is mandatory for PC power supplies, for example. Only the active power is counted and paid for, but the grid has to bear/transport the reactive power, too.

I think, as seen in one of my screenshots, the BL0937 fakes sometimes some active power, too. However, this happens very rarely and only briefly.

The "better" of my plugs is directly plugged in a power strip, the worse has a cheap energy counter in between. Tomorrow I will interchange both and conduct a "high power" test with a water kettle and a toaster...

* PFC = Power factor control. In the past, capacitors were used to compensate for the inductive loads of motors and fluorescent lighting (resp. their current limiting chokes).

Is that really an issue? one my socket gets 0.005Wh per hour, that's like 44Wh or 0.05KWh a year, which is basically zero.

@miegapele I don't think that's a real problem, but it's curiosity that drives the world in a positive way. But, okay, sometimes it costs a bit too much time in a way... ;-)

@p.kaczmarek2 I did some tests with my two Action-LSC things, a 1900W kettle and a 100W bulb. I compared them to a very carefully calibrated NOUS A1T (that also uses a BL0937) and a cheap energy meter that is supposed to be accurate to 1% (V) to 2% (W) - if one can believe it.

The NOUS with Tasmota displays all values somewhat less quickly and more smoothly. But I have managed to calibrate both LSCs to tighter tolerances. However, both remain somewhat differently "fidgety" as far as the reactive and apparent powers and the power factor are concerned. At the NOUS without load all figures remain steadily at zero.

But: When I connected only the LSC sockets to the NOUS, I had a suspicion: The NOUS also began to sporadically display reactive/apparent power and changing PF values of comparable magnitudes. Could it be that some connector brands also measure their own SPS's power consumption?

Unfortunately I don't feel like opening the LSCs again and examining the circuit because of the fiddling with that somewhat special housings. Maybe @cicciobastardo can get his plugs open more easily and have a look?

Thanks so far for the support with "Jugend forscht" as we say in Germany when older men spend their time with old or new technical toys...

Have all a nice and curious weekend

The spurious power is probably just noise, due to the way these devices work, I think tuya hides any power below 4w to deal with it.

Also, Power factor and reactive power is calculated and not intended to be correct. Data for calculation is not read at the same time. BL0937 is not a precision measuring device. It's for approximation to like 5%. But that is good enough for most use cases.

There's one more thing:

I've had a lot of "wriggling" digits in front of my eyes today - an energy meter, two different multimeters and the displays of three IoT plugs on my notebook and smartphone.

Somehow it occurred to me that we will be dealing with differently "polluted" power grids - far from any clean sine wave - and that each digital measuring device will try to somehow integrate what it "sees" across possibly very crooked waveforms and each in different short time periods. More or less talented... (the LSC was 6.95€, four NOUS A1T 50€)

The truth of "true RMS" always will be relatively relative

I found another smart plug with BL0937
This smart plug uses ESP 02s chip so I flash Tasmota on the chip.

Well the problem is always present:
- power on without load and the measurement is 1.4 W





I calibrate with the same 60W lamp:
the result in phantom watts are similar, but I see that the phantom current and the power factor during phantom power is different
(?)

but for the "main" PhantomPower problem I think the cause is the BL0937 chip that makes the measurement
(maybe the watts consumption of the BL0937 chip itself? or the consumption of the smart plug?)

I search other plug to test, but maybe this thread can be closed

Added after 1 [minutes]:

p.kaczmarek2 wrote:

Maybe, as a workaround, I could introduce a configurable command like:

Code: text Expand Select all Copy to clipboard

MinPower 0.5

which will effectively cull all power values less than the given value?

this command exists in the last firmware?

@cicciobastardo

With two NOUS A1T/BL0937 on Tasmota 13.4.0 I surprisingly don't have this problem (as mentioned above). They remain rock solid at zero. However, I have now found out by experimenting with resistors that power measurements below 2 watts are not really reliable, even with the NOUSs.



One of the two LSC plugs (the "worse" one) has even been very erratic in the last hours: it has conjured up 275 W/h from somewhere, 220 of which in just one minute. For some unknown reason, it was switched on when I got there but could no longer be reached via WiFi. I should probably keep an eye on that?



The connected appliance is a tumble dryer, but it must have been in STBY mode (real mains switches have become very rare...)

So I think, too, the quality of the BL0937 is not particularly good and possibly fluctuating and it therefore depends very much on the material circuit design and the software how "nicely" the measured values fit into reality.